Validation of Installation Methods for CSST Gas Piping to · PDF fileValidation of Installation Methods for CSST Gas Piping to Mitigate Indirect Lightning Related Damage Page 1 Executive

GTI PROJECT NUMBER 21323

Validation of Installation Methods for CSST Gas Piping to Mitigate Indirect Lightning Related Damage Reporting Period: April 23, 2012 through January 31, 2013

Report Issued: May 3, 2013

Prepared for: NFPA 54 Technical Committee GTI Project Manager: Andrew Hammerschmidt R&D Director, Infrastructure Sector 847-768-0686 [email protected]

GTI Technical Contact: Christopher J. Ziolkowski R&D Manager, Sensors and Automation 847-768-5549 [email protected]

Gas Technology Institute 1700 S. Mount Prospect Rd. Des Plaines, Illinois 60018 www.gastechnology.org

FINAL REPORT

Validation of Installation Methods for CSST Gas Piping to Mitigate Indirect Lightning Related Damage Page i

Legal Notice

This information was prepared by Gas Technology Institute (“GTI”) for the CSST Project Sponsors

Neither GTI, the members of GTI, the Sponsor(s), nor any person acting on behalf of any of them:

a. Makes any warranty or representation, express or implied with respect to the accuracy, completeness, or usefulness of the information contained in this report, or that the use of any information, apparatus, method, or process disclosed in this report may not infringe privately-owned rights. Inasmuch as this project is experimental in nature, the technical information, results, or conclusions cannot be predicted. Conclusions and analysis of results by GTI represent GTI's opinion based on inferences from measurements and empirical relationships, which inferences and assumptions are not infallible, and with respect to which competent specialists may differ.

b. Assumes any liability with respect to the use of, or for any and all damages resulting from the use of, any information, apparatus, method, or process disclosed in this report; any other use of, or reliance on, this report by any third party is at the third party's sole risk.

c. The results within this report relate only to the items tested.

Validation of Installation Methods for CSST Gas Piping to Mitigate Indirect Lightning Related Damage Page ii

Table of Contents

Page Legal Notice ................................................................................................................................. i

Table of Contents ........................................................................................................................ ii

Table of Figures ......................................................................................................................... iii

List of Tables ............................................................................................................................. iv

Executive Summary ................................................................................................................... 1

Introduction ................................................................................................................................ 4

Background ............................................................................................................................ 4

Experimental Approach .............................................................................................................. 6

Parametric Testing ................................................................................................................. 8

Simulation of Selected Scenarios ........................................................................................... 9

Model Validation Testing ........................................................................................................ 9

Data Analysis and Reporting .................................................................................................. 9

Results and Discussions ...........................................................................................................10

Basic Parameter Measurements ...........................................................................................10

High Current Damage Tolerance Testing ..............................................................................13

Initial Simulation of Direct Bonding ........................................................................................17

Laboratory Validation of Simulation Results ..........................................................................19

Analysis of Results ................................................................................................................22

Conclusions ..............................................................................................................................27

List of Acronyms .......................................................................................................................29

Appendices ...............................................................................................................................30

A – NFPA Standards Council Decision #10-2........................................................................30

B - Validation of Installation Methods for CSST – Phase 2 v2 Proposal, November 2011 .....30

C - CSST GAS PIPE LIGHTNING HIGH VOLTAGE AND HIGH CURRENT TESTS .............30

D – Initial Simulations by PowerCET .....................................................................................30

E - VALIDATION OF INSTALLATION METHODS FOR CSST - REPORT ............................31

Validation of Installation Methods for CSST Gas Piping to Mitigate Indirect Lightning Related DamagePage iii

Table of Figures

Page Figure 1 – Simulation Scenario 1 ............................................................................................... 9

Figure 2 – Typical Test Article from Manufacturers ...................................................................10

Figure 3 – Typical Impedance Measurement Set Up .................................................................11

Figure 4 – High Voltage Withstand Test Setup ..........................................................................12

Figure 5 – Typical 8x20 µS Current and Charge Waveforms ....................................................13

Figure 6 – Arc Damage Tolerance Test Set-Up .........................................................................14

Figure 7 – Typical Damage from 8x20 µS Current Pulse...........................................................14

Figure 8 – Damage Tolerance Data for 8x20 µS Current Pulse Series ......................................15

Figure 9 - Typical Damage from 10x350 µS Pulse at 5 kA ........................................................15

Figure 10 - Damage Tolerance Data for 10x350 and 8x20 µS Pulses .......................................16

Figure 11 – Typical SPICE Simulation Model ............................................................................17

Figure 12 – Scenario 2 with Manifold Bond ...............................................................................17

Figure 13 – Scenario 3 Simulation of Current Entering Through Electric Service ......................18

Figure 14 – Arc Charge versus Peak Current for Simulations ...................................................19

Figure 15 – Measured and Simulated Cases Compared ...........................................................19

Figure 16 – Model Validation Test Set Up .................................................................................21

Figure 17 – Validation test 0.5 inch CSST with 32 foot bond wire .............................................21

Figure 18 – Simulation of 0.5 inch CSST with 32 foot bond wire ...............................................22

Figure 19 – Model Validation Data versus Simulations ..............................................................22

Figure 20 – All Data Sets Charge versus Current ......................................................................23

Figure 21 – All Data Sets Expanded Scale ................................................................................23

Figure 22 – Arc Charge versus Duration for All Data .................................................................25

Figure 23 – Criterion Proposed by PowerCET ..........................................................................25

Figure 24 – Log Arc Charge versus Duration ............................................................................26

Validation of Installation Methods for CSST Gas Piping to Mitigate Indirect Lightning Related DamagePage iv

List of Tables

Page Table 1 – Test Program from SEFTIM Proposal Phase 2-v2 (Nov. 2011) page 15 ..................... 7

Table 2 – Charged Delivered by Waveform and Current Level ................................................... 8

Table 3 – Impedance Properties of CSST .................................................................................11

Table 4 – Average Dielectric Breakdown Voltage by Manufacturer ...........................................12

Table 5 – Arc Time-Charge Products ........................................................................................26

Validation of Installation Methods for CSST Gas Piping to Mitigate Indirect Lightning Related Damage Page 1

Executive Summary

The testing described in this report is part of the project originally proposed to the NFPA Research Council by SEFTIM to validate the effectiveness of a direct bond of CSST to earth ground for protection against the effects of indirect lightning strikes of corrugated stainless steel tubing (CSST) used for the delivery of fuel gas in buildings. The testing plan is designed to address the following points:

1. Validate whether or not bonding of CSST is an adequate solution to lightning exposure problem. 2. If bonding is the solution, validate how bonding should be done. 3. If bonding is the solution, validate the size of the bonding jumpers. 4. Determine if bonding should be done at a location or locations other than where the gas pipe

enters the building. 5. Determine if alternate methods can be used for safe installation, i.e., separation from other

equipment.

There are two areas that the testing plan explicitly does not address.

The sustained conduction of power line fault current by CSST is outside of the scope of this project; this condition has been shown to cause perforation in prior studies. This issue is properly addressed by circuit protection devices that detect the flow of fault current and disconnect it at the source.

Direct lightning strikes are outside of the scope of this project. While direct strikes may be simulated, there is no means to replicate these events in a laboratory. It will be shown that the focus of this project was both simulation and laboratory verification of the simulations. It is also the case that nearby strikes that induce currents in various residential structures are far more numerous than direct strikes, providing further motivation to deal with this category of event first.

Testing Plan

The proposed test plan consisted of three primary stages. First, physical testing of CSST will establish basic material parameters to use in a SPICE model. The resistance, capacitance, and self-inductance of CSST from multiple CSST manufacturers will be measured. The dielectric breakdown of the CSST jacket material will be measured. The magnitudes of the arc charges and currents required to perforate the CSST will be established.

Next, simulations of various scenarios of CSST and direct bond conductor lengths will be performed. The scenarios are chosen to be representative of piping and electrical systems typical of residential construction. These simulation results will be used as first approximations of lengths of CSST and bond conductor that will provide adequate protection from lightning exposure.

Finally, physical tests of selected scenarios will performed to verify that the simulation model produced reasonable results. Scenarios that involve extreme lengths of CSST or direct bond conductor cannot be physically tested due to the limitations of real world laboratory equipment. Scenarios that can be replicated in the laboratory will be identified and executed. Good correlation between the SPICE model results and the verification tests will enable predictions of the adequacy of specific lengths of bond conductors.

Initial Laboratory Testing

The initial laboratory testing was carried out in the facilities of Lightning Technologies Incorporated in Pittsfield, Massachusetts. The initial round of testing established several parameters: the inductance per meter and resistance per meter of various types of CSST were measured. The breakdown or “withstand” voltage of the CSST insulating jacket was measured. The amount of arc charge required to produce


physical damage to the CSST was also established. In the initial testing, samples of CSST from four manufacturers in both 1” and ½” diameter were tested.

The initial tests were carried out with both the 8x20 microsecond and the 10x350 microsecond current waveforms injected into CSST through an intentional arc. No direct bond conductor was used in these tests as their purpose was to determine the amount of arc charge required to perforate the CSST. A standardized method of pre-puncturing the insulating jacket and using a fine wire to initiate the arc was developed during these tests. In no instance was the 8x20 waveform able to cause significant damage to the CSST product. Also, no significant damage was caused by the 10x350 waveform at currents below 1 kA. These waveforms caused an arc flash and some surface damage to the CSST but no perforation. Only at the 5 kA and 10 kA levels did the 10x350 waveform cause an actual perforation in the CSST wall.

An additional finding was that there was little variation in these results from manufacturer to manufacturer. There was no need to test samples from all manufacturers; only to vary the diameter and length of the CSST.

Simulation Results

The simulations were based on the well-established SPICE computer model for electrical circuits. The values for resistance, inductance, and dielectric breakdown determined in the initial testing were factored into the model, as were the observed arc charge values required to damage the CSST. All the simulations made use of a 10x350 waveform at 10 kA as this corresponds to the most severe test levels.

The results of the model simulations were that CSST with no direct bond would burn through under these conditions. Further, a 6AWG copper bonding conductor of up to 164 feet (50 m) in length prevented burn through of the CSST. If the bonding conductor was reduced to 8m in length, arcing of any type was also suppressed. These results are for the worst case scenario, where the only ground bond wire is applied to the gas line where it enters the residence and the current is injected through the gas line. Other scenarios involving additional grounding at manifolds, appliances, and injection of current through the appliance gave results with slightly lower arc charges delivered in those instances where any arcing occurred.

Verification Testing

Follow up testing was performed at the LTI facility to further verify the predictions of the model. For reasons of practicality, the lengths of the CSST and of the 6AWG copper bond wire were limited to roughly 4.5 meters. This limitation is imposed by the current source available: it was not practical to provide 10 kA of current for the 10x350 duration if the lengths went beyond this. The currents monitored were: the total current into the test set up, the current passing through the CSST, and the current passing through the bond wire.

Model simulations were run for these adjusted lengths and the partition of the current predicted. The tests were run and the actual current waveforms recorded. The agreement for the total arc charge available between the simulations and the physical testing were within 10% of one another.

Another aspect of the verification testing was to re-measure the inductance and resistance per meter of CSST product. This was ultimately done by making two-turn loops of the CSST that were suspended about 6 feet above the floor. The measurements were performed with a very sensitive LRC meter. The inductance measured for the CSST was very close to that theoretically predicted for a smooth walled tube of roughly the same diameter. This was performed for both the 1” and ½” diameter product.

Conclusions

The connection of a 6AWG copper direct bond conductor between the CSST and earth ground diverts enough of the energy to prevent perforation over a wide range of conditions. Simulations of direct bonds of up to 164 feet (50 m) did not indicate perforation with inputs of 10x350 µS at 10 kA.


Simulation also indicates the addition of manifold grounds, appliance grounds, or any additional direct bonds further dilute the energy available to damage the CSST. The addition of a direct bond to gas manifold can provide a further 20% reduction in the current available.

In the absence of any direct bond conductor, CSST can be perforated with current represented by a 10x350 µS waveform carrying 5 kA or more, representative of a near lightning strike.

The simulation model and the verification tests were within 10% agreement for those scenarios where practical physical tests could be achieved. This is easily within the bounds of experimental error.

At no time during the testing, initial or verification, was any corrugation to corrugation arcing observed. As noted above, the measured inductance value for the CSST was very close to that predicted for smooth walled tubing. The only arcing observed was at the point where the coating was intentionally punctured and the initiator wire placed to encourage the arc.


Introduction

The motivation for this work was to provide the National Fire Protection Agency (NFPA) with sufficient data to make an informed decision as to the efficacy of direct bonding for Corrugated Stainless Steel Tubing (CSST) as a means of preventing lightning damage.

Background In 2009, the NFPA Standards Council became aware of concerns with the adequacy of the ground bonding provisions contained in NFPA 54, National Fuel Gas Code, for CSST in gas piping systems. In that Decision, the Council noted that the record before it revealed both jurisdictional and, more importantly, potential technical issues that called for further attention within the standards development process going forward. The technical issues involved whether the bonding requirements in NFPA 54 for protecting CSST against lightning related damage had been adequately substantiated. A Council Task Group was formed to gather information and make recommendations to the Council on CSST.

The Council Task Group reported back to the Council in a report dated February 11, 2010. The Council‟s consideration of this report is set forth in Standards Council Decision #10-2, attached as Appendix A. As more fully described in that Decision #10-2, the task group reported a lack of technical substantiation sufficient to ascertain whether the existing bonding requirements in NFPA 54 provided adequate protection from lightning induced surges. Concerned with the lack of technical substantiation, the Council Task Group concluded that a research program was necessary to “identify safe methods for the installation of CSST to protect against lightning induced failure with consequent gas leakage.”

After review, the Council agreed with the Council Task Group that CSST would need to receive further attention in the standards development process going forward. To assist the NFPA 54 Technical Committee with input and expertise concerning the lightning-related safety issues related to CSST, the Council also directed that an NFPA 54 CSST Task Group be formed containing expertise from members of the Technical Committees responsible for NFPA 54, NFPA 70®, National Electrical Code®, NFPA 780, Standard for the Installation of Lightning Protection Systems, and from other appropriate organizations such as those that certify or develop product standards related to CSST. More importantly, the Council directed that the CSST industry or others advocating the continued use of CSST in gas piping systems should validate the safe use of the product through independent third-party validated research and testing that can be reviewed and evaluated by standards developers in a timely way.

On this point, Decision #10-2 states, in greater detail as follows:

Over the next full revision currently scheduled to be in the Annual 2014 revision cycle, the industry or others advocating the continued use of CSST in gas piping systems shall validate the safe use of the product through independent third-party validated research and testing that can be reviewed and evaluated by standards developers in a timely way.

To assist in meeting the requirements of the Standards Council, Decision #10-2, a project meeting was organized by the Fire Protection Research Foundation (FPRF), in March of 2010. The project meeting included experienced members from key stakeholder areas, NFPA staff, NFPA 70, 54, 780, manufacturers, NAHB, and insurance. The meeting led to the framework of the project, a project scope, and a preliminary work plan. A project technical panel was assembled by FPRF, June 2010.

In July 2010 an engineering firm in the lightning area, SEFTIM, was selected by the technical panel. SEFTIM‟s first task was to complete a literature review and develop a gap analysis to inform a future research project designed to validate installation methods for CSST gas piping to mitigate damage due to lightning events. This initial engineering review and gap analysis work is referred to as „Phase I‟ of the project.


The SEFTIM report Validation of Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage (Phase I) was completed and distributed to stakeholders in April 2011. The executive summary of the report concludes with reference to the beneficial role of bonding metallic systems, but that there is a lack of sufficient information to validate installation methods of CSST gas piping to mitigate damage due to lightning events. The summary concludes with the need to perform a targeted testing program to gain greater information as a proposed Phase II of the project.

The project technical panel accepted SEFTIM‟s recommendation, and also selected SEFTIM to produce the Phase II Test Plan. SEFTIM produced the test plan, Validation of Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage, Phase II, Proposal V2 (November 2011), and this test plan was also accepted by the project technical panel.

In April 2012, the sponsors of the project, selected Gas Technology Institute (GTI), to manage the testing phase of the project as laid out in the SEFTIM Test Plan Phase II V2 (November 2011).

The Standards Council accepted this approach in Decision #12-15, August 2012, and also reminded the sponsors that the testing be carried out per the accepted testing plan, SEFTIM Test Plan Phase II V2 (November 2011).

Referring to this test plan:

“The Council believes that this test plan must be carried out in order to meet the intent of Decision #10-2.

Continuing with further clarifying remarks in #D12-15:

“The Phase II Test plan need not be conducted by the Research Foundation. It should however, be conducted or managed by a reputable independent, third party testing laboratory or similar entity which undertakes to conduct the testing as set forth in the Phase II Test Plan. In implementing the Phase II Test Plan, there will undoubtedly be a need to work out certain details of how the tests are to be conducted, and judgments about those details will invariably be called for by the independent entity that is chosen to implement the testing. This is to be expected and is acceptable so long as the independent entity makes those judgments and undertakes to do so in a manner that is consistent with the intent and purpose of the Phase II Test Plan.”

In October 2012, GTI presented a project update and initial findings to the NFPA54 Technical Committee. GTI has conducted further testing through December of 2012, and simulations in 2012 and early 2013. This report provides a summary of the execution of this testing program and its results.


Experimental Approach

The overall goal was to determine the efficacy of direct bonding of Corrugated Stainless Steel Tubing (CSST) as a means of preventing damage from nearby lightning strikes. The following is the specific language from decision 10-2:

Concerned with the lack of technical substantiation, the CSST Task Group concluded that a research program was necessary to "identify safe methods for the installation of CSST to protect against lightning induced failure with consequent gas leakage." The CSST Task Group report identified, among the areas that should be addressed, the following:

Validate whether or not bonding of CSST is an adequate solution to lightning exposure problem. If bonding is the solution, validate how bonding should be done. If bonding is the solution, validate the size of the bonding jumpers. Determine if bonding should be done at a location or locations other than where the gas pipe

enters the building. Determine if alternate methods can be used for safe installation, i.e., separation from other

equipment.

The data obtained from this work should provide a basis for defining an engineering solution for the grounding of CSST so as to prevent lightning damage. As described in the Executive Summary, the project work was divided into several distinct tasks that were to be executed sequentially. These tasks consisted of the following:

Parametric Testing that was intended to verify the basic physical attributes of the CSST in the laboratory

Simulation of several bonding conductor and lightning strike scenarios using the parametric data from the previous task to set up the computer model used for simulation

Validation Testing of several specific scenarios in the laboratory to verify that the simulation model does provide accurate results

Conclusions and Recommendations on the efficacy of ground bonding of CSST based on the results of laboratory testing and computer simulation

The overall test approach shown in Table 1 follows that laid out by SEFTIM in their proposal “Validation of Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage – Phase 2, v2 (November 2011)”, included with this report as Appendix B. The one exception to the test matrix below is that AC fault currents were not included in this scope of work; the focus of the subject project was lightning induced transients. The parametric and follow-up laboratory testing was performed by Lightning Technology Incorporated (LTI) division of National Testing Services. The simulation modeling of various bonding scenarios and lightning waveforms was performed by PowerCET Corporation. Project management and technical oversight was provided by GTI.


Table 1 – Test Program from SEFTIM Proposal Phase 2-v2 (Nov. 2011) page 15


Parametric Testing The initial task was to perform parametric tests that characterized the physical properties of the CSST tubing. This phase of the work was carried out by National Testing Services at the Lightning Technologies Incorporated (LTI) facility in Pittsfield, Mass. LTI has provided a thorough and comprehensive report on how this testing was carried out along with the raw data; this is included as Appendix C. This report will be cited extensively during the following discussion.

There were two aspects to this testing: first to capture physical data that could be used to improve the simulation model accuracy and second to determine if the physical parameters varied amongst CSST manufacturers. Samples of CSST piping with representative fittings were obtained from four manufacturers in ½” and 1” diameters. These samples were subjected to the following tests:

The resistance, capacitance, and inductance per meter of CSST were measured with respect to a ground plane.

The dielectric breakdown of the jacketing material was tested by placing CSST in proximity to a ground plane immersed in oil and incrementally increasing the voltage level until breakdown occurred.

The energy required to burn through the CSST was established by intentionally establishing an arc and incrementally increasing the energy until burn through occurred.

The practical details of the test plan were agreed upon by GTI, the CSST manufacturers, SEFTIM, and NTS. The following steps describe the initial parametric testing to obtain electrical characteristics and parameter data on the various CSST products.

Tests to be performed on, 1/2” OD, non-conductive, dielectric jacketed CSST samples of 1m to 2m in length, terminated with manufacturer specific end fittings to be adapted to standard black iron pipe threaded fittings for ease of laboratory attachments to transient generator return (for high current tests).

Impedance measurements consisting of per unit length values for CSST self-inductance, parasitic shunt capacitance and DC resistance (all with respect to a 1m distant ground plane below CSST).

Determine the dielectric strength of the CSST insulating sheaths using the standard 1.2μsx50μs voltage impulse. Potential levels needed are on the order of 25~35kV, peak. Potential applied with respect to flat, grounded electrode flush with CSST outer jacketing.

Repeat the high current, charge transfer tests of the various non-conductive jacketed CSST products.

o The jacket was pre-punctured in a repeatable fashion using the hot tip of a soldering iron o A 38 AWG wire was placed between the electrode and the CSST wall to reliably initiate

the arc Charge transfer testing used the following standardized impulse waveforms and current levels as

shown in Table 2. Measured parameters for the tests would be applied currents, including the associated charge

delivered, and degree of CSST wall perforation/melt-through.

8μs x 20μs 10μs x 350μs

1kA 0.0175C 0.498C

5kA 0.0873C 2.49C

10kA 0.1746C 4.98C

Table 2 – Charged Delivered by Waveform and Current Level


Simulation of Selected Scenarios This portion of the work would use the measured parametric, electrical characteristic data of various manufacturers‟ CSST lines as inputs into a lumped parameter, PSPICE based circuit model.

Analytical model work was based on several scenarios suggested in the SEFTIM Phase 2 proposal such as shown in Figure 1.

The modeling was actually created/performed by M. Stringfellow of PowerCET. The circuit model is then used to provide predictions of voltage rises and/or current divisions at

various locations within a modeled „CSST installation‟ system, and to serve as basis as an engineering tool towards determining the recommended direct bonding to ground provisions for CSST residential installations.

Figure 1 – Simulation Scenario 1

Model Validation Testing This portion of the work would validate the circuit model via a simplified CSST arc entry test arrangement. A validated model will be a valuable tool in determining the best practices for direct bonding of CSST installations.

Second series of experimental tests on sections of various CSST lines would be used to validate the results of the analytical circuit model. These tests would consist of recreating a simple CSST to earth ground direct bond arrangement (already modeled), imagined to include representations of the grounding wire, grounding rod, and rod to earth impedances as paths back to the transient generator current return.

Measured parameters for the tests would be applied currents, voltage rise measurements at grounding elements and degree of CSST wall perforation/melt-through. Only one side of the CSST would likely be grounded for this test, the recommendation being that whatever physical situation was modeled be reproduced in the test lab for purposes of validations of the model results.

Data Analysis and Reporting The data generated by LTI and PowerCET in the preceding tasks will evaluated by GTI to verify that the model results are reasonable. The validated model can be used to arrive at recommending direct bonding configurations for CSST service entrance grounding best practice. The resulting report will be further evaluated by a panel of industry advisors to verify that the issues have been thoroughly addressed.

Validation of Installation Methods for CSST Gas Piping to Mitigate Indirect Lightning Related DamagePage 10

Results and Discussions

The following section will provide a high-level synopsis of each aspect of the test program, both experimental and simulation. The detailed supporting data is provided in the reports written by LTI and by PowerCET, which are provided as appendices. What can be drawn from the data will be discussed in detail in the “Conclusions” section. The conclusions at a high level are:

1. The physical properties of the CSST product are reasonably repeatable from sample to sample and across manufacturers.

2. The quantitative level of current and charge transfer that is required to burn through the CSST wall was established during this testing.

3. The simulation model gave results indicating that direct bonding the CSST at the gas service entrance with a 6AWG conductor providing an adequate drain to keep the current levels well below the established burn through level over a wide range of conductor lengths.

4. Further laboratory testing scenarios with side by side simulations showed that the two methods were in good agreement, with a worst case variance of 5%.

Basic Parameter Measurements The physical properties of samples of CSST tubing from four manufacturers were measured both to inform the simulation model and to verify uniformity across manufacturers. The test articles provided by the manufactures consisted of 2 meter lengths of CSST tubing already made up with the end connectors as shown in Figure 2. These were provided in both 0.5 inch and 1 inch diameters. Also provided were typical direct bonding clamps as would be applied to black iron pipe (BIP) immediately adjacent to the point where the CSST and BIP are joined.

Figure 2 – Typical Test Article from Manufacturers

The first round of physical tests was designed to capture the typical resistance, inductance, and capacitance per meter of CSST product. The typical set up was to suspend one of the provided test articles above a grounded surface as shown Figure 3. A set of precision meters were hooked to the test articles and the values recorded. The detailed procedures for calibrating this set up and the methodology for the measurements can be found on pages 8 through 18 of Appendix C.


Figure 3 – Typical Impedance Measurement Set Up

The LTI measured values for the various impedance parameters for the CSST test articles were normalized to a per unit length and recorded. The values are given in Table 3. Section 6.2, page 10 of Appendix C provides insight as to the reasons for the 0.5 inch tubing from Manufacturer B exhibiting higher resistance than that from other manufacturers. In short, the test articles were provided with the brass end fittings already applied; these were tested by LTI as received. The fitting to tubing resistance is also included in the measurement and, in this instance, added resistance. Were the tubing measured without the end fittings, the bulk properties of the CSST would be more uniform.

Table 3 – Impedance Properties of CSST

In addition to the impedance components of resistance, inductance, and capacitance the dielectric breakdown voltage of the CSST jacket material was also measured. This measurement required the use of a specialized high voltage generator capable of raising the CSST test items to a potential in excess of 30kV. The high potential is developed between the test item and a grounded metal plate. In order to prevent spurious arc flash over from occurring, the section of the CSST under test and the metal plate were submerged in an oil bath. The test sample is inserted into the oil bath and connected to the HV generator; the potential applied to the sample is gradually increased until a breakthrough of the coating is achieved. The sample is then repositioned so that an undamaged portion of the coating is adjacent to the


ground plate and the procedure repeated. A total of ten breakthroughs were accomplished for each size and manufacturer to provide statistical significance. The set-up is shown schematically in Figure 4 and the detailed procedure with photographs can be found on pages 12 through 24 of Appendix C.

Figure 4 – High Voltage Withstand Test Setup

The general findings of the dielectric breakdown testing are given in Table 4. The data is a composite of 10 tests performed for each manufacturer and CSST diameter. AS can be seen from the data, the dielectric breakdown voltage is fairly uniform with the exception of manufacturer B. This variance is caused by a thicker polyethylene jacket than that found on the other three manufacturers, so the result is not surprising. The thicker jacket material probably contributed to the higher contact resistance between the tubing and its end fittings that was seen in the previous discussion of resistance per unit length. The supporting data for all of the test runs is provided on pages 24 through 57 of Appendix C.

Table 4 – Average Dielectric Breakdown Voltage by Manufacturer

Size/Mfg. A B C D

0.5 inch 32.9 kV 55.5 kV 30.6 kV 30.8 kV

1.0 inch 35.8 kV 60.6 kV 35.8 kV 37.8 kV


High Current Damage Tolerance Testing A series of tests were carried out by LTI to quantify the conditions that are required to cause damage to CSST product. For this series of tests, an arc was intentionally created at the wall of the CSST and all of the available current was passed through this arc. This test was performed for multiple samples of CSST from each manufacturer and each diameter for a total of 175 test pulses. The general experimental set up and calibration for this test series is detailed on pages 57 through 64 of Appendix C.

A typical 8x20 µS waveform is shown in Figure 5. The double exponential waveform is the standard for lightning testing and some discussion is in order. The first number expresses the number of microseconds required to reach the peak test current, the second number, the microseconds required for the current to decay to 50% of its peak value. The upper of the two figures shows the current level as a function of time; the lower figure is the total charge delivered by the waveform which is found by integrating the current over time. From this discussion, it can be seen that a peak current must also be specified to completely describe the test waveform. The other standard waveform that is often used for lightning testing is 10x350 µS with a specified peak current. There is discussion in the literature (ICLP2004-74) of the 100x1000 µS waveform but it is not widely adopted as a test standard for two reasons: there is not sufficient observational evidence that many real strikes fit this envelope and it is difficult to produce this waveform in the laboratory at any great current level.

Figure 5 – Typical 8x20 µS Current and Charge Waveforms

Two series of tests were run: one using the 8x20 µS waveform and a second using the 10x350 µS waveform. Each waveform was run at three different current levels: 1 kA, 5 kA, and 10 kA. The complete damage tolerance test results, photographs, and numerical data can be found on pages 64 through 78 of Appendix C. A high level synopsis of the results follows.

The arc was created at a specific point on the test sample in the following manner. First, a small hole was placed in the polyethylene jacket using the tip of a heated soldering iron, exposing the metal. A pointed electrode is placed near to the opening and a short length of 38AWG wire is placed between the CSST and the electrode. This “initiator” wire is vaporized almost instantly by the current pulse but establishes a bridge of ionized air that directs the arc where it is required. The pointed electrode is connected to the output of the current generator and the far end of the CSST is connected to the generator ground, completing the circuit. This arrangement is shown schematically in Figure 6.


Figure 6 – Arc Damage Tolerance Test Set-Up

The initial series of 8x20 µS waveform tests did not produce any perforations of the CSST samples at the 1 kA, 5 kA, or 10 kA current levels. There was, however, some discoloration and burn marks around the arc entry point. The test series was repeated using the 10x350 µS waveform with a similar result for the 1 kA current level: discoloration and small scale melting without perforation. Figure 7 shows the damage typical of these arc flashes that do not result in perforations. Figure 8 summarizes the data for the entire 8x20 test series as charge delivered versus peak current. The three current levels (1kA, 5kA, 10kA) are clearly visible as is the linear relationship between them.

Figure 7 – Typical Damage from 8x20 µS Current Pulse


Figure 8 – Damage Tolerance Data for 8x20 µS Current Pulse Series

At the 5 kA and the 10 kA current levels the 10x350 µS waveform did consistently produce perforations in the CSST wall. At this level, each test delivered 2.5 of 5.0 Coulombs of charge respectively. A typical perforation caused by these conditions is shown in Figure 9. The cumulative data for all of the test pulses is shown in Figure 10. In this representation the upper line of data spans all of the 10x350 test pulses: the red points represent perforations (P) and the blue points, no perforation (N). The 8x20 data points are included in green in order to provide scale perspective of the testing regime. This data begins to outline the safe operating zone for CSST subjected to lightning induced transient current.

Figure 9 - Typical Damage from 10x350 µS Pulse at 5 kA


Figure 10 - Damage Tolerance Data for 10x350 and 8x20 µS Pulses


Initial Simulation of Direct Bonding Using the data from the physical measurements performed by LTI, a set of initial simulations was performed by PowerCET. Figure 11 shows the set up for a Simulation Program with Integrated Circuit Emphasis (SPICE) model for the Scenario 1 simulations. These simulations were provided to GTI by PowerCET as a PowerPoint presentation. The entire presentation is available as Appendix D.

Figure 11 – Typical SPICE Simulation Model

Three different scenarios were simulated in this first round of testing. The basic scenario consists of a direct bond conductor attached to the gas system at the point of transition between BIP and CSST; this scenario assumes that the current pulse enters the residence through the gas line, as shown in Figure 1. Two other scenarios were also modeled: one wherein the CSST is run through a manifold that is also grounded through a 14AWG copper conductor (Figure 12) and a final scenario in which the transient current enters the residence through the appliance electrical connection (Figure 13).

Figure 12 – Scenario 2 with Manifold Bond


Figure 13 – Scenario 3 Simulation of Current Entering Through Electric Service

The following are the salient points of these simulations:

All three simulated cases were modeled with the 10x350 µS waveform at 10 kA, representing a severe lightning event, but short of a direct strike.

In all three simulations one case was run with no direct bond applied at the gas service entrance; in all cases a perforation of the CSST was indicated when the direct bond was absent.

Additional test cases were run with varying lengths of 6AWG direct bonding conductor applied at the gas service entrance; no perforation of the CSST was indicated in any case where a direct bond was connected.

For this initial set of simulations the bond length was varied from 13 feet to 98 feet (4m to 30m). Later simulations were run out to 164 feet after validation testing. In no case where a bond was attached was perforation indicated.

The addition of the manifold bond created a 20% decrease in the arc charge passed but also had the side effect of prolonging the duration of the arc. Later analysis shows that these two effects may cancel one another.

For perspective, the addition of a 98 feet length of 6AWG direct bond versus no direct bond decreases the arc charge passed by 1700% in the same simulation sequence.

The addition of the manifold bond made the decrease 2200% versus the unbonded case.

The cumulative data for all simulations of CSST with direct bond conductors is given in Figure 14. The right-most point in the data set represents a simulated bond length of 164 feet (50m) for a Scenario 1 case. The data set also includes the Scenario 2 and 3 cases which introduce some scatter into the data. The simulation data is then overlaid with the damage tolerance testing data in Figure 15 as the purple points. The simulations that did indicate perforations would fall on the same trend-line as the 10x350 P (perforation) series. This gives some perspective on how the addition of the direct bond affects the behavior of the system. Displaying the data in this fashion provides guidance as to which region of this data space constitutes a safe operating area for bonded CSST.


Figure 14 – Arc Charge versus Peak Current for Simulations

Figure 15 – Measured and Simulated Cases Compared

Laboratory Validation of Simulation Results A further series of tests were carried out by LTI to verify that the results of the simulation model did indeed provide an accurate prediction of real world conditions. Several simplified Scenario 1 test cases were set up in the LTI laboratory and compared with the corresponding simulation. The detailed description of the set up and execution of these tests with resultant data is given on pages 78 through 97 of Appendix C. In addition to the data reporting provided by LTI, Appendix E from PowerCET provides an analysis of the follow-up testing. These tests were witnessed by personnel from PowerCET and GTI, so the feedback on agreement with simulation model was immediate.

Another issue that was dealt with during the validation testing was measurements of CSST inductance. This was done in order to address concerns that either the corrugation geometry or the magnetic


properties of the stainless steel causes CSST to have unusual inductance characteristics. There have been instances where adjacent corrugations show arc damage, leading to the conjecture that the arc can travel “ridge to ridge”. A much more likely scenario is that consecutive arcs struck from an external conductor as described in Appendix C, page 65: the raised portion has a sharper curvature and hence greater electric field with respect to external conductors.

As noted earlier in the report, LTI measured the inductance of CSST by the method called out in the SEFTIM proposal: suspending it a known distance above a ground plane. While this method is theoretically sound, it is difficult to implement in the laboratory. The return current at either end of the CSST sample must be shielded in order for it not to interfere with the measurement. The shielding may interfere with the measurement by bringing a portion of the ground plane close to the item under test. An alternative measurement method was suggested by PowerCET that minimized these sources of error.

The CSST inductance was measured during the validation testing by an alternative method of creating two-turn coils of CSST with a know diameter. Coils of 0.5 inch and 1inch diameter CSST were made and suspend on wooden stands 5 feet from the floor to isolate them from the environment. The inductance of these coils was then measured with a sensitive LCR meter. The inductance values found were nearly identical with those of straight walled tubing of the same OD. These inductance values with then used in subsequent simulations with good effect, as will be seen. These observations do not support the premise that CSST has unusual inductance properties due to geometry or material. Pages 28 through 30 of Appendix E provide a technical discussion of this measurement method.

There has also been discussion of the CSST providing a “waveguide” for the transport of high frequency electromagnetic energy. If one examines the specifications for commercially available circular waveguides several interesting facts emerge. Low-loss commercial waveguides have very high tolerances for uniformity and eccentricity. Stated another way, the optimum waveguide will have a constant cross sectional area and be perfectly round. There is also a high tolerance on the flatness of the interior finish for commercial waveguides. Given that CSST undergoes several changes of cross section per inch, it will provide a poor waveguide at best.

Not all the simulated cases of the 10x350 µS waveform at 10 kA can be reproduced in the laboratory. The limitation is governed by the internal impedance and charge storage capacity of the current pulse generator in combination with the impedance of the CSST and grounding system. As the lengths of the CSST and direct bond increases, higher voltages are required to maintain the waveform shape. The current generator in use has an upper operating limit of 36 kV and a current output limited by the load impedance. The series of verification tests were run with CSST sections 15 feet in length and varying lengths of direct bond conductor. Figure 16 shows the typical set up for this test series. The current pulse is launched from the generator in the background; the CSST and bond wire under test extend into the foreground.


Figure 16 – Model Validation Test Set Up

Figure 17 and Figure 18 are illustrative of the results of the arc charge and current validation testing; they provide the measured and the predicted results of a test scenario respectively. In this particular instance a 32 feet (10m) bond wire was attached to the test section of CSST. In all instances, the amount of measured charge transferred through an arc (if one were present) agreed to within 10% of the simulation predictions. Figure 19 shows some of the validation test data overlaid with the simulations. The figure shows that the all the data fits the linear trend predicted by the model. The laboratory testing was able to test some regions of higher charge and current than the initial simulations without producing any perforations of the CSST. From this finding it is reasonable to assume that extrapolating the model to direct bond lengths beyond 100 feet (30 m) would still give good results.

Figure 17 – Validation test 0.5 inch CSST with 32 foot bond wire


Figure 18 – Simulation of 0.5 inch CSST with 32 foot bond wire

Figure 19 – Model Validation Data versus Simulations

Analysis of Results In Appendix E, PowerCET asserts the premise that the likelihood of damage to the CSST can be predicted by both the amount of charge available to the arc and by the duration of the arc in time. The addition of a direct bond conductor provides an additional path to drain the charge from the CSST, shortening the duration of any arc or eliminating altogether. When the duration is shortened, the total charge passed through the arc is also reduced. GTI‟s analysis of the data, both measured and simulated, supports this premise.

Figure 20 shows the composite of the various data sets plus an “Additional” set of damage tolerance tests that were run during the validation tests. All of these additional tests resulted in perforations. A discussion of these tests can be found on page 10 of Appendix E; the raw data at the end of Appendix C.


Figure 20 – All Data Sets Charge versus Current

Figure 21 shows this data on an expanded scale. At first glance one sees a very close boundary between the “Additional” perforations that were generated during validation testing and some of the no perforation cases of the initial testing. At least one of the validation test cases that did not produce a perforation is close to the boundary. Also note that the simulated cases of direct bonding are generally far from this boundary.

Figure 21 – All Data Sets Expanded Scale


There is an alternative way of viewing the data that is more fruitful: it is possible to extract a rough duration of the arc from the data by the following means. The current through the arc, defined in Coulombs per second, gives a rate at which the charge passes through the arc. From many of the measurements and simulations we also have the total charge dissipated by the arc. We can scale the average current through the arc using 0.5*I-peak; this scaling is implicit in the double exponential waveform as described on page 4 of Appendix B (SEFTIM proposal). The approximate duration of the arc in microseconds is given by the following, where Q is the number of Coulombs passed by the arc and the peak current is given in KiloAmps.

If we now plot the data series in terms of arc charge versus the arc duration, there is a much clearer separation of the region in which the bonded CSST can operate from that where there is demonstrable damage. Figure 22 shows all the data plotted in this fashion. All of the simulated bonding conditions and the bond validation test data have arc durations of 200 µS or less whereas all the experimental perforations to date have arc durations in excess of 900 µS.

Figure 23 shows a damage criterion proposed by PowerCET that accounts for both the charge and duration of an arc that substantially agrees with our observed data. It implies that there is a value of Q*t above which damage can be expected. PowerCET cites studies for aircraft skin that indicate that a value of Q*t = 300 would be a reasonable estimate of this boundary condition. Figure 24 shows the test data on a log scale similar to the previous figure; the boundary line Q*t=300 is overlaid on the data sets. Clearly all the experimentally observed perforations are above this line. Further work could refine the position of the boundary but the trend is clearly shown.

A set of Scenario 1 simulations are given in Table 5 with arc charge time products calculated. The simulations provide the peak currents and arc charges out to a bond length of 198 feet (60m). At this length, the Q*t product is roughly 70. As we have seen from the experimental results, the boundary where perforation begins is in the vicinity of Q*t=300. This would indicate that there is reasonable latitude that one can exercise when choosing the bond conductor length.

It is also a corollary that perforation can be expected where the values of t or of Q become extreme; this is not a surprising result. It has been shown in previous studies that sustained (long t) fault currents from AC power lines will cause damage. Direct lightning strikes can provide 50 kA (large Q) or more in pulses of short duration. Both of these scenarios are outside of the scope of this work. Future work could refine our quantitative knowledge of the Q*t conditions required to cause perforation.


Figure 22 – Arc Charge versus Duration for All Data

Figure 23 – Criterion Proposed by PowerCET


Figure 24 – Log Arc Charge versus Duration

Table 5 – Arc Time-Charge Products

Bond

feetI-peak Q-arc t -µS Q*t

7 270 0.01 89 1

16 740 0.03 81 2

33 1500 0.07 93 7

49 2160 0.10 93 9

66 2770 0.14 101 14

82 3280 0.18 110 20

98 3720 0.22 118 26

131 4470 0.28 125 35

164 5060 0.35 138 48


Conclusions

The project work was planned and executed in order to address the following points:

1. Validate whether or not bonding of CSST is an adequate solution to lightning exposure problem. 2. If bonding is the solution, validate how bonding should be done. 3. If bonding is the solution, validate the size of the bonding jumpers. 4. Determine if bonding should be done at a location or locations other than where the gas pipe

enters the building. 5. Determine if alternate methods can be used for safe installation, i.e., separation from other

equipment.

The overall conclusion is that direct bonding of CSST to earth ground clearly limits the amount of charge available on the CSST during nearby lightning strikes versus the unbonded condition. Limiting the charge available was shown to prevent perforation of the CSST in all the simulated and observed cases that made use of a ground bond. Simulations of direct bonds of up to 164 feet (50 m) did not indicate perforation.

Perforation of the CSST was observed only in those instances where there was no direct bonding whatsoever. Even in the unbonded cases, a current on the order of 5 kA with a waveform of 10x350 µS was required to cause perforation. The amount of charge delivered in these cases is severe, representative of a lightning strike near the home.

Direct bonding of secondary gas manifolds did provide incremental benefit. Simulations indicate that the total available arc charge was reduced by an additional 20% when a manifold bond was used in conjunction with a primary bond.

The data measured and simulations based on this data indicate that two conditions must be satisfied before a perforation can take place on the CSST: there must be sufficient charge present and the duration of an arc must be long enough. Without these two being satisfied, it is improbable that the metal can achieve a high enough temperature for perforation to occur. The critical value appears to be in the vicinity of Q*t=300 where t is in microseconds and Q is in Coulombs. Further work could refine this value.

Direct bonding substantially shortens the duration (t) of arcing, or eliminates it entirely, by providing alternative paths to dissipate the available charge, thus removing one of the necessary conditions for perforation. A simulated direct bond with a length of 164 feet provides a Q*t of roughly 48.

The concept that there is a value of Q*t above which damage can be expected is supported by the observed data that a long duration AC power fault current carried on CSST can lead to arc perforations. It would also indicate that a direct lightning strike may carry enough charge to cause damage; even though the duration is brief a large number of Coulombs (Q) may be transferred. This conclusion indicates that, while direct bonding will provide very broad protection from lightning damage, it does have limits.

The issue of CSST exhibiting unusual inductance properties due to its geometry or material properties was examined. The self-inductance of CSST was measured by two different methods, by suspending samples of CSST over a ground plane and by making coils of the CSST. In both methods the measured inductance values were very close to that predicted for a straight walled conductor of the same OD. The coil method is preferred as it minimizes sources of experimental error. These inductance values were used in the simulation models to predict arc current and charge. The accuracy of these models were also verified as part of this testing. There is no real evidence to support the premise of unusual inductive properties.

These findings were developed through a combination of experimental work and simulations. Initial experimental work measured baseline parameters of CSST samples from several manufacturers and


verified there was little variation from manufacturer to manufacturer. Using this data, simulations of a number of CSST and ground conductor configurations were carried out. A subset of the simulated configurations were set up in a laboratory and tested for verification. The agreement between the simulated and tested cases is described in the following excerpt from the PowerCET Final Report (Appendix E). Configurations involving extreme lengths of CSST and direct bond ground conductor could not be tested in the laboratory due to the real limitations of the current pulse generator.

All simulations produced predicted waveforms in the various paths that were close to those measured in the laboratory. Simulated peak current magnitudes differed by between 1% and 4% of those measured. Predicted waveforms were within about 5% of those measured, with the largest discrepancy being on the wave tail. These minor discrepancies occurred in cases where it was difficult to measure the inductance of the circuit in the presence of parallel conductors. In all cases, the simulated arc waves resulted in calculated charge transfer within 10% of that measured in the laboratory tests. These results are well within experimental error and quite sufficient to validate the simulation models for their intended purpose.

Given the finding that direct bonding of CSST can prevent perforation by dissipating energy; the concept of separating the CSST from other facilities must be re-examined. The typical residential HVAC systems bring gas, electricity, and sometimes water facilities together at a single node. While it would be possible (in new construction) to separate the facilities leading up to a furnace or boiler, maintaining the separation at the HVAC unit has practical problems. As engineering control to provide lightning protection, direct bonding is much more straightforward to implement.


List of Acronyms

Acronym Description

LTI Lightning Technologies Incorporated

CSST Corrugated Stainless Steel Tubing

BIP Black Iron Pipe

NFPA National Fire Protection Agency

SPICE Simulation Program with Integrated Circuit Emphasis


Appendices

A – NFPA Standards Council Decision #10-2 This decision document was produced by the NFPA Standards Council that deals with jurisdictional and technical aspects of CSST installation practice. In the technical realm calls for work to be performed that will lead to the technical substantiation of direct bonding practice for CSST that will address the following points.

Validate whether or not bonding of CSST is an adequate solution to lightning exposure problem. If bonding is the solution, validate how bonding should be done. If bonding is the solution, validate the size of the bonding jumpers. Determine if bonding should be done at a location or locations other than where the gas pipe

enters the building. Determine if alternate methods can be used for safe installation, i.e., separation from other

equipment.

B - Validation of Installation Methods for CSST – Phase 2 v2 Proposal, November 2011 This proposal document “Validation of Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage – Phase 2 v2 (November 2011)” was produced by SEFTIM. This proposal was developed after the completion of the Phase 1 Study, a comprehensive review of the literature in this area along with case studies of damage incidents. The proposal describes a testing methodology that was substantially followed for this project. The proposal also identifies Lightning Technologies Incorporated and PowerCET as entities capable of performing portions of the test program.

C - CSST GAS PIPE LIGHTNING HIGH VOLTAGE AND HIGH CURRENT TESTS This technical report “CSST Gas Pipe Lightning High Voltage and High Current Characterization Tests and Model Validation Tests” document was produced by Lightning Technology Incorporated. The subject matter of this report is the laboratory testing of the CSST product to characterize its physical properties and the follow-up testing of selected configurations of direct bonds to verify the accuracy of the simulation model. This testing collected information on the resistance, inductance, and capacitance per meter of the CSST product. The damage resistance of CSST to varying levels of arc current was quantified. Finally, selected configurations of CSST and direct bond ground conductor were subjected to arc currents and the results compared to simulations. The report gives a detailed account of the measurement methodology, instrumentation used, and a full catalog of the data collected.

D – Initial Simulations by PowerCET This slide deck was produced by PowerCET for GTI. It contains a high level synopsis of the parametric measurements taken by LTI as a precursor to the simulations work by PowerCET. The presentation contains the results of a series of simulations produced by PowerCET after the basic parameter measurements were completed by LTI and factored into the simulation model. The simulations indicate that a 6AWG copper direct bonding conductor applied at the point where the gas service enters the residence can prevent CSST perforations caused by lightning induced arcing. This model was used to propose scenarios for further testing at LTI for the purposes of validating the accuracy of the simulation model.


E - VALIDATION OF INSTALLATION METHODS FOR CSST - REPORT This draft report document “Validation of Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage” was produced by PowerCET. It covers the follow up testing performed at LTI to validate the simulation model. A number of laboratory tests were set up to allow direct comparison of Scenario 1 simulations to measured results. The practical limitations of laboratory equipment did not allow for 30m lengths of CSST and bond wire to be tested; the current pulse generators cannot provide the requisite 10x350 waveform to that great of load impedance. The CSST length was standardized to 15 feet (4.5m) and the bonding conductor was tested in lengths from 1m to 16m. There was good agreement between the model predications and the experimental results. In all instances the total charge transferred through the arc, predicted and measured, agreed within 10%.

This report also contains a detailed treatment of the measurement of CSST inductance. The method of inductance measurement preferred by PowerCET is to form closed loop coils of CSST. This approach minimizes several sources of experimental error while making the measurement. The measurements of CSST inductance carried out in this manner provided results very close to those for a smooth walled tube of the same OD calculated using a standard (Rosa) formula. The results of the calculated and measured inductances for CSST also agreed within 10%.

*NOTE: Participants in NFPA’s codes and standards making process should know that limited review of this decision may be sought from the NFPA Board of Directors. For the rules describing the available review and the method for petitioning the Board for review, please consult section 1-7 of the NFPA Regulations Governing Committee Projects (Regs.) and the NFPA Regulations Governing Petitions to the Board of Directors from Decisions of the Standards Council. Since this Council decision is not related to the issuance of a document as referenced in 1.7.2 of the Regs., notice of the intent to file such a petition must be submitted to the Clerk of the Board of Directors within a reasonable time period from the availability of this decision.

SC #10-3-20 D#10-2

Amy Beasley Cronin Secretary, Standards Council 23 June 2010 To: Interested Parties Subject:

Standards Council Decision (Final): D#10-2 Standards Council Agenda Item: SC#10-3-20 Date of Decision*: 3 March 2010

Action following Report of the Council Task Group on CSST Dear Interested Parties: At its meeting of 2 March 2010, the Standards Council considered an appeal on the above referenced matter. Attached is the final decision of the Standards Council on this matter. Sincerely,

Amy Beasley Cronin Secretary, NFPA Standards Council c: D. Berry, M. Brodoff, L. Fuller, M. Earley, D. Roux, T. Lemoff, J. Moreau-Correia, C. Henderson Members, TC on Lightning Protection (LIG-AAA) Members, TC on National Fuel Gas Code (NFG-AAA) Members, NEC Code Making Panel 5 (NEC-P05) Members, TCC on National Electrical Code (NEC-AAC) Members, NFPA Standards Council (AAD-AAA)

Validation of CSST Installation Methods Appendix A

SC#10-3-20 Page 1 of 4 D#10-2

Standards Council Decision (Final): D#10-2 Standards Council Agenda Item: SC#10-3-20 Date of Decision*: 3 March 2010

Action following Report of the Council Task Group on CSST This Standards Council decision sets forth the Standards Council’s conclusions and directives following its receipt and consideration, at its March 2010 meeting, of a report submitted by a Council task group on issues concerning bonding and other lightning-related safety issues affecting corrugated stainless steel tubing (CSST) in gas piping systems. Background In August of 2009, the Standards Council considered a proposed TIA to the 2008 edition of NFPA 70®, National Electrical Code® (NEC), to specify requirements concerning the bonding of corrugated stainless steel tubing (CSST) in gas piping systems. The TIA was proposed by the submitter as the appropriate means of protecting CSST against damage that could be caused if the system is energized due to a lightning strike. The submitter pointed out that a similar (though not identical) bonding provision had been added to the 2009 edition of NFPA 54, National Fuel Gas Code (NFPA 54), and he suggested that a TIA was necessary for correlation and consistency between NFPA 54 and the NEC. The Council declined to issue the TIA since the TIA had been soundly defeated in the balloting of the responsible panel. See Standards Council Decision #09-18 (Agenda Item SC#09-8-16[d], August 6, 2009). In doing so, however, the Council noted that the record before it revealed both jurisdictional and potential technical issues that called for further attention within the standards development process going forward. First, as to the jurisdictional issue, the Council noted that questions had been raised regarding whether the issue addressed by the proposed TIA was properly within the scope of the NEC. Specifically, the Council noted:

In the balloting on the TIA and elsewhere in the record, it has been observed that the scope of the NEC is the practical safeguarding of persons and property from hazards arising “from the use of electricity,” see NEC at 90.1(A), and it has been suggested that a provision, such as the proposed TIA, addressed to the hazards arising from lightning rather than from human use of electricity, is not within the scope of the NEC. (Decision #09-18 at p. 2)

Secondly, the Council noted that in addition to jurisdictional/scope concerns, the balloting on the TIA raised questions regarding whether the proposed bonding requirements for CSST had been adequately substantiated:


SC#10-3-20 Page 2 of 4 D#10-2

Whether or not the NEC has lightning protection within its scope, Panel 5 has expertise on issues of grounding and bonding. Concerns have been raised by some panel members in the balloting and elsewhere as to whether the bonding requirements proposed for the NEC in the TIA and which, in similar form, are currently contained in NFPA 54 have been adequately substantiated. Although the Technical Committee on Lightning Protection was consulted, it was also stated that no correlation or input from Panel 5 was sought by the Technical Committee responsible for NFPA 54 when it considered and accepted the proposal for bonding of CSST now contained in NFPA 54. (Decision #09-18 at p. 2)

The Council concluded that there ought to be a review and study of both the jurisdictional/scope issues and the technical questions concerning bonding or other lightning-related technical issues affecting CSST in gas piping systems:

The Council believes that these issues are deserving of study both for the purpose of assisting the Council in fulfilling its responsibilities to assign scopes and coordinate and oversee the activities of the various NFPA committee projects and also for the benefit of the technical committees that have or should play a role in reviewing the technical issues relating to CSST. (Decision #09-18 at p. 2)

To conduct this review, the Council designated Council Member Farr to appoint and chair a task group made up of members from NEC Panel 5, the technical committees responsible for NFPA 54 and NFPA 780, and any other relevant technical committees. This group is hereafter referred to as “the CSST Task Group”. The Council charged this task group as follows:

The CSST Task Group is requested to provide the Council with a review and analysis of the jurisdictional and technical issues relating to lightning and CSST in gas piping systems, to identify and discuss any technical issues that need to be addressed, to identify potential research or data needs, and to identify which technical committee or committees should play a role in addressing the technical issues and what that role should be. The CSST Task Group’s report should include its recommendations as to steps that should be taken so that any issues can be further addressed, if necessary, within the standards development process.

The CSST Task Group was subsequently formed and, after conducting its work, has now submitted its report to the Council. Conclusions The Council has now reviewed the report and, the CSST Task Group’s work now being complete, the Council has discharged the CSST Task Group with thanks. In the remainder of this decision, the Council sets forth and discusses its conclusions, based on the recommendations of the CSST Task Group and a review of the entire record. Jurisdiction. On the jurisdictional issue, the CSST Task Group noted in its report that lightning protection was generally outside the scope of the NEC and that the Technical Committee on Lightning Protection addresses the installation of lightning protection


SC#10-3-20 Page 3 of 4 D#10-2

systems and deals with gas piping only as it may be part of a lightning protection system. The CSST Task Group, therefore, recommended that the jurisdiction of bonding for lightning protection of gas piping reside with the Technical Committee on the National Fuel Gas Code. NFPA 54, National Fuel Gas Code, is the document that addresses the safe installation of fuel gas piping systems and currently contains bonding requirements. The Council concludes that, based on the recommendation of the CSST Task Group, the Technical Committee on the National Fuel Gas Code should have the jurisdiction over requirements for the bonding of fuel gas piping systems, including CSST. As discussed further below, CSST will need to receive further attention in the standards development process going forward. So as to ensure that the Technical Committee on the National Fuel Gas Code receives input and expertise concerning the lightning-related safety issues related to CSST from other relevant projects and sources, the Council directs the Technical Committee on the National Fuel Gas Code to create a task group to address the CSST issues (hereafter referred to as the NFPA 54 CSST Task Group), drawing on the expertise, as appropriate, of the members of NFPA 70, National Electrical Code®, NFPA 780, Standard for the Installation of Lightning Protection Systems, and from other appropriate organizations such as those that certify or develop product standards related to CSST. This new NFPA 54 CSST Task Group should be for the purpose of studying the issues and providing input to the Technical Committee on the National Fuel Gas Code and others on the safety and use of CSST. Without limitation, such input may include recommendations concerning the scope or content of any necessary research or testing, recommendations for revisions to NFPA 54, review and comment on any Proposals and Comments under consideration, and recommendations concerning relevant questions such as whether or to what extent listing requirements or product standards developers should play a role in addressing lightning-related safety of the CSST product. Technical Substantiation. On the technical lightning safety issues surrounding CSST, the CSST Task Group reported that it had sought information on the research that supports the current CSST bonding requirements of NFPA 54, including any research performed by or on behalf of any manufacturers. The reports received were of limited value and as stated in the CSST Task Group report provided to the Council "did not provide enough information for the CSST Task Group to ascertain that the proposed bonding remedy will provide adequate protection from lightning induced surges.” In addition, the CSST Task Group noted limited anecdotal reports concerning failures where the bonding of the installation may have complied with the current edition of NFPA 54. The CSST Task Group cautioned that the lack of detailed information or incident reports made assessment of these anecdotes impossible. Concerned with the lack of technical substantiation, the CSST Task Group concluded that a research program was necessary to "identify safe methods for the installation of CSST to protect against lightning induced failure with consequent gas leakage." The CSST Task Group report identified, among the areas that should be addressed, the following:

Validate whether or not bonding of CSST is an adequate solution to lightning exposure problem.


SC#10-3-20 Page 4 of 4 D#10-2

If bonding is the solution, validate how bonding should be done. If bonding is the solution, validate the size of the bonding jumpers. Determine if bonding should be done at a location or locations other than where

the gas pipe enters the building. Determine if alternate methods can be used for safe installation, i.e., separation

from other equipment.

The CSST Task Group’s conclusion that there is inadequate substantiation regarding the safe use of CSST echoes the previously expressed concerns that prompted the Council to form the task group in the first place. See Standards Council Decision #09-18 (Agenda Item SC#09-8-16[d], August 6, 2009). Because so little information was provided to the task group, it is unclear whether and to what extent a problem exists. The paucity of the submissions to the task group, however, confirms the Council’s view that the concerns that have been raised about CSST should be addressed and resolved. After review of the CSST task group report and other information available to it, the Council agrees that further research must be produced to technically substantiate whether and, if so, how and in what conditions CSST can be safely used, with respect to lightning, in gas piping systems. Over the next full revision currently scheduled to be in the Annual 2014 revision cycle, the industry or others advocating the continued use of CSST in gas piping systems shall validate the safe use of the product through independent third-party validated research and testing that can be reviewed and evaluated by standards developers in a timely way. Without prescribing who would be most appropriate to organize or conduct this independent research, the Council notes that the NFPA 54 CSST Task Group may be useful in providing input into the scope of research necessary to allow standards developers to establish adequate provisions concerning CSST. In addition, the Council's CSST Task Group noted that the Fire Protection Research Foundation is discussing the possibility of undertaking a research program related to CSST and lightning protection. The Research Foundation frequently can play a useful role in identifying research needs or in conducting research. The Standards Council, however, wishes to emphasize that it is primarily for the participants in the NFPA standards development process to fund and produce the technical substantiation necessary to support the technical content of codes and standards. See, e.g., Standards Council Decision #00-22 at p. 5 (SC#00-60, July 20, 2000); Standards Council Decision #00-30 (SC#00-60, October 6, 2000). Whether through the auspices of the Research Foundation or through other means, it is incumbent upon the manufacturers or others promoting the use of CSST in gas piping systems to provide independently validated and reliable technical substantiation demonstrating that CSST can be safely used. If such substantiation is not provided, the Technical Committee on the National Fuel Gas Code must consider prohibiting the use of CSST in NFPA 54, National Fuel Gas Code. In addition, should the issues not be reasonably addressed by the end of the next full revision cycle, Annual 2014, the Council may take action as it deems appropriate up to and including the prohibition of the use of CSST in NFPA 54.


49, RUE DE LA BIENFAISANCE - 94300 VINCENNES SA. AU CAPITAL DE 152 449 € – RCS CRETEIL B 316 719 855

SIRET 316 719 855 00025 – CODE APE 742 C CERTIFIÉE ISO 9001

X54‐2 Phase II V2 CSST & lightning. Proposal Page 1 / 29

Validation of Installation Methods for

CSST Gas Piping to Mitigate Lightning Related Damage

Part II

Proposal V2

November 2011

Validation of CSST Installation Methods Appendix B


Content: 1. RESEARCH OBJECTIVE ..................................................................................................................................... 3

2. PROJECT TASKS: .............................................................................................................................................. 3

a. Literature and industry search to determine if relevant test or modeling data exists: ................................. 3

b. Sampling plan to ensure a cross section of the generic product available in the U.S. marketplace is included in the testing program: ........................................................................................................................ 4

c. Identification of candidate grounding/bonding methods to be evaluated: ................................................... 6

d. Written research (testing and/or modeling) plan outlining methodology, experimental set‐up, and evaluation criteria to fill the gaps identified in item a) to evaluate effectiveness: ............................................ 6

e. Implementation of this plan: ........................................................................................................................ 16

f. Written Phase II report. ................................................................................................................................. 18

3. PRICING : ....................................................................................................................................................... 19

4. THE TEAM : ................................................................................................................................................... 23



1. RESEARCH OBJECTIVE The goal of this Phase II project is to implement the testing recommendations from the Phase I study. The outcome will be an evaluation of the effectiveness of grounding and bonding methods in the different hazard scenarios outlined in the Phase I study.

2. PROJECT TASKS:

• a. Literature and industry search to determine if relevant test or modeling data exists;

• b. Sampling plan to ensure a cross section of the generic product available in the U.S. marketplace is included in the testing program;

• c. Identification of candidate grounding/bonding methods to be evaluated; • d. Written research (testing and/or modeling) plan outlining a methodology,

experimental set-up, and evaluation criteria to fill the gaps identified in item a) to evaluate effectiveness;

• e. Implementation of this plan; • f. Written Phase II report.

a. Literature and industry search to determine if relevant test or modeling data exists:

A quick additional literature search will be performed as most of it has already been reviewed for Phase I. We are aware of new reports published and they will be studied to identify potential new information. Contacts will be made with CSST manufacturers and with research parties (mainly PowerCet -Dr. Michael Stringfellow that has produced new reports since Part I report has been published) to access existing data. Any existing test could help reducing time for test and thus pricing, provided test method is consistent with what has been proposed in Phase I report and test report are made available in full. Any reviewed report that is considered for possible revision of the test plan need be provided to the Technical Panel members for peer review. As Standards Council Decision FD10-3-20 (D10-2) in its last paragraph states that “… it is incumbent upon the manufacturers or others promoting the use of CSST in gas piping systems to provide independently validated and reliable technical substantiation demonstrating that CSST can be safely used.”



b. Sampling plan to ensure a cross section of the generic product available in the U.S. marketplace is included in the testing program:

Based on preliminary discussions, it appears necessary to test all US based CSST products so as to not show any product bias. It may appear during preliminary tests that some products are very similar. In such a case, some tests could be skipped. In particular it seems that differences mainly exist on CSST jacket. As in most of the tests, the CSST jacket is pre-punctured, this difference may not lead to different results. However, the decision to skip some tests will not be done without approval of the project technical panel. 4 manufacturers are concerned:

Manufacturers CSST Product • Omega Flex Tracpipe • Titeflex Corp. Gastite • Ward Manufacturing Inc. Wardflex • Tru-Flex Metal Hose Corp. Pro-Flex

For impulse tests (tests N°1), we will use: - two sizes of CSST only (mainly ½-inch diameter and one of : ¾ or 1-inch diameter – to

be determined when sampling plan is performed) - 3 samples for each test in order to show repeatability of test results (this is basic standard

procedure). - 2 waveshapes : 8/20 and 10/350 (this waveshape is defined also with three parameters

peak value Ipeak, charge C and specific energy W/R with values Q = I/2 (I in kA, Q in C) and W/R = Q² (W/R in kJ/ohm, Q in As) : i.e. Ipeak= 10 kA Q = 5 As and W/R 25 kJ/ohm)

- 1 magnitude (10 kA, see later in the text) - 3 bonding conductor positions (b, b1 and b2)

So this means that for each test and each manufacturer (outside of investigation tests and tests needed for generator settings) we will perform: 2 size x 3 samples x 2 type of impulses tests x 3 bonding positions = 36 shots. And then a total of 36 x 4 = 144 impulse tests (tests N°1). In addition, the same impulse tests need to be performed on a steel or copper tube (with only one diameter) for comparison sake. So this means 1 diameter, 3 samples, 2 waveshapes, 1 magnitude, one bonding location (end of the sample) : 6 impulses tests on steel or copper tube (preferentially steel) When the most severe configuration is found (based on both diameter and bonding conditions), impulse current with values 5 kA and 1 kA for both waveshapes 8/20 and 10/350 need to be performed as well (2 waveshapes on 3 samples each time, only



one CSST diameter : the most severe one). So this means 12 additional impulse tests. For power follow current tests (tests N°2), we will use using : - two sizes of CSST only (mainly ½-inch diameter and one of : ¾ or 1-inch diameter – to

be determined when sampling plan is performed) - 3 samples for each test in order to show repeatability of test results (this is basic standard

procedure). - 2 values of power follow current associated to 2 different application time.

So this means that for each test and each manufacturer (outside of investigation tests and tests needed for generator settings) we will perform: 2 size x 3 samples x 2 type of power follow current tests = 12 tests. And then a total of 12 x 4 = 48 power follow current tests (tests N°2). In addition, the same power follow current tests need to be performed on a steel or copper tube (with only one diameter) for comparison sake. We assume that all test (CSST) materials will be provided free of charge. Note that sample is changed at each test even if it seems not damaged. After each test a damage evaluation is made by visual inspection. If there is any doubt on damage on CSST wall, a damage procedure will be used (see later in this text). CSST brand will be named A to D in test report and final report. Only laboratory and SEFTIM will know what the relation between letter and manufacturer brand is. Coding for tests and samples will be the following First letter : P (preliminary), I (impulse test), F (power follow current test) Second block : impulse magnitude in kA or power follow current magnitude in kA or 1, 2, 3 for preliminary tests Third block : CSST diameter in inch or s for steel tube Fourth block : sample number n (1, 2, 3 or x1, x2 etc for investigation tests and generator setting tests) Fifth block : waveshape (20 for 8/20 and 350 for 10/350) or duration of power follow current setting in s or B/S (bended or straight) for preliminary tests Sixth block : CSST brand (A, B, C, D) or nothing for steel tube Seventh block : position of bonding conductor for impulse test (b, b1, b2) or nothing for other tests For example : I_10_1/2_1_350_A_b1 : would be the 10 kA 10/350 impulse test performed on



sample N°1, 1/2 inch diameter, for product A, position b1 for bonding conductor F_1_1/2_3_10_D : would be the 1 kA 10 s power follow current test performed on sample N°3, 1/2 inch diameter, for product D

c. Identification of candidate grounding/bonding methods to be evaluated: Basically, the first and main candidate is the one included in the manufacturer data sheet and included in NFPA 54: # 6 AWG bonding conductor. As we don’t consider direct strikes in the present study, a bigger bonding conductor size will not change the result significantly. What is important is bonding conductor length. However, in practice bonding conductor length is an important parameter that will be studied here. Based on generator limitations we will test the more severe configuration (based on the longest possible bonding conductor due to generator capability) and derive more conclusions from simulations (see below).

d. Written research (testing and/or modeling) plan outlining methodology, experimental set-up, and evaluation criteria to fill the gaps identified in item a) to evaluate effectiveness: All impulse tests need to be performed with surge current generators and not combination wave generators. Simulations are needed to show what voltages are generated based on bonding conductor length and possible lightning currents given from the standard database. We will use IEC 62305-1 as a basis for determination of lightning currents expected in field. This will be compare to what laboratory can do due to lab capability, CSST impedance and bonding conductor length. Bonding conductors located at the entrance may not be enough if the bonding conductor is too long and multiple bonding may then help reduce the voltage. --> this conclusion could be derived from a limited number of tests configurations thanks to simulations Tests should be made to check the ability of CSST to withstand power fault current. Some tests have already been performed and if tests reports are provided some tests may be skipped and cost reduced accordingly. Tests should be performed to identify the impedance (mainly inductance) of CSST per m. Tests to determine CSST impedance should incorporate maximum bending radius as given in technical brochures if this is applicable due to laboratory constraints. This impedance will serve as a basis for simulations. Tests should be performed with 8/20 impulses (representing induced surges) and 10/350 impulses (representing partial direct lightning).



Magnitude for these tests will be derived from international lightning protection standards (IEC 62305-1):

Expected surge overcurrents due to lightning flashes

LPL

Low voltage systems

Flash to the service

Flash near the service

Near to, or on the structure

Source of damage S3

(direct flash)

Waveform: 10/350 μs

(kA)

Source of damage S4

(indirect flash)

Waveform: 8/20 μs

(kA)

Source of damage S1 or S2 (induced

current only for S1)

Waveform: 8/20 μs

(kA)

III-IV 5 2,5 0,1

I-II 10 5 0,2

Maximum amplitude given in this standard will be used for our tests; 10 kA for both 10/350 and 8/20 waveshapes. Additionally, for the worse conditions found during the tests and for the diameter of CSST leading to the more severe result, 3 tests at 5 kA and 3 tests at 1 kA will also be performed). If necessary, results for steeper surges (second impulse of return strike current) may be obtained from simulations.



Preliminary tests (P): Purposes:

- measure the impedance of CSST for simulations purposes (mainly R and L) - determine the high frequency behavior of CSST when straight and check if this

behavior changes with minimum bending radius. - allows computer simulation to be made with various configurations. - determine the withstand voltage of the jacket of each of the samples

First preliminary test: P_1_1 or 2_1_S_brand The impedance is first measured with an impedance meter on a single sample 1 m long. This will give the necessary parameters for simulation purposes. The sample will be straight and located 1 m above a grounded metal plate. We will obtain R, L and a stray capacitance C per m. 8 elementary tests at total Same test (P_1_s_1_S) is then made for comparison purpose on a steel tube with same outside diameter (only one diameter) and same thickness for comparison purpose. This tube to be provided by the laboratory. 1 elementary tests at total Second preliminary test: P_2_1 or 2_n: 1, 2 and 3_S_brand and P_2_1 or 2_n: 1, 2 and

3_B_brand It has been observed multiple holes in a few occasions and effect of corrugated surface has been discussed. To try observing such a phenomenon a steep front current (1 µs or less) is injected on a significant length of CSST (> 1m, depending on what laboratory can provide) The CSST samples should be first configured straight then with minimum bending radius allowed by the manufacturer. CSST is located 1 m above a grounded plate as above. A current generator injecting a 10 kA impulse with a front lower or equal to 1 µs should be used. Voltage at extremities of CSST sample is measured and compared between the two configurations. This is compared also to the impedance obtained from previous tests with impedance meter. 48 elementary tests at total Third preliminary test: P_3_1 or 2_1_S_brand



A last test is needed: determine the voltage withstand of CSST. The withstand voltage will be obtained with a 1.2/50 waveshape from a voltage generator in a CSST-plate configuration. A CSST sample 1 m or shorter long with be located on a grounded metal plate. Generator is applied between CSST and plate. The withstand voltage is determined by increasing the voltage from a starting point (either given by the manufacturer or based on the laboratory experience) and increased by steps of 10% of estimated withstand voltage (and not less than 100 V) until a breakdown occurs. This tests is repeated 10 times (10 locations on the same sample) and the withstand voltage will be assumed to be the mean value of the 10 tests. 80 elementary tests at total



First test with test layout described below (surge conditions): What is known:

- From previous tests reports already published we know that CSST jacket can be punctured if the voltage generated between CSST and an adjacent electrode is high enough. In fact this is related to the insulating capability of the CSST jacket more than on CSST design. In field there are two possible occurrences for this situation:

o either the CSST is subjected a high voltage due to lightning transient conditions

o or the adjacent electrode is subjected to a high voltage due to lightning transient conditions. This electrode may be an electrical conductor experiencing a surge.

- When CSST itself is subjected to a high voltage due to a surge current propagation, this voltage depends mainly on the length of bonding conductor (its impedance in fact) as well as CSST impedance and the magnitude of surge current.

Tests to determine the jacket insulation withstand have been described in preliminary tests section. We then assume that CSST is possibly damaged and this will be then obtained by pre puncturing the CSST jacket (pin hole) and making the CSST wall in direct contact with a thin fuse wire (diameter of fuse wire to be determined to be as quick as possible under testing conditions). The fuse wire should be around 1 to 5 cm depending on generator output voltage (alternatively if the output voltage of the generator is high enough, we can use an electrode located above the pin hole to initiate the arc without fuse link). NFPA 54 CSST Joint Task Group recommended that the testing be performed with the insulation removed. This pin hole method will cover the same need and seems more realistic. What we want to know is if at a given current fixed by standardized values the arc will be created and can remain long enough to create damages. Also position of bonding conductor will be changed to see when the arc can self extinguish. Purposes:

- demonstrate the influence of the CSST impedance for all types of lightning stresses (8/20 and 10/350).

- check the effect of the arc on CSST depending on current, waveshape, CSST length and bonding conductor length

- determine maximum values for bonding conductor length without additional bonding, based on tests combined to simulations



Test configuration: We will use a 10/350 current generator (indirect strike) and a 8/20 current generator (induced strike). As explained above, tests with a test generator capable of generating steeper front of wave (typical of the second impulse in a multiple impulse lighting strike), might be a challenge since the lengths of CSST and #6 AWG bonding conductor will probably have a major influence on the generator output. Results could be obtained instead by computer simulations combined with laboratory tests. For all the tests, computer simulations will allow to expand the results. Based on results with two waveshapes at position b, b2 and b1 and depending on effect of arc on CSST as well as possible self extinguishment of the arc, simulations will determine which practical cases are covered by the #6 AWG bonding conductor and what should be its maximum length to be efficient.

Fuse link + 1m # 6 AWG bonding conductor

Generator ouput connected to the fuse link + bonding conductor and to the end of the CSST sample (position b) and move to position b1 and b2 until the arc self extinguish

Generator ground

This length (a) should be as long as generator allows, driving a significant current in the fuse link to create and arc. Typical value is 1m .

This length (b) should be a few meters long (at least 2 times a) – position b1 is 1,5 times a and b2 1,25 times a

#6 AWG bonding conductor 2 m long

b2 b1



The arc will be created by a fuse link between CSST and generator electrode in order to create an arc at each shot. For that purpose the CSST jacket will be pinholed with an appropriate mean, without puncturing the CSST wall (this may not be necessary if the generator open circuit voltage is high enough). The fuse link will then be in direct contact with CSST wall. When current will flow the fuse link will melt and an arc will be created. If the bonding conductor is efficient the arc will self extinguish (depends on sharing of current between bonding conductor and arc path) and damage will be limited if any. This test procedure is less dependent on testing conditions and generator capability. Damage will be defined by a leak at CSST wall to be checked by the following method: pressurize the pipe to working pressure (10-15 psi) with soapy water solution that will form bubble if lead is detected. Laboratory or manufacturers to provide the needed equipment to be available at test lab for the whole test period.



Test numbering will then be: I_10_1 or 2_n: 1, 2 and 3_20_brand A, B, C and D_bond b, b1 or b2

: 10 kA 8/20 wave tests on CSST 72 elementary tests at total

I_5__n: 1, 2 and 3_20_brand A, B, C and D

: 5 kA 8/20 wave tests on CSST on worse conditions 12 elementary tests at total



I_10_s_n: 1, 2 and 3_20

: 10 kA 8/20 wave tests on metal tube 3 elementary tests at total

I_10_1 or 2_n: 1, 2 and 3_350_brand A, B, C and D_bond b, b1 or b2

: 10 kA 10/350 wave tests on CSST 72 elementary tests at total





I_10_s_n: 1, 2 and 3_350

: 10 kA 10/350 wave tests on metal tube 3 elementary tests at total



Second, a test should be performed to inject power fault current in CSST wall at a supposed arc location. Purposes

- check ability of CSST to withstand small power fault current for a long time - check ability of CSST to withstand higher power fault current for a smaller time - try to reproduce what is observed in field

Values for time and fault current magnitude should be based on protection rules existing in typical residential power installations including a low current power fault ac source for the longest time given by protection means existing in practice and a high current power fault ac source (associated of course to a smaller time due to quicker reaction of the protection means). For that test, the fuse wire needs to be in contact with the CSST wall and the CSST jacket punctured before starting the test as previously. This will simulate the effect of a surge coming from the power system and jumping to CSST (and then damaging the CSST jacket) leaving a power fault current to flow to ground through CSST. Bonding conductor position will be only b for that test. There is no surge superimposed on the power follow current. The power follow current generator is directly connected between the fuse wire and the bonding conductor. Various experts will be consulted to know the more appropriate values (Ray Hill, Joe Koepfinger for example). For the time being the got the following information: A 100 A entrance panel short circuit rating is 10,000 Calculations are available that contains information regarding single phase distribution transformer sizes and available short current at the entrance panel. Test numbering will then be: F_hi_1 and 2_n: 1, 2 and 3_lt_brand A, B, C and D

: high power follow current low duration tests on CSST

24 elementary tests at total F_li_1 and 2_n: 1, 2 and 3_ht_brand A, B, C and D

: low power follow current long duration tests on CSST

24 elementary tests at total F_hi_s_n: 1, 2 and 3_lt

: high power follow current low duration tests on metal tube

3 elementary tests at total F_li_s_n: 1, 2 and 3_ht

: low power follow current long duration tests on metal tube

3 elementary tests at total



Tests are summarized in the following table.

Test Number

Quick description of test

How to apply the test

Main knowledge that will be gained from the test

Preliminary Steep front surge current test Insulation withstand test Impedance meter

Directly on CSST length either straight or bended CSST –plate configuration Directly on CSST

• determine the high frequency behavior of CSST

• allow computer simulation to be made

• determine the withstand voltage of the jacket

1 Main test On complete configuration with bonding conductor and arc created by a fuse link

• demonstrate the influence of the CSST impedance for all types of lightning stresses.

• check the effect of the arc on CSST depending on current, waveshape, CSST length and bonding conductor length

• determine maximum values for bonding conductor length without additional bonding, based on tests combined to simulations

2 Power fault current test

Inject a power fault current to CSST wall through an arc created by a fuse link

• check ability of CSST to withstand small power fault current for a long time and higher power fault current for a smaller time

Proposed tests summary



e. Implementation of this plan:

e.1 Preliminary tests See description above. A test report is provided

e.2 Test N°1 : lightning effect The tests are made according to what has been described above. A test report is provided

e.3 Test N°2 : power follow current effect The tests are made according to what has been described above. A test report is provided

e2 Simulations Based on preliminary tests, simulations need to be made to show:

- influence of bonding: o one bonding or multiple bonding effect o admissible length of bonding conductor

- influence of current magnitude (10 kA): o frequency of surge : 8/20, 10/350 and 1 µs front o magnitude of current (most tests will be performed at 10 kA but some will

be performed at 1 ka and 5 kA).

Basically, at this stage, we will extend or validate the result of what has been tested in Test N°1 to other cases using simulation tools. The outcome of the simulation will be to determine the maximum bonding conductor length for each configuration (tests configurations defined above and practical configurations defined below) as well as give recommendations for distance between bonding conductor(s) and adjacent metal parts. Simulations are made in two steps: Step 1 : simulate the tested configurations Purpose is then to validate the model and fixed a few parameters Step 2 : simulate practical cases based on the model defined at step 1 Purpose is then to expand the test results to other cases and especially more practical cases, without limitations inherent to any test laboratory as soon as we want to inject surge currents in long samples



Simulation should represent :

- current generator with various steepness (8-10 µs, 1µs for example) injecting current directly on one end of CSST. Tests with steep front are difficult to make on long length in laboratories.

- various lengths of CSST with impedance based on preliminary test (typically 10 m and 30 m long)

- distance between generator input and CSST bonding : ranging from 1 m to half of CSST length)

- length of bonding #6 AWG (ranging from 1 m to 10 m) - multiple bonding (one at 1 m from CSST end and another at half of CSST

length) This will allow determining the maximum voltage on CSST generated for various configurations and expand the test results. As jacket impulse voltage withstand will be known after preliminary test, as soon as we know the voltage based on simulations, we are able to give requirement regarding maximum bonding length for single or multiple bonding conductors configuration.



Following practical schemes needs to be simulated:

f. Written Phase II report. A complete report including tests reports and simulations reports will be provided with conclusions and proposals.



3. PRICING : This includes all tasks listed above as well as testing on 4 manufacturer’s products. The price for each test is given in order to better focus effort on what is possible in a given budget. It should be noticed that a few tests may probably be skipped if CSST products are similar from on manufacturer to another one. This will only be done after agreement of concerned parties. See also the note below the table. Responsibility for providing and transportation of test items to the laboratory remain with each manufacturers. The price given by laboratory is based on tests management. Manufacturer willing to participate to the tests, should do so without modifying the test program. The price below will be finalized when the laboratory has given a fixed price for all items. Laboratory has provided the following remarks : "Lightning Technologies is pleased to provide this quotation for testing of various CSST configurations. The quotation covers our best estimate of time and materials required to perform the tests and is considered a firm price for the specific number of days of testing and engineering support. … Please note that this estimate is based on the level of effort presently thought necessary to conduct the desired tests and prepare an optional report. It does not, however, constitute a guarantee that the desired tests can be conducted within this level of effort, as unforeseen circumstances occasionally necessitate additional work to complete tests. If this should occur, the details will be discussed with SEFTIM, and either a change in test plan or level of funding will be agreed upon before proceeding.



Items $ per unit $ x 4 Sub TotalFix price including all tasks aboveand not listed below with a meeting in NFPA premises as well as presentation of final results. This includes SEFTIM preparation of tests and simulations but SEFTIM time dedicated to tests and simulations is included in each line below.

24 621 24 621

Price for preliminary test1 for each brand including SEFTIM attendance

896 27 272

Same multiplied by 4 manufacturers

3584


3384


13536


2538


10152

Price for preliminary test1 on steel or copper tube

0,00

Price for test 1 @ 10 kA 10/350 for each brand including SEFTIM attendance

6960 73 880


27840


825


3300


825


3300

Price for test 1 @ 10 kA 10/350 on steel or copper tube

0



Items $ per unit $ x 4 Sub TotalPrice for test 1 @ 10 kA 8/20 for each brand including SEFTIM attendance

6960


27840


825


3300


825


3300

Price for test 1 @ 10 kA 8/20 on steel or copper tube

0

Test engineering 5000

Price for test 2 with high power follow current for each brand including SEFTIM attendance

2613 25 904


10452

Price for test 2 with low power follow current for each brand including SEFTIM attendance

2613


10452

Price for test 2 with high power follow current on steel or copper tube

0

Price for test 2 with low power follow current on steel or copper tube

0

Test engineering 5000

Price for simulations step 1 including technical discussions with SEFTIM

8 450,00 16 900

Price for simulations step 2 including technical discussions with SEFTIM

8 450,00



Items $ per unit $ x 4 Sub Total

Total 168 577

Note: it seems, based on preliminary discussions, that the various CSST products under test are rather similar in structure. One of the main differences is the jacket but in our tests the jacket will be punctured. So this means that the difference between one product and the other one will be limited to the steel tube (shape, thickness). In addition, whatever are the differences between one product to another one, at the end there will be no difference between the tests results as the tests will be anonymous. So this means that what will be the most severe result, will be considered as the result valid for all CSST products. Based on these assumptions, it is proposed below another pricing where instead of testing 3 samples per brand, we tests 4 samples (one sample for each brand) so this means that for each test the number of tests is reduced from 12 to 4 only. This of course doesn't apply to the preliminary tests. Based on this, the pricing below has been estimated (if this approach is interesting a new fix price will be established in cooperation with the laboratory). For the time being it is only an estimation to help taking a decision. As a matter of fact, moving from one brand to another one at each test, especially if the manufacturers want to participate, may reduce a bit the money saving. An estimated pricing reduction would be around 25% of total project.



4. THE TEAM : Team leader: Alain Rousseau (Company SEFTIM). He has been in charge of project Phase I. He will :

- coordinate the effort - define test program - be present for the tests - allow interface between test and simulations to expand as much as possible

the test results from a limited number of tests - define needed simulations - produce a global report including tests reports and simulations reports.

Alain has a large experience in testing with 8/20, 10/350 and long duration waves (tests on building elements including bridge stay cable, on SPDs, on varistors, on aluminotheric links and other connectors). He has also a long experience on simulations for electrical networks and earthing systems. He will be assisted by SEFTIM engineers having additional experience in these two fields. SEFTIM has some internal simulations capabilities that may be used to add additional cases if needed. He may also be assisted by François Buret Ampère Laboratory - ECL. A French university with whom we cooperate on a regular basis for simulations. François is professor with long experience in simulations as well as in HV testing techniques.



Simulations will be performed by : Dr. Michael Stringfellow - PowerCET Corporation. Michael has already experience with CSST simulations and this will help the project to go faster. PowerCET Corporation 10461 E. Quartz Rock Road Scottsdale, AZ 85255 [email protected] (480) 556-9510 PowerCET Corporation has carried out many similar modeling projects for Fortune 500 companies and major corporations. The company has specific experience with both laboratory testing and simulating the flow of transients on electrical and other services in buildings, including gas lines. Full details for PowerCET Corporation’s can be found at http://www.powercet.com. Dr. Stringfellow is very well known in the lightning academic community and his work on surge in HV systems is very documented. This is directly related to what he will do for this project. He has used several simulation programs over the years, but has always used SPICE-based programs. SPICE is an acronym for Simulation Program with Integrated Circuit Emphasis that was originally developed by the University of California at Berkeley for simulating the behavior of complex circuits. This is also what SEFTM uses for its simulations. Although not originally intended for simulating power circuits, it has proved very good for this, especially for transient analysis. This program has been made freely available to the community, so that a number of commercial products have been based on it. For many years, Michael mostly used Spectrum Software's Micro-CAP, a SPICE-based program that has been available for nearly 20 years. When he was Chief Scientist at EFI Electronics in Salt Lake City in the 1990's, he developed non-linear SPICE models for simulating varistors, spark gaps and suppressor diodes. A couple of papers from those days are referenced below. The predicted behavior of varistors separated by power system conductors was validated by many lab tests and comparisons between measurement and simulation is given in the Coordination paper listed below. Coordination of Surge Suppressors in Low-Voltage AC Power Circuits. Session 6B. Lightning Surges in Low-Voltage AC Power Systems More recently, he has evaluated other commercial SPICE programs and is currently using an early version of Micro-CAP as well as 5Spice for his lightning simulations. These are more than adequate for most simulations. He uses lumped equivalents of conductors and transmission lines (inductance, resistance and capacitance) and has used both time- and current-dependent models



for earth resistance. He can simulate non-linear elements, including gas tubes, diodes and varistors. Michael has access to a surge generator in his own lab that he could use for validation purposes of his simulation model. Michael gave the following information for this project: Ability to simulate transient currents and voltages developed when lightning currents of defined waveshapes are applied to specified locations on the gas supply and electricity services of buildings. The simulations will include the following parameters: • Gas lines, power lines, water lines, grounding and bonding conductors will be represented by lumped inductors, resistors and capacitors, based on values determined both theoretically and from laboratory testing. • Lightning surges will be represented by standard current waveshapes of specified risetime and decay time and specified source impedance. • The earth connection of ground rods will be represented by a time-invariant resistor, unless otherwise specified. Output from the simulation will include the following: • Distribution of current, including amplitude and waveshape, among the gas pipe, grounds and other services modeled. • Distribution of voltages, including amplitude and waveshape, between specified locations • Available surge energy at arc paths involving CSST pipes



Tests will be performed by LTI. This lab has already experience with CSST tests and this will help the project to go faster. LTI. Mr. Ed Rupke Lightning Technologies Inc. 10 Downing Industrial Parkway Pittsfield, MA 01201-3890 [email protected] (413) 499-2135 LTI fulfill the following requirements

• Supplier shall have a wide selection of voltage and current generator capabilities. These capabilities shall be able to simulate basic industry standard waveforms along with the flexibility to produce non-standard waveforms with fast rates of change.

• Supplier shall be familiar with and have capabilities to generate the waveforms utilized for CSST qualification.

• Laboratory shall have recognized industry accepted quality procedures or accreditation.

Below is a short presentation of LTI capabilities. Should LTI be unable to perform the needed tests, another lab will be looked for.


49, RUE DE LA BIENFAISANCE - 94300 VINCENNES SA. AU CAPITAL DE 152 449 € – RCS CRETEIL B 316 719 855

SIRET 316 719 855 00025 – CODE APE 742 C CERTIFIÉE ISO 9001







This test report must not be used by the client product certification, approval or endorsement by NAVLAP, NIST or any agency of the Federal Government

LT-13-3671

TEST REPORT:

CSST GAS PIPE LIGHTNING HIGH VOLTAGE AND

HIGH CURRENT CHARACTERIZATION TESTS AND

MODEL VALIDATION TESTS

Report by: Approved by: T. P. Zeik M. M. Dargi

For

Gas Technology Institute

1700 South Mount Prospect Rd Des Plaines, IL 60018

Purchase Order No. S394

Tests by: Witnesses: T. P. Zeik Dr. M. Stringfellow P. P. Saldo (PowerCET) G. A. Crochiere C. Ziolkowski (GTI) Test dates: References: 31 August - 28 September 2012 PR017862 4-6 December 2012 DB 378, pp 54-100 DB 389, pp 1-6, 30-34

18 March 2013

Lightning Technologies, an NTS Company

10 Downing Industrial Parkway Pittsfield, MA 01201-3890

U.S.A.

Validation of CSST Installation Methods Appendix C

Lightning Technologies, an NTS Company Test Report LT-13-3671, Rev. (-) 10 Downing Industrial Parkway 18 March 2013 Pittsfield, MA 01201 Page ii of vi

TABLE OF CONTENTS

SECTION NO. PAGE NO. 1.0 INTRODUCTION .................................................................................................. 1

2.0 TEST SUMMARY ................................................................................................. 2

3.0 TEST ARTICLES .................................................................................................. 3

4.0 REFERENCE DOCUMENTS ............................................................................... 4

5.0 HIGH VOLTAGE AND HIGH CURRENT CALIBRATIONS .................................. 5

5.1 Test Measurement Equipment ................................................................... 5 5.2 Ambient Test Conditions ............................................................................ 6

6.0 CSST ELECTRICAL AND LIGHTNING CHARACTERIZATION TESTS .............. 7

6.1 Per Unit Length Electrical Characterization Measurement Setup .............. 8

6.2 Per Unit Electrical Characterization Measurement Test Results .............. 10 6.3 High Voltage Impulse Withstand/Breakdown Test Setup ......................... 12

6.4 High Voltage Impulse Withstand/Breakdown Test Results ...................... 18 6.5 High Current Damage Tolerance Test Setup ........................................... 57

6.6 High Current Damage Tolerance Test Results ........................................ 64

7.0 CSST HIGH CURRENT MODEL VALIDATION TESTING ................................. 78

7.1 High Current Model Validation Test Setup ............................................... 79 7.2 High Current Model Validation Test Results ............................................ 84

APPENDIX A – GTI Meeting Minutes................................................................. A-1 – A-6

APPENDIX B – Raw Data Oscillograms – High Voltage Impulse Withstand/Breakdown Tests .................................................. B-1 – B-XX

APPENDIX C – Raw Data Oscillograms – High Current Damage Tolerance Tests ..................................................................... C-1 – C-XX

APPENDIX D – Raw Data Oscillograms – High Current Model Validation Tests ..................................................................... D-1 – D-XX


Lightning Technologies, an NTS Company Test Report LT-13-3671, Rev. (-) 10 Downing Industrial Parkway 18 March 2013 Pittsfield, MA 01201 Page iii of vi

LIST OF FIGURES

FIGURE NO. PAGE NO. Figure 1 – Typical CSST Test Sample, 2-meters long (1” Dia. Size Shown) ................... 3

Figure 2 – Setup Diagram for CSST DC Resistance Measurements .............................. 8

Figure 3 – Photograph of Setup used for CSST DC Resistance, Self Inductance and Stray Capacitance Measurements ................................ 9

Figure 4 – Setup Diagram for CSST Self Inductance and Capacitance Measurements 10

Figure 5 – Typical Inductance Measurement of 2-Meter CSST Sample........................ 11

Figure 6 – Setup Diagram for CSST High Voltage Impulse Withstand/ Breakdown Tests ......................................................................................... 13

Figure 7 – Overall CSST High Voltage Impulse Withstand/ Breakdown Test Setup using LRC Circuit Generator ................................... 14

Figure 8 – Oil Bath for the CSST High Voltage Impulse Withstand/ Breakdown Test Setup using LRC Circuit Generator ................................... 15

Figure 9 – Overall CSST High Voltage Impulse Withstand/ Breakdown Test Setup using Marx Generator ............................................. 15

Figure 10 – Oil Bath for the CSST High Voltage Impulse Withstand/ Breakdown Test Setup using Marx Generator ............................................ 16

Figure 11 – Typical 1.2 µs x 50 µs Impulse Voltage Generator Calibration Oscillograph, Overall ................................................................ 16

Figure 12 – Typical 1.2 µs x 50 µs Impulse Voltage Generator Calibration Oscillograph, t10-90% .................................................................. 17

Figure 13 – Typical 1.2 µs x 50 µs Impulse Voltage Generator Calibration Oscillograph, t50% Decay .............................................................. 17

Figure 14 – Reaction of CSST Dielectric Jacket of Manufacturer B to Insulating Transformer Oil Bath................................. 19

Figure 15 – Puncture in a CSST Dielectric Jacket during High Voltage Impulse Withstand Tests ....................................................... 20

Figure 16 – Oscillograph of a High Voltage Impulse Withstand Test during a Puncture of the CSST Dielectric Jacket ........................................ 20

Figure 17 – General Test Setup for the CSST High Current, Damage Tolerance Tests ........................................................................... 59


Lightning Technologies, an NTS Company Test Report LT-13-3671, Rev. (-) 10 Downing Industrial Parkway 18 March 2013 Pittsfield, MA 01201 Page iv of vi

LIST OF FIGURES

FIGURE NO. PAGE NO. Figure 18 – CSST High Current, Damage Tolerance Calibrations

(8 µs x 20 µs Waveform) ............................................................................ 59

Figure 19 – CSST High Current, Damage Tolerance Test Setup (8 µs x 20 µs Waveform) ............................................................................ 60

Figure 20 – Close-up of Electrode-Initiating Wire Arrangement .................................... 60

Figure 21 – CSST High Current, Damage Tolerance Calibrations (10 µs x 350 µs Waveform) ........................................................................ 61

Figure 22 – 8 µs x 20 µs Waveform Calibration, Total Charge (10 kA, 0.22 Coulombs Shown) .................................................................. 61

Figure 23 – 8 µs x 20 µs Waveform Calibration, t10-90% ................................................. 62

Figure 24 – Typical 8 µs x 20 µs Waveform Calibration, t50% Decay ................................. 62

Figure 25 – 10 µs x 350 µs Waveform Calibration, Total Charge (10 kA, 0.5 Coulombs Shown) .................................................................... 63

Figure 26 – 10 µs x 350 µs Waveform Calibration, t10-90% ............................................. 63

Figure 27 – 10 µs x 350 µs Waveform Calibration, t050% Decay ....................................... 64

Figure 28 – Typical 1” CSST Sidewall Damage from 8 µs x 20 µs Waveform, 1 kA/ 0.026 Coulombs ................................................................................ 66

Figure 29 – Typical 1” CSST Sidewall Damage from 8 µs x 20 µs Waveform, 5 kA/ 0.11 Coulombs .................................................................................. 66

Figure 30 – Typical 1” CSST Sidewall Damage from 8 µs x 20 µs Waveform, 10 kA/ 0.23 Coulombs ................................................................................ 67

Figure 31 – Typical 0.5” CSST Sidewall Damage from 8 µs x 20 µs Waveform, 1 kA/ 0.026 Coulombs ................................................................................ 67

Figure 32 – Typical 0.5” CSST Sidewall Damage from 8 µs x 20 µs Waveform, 5 kA/ 0.11 Coulombs .................................................................................. 68

Figure 33 – Typical 0.5” CSST Sidewall Damage from 8 µs x 20 µs Waveform, 10 kA/ 0.23 Coulombs ................................................................................ 68

Figure 34 – Typical 1” CSST Sidewall Damage from 10 µs x 350 µs Waveform, 1 kA/ 0.5 Coulombs .................................................................................... 69


Lightning Technologies, an NTS Company Test Report LT-13-3671, Rev. (-) 10 Downing Industrial Parkway 18 March 2013 Pittsfield, MA 01201 Page v of vi

LIST OF FIGURES

FIGURE NO. PAGE NO. Figure 35 – Typical 1” CSST Sidewall Damage from 10 µs x 350 µs Waveform,

5 kA/ 2.4 Coulombs .................................................................................... 69

Figure 36 – Typical 1” CSST Sidewall Damage from 10 µs x 350 µs Waveform, 10 kA/ 4.9 Coulombs .................................................................................. 70

Figure 37 – Typical 0.5” CSST Sidewall Damage from 10 µs x 350 µs Waveform, 1 kA/ 0.5 Coulombs .................................................................................... 70

Figure 38 – Typical 0.5” CSST Sidewall Damage from 10 µs x 350 µs Waveform, 5 kA/ 2.4 Coulombs .................................................................................... 71

Figure 39 – Typical 0.5” CSST Sidewall Damage from 10 µs x 350 µs Waveform, 10 kA/ 5.0 Coulombs .................................................................................. 71

Figure 40 – CSST High Current Model Validation Test Setup, Plan View ..................... 80

Figure 41 – CSST High Current Model Validation Calibration Setup............................. 81

Figure 42 – CSST High Current Model Validation Test Setup ....................................... 81

Figure 43 – Transient Generator Output Node at BIP Manifold (High Current Model Validation Tests) ........................................................ 82

Figure 44 – Typical 10 µs x 350 µs Waveform Calibration, t50% Decay ............................. 83

Figure 45 – Typical 10 µs x 350 µs Waveform Calibration, t10-90% ................................. 83

Figure 46 – Typical Applied 10 µs x 350 µs Current Impulse, 10.2 kA peak (Test Nos. 11-18)........................................................................................ 86

Figure 47 – CSST and Bond Wire Currents, 4.5 m, 1” CSST in Parallel with 1.0 m, 6 AWG Bonding Conductor (Test No. 16) ................... 86

Figure 48 – Transient Currents in 4.5 m, 1” CSST in Parallel with 1.0 m, 6 AWG Bonding Conductor, 10.2 kA peak (Test No. 18) ............................ 87

Figure 49 – Transient Currents in 4.5 m, 1” CSST in Parallel with 6.0 m, 6 AWG Bonding Conductor, 10.2 kA peak (Test No. 14) ............................ 87

Figure 50 – CSST and Bond Wire Currents, 4.5 m, 1” CSST in Parallel with 10.0 m, 6 AWG Bonding Conductor, 10.2 kA peak (Test No. 38)............................... 88

Figure 51 – CSST and Bond Wire Currents, 4.5 m, 1” CSST in Parallel with 16.0 m, 6 AWG Bonding Conductor, 10.2 kA peak (Test No. 39)............................... 88


Lightning Technologies, an NTS Company Test Report LT-13-3671, Rev. (-) 10 Downing Industrial Parkway 18 March 2013 Pittsfield, MA 01201 Page vi of vi

LIST OF FIGURES

FIGURE NO. PAGE NO. Figure 52 – Typical Arc Entry to CSST Test Setup ....................................................... 90

Figure 53 – 1” CSST Arc Entry Damage, 947 A/ 0.467 C (Test No. 19)........................ 90



Figure 56 – 1” CSST Arc Entry Damage, in Parallel with 16 m Bond Conductor, 8.3 kA/ 0.815C (Test No. 39) ................................. 92

LIST OF TABLES

TABLE NO. PAGE NO. Table 1 – CSST Electrical and Lightning Parametric Test Measurement Equipment ...... 5

Table 2 – High Current Model Validation Test Measurement Equipment ........................ 6

Table 3 – Ambient Test Conditions – Electrical and Lightning Parametric Tests ............ 6

Table 4 – Ambient Test Conditions – CSST Model Validation ........................................ 7

Table 5 – Resistance, Inductance and Capacitance Measured on 2-Meter CSST Samples .............................................................................................. 12

Table 6 – CSST per Unit Length Resistance, Inductance and Capacitance ................. 12

Table 7 – High Voltage Impulse Withstand Test Summary ........................................... 21

Table 8 – High Voltage Impulse Withstand Test Log..................................................... 24

Table 9 – High Current, Damage Tolerance Test Summary ......................................... 72

Table 10 – Electrical Bonding Resistances of Model Validation Test Setup .................. 80

Table 11 – Model Validation Test Summary .................................................................. 93

Table 12 – Additional Damage Tolerance Tests on CSST Wall Burn-through vs. Total Charge Delivered........................................................................... 97


Lightning Technologies, an NTS Company Test Report LT-13-3671, Rev. (-) 10 Downing Industrial Parkway 18 March 2013 Pittsfield, MA 01201 Page 1 of 97

1.0 INTRODUCTION This document contains the results of per unit length characterization tests, high

voltage breakdown tests, and high current damage tolerance tests performed on Corrugated Stainless Steel Tubing (CSST) gas pipe produced by four different manufacturers in support of the Gas Technology Institute (GTI) proposal Validation of Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage, Part II, Final Proposal (V3), dated August 2012 (Ref. 1). This test proposal was amended as a result of a meeting held during the period of 30-31 July 2012 at Lightning Technologies (LTI) in Pittsfield, MA attended by representatives of GTI, three of the four CSST manufacturers, PowerCET and LTI, with the finalized test plan being outlined in the GTI Meeting Minutes (Appendix A). The amended proposal’s plan of action has four steps:

1. Obtain electrical characteristics of CSST and lightning voltage and current

tolerance parametric data on CSST products from four manufacturers.

2. Use this parametric and electrical characteristic data from the various CSST samples as inputs into a lumped-parameter, PSpice based circuit model.

3. Validate the circuit models via simplified CSST arc entry test arrangements.

4. Use the validated models to arrive at recommendations for electrical bonding configurations for CSST service entrance grounding for a variety of installation limitations and situations.

This Report presents the results of laboratory testing in support of Step 1 above, and contains data and results of model validation tests performed in support of PowerCET’s analytical circuit simulations of CSST gas pipe in parallel with electrical bonding conductors (Step 3).

The tests were performed in accordance with the plan outlined in the GTI

Meeting Minutes in Appendix A and using standardized, lightning test impulse waveforms per IEC 61000-4-5 (Ref. 2). The tested CSST samples were provided by four manufacturers: Titeflex Inc., Omega Flex Inc., Ward Manufacturing Inc. and Tru-Flex Metal Hose Corp. Both 0.5-inch and 1.0-inch CSST sizes from each manufacturer were evaluated.

The tests were performed by G. A. Crochiere, P. P. Saldo and T. P. Zeik of

Lightning Technologies in Pittsfield, MA over the course of two separate periods: 30 August - 28 September 2012 for the CSST per unit characterizations, high voltage breakdown and high current damage tolerance tests, and 4 - 6 December 2012 for the model validation tests. The latter tests were witnessed by Dr. M. Stringfellow of PowerCET and C. Ziolkowski of GTI. The test results contained in this report relate only to the test items/part numbers tested and to the specific test conditions and arrangements.



2.0 TEST SUMMARY The laboratory tests in support of the GTI Project 21323 and technical proposal

Validation of Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage (Ref. 1) constituted four separate test arrangements and objectives.

The first test set was a collection of basic electrical properties of the various

CSST products, specifically end-to-end DC resistance, self inductance and parasitic capacitance, in support of defining per unit length values for use in analytical models. The resulting average resistances per meter for all CSST samples were 4.54 mΩ/m for the 1.0” diameter products, and 10.42 mΩ/m for the 0.5” diameter products. The average parasitic capacitance of all CSST products (both diameters), with respect to a 1-meter distance equipotential plane and at 10 kHz, was 0.8 pF/m. The average recorded self inductance as measured 1-meter above a grounded plane and at 10 kHz, was 2.0 µH/m for the 1” diameter CSST and 2.2 µH/m for the 0.5” diameter CSST products. All four manufacturers had very similar per unit electrical characteristics, which is not surprising given that all eight types of CSST measured have similar material and physical geometries.

The second test set was the characterization of the impulse voltage withstand of

non-conductive jacketed CSST products to quantify the dielectric strength of the CSST jacket insulation. This was accomplished with the power utility industry’s standardized Basic Insulation Level (BIL) waveform (1.2 µs x 50 µs) as defined in IEC 61000-4-5 (Ref. 2) and IEEC C62.41-1991 (Ref. 3). Because of the inherently statistical nature of electrical breakdown of dielectrics, both CSST sizes for each manufacturer had a total of ten data points recorded, for a total of 80 tests. For all four CSST manufacturers, the 0.5” diameter products were observed to breakdown at lower peak impulse voltages than the 1” diameter products. The highest and lowest breakdown voltages for the 0.5” diameter products were 55.5 kV and 30.6 kV, respectively, averaging 38.2 kV, with Manufacturer B having the highest observed withstand voltage levels. The highest and lowest breakdown voltages for the 1” diameter product were 60.6 kV and 35.8 kV, respectively, averaging 42.5 kV, again with Manufacturer B having the highest observed withstand voltage levels.

The third set of tests was comprised of damage tolerance evaluations of the

CSST gas piping using the telecommunications and power utility industries’ standardized lightning current waveforms (Ref. 2, IEC 61000-4-5, 8 µs x 20 µs and Ref. 4, ITU-T K.44, Series K 10 µs x 350 µs test impulses). These tests currents were used in conjunction with an arc entry method to evaluate the CSST metal wall burn through tolerance as related to the total electrical charge deposited. For both 0.5” and 1” diameter CSST products from all four manufacturers, the 8 µs x 20 µs current waveform applied at peak currents of 1 kA, 5 kA and 10 kA produced no punctures of the CSST metal sidewalls. The tests using the 10 µs x 350 µs current waveform applied at a peak current of 1 kA also did not produce punctures of any of the CSST metal sidewalls for all four manufacturers (both 0.5” and 1” diameters). However, for the tests with the 10 µs x 350 µs waveform applied at peak currents of 5 kA and 10 kA, burn through and punctures consistently occurred in each manufacturers’ 0.5” and 1” diameter CSST



products. (Leak/pressure proof tests were not performed after testing at LTI for any of these high current damage tolerance tests.)

The final test set was a controlled test arrangement using the 10 µs x 350 µs

standardized lightning current waveform again, but conducted into the CSST products placed in parallel with various lengths of 6 AWG bonding conductors. The objective of this series of tests was to provide empirical data for direct comparisons to circuit models developed by Dr. M. Stringfellow of PowerCET. For these tests, a single manufacturer’s 0.5” or 1” diameter product was used, put in parallel with bonding conductors of different lengths, and the current division and resulting waveforms recorded for both conduction paths. The CSST products were tested at 4.5 m of length. The samples had either an arc entry type of return to the transient generator ground, or were directly grounded via a short jumper, with the 10 µs x 350 µs generator output at a common node of a black iron pipe manifold interface for both the CSST and bond conductor. The measured current amplitudes and waveforms in the CSST and the parallel bonding conductor compared well with the circuit model predicted waveforms for each of the scenarios tested. Longer bonding conductors resulted in more of the transient currents being directed to the CSST pipe.

3.0 TEST ARTICLES

The samples tested were several varieties of corrugated stainless steel tubing

(CSST) made of ASTM A240 type 304 stainless steel, having a wall thickness of 0.010” and covered with a non-conductive polyethylene jacketing. Each CSST test sample had the appropriate manufacturers’ 0.5” or 1” diameter C360 brass fitting installed on both ends. Figure 1 shows a typical CSST test article.

Figure 1 – Typical CSST Test Sample, 2-meters long (1” Dia. Size Shown)

Polyethylene Jacket

Brass End Fitting



For the per unit length measurements, high voltage impulse withstand tests and high current damage tolerance tests, each CSST test sample was approximately two-meters in length. For the model validation tests, the sample length was 4.5-meters. The model validation test series used only one representative manufacturer’s CSST product in addition to a black iron pipe (BIP) manifold with a UL Listed 467 Bronze bonding clamp attached as the termination method for the 6 AWG bonding conductor. This manufacturer supplied an appropriate adapter for the CSST to BIP manifold, which allowed the BIP manifold to serve as a common electrical node for the current division tests.

All CSST samples were prepared by the respective manufacturers by cutting to

length and installing the respective end fittings prior to shipment to the LTI lightning laboratory and were tested in the condition in which they were received. Note that some of the samples tested arrived with mechanical punctures to their polyethylene jackets. These areas were avoided during the high voltage withstand tests.

The CSST manufacturers Titeflex, Inc. (‘Gastite’ product), Omega Flex, Inc.

(‘TracPipe’ product), Ward Manufacturing, Inc. (‘WardFlex’ product), and Tru-Flex Metal Hose Corp. (‘Pro-Flex’ product) all provided CSST gas pipes in both the 0.5 and 1” fitting sizes for use as destructive test samples. To keep the test results anonymous, the above manufacturers’ CSST products tested are referenced throughout this report by the letters A, B, C and D, with the relation of assigned letter to specific manufacturer provided to GTI under separate cover.

4.0 REFERENCE DOCUMENTS

1. Gas Technology Institute, Technical Proposal: Validation of Installation

Methods for CSST Gas Piping to Mitigate Lightning Related Damage, Part II, Final Proposal (V3), August 2012

2. International Electrotechnical Commission, IEC 61000-4-5: Electromagnetic Compatibility (EMC), Part 4-5: Testing and Measurement Techniques – Surge Immunity Test, Second Edition, November 2005

3. IEEE Standards Association, C62.41-1991: IEEE Recommended Practice for Surge Voltages in Low-Voltage AC Power Circuits, 1991

4. International Telecommunication Union, ITU-T K.44: Resistibility Tests for Telecommunication Equipment Exposed to Overvoltages and Overcurrents – Basic Recommendation, April 2008

5. International Standards Organization, ISO/IEC 17025:2005: General Requirements for the Competence of Testing and Calibration Laboratories, 2005

6. American National Standards Institute, ANSI/NCSL Z540.1-1994: Requirements for the Calibration of Measuring and Test Equipment, 1994



7. CRC Press, The Electrical Engineering Handbook, R.C. Dorf, 1993

5.0 HIGH VOLTAGE AND HIGH CURRENT CALIBRATIONS

5.1 Test Measurement Equipment All measurement equipment furnished by LTI is calibrated by a commercial

calibration agency in accordance with the requirements of the second edition of ISO/IEC 17025:2005 (Ref. 5) and/or ANSI/NCSL Z540-1-1994 (Ref. 6) using standards traceable to the National Institute of Standards and Technology. Table 1 lists the measurement equipment utilized for the CSST per unit length characterizations, high voltage withstand tests and the high current damage tolerance tests performed during the period of 31 August - 28 September 2012. Table 2 shows the measurement equipment used for the high current model validation tests performed during the period of 4 - 6 December 2012.

Table 1 – CSST Electrical and Lightning Parametric Test Measurement Equipment

Manufacturer Equipment Model No. Serial No. Calibration Date

Calibration Due Date

31 August – 13 September 2012

Valhalla Micro-Ohmmeter 4300B 32-1517 23 Nov 11 23 Nov 12

LTI Resistive Divider 1976.5 0713-2 14 Jul 11 14 Jul 13

Tektronix High Voltage Probe P6015A B058212 10 Feb 12 10 Feb 13

LeCroy Digital Storage Oscilloscope 24Xs 15079 17 Jan 12 17 Jan 13

Fluke Digital Multimeter 87V 88450583 25 Jul 12 25 Jul 13

Pearson

Attenuator A10 099670 2 Jul 12 2 Jul 13

Current Transformer 4160 081130 23 Sep 11 23 Sep 12

080111 19 Sep 12 19 Sep 13

Agilent LRC Meter 4263B MY40100325 23 May 12 23 May 13



Table 2 – High Current Model Validation Test Measurement Equipment

Manufacturer Equipment Model No. Serial No. Calibration Date

Calibration Due Date

19 September – 6 December 2012

BCD Milliohm Meter M1 DC001158 5 Apr 12 5 Apr 13

Fluke High Voltage Probe 80K40 74450015 15 Aug 12 15 Aug 13

LeCroy Digital Storage Oscilloscope

6050A 17524 24 May 12 24 May 13

WS24XS 0304M15079 17 Jan 12 17 Jan 13

Pearson

Attenuator A10 099664 20 Jan 12 20 Jan 13

Current Transformer 4160

080861 16 Jul 12 16 Jul 13

081130 16 Oct 12 16 Oct 13

080111 19 Sep 12 19 Sep 13

3525 79654 19 Sep 12 19 Sep 13

5.2 Ambient Test Conditions

The CSST per unit length electrical characterizations tests, high voltage

withstand tests and the high current damage tolerance tests were performed within the following ambient conditions listed in Table 3. The CSST high current model validation tests were performed within the following ambient conditions listed in Table 4.

Table 3 – Ambient Test Conditions – Electrical and Lightning Parametric Tests

Test Date Test Type Temperature (˚F)

Relative Humidity

(%) Pressure

(inches Hg)

31 Aug 12 Per Unit Measurements 70.7 51.9 29.2

4 Sep 12 High Voltage Withstand 77 59 29.1

5 Sep 12 High Voltage Withstand 78.2 57 29

6 Sep 12 High Voltage Withstand 71 56.5 28.9

10 Sep 12 High Voltage Withstand 69.6 50 29.1

11 Sep 12 High Voltage Withstand 65.6 39.2 29.3

12 Sep 12 High Voltage Withstand 79.5 38.5 Not Recorded



Table 3 – Ambient Test Conditions – Electrical and Lightning Parametric Tests (Continued)


Relative Humidity

(%) Pressure

(inches Hg)

13 Sep 12 High Voltage Withstand 52.1 67.4 30.4

19 Sep 12 High Current Damage 71 51 28.9

20 Sep 12 High Current Damage 64.4 41.6 28.9


25 Sep 12 High Current Damage 74 28 29.2




4 Dec 12 High Current Model Validations 68.3 36 30.1

5 Dec 12 High Current Model Validations 72.5 27.5 30

6 Dec 12 High Current Model Validations 69.4 19 30.2

Table 4 – Ambient Test Conditions – CSST Model Validation


Relative Humidity

(%) Pressure

(inches Hg)

4 Dec 12 High Current Model Validations 68.3 36.0 30.1



6.0 CSST ELECTRICAL AND LIGHTNING CHARACTERIZATION TESTS

Before beginning the lightning high voltage withstand and high current damage

tolerance tests, each type of CSST test sample from all four manufacturers had its basic electrical properties measured. Following this, the high voltage, impulse withstand tests were performed on ten samples of each type of CSST being evaluated. The high current damage tolerance tests were performed last on three samples of each type of CSST being evaluated.



6.1 Per Unit Length Electrical Characterization Measurement Setup Each 2-meter long CSST sample had its end to end DC resistance, self-

inductance and parasitic capacitance recorded, each with respect to a 1-meter distant ground plane, using either a 4-wire bond milliohm meter or an impedance bridge. Each CSST test sample also had its physical total length measured (not including its brass end fittings). The CSST sample under test was supported by a pair of wooden stanchions atop a large copper ground plane to provide a 1-meter separation to the current return of the LRC meter.

For the DC resistance measurements, a Valhalla Model 4300B 4-wire milliohm meter was used to measure across the entire CSST sample, including any electrical bonding resistances from the brass end fittings to CSST interfaces. These resistance measurements all used 1 A as the driving current. For reference, the bulk conductivity of 304 stainless steel is about 1.4 x10⁶ S-m. Figure 2 shows the DC resistance measurement setup schematic, and Figure 3 is a photograph of a typical test setup. This arrangement was used for resistance measurements of all CSST samples.

Figure 2 – Setup Diagram for CSST DC Resistance Measurements

CSST under Measurement

Wooden Supports

Ground Plane

Ground Plane Extensions



Figure 3 – Photograph of Setup used for CSST DC Resistance, Self Inductance and Stray Capacitance Measurements

For the inductance measurements, an Agilent Technologies Model 4263B LRC bridge was driving a 10 kHz sinusoidal current at approximately 40 mA rms through a straight 2-meter section of CSST supported 1-meter above a ground plane. Copper sheet extensions of the ground plane were used to bring the current return for the LRC meter to the CSST sample ends in order to reduce parasitic inductance in the measurement leads. Before measuring any CSST samples, a 0.5” diameter, 2-meter long copper pipe was measured in the test fixture arrangement at 4.4 µH. The shunt capacitance for each CSST sample fixed 1-meter above a ground plane was also measured using the Agilent LRC bridge, set at the same 10 kHz signal. Prior to measuring the capacitance, the LRC meter’s measurement and drive leads had open circuit corrections performed to zero the stray capacitance of the leads themselves.

Figure 4 is a schematic of the inductance and capacitance measurement setups

and Figure 3 is a photograph of a typical test setup. This arrangement was used for inductance and capacitance measurements of all CSST samples.



Figure 4 – Setup Diagram for CSST Self Inductance and Capacitance Measurements

6.2 Per Unit Electrical Characterization Measurement Test Results Because all of the CSST test articles are constructed of the same materials (304

Stainless Steel with a polyethylene jacket) and are of comparable dimensions and wall thicknesses, the samples had very similar per unit length electrical characteristics.

The average DC resistance for all four manufacturers’ 0.5” and 1” diameter sized

CSST was 10.42 mΩ/m and 4.54 mΩ/m, respectively. However, it was noticed that the resistance measurement from the 0.5” CSST of Manufacturer ‘B’ was nearly three times as high as the other three 0.5” samples at 19.9 mΩ/m. This difference is likely due to the additional bonding resistances between the end couplers and the type ‘B’ CSST, and if this is the case, then the average DC resistance of the 0.5” size CSST is a more reasonable 7.26 mΩ/m, or roughly double that of the 1” diameter CSST resistance. This value is more intuitive because the 0.5” diameter CSSTs have about half as much cross sectional area as the 1” diameter CSSTs and would be expected to have roughly twice the resistance; resistance is inversely proportional to the end area of a conductor. These resistance measurements result from 1 A of driven current from the four wire bond meter.

All four manufacturers had very similar per unit length inductances. The average

self inductance for all four manufacturers’ 0.5” diameter CSSTs was 2.0 µH/m, and for all 1” diameter CSSTs was 2.2 µH/m. These inductance values are at 10 kHz and with respect to a ground plane 1-meter below the CSST. Figure 5 shows a photograph of a typical CSST inductance measurement from the LRC meter. Note that these self inductance values are due to less than 60 milliamps of driven current from the LRC meter.



Figure 5 – Typical Inductance Measurement of 2-meter CSST Sample

Initially, the total inductance values for each CSST sample measured were

erroneously zeroed out by subtracting the copper tube calibration fixture inductance of 4.4 µH from each measured inductance of the 2-meter long CSST samples, leading to inductance values roughly 4x too low being recorded in the data book. This calculation has been corrected in Tables 5 and 6 below.

All four manufacturers had very similar per unit length stray shunt capacitance.

The average capacitance per meter for all four manufacturers’ 0.5” and 1” diameter CSST was 0.39 pF/m and 0.43 pF/m, respectively. These capacitance values are at 10 kHz and with respect to a ground plane 1-meter below the samples.

Table 5 lists the DC resistance, inductance and capacitance values recorded for

the 2-meter CSST test samples. Table 6 reports these values in per unit length (per meter) format.

LRC Bridge



Table 5 – Resistance, Inductance and Capacitance Measured on 2-meter CSST Samples

Test No.

CSST Manufacturer

CSST Diameter (Inches)

DC Resistance

(mΩ)

Self Inductance at 10 kHz

(µH)

Shunt Capacitance

at 10 kHz (pF)

LRC Bridge Measurements on 2-meter Long Sections of CSST

1 A

0.5 136.8 4.22 0.82

2 1 87.1 3.93 0.98

3 B

0.5 385.7 4.22 0.8

4 1 91.7 3.94 0.85

5 C

0.5 137.9 4.4 0.7

6 1 88.7 3.85 0.71

7 D

0.5 144.8 4.19 0.67

8 1 85.14 3.87 0.79

Table 6 – CSST per Unit Length Resistance, Inductance and Capacitance

Test No.

CSST Manufacturer

CSST Diameter (Inches)

DC Resistance

(mΩ/m)

Self Inductance at 10 kHz

(µH/m)

Shunt Capacitance

at 10 kHz (pF/m)

1 A

0.5 7.13 2.39 0.428

2 1 4.33 2.42 0.487

3 B

0.5 19.91 2.37 0.413

4 1 4.75 2.52 0.446

5 C

0.5 7.29 2.47 0.37

6 1 4.72 2.63 0.378

7 D

0.5 7.35 2.34 0.34

8 1 4.35 2.52 0.404

6.3 High Voltage Impulse Withstand/Breakdown Test Setup

The high voltage, impulse withstand, and breakdown tests on the CSST gas pipe samples were performed per the method outlined in the GTI Proposal Validation of



Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage (Ref. 1), testing ten samples of each manufacturers’ 0.5” and 1” diameter CSST. The high voltage tests involved raising the potential of the CSST metal tubing using an impulse voltage with respect to a grounded metal plate in contact with the CSST test sample (see Figure 6). This arrangement allowed the outer polyethylene jacket dielectrics to be stressed with the resulting electric fields generated between the CSST tube wall and the grounded plate. Progressively higher impulse voltages were then applied to determine the minimum potential at which dielectric breakdown of the jacket occurred. A breakdown through the jacket grounded the impulse voltage generator, establishing a conductive path from the CSST tube wall to the ground reference of the high voltage generator. If a breakdown did not occur when the high voltage generator was discharged, then a “withstand” resulted, meaning that the full voltage impulse was insufficient to cause the dielectric jacketing to fail.

Figure 6 – Setup Diagram for CSST High Voltage

Impulse Withstand/Breakdown Tests

The high voltage test setup used 2-meter long CSST samples which were oriented so that a 10 cm length of the CSST was in contact with a grounded metal plate. One end of the CSST test sample, well away from the area under test, was attached to the output of a high voltage, impulse generator producing the 1.2 µs x 50 µs waveform of IEC 61000-4-5 (Ref. 2) and IEEC C62.41-1991 (Ref. 3). The other end, near to the area under test, was left open circuit. The 10 inch section of the CSST being evaluated, along with the grounded metal plate, were submerged in a plastic tub containing transformer oil (mineral oil). This “oil bath” was devised to eliminate surface flashover of the CSST jacket which was observed during initial tests. Flashover occurs when the

CSST

Wooden Support

Ground Plane

High Voltage Generator

Grounded Metal Plate

Transformer Oil Bath

Plastic Tub

Open Circuit End of CSST



electrical breakdown takes place in the air surrounding the surface of the CSST’s dielectric jacket, bypassing the jacket to the metal plate area under evaluation. With the setup submerged in oil, meaningful test data could be collected. Each tested CSST sample was given slight downward pressure to maintain a consistent interface between the CSST jacket and the grounded metal plate during the high voltage tests. Some of the 2-meter samples were rotated 180˚ to test the opposite end of the CSST sample in order to avoid flaws in its jacket.

Because of the voltage levels required to achieve electrical breakdown through some of the samples’ jackets, two different high voltage impulse generators were required. Initially an LRC tank circuit was used which had a peak voltage of 42-45 kV. A second Marx type 100 kV generator was used to test those CSST samples which had impulse dielectric strengths higher than 45 kV. Both test setups generated the same 1.2 µs x 50 µs waveform and used either a high voltage probe or voltage resistive divider to record the applied/resulting impulse voltage on a digital storage oscilloscope (DSO). A current transformer (CT) probe was used to record the resulting current when a puncture occurred, which was also recorded on the DSO. The magnitude of this current is almost entirely dependent on the high voltage impulse generator configuration, and not a function of the CSST sample under test.

Figures 7 and 8 show the high voltage test setup using the LRC tank circuit capable of 45 kV. Figures 9 and 10 show the high voltage test setup using the 100 kV Marx high voltage apparatus. Figures 11 through 13 show oscillographs of typical open circuit voltage (Voc) calibrations for the 1.2 µs x 50 µs waveform.

Figure 7 – Overall CSST High Voltage Impulse Withstand/Breakdown Test Setup using LRC Circuit Generator

CSST to Impulse Generator

Connection

1000:1 Voltage Probe

LRC Tank Circuit Voltage Impulse

Generator

Generator Ground

Wooden Stand

Current Probe at Generator Output

CSST Oil Bath with

Grounded Metal Plate Digital

Multimeter



Figure 8 – Oil Bath for the CSST High Voltage Impulse Withstand/ Breakdown Test Setup using LRC Circuit Generator

Figure 9 – Overall CSST High Voltage Impulse Withstand/ Breakdown Test Setup using Marx Generator


Generator Ground


Submerged Section of CSST, Flush with Plate

100 kV Marx High Voltage Impulse

Generator

CSST to Impulse

Generator Connection

Resistive Divider Voltage Probe Generator

Ground

Wooden Stand

Current Probe

Oil Bath



Figure 10 – Oil Bath for the CSST High Voltage Impulse Withstand/Breakdown Test Setup using Marx Generator

Figure 11 – Typical 1.2 µs x 50 µs Impulse Voltage Generator Calibration Oscillograph, Overall


Generator Ground


Submerged Section of

CSST, Flush with Plate



Figure 12 – Typical 1.2 µs x 50 µs Impulse Voltage Generator Calibration Oscillograph, t10-90%

Figure 13 – Typical 1.2 µs x 50 µs Impulse Voltage Generator Calibration Oscillograph, t50% Decay



6.4 High Voltage Impulse Withstand/Breakdown Test Results The range of average dielectric breakdown voltages using the 1.2 µs x 50 µs

voltage impulse waveform for the CSST gas pipe samples was 33 kV to 60.5 kV, depending on the pipe size. The 0.5” diameter products were observed to break down at lower peak impulse voltages than the 1” diameter products from all four manufacturers. The highest and lowest breakdown voltages for the 0.5” diameter products were 55.5 kV and 30.6 kV, respectively, averaging 38.2 kV, with Manufacturer B having the highest withstand voltage levels observed. The highest and lowest breakdown voltages for the 1” diameter products were 60.6 kV and 35.8 kV, respectively, averaging 42.5 kV, with Manufacturer B having the highest withstand voltage levels observed. For each CSST type tested, a total of ten tests from each manufacturer were performed to obtain a reasonable statistical average breakdown voltage level for each CSST type/brand.

Punctures in the CSST samples’ polyethylene jackets were observed as small

pin holes, with a small amount of soot around the puncture point. This type of damage is typical for high voltage, dielectric strength tests because voltage impulse test generators are circuits with inherently high source impedance, and produce low currents (<1 kA).

As mentioned in Section 6.3, the impulse high voltage withstand/ breakdown

tests were performed in a transformer oil bath to prevent flashover in the air outside the jacket surface. This was necessary to produce dielectric strength data for each CSST samples’ jacket material and thickness. Additionally, during the impulse high voltage withstand/breakdown tests, care was taken to test damage- and puncture-free sections of the CSST jacket. Some test samples from all four manufacturers arrived at the test laboratory with small portions of the jacket cracked, split, or torn. These areas were avoided during the high voltage tests. In an actual CSST installation, it should be expected that the CSST jacket could be compromised by similar mechanical damage from installation and/or handling, which would adversely affect the dielectric strength levels observed in the tests. Damage to the CSST insulating jacket, in conjunction with differences in impulse dielectric strength of the air volume around the CSST and its polyethylene jacket material, could result in peak impulse voltages lower than those observed in the high voltage tests in this report.

The high voltage test results summarized herein evaluated only the voltage

impulse withstand characteristics of the CSST samples’ dielectric jackets with respect to a minimum gap to an oppositely charged flat plate. In practice, the impulse withstand voltage of any CSST gas piping depends upon many factors: the rise time and duration of the transient, difference in potential, polarity, orientation and proximity of the CSST with respect to nearby conductors, the atmospheric conditions of the surrounding air, mechanical damage to the jacket material, voids within the dielectric, thickness of the dielectric, temperature, etc. Polyethylene has a wide range of dielectric strengths quoted in various handbooks, all dependent upon the density type (low, medium, or high). For example, the CRC Press’ Electrical Engineering Handbook (Ref. 7) shows an average AC dielectric strength of around 24 kV/mm for all three types of polyethylene.



One of the manufacturers’ CSST had a polyethylene jacket which, for some test articles, reacted adversely to the transformer oil bath. When a split or crack in the article’s jacket was exposed to the transformer oil, it would start to dissolve and the tear or split would grow. Such CSST samples were repositioned until a “non-reactive” section was found or a different, undamaged CSST sample was tested. Figure 14 shows one of these dissolved CSST sections.

Figure 14 – Reaction of CSST Dielectric Jacket of Manufacturer B to Insulating Transformer Oil Bath

Figure 15 shows a typical high voltage post-test pinhole puncture. Breakdown locations were identified on every high voltage CSST sample tested. Figure 16 shows a typical oscillograph resulting from dielectric breakdown and puncture of a CSST jacket. In the figure, the applied voltage (yellow trace) can be seen to collapse towards zero at the time of breakdown. At the same time, the applied current (magenta trace) begins to conduct and rises above zero. Table 7 summarizes the high voltage impulse test results performed on the four manufacturers’ 0.5” and 1” diameter CSST samples. Table 8 contains high voltage impulse test results, including average breakdown voltages for each type/brand tested. The DSO oscillographs for all of the high voltage tests are in Appendix B of this report.



Figure 15 – Puncture in a CSST Dielectric Jacket during High Voltage Impulse Withstand Tests

Figure 16 – Oscillograph of a High Voltage Impulse Withstand Test during a Puncture of the CSST Dielectric Jacket

Voltage Collapse at Electric Breakdown of

Dielectric Jacket

Resulting Current Due to Electric Breakdown

Pin Hole Puncture from High Voltage Breakdown in

Dielectric Jacket



Table 7 – High Voltage Impulse Withstand Test Summary

Sample Manufacturer

Sample Number

Peak Voltage

(kV)

Time to Breakdown

(µs)

Average Peak

Voltage (kV)

Sample Diameter: 0.5 Inches

A

1 31.25 1.71

32.93

33.6 1.93

2 33.28 4.76

32.32 1.15

3 34.88 3.84

32.15 2.03

4 35.01 2.26

31.36 1.51

5 34.76 2.18

30.71 2.01

B

1 57.86 2.28

55.45

58.56 2.74

2 57.39 5.76

44.94 15.94

3 59.7 2.41

61.32 2.00

4 53.97 4.08

52.83 2.41

5 50.21 4.6

57.69 2.24

C

1 27.37 1.47

30.6

32.95 1.76

2 32.14 6.68

31.52 4.81

3 31.97 2.2



Table 7 – High Voltage Impulse Withstand Test Summary (Continued)

Sample Manufacturer

Sample Number

Peak Voltage

(kV)

Time to Breakdown

(µs)

Average Peak

Voltage (kV)

Sample Diameter: 0.5 Inches

C

3 28.9 6.39

30.6 4

29.95 7.00

33.97 3.2

5 28.16 6.05

29.04 2.44

D

1 33.4 2.4

33.77

34.35 5.41

2 32.82 3.81

33.00 2.71

3 36.09 1.71

34.82 2.7

4 33.49 2.62

33.7 2.8

5 31.5 5.05

34.5 2.41

Sample Diameter: 1 Inch

A

1 38.43 2.49

35.76

34.24 5.2

2 37.96 1.84

35.35 2.65

3 36.13 7.46

34.51 0.98

4 39.96 3.14

33.27 6.15




Sample Manufacturer

Sample Number

Peak Voltage

(kV)

Time to Breakdown

(µs)

Average Peak

Voltage (kV)

5 33.87 4.65


A 5 33.85 4.35 35.76

B

1 65.21 1.82

60.62

2 61.44 5.09

59.95 7.08

3 51.2 4.57

66.18 1.99

4 65.72 4.15

52.92 25.1

5 66.83 2.99

63.34 1.33

6 53.37 0.66

C

1 36.16 2.04

35.79

36.36 1.9

2 36.76 2.18

3 36.92 4.11

4 35.83 3.73

35.49 3.99

5 33.54 3.38

34.24 2.00

6 36.45 3.48

7 36.1 3.45

D 1

43.64 2.45

37.82 37.53 1.22

2 35.04 1.64




Sample Manufacturer

Sample Number

Peak Voltage

(kV)

Time to Breakdown

(µs)

Average Peak

Voltage (kV)

38.68 3.17


D

3 37.39 1.19

37.82

36.98 1.86

4 40.00 1.48

36.16 1.75

5 35.77 1.43

36.97 1.54

Table 8 – High Voltage Impulse Withstand Test Log

Test No.

CSST Mfr./ Sample No.

CSST Dia.

(Inches)

Generator Charge Voltage

(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

4 September 2012

Calibrations of 1.2 µs x 50 µs Voltage Impulse Generator; Jacketed CSST Samples Flush to Grounded Metal Plate; Impulse Potential Applied to CSST End

with Current Return Referenced to Grounded Plate

1 Calibration 13 -- 1.39 + Voc = 10 kV, t50% decay = 55.5 µs

2 Calibration 13 -- 6.2 + Isc = 7.29 kA,

t50% decay = 16 µs; Total Charge = 0.12 C

CSST Manufacturer/Sample No. C/1, 0.5” Diameter

3 C/1 0.5 19.5 15 >90 + Withstand

4 C/1 0.5 21.5 16.8 >90 + Withstand

5 C/1 0.5 23.3 18.1 >90 + Withstand

6 C/1 0.5 25.1 19.9 >90 + Withstand

7 C/1 0.5 26.9 21.6 >90 + Withstand

8 C/1 0.5 28.7 22.9 >90 + Withstand

9 C/1 0.5 30 23.7 2 + Breakdown at CSST W/O Puncture (Off End)



Table 8 – High Voltage Impulse Withstand Test Log (Continued)

Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

5 September 2012

Calibrations of Modified 1.2 µs x 50 µs Voltage Impulse Generator with Higher Source Impedance

10 Calibration 17.4 15.5 1.35 + Isc = 842 A at 7 µs x 13 µs;

t50% decay = 55.3 µs


11 C/1 0.5 22.3 20 >90 + Withstand

12 C/1 0.5 22.4 20.3 >90 + Withstand

13 C/1 0.5 22.5 20.3 >90 + Withstand

14 C/1 0.5 23 20.5 >90 + Withstand

15 C/1 0.5 23.4 20.8 >90 + Withstand

16 C/1 0.5 24.1 21.7 >90 + Withstand

17 C/1 0.5 24.3 21.8 >90 + Withstand

18 C/1 0.5 24.6 22.3 >90 + Withstand

19 C/1 0.5 24.9 22.6 >90 + Withstand

20 C/1 0.5 25.2 22.5 >90 + Withstand

21 C/1 0.5 25.4 23 >90 + Withstand

22 C/1 0.5 25.5 23.1 >90 + Withstand

23 C/1 0.5 25.8 23.4 >90 + Withstand

24 C/1 0.5 25.9 23.4 >90 + Withstand

25 C/1 0.5 26.1 23.8 >90 + Withstand

26 C/1 0.5 26.4 23.8 >90 + Withstand

27 C/1 0.5 26.7 24 >90 + Withstand

28 C/1 0.5 26.8 24.4 >90 + Withstand

29 C/1 0.5 27.1 24.4 2.2 + Breakdown, Surface

Flashover; No Puncture

30 C/1 0.5 26 24.7 >90 + Withstand

31 C/1 0.5 26.2 25 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

5 September 2012


32 C/1 0.5 26.4 25 >90 + Withstand

33 C/1 0.5 26.8 25.5 >90 + Withstand

34 C/1 0.5 26.9 25.7 >90 + Withstand

35 C/1 0.5 27.2 25.8 >90 + Withstand

36 C/1 0.5 27.8 26.9 >90 + Withstand

37 C/1 0.5 28.1 26.6 >90 + Withstand

38 C/1 0.5 28.5 27.3 >90 + Withstand

39 C/1 0.5 28.5 27.4 >90 + Withstand

Added Transformer Oil Bath at Location of CSST at Grounded Plate to Mitigate Flashover at Unwanted Surfaces

40 C/1 0.5 28.5 26.2 >90 + Withstand

41 C/1 0.5 29.7 27.2 >90 + Withstand

42 C/1 0.5 30.5 27.9 >90 + Withstand

43 C/1 0.5 31 28.7 >90 + Withstand

44 C/1 0.5 31.8 28.7 2.2 + Breakdown, Flashover at End, No Puncture

45 C/1 0.5 31.8 28.5 >90 + Withstand

46 C/1 0.5 32 27.4 6.4 + Breakdown, Flashover at End, No Puncture

Rotated CSST Sample to Test Opposite End; Approx. 10 cm Section Submerged in Oil Bath

47 C/1 0.5 29 26.6 >90 + Withstand

48 C/1 0.5 29.4 26.8 >90 + Withstand

49 C/1 0.5 29.8 27.3 >90 + Withstand

50 C/1 0.5 30.2 27.5 >90 + Withstand

51 C/1 0.5 30.7 28 >90 + Withstand

52 C/1 0.5 31 27.8 >90 + Withstand

53 C/1 0.5 31.3 29 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

5 September 2012


54 C/1 0.5 32 29 >90 + Withstand

55 C/1 0.5 32.5 29.4 >90 + Withstand

56 C/1 0.5 33 30.1 >90 + Withstand

57 C/1 0.5 33.5 30.6 >90 + Withstand

58 C/1 0.5 34 30.8 >90 + Withstand

59 C/1 0.5 34.5 31.6 >90 + Withstand

60 C/1 0.5 35 31.8 >90 + Withstand

61 C/1 0.5 35.6 33 2.9 + Puncture of CSST Jacket, 1.76 kA


62 C/2 0.5 33.5 30.2 >90 + Withstand

63 C/2 0.5 34 30.5 >90 + Withstand

64 C/2 0.5 34.5 31.3 >90 + Withstand

65 C/2 0.5 35 31.6 >90 + Withstand

66 C/2 0.5 35.5 32 >90 + Withstand

67 C/2 0.5 36 32.6 >90 + Withstand

68 C/2 0.5 36.5 32.7 >90 + Withstand

69 C/2 0.5 37 32.1 6.7 + Puncture of CSST Jacket, 1.73 kA


70 C/2 0.5 35 31.7 >90 + Withstand

71 C/2 0.5 35.5 32.4 >90 + Withstand





Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

6 September 2012


73 C/3 0.5 35 32 2.2 + Puncture of CSST Jacket, 1.68 kA




75 C/4 0.5 30 27 >90 + Withstand

76 C/4 0.5 30.5 27.7 >90 + Withstand

77 C/4 0.5 31 27.9 >90 + Withstand

78 C/4 0.5 31.5 28.3 >90 + Withstand

79 C/4 0.5 32 28.8 >90 + Withstand

80 C/4 0.5 32.5 29.2 >90 + Withstand

81 C/4 0.5 33 29.5 >90 + Withstand

82 C/4 0.5 33.5 30.2 >90 + Withstand

83 C/4 0.5 34 30.6 >90 + Withstand

84 C/4 0.5 34.5 30.9 >90 + Withstand

85 C/4 0.5 35 31.4 >90 + Withstand

86 C/4 0.5 35.5 30 7 + Puncture of CSST Jacket, 1.6 kA


87 C/4 0.5 31 28.8 >90 + Withstand

88 C/4 0.5 31.5 28.7 >90 + Withstand

89 C/4 0.5 32 29.1 >90 + Withstand

90 C/4 0.5 32.5 29 >90 + Withstand

91 C/4 0.5 33 30 >90 + Withstand

92 C/4 0.5 33.5 30.5 >90 + Withstand

93 C/4 0.5 34 30.9 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

6 September 2012


94 C/4 0.5 34.5 31.2 >90 + Withstand

95 C/4 0.5 35 31.3 >90 + Withstand

96 C/4 0.5 35.5 31.8 >90 + Withstand

97 C/4 0.5 36 32.5 >90 + Withstand

98 C/4 0.5 36.5 32.5 >90 + Withstand

99 C/4 0.5 37 33.3 >90 + Withstand

100 C/4 0.5 37.5 34 >90 + Withstand



102 C/5 0.5 32 29.4 >90 + Withstand

103 C/5 0.5 32.5 28.2 6.1 + Puncture of CSST Jacket, 1.53 kA



CSST Manufacturer/Sample No. C/1, 1” Diameter

105 C/1 1 20 18.9 >90 + Withstand

106 C/1 1 21 19.2 >90 + Withstand

107 C/1 1 22 20.3 >90 + Withstand

108 C/1 1 23 21.4 >90 + Withstand

109 C/1 1 24 22.2 >90 + Withstand

110 C/1 1 25 23.2 >90 + Withstand

111 C/1 1 26 24.6 >90 + Withstand

112 C/1 1 27 25.2 >90 + Withstand

113 C/1 1 28 25.8 >90 + Withstand

114 C/1 1 29 26.9 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

6 September 2012


115 C/1 1 30 27.3 >90 + Withstand

116 C/1 1 31 28.4 >90 + Withstand

117 C/1 1 32 29.8 >90 + Withstand

118 C/1 1 32.5 29.9 >90 + Withstand

119 C/1 1 33 30.7 >90 + Withstand

120 C/1 1 33.5 30.5 >90 + Withstand

121 C/1 1 34 31.2 >90 + Withstand

122 C/1 1 34.5 31.9 >90 + Withstand

123 C/1 1 35 32.2 >90 + Withstand

124 C/1 1 35.5 32.6 >90 + Withstand

125 C/1 1 36 33.1 >90 + Withstand

126 C/1 1 36.5 33.5 >90 + Withstand

127 C/1 1 37 34 >90 + Withstand

128 C/1 1 37.5 34.5 >90 + Withstand

129 C/1 1 38 35 >90 + Withstand

130 C/1 1 38.5 35.5 >90 + Withstand

131 C/1 1 39 36.2 2 + Puncture of CSST Jacket, 1.97 kA


132 C/1 1 30 28.1 >90 + Withstand

133 C/1 1 30.5 27.9 >90 + Withstand

134 C/1 1 31 28.2 >90 + Withstand

135 C/1 1 31.5 28.6 >90 + Withstand

136 C/1 1 32 29.1 >90 + Withstand

137 C/1 1 32.5 29.7 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

6 September 2012


138 C/1 1 33 30.1 >90 + Withstand

139 C/1 1 33.5 30.3 >90 + Withstand

140 C/1 1 34 31.3 >90 + Withstand

141 C/1 1 34.5 31.4 >90 + Withstand

142 C/1 1 35 31.9 >90 + Withstand

143 C/1 1 35.5 32.4 >90 + Withstand

144 C/1 1 36 33.1 >90 + Withstand

145 C/1 1 36.5 33.3 >90 + Withstand

146 C/1 1 37 33.9 >90 + Withstand

147 C/1 1 37.5 34.2 >90 + Withstand

148 C/1 1 38 35.7 >90 + Withstand

149 C/1 1 38.5 35.8 >90 + Withstand

150 C/1 1 39 35.8 >90 + Withstand

151 C/1 1 39.5 36.4 2.5 + Puncture of CSST Jacket, 1.9 kA


152 C/2 1 33 30.7 >90 + Withstand

153 C/2 1 34 31.3 >90 + Withstand

154 C/2 1 34.5 32.4 >90 + Withstand

155 C/2 1 35 33 >90 + Withstand

156 C/2 1 35.5 33.3 >90 + Withstand

157 C/2 1 36 33.6 >90 + Withstand

158 C/2 1 36.5 34.2 >90 + Withstand

159 C/2 1 37 34.7 >90 + Withstand

160 C/2 1 37.5 35 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

6 September 2012


161 C/2 1 38 35 >90 + Withstand

162 C/2 1 38.5 36.2 >90 + Withstand

163 C/2 1 39 36.1 >90 + Withstand

164 C/2 1 39.5 36.7 >90 + Withstand

165 C/2 1 40 37.1 >90 + Withstand

166 C/2 1 40.5 37.4 >90 + Withstand

167 C/2 1 41 37.4 >90 + Withstand

Limitation of Impulse Generator Reached. Rotated CSST Sample to Test Opposite End; Approx. 10 cm Section Submerged in Oil Bath

168 C/2 1 35 32.9 >90 + Withstand

169 C/2 1 35.5 32.9 >90 + Withstand

170 C/2 1 36 33.9 >90 + Withstand

171 C/2 1 36.5 33.2 >90 + Withstand

172 C/2 1 37 34.3 >90 + Withstand

173 C/2 1 37.5 35.2 >90 + Withstand

174 C/2 1 38 34.9 >90 + Withstand

175 C/2 1 38.5 35.8 >90 + Withstand

176 C/2 1 39 36.2 >90 + Withstand



178 C/3 1 35 32.9 >90 + Withstand

179 C/3 1 35.5 33.4 >90 + Withstand

180 C/3 1 36 33.5 >90 + Withstand

181 C/3 1 36.5 34.3 >90 + Withstand

182 C/3 1 37 34.6 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

6 September 2012


183 C/3 1 37.5 35.4 >90 + Withstand

184 C/3 1 38 35.4 >90 + Withstand

185 C/3 1 38.5 35.8 >90 + Withstand

186 C/3 1 39 36 >90 + Withstand

187 C/3 1 39.5 36.9 >90 + Withstand

188 C/3 1 40 37.1 >90 + Withstand

189 C/3 1 40.5 37.4 >90 + Withstand

190 C/3 1 41 38 >90 + Withstand


191 C/3 1 37 33.7 >90 + Withstand

192 C/3 1 37.5 34.5 >90 + Withstand

193 C/3 1 38 34.5 >90 + Withstand

194 C/3 1 38.5 34.8 >90 + Withstand

195 C/3 1 39 35.3 >90 + Withstand

196 C/3 1 39.5 36.1 >90 + Withstand

197 C/3 1 40 37.2 >90 + Withstand

198 C/3 1 40.5 37.4 >90 + Withstand

199 C/3 1 41 36.9 4.1 + Puncture of CSST Jacket, 1.96 kA


200 C/4 1 37 34.1 >90 + Withstand

201 C/4 1 37.5 34.7 >90 + Withstand

202 C/4 1 38 35.2 >90 + Withstand

203 C/4 1 38.5 35.8 >90 + Withstand





Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

6 September 2012



205 C/4 1 37 34.3 >90 + Withstand

206 C/4 1 37.5 34.9 >90 + Withstand



208 C/5 1 37 34.6 >90 + Withstand

209 C/5 1 37.5 35 >90 + Withstand

210 C/5 1 38 36.2 >90 + Withstand

211 C/5 1 38.5 36.1 >90 + Withstand

212 C/5 1 39 36.2 >90 + Withstand

213 C/5 1 39.5 36.8 >90 + Withstand

214 C/5 1 40 37.6 >90 + Withstand

215 C/5 1 40.5 37.8 >90 + Withstand

216 C/5 1 41 38.1 >90 + Withstand





10 September 2012


219 C/6 1 35 32.4 >90 + Withstand

220 C/6 1 35.5 33.6 >90 + Withstand

221 C/6 1 36.5 33.3 >90 + Withstand

222 C/6 1 37 35.8 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

10 September 2012


223 C/6 1 37.5 34.5 >90 + Withstand

224 C/6 1 38 35 >90 + Withstand

225 C/6 1 38.5 35.5 >90 + Withstand

226 C/6 1 39 36.2 >90 + Withstand

227 C/6 1 39.5 37.1 >90 + Withstand

228 C/6 1 40 36.8 >90 + Withstand




CSST Manufacturer/Sample No. B/1, 1” Diameter

231 B/1 1 22 19.6 >90 + Withstand

232 B/1 1 23 21.8 >90 + Withstand

233 B/1 1 25 24 >90 + Withstand

234 B/1 1 27 25.5 >90 + Withstand

235 B/1 1 28 26.1 >90 + Withstand

236 B/1 1 29 27.6 >90 + Withstand

237 B/1 1 30 28.5 >90 + Withstand

238 B/1 1 31 29.2 >90 + Withstand

239 B/1 1 32 30 >90 + Withstand

240 B/1 1 33 30 >90 + Withstand

241 B/1 1 34 31.7 >90 + Withstand

242 B/1 1 35 32.8 >90 + Withstand

243 B/1 1 36 33.5 >90 + Withstand

244 B/1 1 37 34.9 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

10 September 2012


245 B/1 1 38 35.5 >90 + Withstand

246 B/1 1 38.5 36.5 >90 + Withstand

247 B/1 1 39 36.4 >90 + Withstand

248 B/1 1 39.5 36.9 >90 + Withstand

249 B/1 1 40 37.7 >90 + Withstand

250 B/1 1 40.5 37.2 >90 + Withstand

251 B/1 1 41 38.1 >90 + Withstand

Limitation of Voltage Probe Reached. Rotated CSST Sample to Test Opposite End; Approx. 10 cm Section Submerged in Oil Bath

252 B/1 1 36 34.3 >90 + Withstand

253 B/1 1 37 34.9 >90 + Withstand

254 B/1 1 38 35.7 >90 + Withstand

255 B/1 1 39 36.9 >90 + Withstand

256 B/1 1 39.5 36.9 >90 + Withstand

257 B/1 1 40 37.4 >90 + Withstand

258 B/1 1 40.5 38.3 >90 + Withstand

259 B/1 1 41 38.6 >90 + Withstand

260 B/1 1 41.5 38.5 >90 + Withstand

261 B/1 1 42 38.5 >90 + Withstand

262 B/1 1 42.5 39.2 >90 + Withstand

Limitation of Voltage Probe (40 kV rating) Reached. Modified Setup to Achieve Slightly Higher Output Voltages by Removing Voltage Probe

263 B/1 1 43 No Data >90 + Withstand

264 B/1 1 43.5 No Data >90 + Withstand





Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

10 September 2012


266 B/1 1 44.5 No Data >90 + Withstand


Limitation of Impulse Generator Reached. Will Re-Test B/1 Samples with Higher Voltage Generator (See Test Nos. 429-443)

CSST Manufacturer/Sample No. A/1, 1” Diameter

268 A/1 1 30 27.8 >90 + Withstand

269 A/1 1 31 28.8 >90 + Withstand

270 A/1 1 32 29.7 >90 + Withstand

271 A/1 1 33 30.7 >90 + Withstand

272 A/1 1 34 31.6 >90 + Withstand

273 A/1 1 35 32.5 >90 + Withstand

274 A/1 1 36 33.9 >90 + Withstand

275 A/1 1 37 35 >90 + Withstand

276 A/1 1 38 35.5 >90 + Withstand

277 A/1 1 39 36.3 >90 + Withstand

278 A/1 1 39.5 36.7 >90 + Withstand

279 A/1 1 40 37 >90 + Withstand

280 A/1 1 40.5 38.6 >90 + Withstand

281 A/1 1 41 38.4 >90 + Withstand

282 A/1 1 41.5 38.6 >90 + Withstand

283 A/1 1 42 39 >90 + Withstand

Limitation of Voltage Probe Reached

CSST Manufacturer/Sample No. A/1, 0.5” Diameter

284 A/1 0.5 32 29.3 >90 + Withstand

285 A/1 0.5 33 30.1 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

10 September 2012


286 A/1 0.5 34 31.3 1.7 + Puncture of CSST Jacket, 1.69 kA


287 A/1 0.5 32 29.4 >90 + Withstand

288 A/1 0.5 33 30.9 >90 + Withstand

289 A/1 0.5 33.5 30.5 >90 + Withstand

290 A/1 0.5 34 31 >90 + Withstand

291 A/1 0.5 34.5 31.2 >90 + Withstand

292 A/1 0.5 35 32.2 >90 + Withstand

293 A/1 0.5 35.5 32.5 >90 + Withstand

294 A/1 0.5 36 33.2 >90 + Withstand

295 A/1 0.5 36.5 33.6 1.9 + Puncture of CSST Jacket, 1.74 kA


296 A/2 0.5 33 29.8 >90 + Withstand

297 A/2 0.5 33.5 29.8 >90 + Withstand

298 A/2 0.5 34 31.3 >90 + Withstand

299 A/2 0.5 35 32.1 >90 + Withstand

300 A/2 0.5 35.5 32.3 >90 + Withstand

301 A/2 0.5 36 32.5 >90 + Withstand

302 A/2 0.5 36.5 33 >90 + Withstand



304 A/2 0.5 33 30.3 >90 + Withstand

305 A/2 0.5 33.5 30.7 >90 + Withstand

306 A/2 0.5 34 31.2 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

10 September 2012


307 A/2 0.5 34.5 31.3 >90 + Withstand

308 A/2 0.5 35 31.1 >90 + Withstand

309 A/2 0.5 35.5 32 >90 + Withstand

310 A/2 0.5 36 33.4 >90 + Withstand

311 A/2 0.5 36.5 33.1 >90 + Withstand



313 A/3 0.5 34 31.6 >90 + Withstand

314 A/3 0.5 34.5 31.9 >90 + Withstand

315 A/3 0.5 35 31.8 >90 + Withstand

316 A/3 0.5 35.5 32.2 >90 + Withstand

317 A/3 0.5 36 33.5 >90 + Withstand

318 A/3 0.5 36.5 32.9 >90 + Withstand

319 A/3 0.5 37 34.2 >90 + Withstand

320 A/3 0.5 37.5 34.5 >90 + Withstand

321 A/3 0.5 38 34.3 >90 + Withstand

322 A/3 0.5 38.5 35.2 >90 + Withstand



324 A/3 0.5 34 30.9 >90 + Withstand

325 A/3 0.5 34.5 31.5 >90 + Withstand

326 A/3 0.5 35 32 >90 + Withstand

327 A/3 0.5 35.5 32.1 >90 + Withstand

328 A/3 0.5 36 32.2 2 + Puncture of CSST Jacket, 1.85 kA




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

10 September 2012


329 A/4 0.5 34 31.1 >90 + Withstand

330 A/4 0.5 34.5 32 >90 + Withstand

331 A/4 0.5 35 32.2 >90 + Withstand

332 A/4 0.5 35.5 32.5 >90 + Withstand

333 A/4 0.5 36 33.4 >90 + Withstand

334 A/4 0.5 36.5 33.3 >90 + Withstand

335 A/4 0.5 37 33.1 >90 + Withstand

336 A/4 0.5 37.5 34.2 >90 + Withstand

337 A/4 0.5 38 34.9 >90 + Withstand

338 A/4 0.5 38.5 35.1 >90 + Withstand

339 A/4 0.5 39 35 2.3 + Puncture of CSST Jacket, 1.92 kA




341 A/5 0.5 34 32 >90 + Withstand

342 A/5 0.5 34.5 31.3 >90 + Withstand

343 A/5 0.5 35 31.4 >90 + Withstand

344 A/5 0.5 36 33.2 >90 + Withstand

345 A/5 0.5 36.5 33.3 >90 + Withstand

346 A/5 0.5 37 33.6 >90 + Withstand

347 A/5 0.5 37.5 34.8 2.2 + Puncture of CSST Jacket, 1.85 kA


348 A/5 0.5 34 32.3 >90 + Withstand

349 A/5 0.5 34.5 30.7 2 + Puncture of CSST Jacket, 1.66 kA




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

10 September 2012

CSST Manufacturer/Sample No. D/1, 0.5” Diameter

350 D/1 0.5 30 28.5 >90 + Withstand

351 D/1 0.5 31.5 29 >90 + Withstand

352 D/1 0.5 33 30.4 >90 + Withstand

353 D/1 0.5 34 31 >90 + Withstand

354 D/1 0.5 35 31.7 >90 + Withstand

355 D/1 0.5 36 33.3 >90 + Withstand

356 D/1 0.5 36.5 33.1 >90 + Withstand

357 D/1 0.5 37 33.4 2.4 + Puncture of CSST Jacket, 1.77 kA


358 D/1 0.5 34 31.4 >90 + Withstand

359 D/1 0.5 35 31.4 >90 + Withstand

360 D/1 0.5 35.5 32.6 >90 + Withstand

361 D/1 0.5 36 32.4 >90 + Withstand

362 D/1 0.5 36.5 33.1 >90 + Withstand

363 D/1 0.5 37 33.7 >90 + Withstand

364 D/1 0.5 37.5 34.5 5.4 + Puncture of CSST Jacket, 1.79 kA


365 D/2 0.5 34 30.4 >90 + Withstand

366 D/2 0.5 35 32.3 >90 + Withstand



368 D/2 0.5 34 31.3 >90 + Withstand

369 D/2 0.5 35 31.4 >90 + Withstand

370 D/2 0.5 36 32.7 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

10 September 2012


371 D/2 0.5 36.5 33 2.7 + Puncture of CSST Jacket, 1.74 kA


372 D/3 0.5 34 30.7 >90 + Withstand

373 D/3 0.5 35 31.9 >90 + Withstand

374 D/3 0.5 36.6 33 >90 + Withstand

375 D/3 0.5 37 33.7 >90 + Withstand

376 D/3 0.5 37.5 33.7 >90 + Withstand

377 D/3 0.5 38 35.1 >90 + Withstand

378 D/3 0.5 38.5 35.3 >90 + Withstand

379 D/3 0.5 39 35.9 >90 + Withstand

380 D/3 0.5 39.5 35.8 >90 + Withstand



382 D/3 0.5 34 31.1 >90 + Withstand

383 D/3 0.5 35 32.1 >90 + Withstand

384 D/3 0.5 36 32.4 >90 + Withstand

385 D/3 0.5 36.5 33.5 >90 + Withstand

386 D/3 0.5 37 33.5 >90 + Withstand

387 D/3 0.5 37.5 34.4 >90 + Withstand



389 D/4 0.5 35 32.3 >90 + Withstand

390 D/4 0.5 36 33.3 >90 + Withstand





Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

10 September 2012



392 D/4 0.5 35 32.2 >90 + Withstand



394 D/5 0.5 34 31.7 >90 + Withstand

395 D/5 0.5 35 31.7 >90 + Withstand



397 D/5 0.5 35 32.2 >90 + Withstand

398 D/5 0.5 35.5 32.8 >90 + Withstand

399 D/5 0.5 36 33.3 >90 + Withstand

400 D/5 0.5 36.5 33.4 >90 + Withstand

401 D/5 0.5 37 34.6 >90 + Withstand


11 September 2012

CSST Manufacturer/Sample No. B/1, 0.5” Diameter

403 B/1 0.5 30 28.5 >90 + Withstand

404 B/1 0.5 31 28.6 >90 + Withstand

405 B/1 0.5 32 29.8 >90 + Withstand

406 B/1 0.5 33 30.3 >90 + Withstand

407 B/1 0.5 34 31.4 >90 + Withstand

408 B/1 0.5 35 32.9 >90 + Withstand

409 B/1 0.5 36 33.6 >90 + Withstand

410 B/1 0.5 37 34.6 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

11 September 2012


411 B/1 0.5 38 35.1 >90 + Withstand

412 B/1 0.5 39 36.4 >90 + Withstand

413 B/1 0.5 39.5 36.4 >90 + Withstand

414 B/1 0.5 40 37.1 >90 + Withstand

415 B/1 0.5 40.5 37.6 >90 + Withstand

416 B/1 0.5 41 37.3 >90 + Withstand

417 B/1 0.5 41.5 38.4 >90 + Withstand

418 B/1 0.5 42 39.3 >90 + Withstand

419 B/1 0.5 42.5 39.4 >90 + Withstand


420 B/1 0.5 38 35 >90 + Withstand

421 B/1 0.5 39 35.2 >90 + Withstand

422 B/1 0.5 40 36 >90 + Withstand

423 B/1 0.5 40.5 37 >90 + Withstand

424 B/1 0.5 41 37.9 >90 + Withstand

425 B/1 0.5 41.5 38.1 >90 + Withstand

426 B/1 0.5 42 37.8 >90 + Withstand

427 B/1 0.5 42.5 39.1 >90 + Withstand

428 B/1 0.5 43 39.1 >90 + Withstand

12 September 2012

Switching to a Higher Output, Impulse Voltage Apparatus: 100 kV Marx Generator at 1.2 µs x 50 µs

429 Calibration -- 46 0.87 + Voc = 57.6 kV t50% decay = 49.7 µs

430 B/1 0.5 40 32.6 >90 + Withstand

431 B/1 0.5 50 40.3 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

12 September 2012


432 B/1 0.5 52.4 42.4 >90 + Withstand

433 B/1 0.5 56 44.9 >90 + Withstand

434 B/1 0.5 60 48 >90 + Withstand

435 B/1 0.5 60 48 >90 + Withstand

436 B/1 0.5 62.4 50.2 >90 + Withstand

437 B/1 0.5 72 57.9 2.3 + Puncture of CSST Jacket, 0.5 kA


438 B/1 0.5 62 49.9 >90 + Withstand

439 B/1 0.5 64.4 52.3 >90 + Withstand

440 B/1 0.5 66.8 54.4 >90 + Withstand

441 B/1 0.5 69.2 55.7 >90 + Withstand

442 B/1 0.5 71 57.3 >90 + Withstand

443 B/1 0.5 72.4 58.6 2.7 + Puncture of CSST Jacket, 0.47 kA


444 B/2 0.5 62 49.9 >90 + Withstand

445 B/2 0.5 64.4 52.6 >90 + Withstand

447 B/2 0.5 66.4 53.2 >90 + Withstand

448 B/2 0.5 68.8 55.5 >90 + Withstand

449 B/2 0.5 71.2 57.6 >90 + Withstand

450 B/2 0.5 72.8 59.1 >90 + Withstand







Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

12 September 2012


453 B/3 0.5 61.4 49.6 >90 + Withstand

454 B/3 0.5 64 51.3 >90 + Withstand

455 B/3 0.5 66.2 53.5 >90 + Withstand

456 B/3 0.5 68.4 54.8 >90 + Withstand

457 B/3 0.5 69.6 56.3 >90 + Withstand

458 B/3 0.5 72 57.6 >90 + Withstand

459 B/3 0.5 73 59.1 >90 + Withstand



461 B/3 0.5 64 51.7 >90 + Withstand

462 B/3 0.5 65 52.8 >90 + Withstand

463 B/3 0.5 66.8 54.4 >90 + Withstand

464 B/3 0.5 68.2 55.4 >90 + Withstand

465 B/3 0.5 70.1 56.3 >90 + Withstand

466 B/3 0.5 71.2 57.9 >90 + Withstand

467 B/3 0.5 72.2 58.1 >90 + Withstand

468 B/3 0.5 73 58.9 >90 + Withstand

469 B/3 0.5 74 59.6 >90 + Withstand

470 B/3 0.5 75.2 61.3 2 + Puncture of CSST Jacket, 0.64 kA


471 B/4 0.5 64.8 52.2 >90 + Withstand

472 B/4 0.5 66.4 53.4 >90 + Withstand

473 B/4 0.5 68 54 4.1 + Puncture of CSST Jacket, 0.45 kA





Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

12 September 2012


474 B/4 0.5 62.4 50.4 >90 + Withstand



476 B/5 0.5 60.2 48.3 >90 + Withstand

477 B/5 0.5 62.6 50.8 >90 + Withstand



479 B/5 0.5 56.8 45.7 >90 + Withstand

480 B/5 0.5 58.8 47.5 >90 + Withstand

481 B/5 0.5 60 48.9 >90 + Withstand

482 B/5 0.5 62 50.1 >90 + Withstand

483 B/5 0.5 64 50.8 >90 + Withstand

484 B/5 0.5 66 53.1 >90 + Withstand

485 B/5 0.5 68 54.7 >90 + Withstand

486 B/5 0.5 70 55.7 >90 + Withstand

487 B/5 0.5 72 57.7 2.2 + Puncture of CSST Jacket, 0.45 kA


488 A/1 1 46 38.4 2.5 + Puncture of CSST Jacket, 0.35 kA




490 A/2 1 40 32.6 >90 + Withstand

491 A/2 1 42 33.8 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

12 September 2012


492 A/2 1 43 34.2 >90 + Withstand

493 A/2 1 44 35.7 >90 + Withstand

494 A/2 1 45 36.5 >90 + Withstand

495 A/2 1 46 39 >90 + Withstand

496 A/2 1 47 38 1.8 + Puncture of CSST Jacket, 0.27 kA


497 A/2 1 42 33.3 >90 + Withstand



499 A/3 1 40 31.7 >90 + Withstand

500 A/3 1 42 33.3 >90 + Withstand

501 A/3 1 44 35 >90 + Withstand

502 A/3 1 46 36.7 >90 + Withstand



504 A/3 1 40 32 >90 + Withstand

505 A/3 1 42 33.3 >90 + Withstand

506 A/3 1 44 34.7 >90 + Withstand

507 A/3 1 46 34.5 1 + Puncture of CSST Jacket, 0.38 kA


508 A/4 1 40 31.8 >90 + Withstand

509 A/4 1 42 33.9 >90 + Withstand

510 A/4 1 44 34.9 >90 + Withstand

511 A/4 1 46 36.9 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

12 September 2012


512 A/4 1 48 39 >90 + Withstand

513 A/4 1 50 40 3.1 + Puncture of CSST Jacket, 0.35 kA


514 A/4 1 40 32.3 >90 + Withstand

515 A/4 1 42 33.5 >90 + Withstand



517 A/5 1 40 33 >90 + Withstand

518 A/5 1 42 34.2 >90 + Withstand



520 A/5 1 40 32.2 >90 + Withstand

521 A/5 1 42 34.3 >90 + Withstand


CSST Manufacturer/Sample No. D/1, 1” Diameter

523 D/1 1 40 32.1 >90 + Withstand

524 D/1 1 42 34.2 >90 + Withstand

525 D/1 1 44 34.7 >90 + Withstand

526 D/1 1 46 37 >90 + Withstand

527 D/1 1 48 38.3 >90 + Withstand

528 D/1 1 50 39.9 >90 + Withstand

529 D/1 1 52 41.8 >90 + Withstand

530 D/1 1 54 43.6 2.5 + Puncture of CSST Jacket, 0.42 kA




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

12 September 2012





532 D/2 1 44 35 1.6 + Puncture of CSST Jacket, 0.37 kA


533 D/2 1 40 31.9 >90 + Withstand

534 D/2 1 42 33 >90 + Withstand

536 D/2 1 44 35.2 >90 + Withstand

537 D/2 1 46 36.6 >90 + Withstand


13 September 2012


539 D/3 1 40 33.1 >90 + Withstand

540 D/3 1 42 34.5 >90 + Withstand

541 D/3 1 44 34.7 >90 + Withstand



543 D/3 1 40 32.6 >90 + Withstand

544 D/3 1 42 33.9 >90 + Withstand

545 D/3 1 44 35.5 >90 + Withstand



547 D/4 1 40 31.7 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

13 September 2012


548 D/4 1 42 34.3 >90 + Withstand

549 D/4 1 44 36.1 >90 + Withstand

550 D/4 1 46 36.4 >90 + Withstand

551 D/4 1 48 39.3 >90 + Withstand



553 D/4 1 40 32.5 >90 + Withstand

554 D/4 1 42 34 >90 + Withstand

555 D/4 1 43 35.7 >90 + Withstand



557 D/5 1 40 32.6 >90 + Withstand

558 D/5 1 42 33.8 >90 + Withstand

559 D/5 1 43 34.4 >90 + Withstand



561 D/5 1 40 32 >90 + Withstand

562 D/5 1 42 33.6 >90 + Withstand

563 D/5 1 44 35 >90 + Withstand



565 B/1 1 36 26.1 >90 + Withstand

566 B/1 1 40 32.3 >90 + Withstand

567 B/1 1 42 33.3 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

13 September 2012


568 B/1 1 44 34.4 >90 + Withstand

569 B/1 1 46 36.2 >90 + Withstand

570 B/1 1 48 33.9 >90 + Withstand

Jacket Insulation of CSST Type ‘B’ is Dissolving in Transformer Oil, Resulting in Exposure of CSST’s Metal

571 B/1 1 50 40.1 >90 + Withstand

572 B/1 1 52 42.5 >90 + Withstand

573 B/1 1 54 44.2 >90 + Withstand

574 B/1 1 56 44.8 >90 + Withstand

575 B/1 1 58 47.3 >90 + Withstand

576 B/1 1 60 48.4 >90 + Withstand

577 B/1 1 62 50.6 >90 + Withstand

578 B/1 1 64 51.7 >90 + Withstand

579 B/1 1 66 53.4 >90 + Withstand

580 B/1 1 68 54.8 >90 + Withstand

581 B/1 1 70 56.9 >90 + Withstand

582 B/1 1 72 57.9 >90 + Withstand

583 B/1 1 74 59.6 >90 + Withstand

584 B/1 1 76 60.9 >90 + Withstand

585 B/1 1 78 62.6 >90 + Withstand

586 B/1 1 80 64 >90 + Withstand

587 B/1 1 82 65.2 1.8 + Puncture of CSST Jacket, 0.74 kA


588 B/2 1 60 48.7 >90 + Withstand

589 B/2 1 62 49.8 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

13 September 2012


590 B/2 1 64 52.6 >90 + Withstand

591 B/2 1 66 54.3 >90 + Withstand

592 B/2 1 68 54.4 >90 + Withstand

593 B/2 1 70 55.1 >90 + Withstand

594 B/2 1 72 57.6 >90 + Withstand

595 B/2 1 74 58.9 >90 + Withstand

596 B/2 1 76 61.3 >90 + Withstand

597 B/2 1 78 62.8 >90 + Withstand



599 B/2 1 60 48.9 >90 + Withstand

600 B/2 1 62 49.4 >90 + Withstand

601 B/2 1 64 51.5 >90 + Withstand

602 B/2 1 66 53 >90 + Withstand

603 B/2 1 68 54.9 >90 + Withstand

604 B/2 1 70 56.1 >90 + Withstand

605 B/2 1 72 57.6 >90 + Withstand

606 B/2 1 74 59.4 >90 + Withstand

607 B/2 1 76 60.9 >90 + Withstand

608 B/2 1 78 62.4 >90 + Withstand

609 B/2 1 80 60 7.1 + Puncture of CSST Jacket, 0.4 kA


610 B/3 1 60 48.3 >90 + Withstand

611 B/3 1 62 50.4 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

13 September 2012


612 B/3 1 64 51.4 >90 + Withstand



614 B/3 1 60 49 >90 + Withstand

615 B/3 1 62 50.2 >90 + Withstand

616 B/3 1 64 51.7 >90 + Withstand

618 B/3 1 66 55.5 >90 + Withstand

619 B/3 1 68 57.3 >90 + Withstand

620 B/3 1 70 58.1 >90 + Withstand

621 B/3 1 72 60.2 >90 + Withstand

622 B/3 1 74 60.9 >90 + Withstand

623 B/3 1 78 62.5 >90 + Withstand

624 B/3 1 80 64.1 >90 + Withstand

625 B/3 1 82 66.2 2 + Puncture of CSST Jacket, 0.67 kA


626 B/4 1 60 48.2 >90 + Withstand

627 B/4 1 62 49.3 >90 + Withstand

628 B/4 1 64 51.5 >90 + Withstand

629 B/4 1 66 54.9 >90 + Withstand

630 B/4 1 68 56.3 >90 + Withstand

631 B/4 1 70 57.6 >90 + Withstand

632 B/4 1 72 59.9 >90 + Withstand

633 B/4 1 74 60.2 >90 + Withstand

634 B/4 1 76 61.1 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

13 September 2012


635 B/4 1 78 62.6 >90 + Withstand

636 B/4 1 79 64.3 >90 + Withstand

637 B/4 1 80 66 >90 + Withstand



639 B/4 1 60 48.3 >90 + Withstand

640 B/4 1 62 50.5 >90 + Withstand

641 B/4 1 64 50.9 >90 + Withstand

642 B/4 1 66 53.2 >90 + Withstand

643 B/4 1 68 54.7 >90 + Withstand

644 B/4 1 70 56.6 >90 + Withstand

645 B/4 1 72 57.9 >90 + Withstand

646 B/4 1 74 60.1 >90 + Withstand

647 B/4 1 76 61.1 >90 + Withstand

648 B/4 1 78 62.7 >90 + Withstand

649 B/4 1 80 63.7 >90 + Withstand

650 B/4 1 82 65.6 >90 + Withstand

651 B/4 1 84 67 >90 + Withstand

652 B/4 1 86 69.3 >90 + Withstand

653 B/4 1 88 70.3 >90 + Withstand

654 B/4 1 90 71.6 25.1 +

Puncture of CSST Jacket, 0.14 kA During

Impulse Decay; Breakdown at 52.9 kV


655 B/5 1 60 48.3 >90 + Withstand




Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

13 September 2012


656 B/5 1 62 49.4 >90 + Withstand

657 B/5 1 64 51 >90 + Withstand

658 B/5 1 66 52.5 >90 + Withstand

659 B/5 1 68 54.8 >90 + Withstand

660 B/5 1 70 56.6 >90 + Withstand

661 B/5 1 72 58.2 >90 + Withstand

662 B/5 1 74 59.3 >90 + Withstand

663 B/5 1 76 60.9 >90 + Withstand

664 B/5 1 78 61.8 >90 + Withstand

665 B/5 1 80 64.3 >90 + Withstand

666 B/5 1 82 65 >90 + Withstand

667 B/5 1 84 66.8 3 + Puncture of CSST Jacket, 0.71 kA


668 B/5 1 60 48.5 >90 + Withstand

669 B/5 1 62 49.9 >90 + Withstand

670 B/5 1 64 51 >90 + Withstand

671 B/5 1 66 52.7 >90 + Withstand

672 B/5 1 68 55 >90 + Withstand

673 B/5 1 70 57.7 >90 + Withstand

674 B/5 1 72 59.7 >90 + Withstand

675 B/5 1 74 61.4 >90 + Withstand

676 B/5 1 76 62.4 >90 + Withstand





Test No.


CSST Dia.

(Inches)


(kV)

Peak Voltage

(kV)

Time at Breakdown

(µs) Polarity

(+/-) Notes/Results

13 September 2012


678 B/6 1 60 48.3 >90 + Withstand

679 B/6 1 62 49.9 >90 + Withstand

680 B/6 1 64 51.3 >90 + Withstand

681 B/6 1 66 52.4 >90 + Withstand

682 B/6 1 68 54.4 >90 + Withstand

683 B/6 1 70 56.2 >90 + Withstand

684 B/6 1 72 57.5 >90 + Withstand

685 B/6 1 74 58.9 >90 + Withstand

686 B/6 1 76 61 >90 + Withstand

687 B/6 1 78 61.4 >90 + Withstand

688 B/6 1 80 63.6 >90 + Withstand

689 B/6 1 82 65 >90 + Withstand


6.5 High Current Damage Tolerance Test Setup

The high current damage tolerance tests on the CSST gas pipe samples were performed in accordance with the Kick-Off Meeting Minutes (Appendix A) amendments to the GTI Proposal Validation of Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage (Ref. 1). Two-meter long CSST samples of each manufacturer’s CSSTs were used in both 0.5” and 1” diameters.

The high current tests involved arcing a transient current of a specific standardized waveshape through a section of CSST and observing the resulting physical effects. A specific area of the CSST sidewall was subjected to the current discharge. The applied transient test currents were returned to the generator via conduction across a short piece of 1” wide nickel-coated copper braid clamped to the remote end of the CSST test sample. No parallel ground paths existed, which meant that the full charge of the applied current waveform was transferred via the arc to the



CSST sidewall. The areas of the metal pipe walls nearest the arc were subjected to the highest current densities and charge transfers and showed significant i²R Joule heating.

With this arrangement, the different CSST types/brands were evaluated for melting and puncture by applying different current waveforms at various peak current levels. The applied transient currents were the power utility standardized lightning current waveform of 8 µs x 20 µs (IEC 61000-4-5; Ref. 2) and the telecommunications industry 10 µs x 350 µs test impulse (ITU-T K.44, Series K; Ref. 4). Three different peak current levels, 1, 5 and 10 kA, were applied to each CSST sample. Three samples of each CSST size (0.5” and 1” diameters) were tested for each of four manufacturers (A-D), resulting in three high current tests per type at each current level. Prior to tests on the CSST samples, each transient generator was calibrated using a 0.5” diameter copper tube as a stand-in CSST. This way, the CSST to ground plane loop area and its mutual inductance could be accounted for in the calibrations of the current transient waveforms, independently of the effects of the CSSTs’ particular materials on these measurements.

Each CSST sample tested was 2-meters in length and positioned on dielectric blocks 9 cm above a large, flat copper ground plane. To ensure that the current entered the CSST metal pipe by an arc, the CSST insulating polyethylene jackets were pre-punctured in a consistent manner. This was necessary because neither of the two high current transient generators used could supply enough voltage to puncture the dielectric jackets and supply the needed peak currents. Therefore, the agreed method was to use the tip of a hot soldering iron to melt each CSST outer jacket just above the desired arc entry location. A sharp-ended rod electrode served as the output conductor attached to the test current generator’s output, and was positioned 0.25” above this pre-punctured area. A fine, 38 AWG wire was used to initiate the electrical breakdown between the electrode tip and the small exposed area of the CSST sidewall at the puncture point. When current was applied, this wire vaporized quickly and provided a “guided” ionized path to the area under test on the gas pipe, further increasing the voltage gradient between the electrode and the area of arc entry on the CSST sidewall.

Figure 17 shows a general diagram of the high current, damage tolerance test setups. Figures 18 and 19 show calibration and test setup photographs of the 8 µs x 20 µs high current generator, and Figure 20 shows a close-up of a typical output electrode to CSST pre-puncture arrangement used for both the 8 µs x 20 µs and 10 µs x 350 µs setups. Figure 21 shows calibration setup photographs of the 10 µs x 350 µs high current generator with a copper calibration pipe in place of the CSST test samples. Figures 22-24 show typical calibration oscillographs for the 8 µs x 20 µs current waveform. Figures 25-27 show typical current calibration oscillographs from the 10 µs x 350 µs current waveform.



Figure 17 – General Test Setup for the CSST High Current, Damage Tolerance Tests

Figure 18 – CSST High Current, Damage Tolerance Calibrations (8 µs x 20 µs Waveform)

CSST

Short, 1” wide Ni-Cu Braided

Strap

Soldered Ground Plane

bracket

Output Electrode

Initiating Wire

Ground Plane

Copper Pipe

Output Electrode

CT Probe

8µs x 20µs Current

Generator

Dielectric Supports



Figure 19 – CSST High Current, Damage Tolerance Test Setup (8 µs x 20 µs Waveform)

Figure 20 – Close-up of Electrode-Initiating Wire Arrangement

Electrode Tip

Initiating Wire

Pre-Puncture of Jacket

Output Electrode

Termination of CSST to Generator Current Return

CSST

CSST

Output Rod Electrode

38 AWG Initiating

Wire Pre-Puncture Area of Jacket



Figure 21 – CSST High Current, Damage Tolerance Calibrations (10 µs x 350 µs Waveform)

Figure 22 – 8 µs x 20 µs Waveform Calibration, Total Charge (10 kA, 0.22 Coulombs Shown)

10 µs x 350 µs Current

Generator

CT Probe

Copper Pipe

Output Electrode



Figure 23 – 8 µs x 20 µs Waveform Calibration, t10-90%

Figure 24 – Typical 8 µs x 20 µs Waveform Calibration, t50% Decay



Figure 25 – 10 µs x 350 µs Waveform Calibration, Total Charge (10 kA, 0.5 Coulombs Shown)

Figure 26 – 10 µs x 350 µs Waveform Calibration, t10-90%



Figure 27 – 10 µs x 350 µs Waveform Calibration, t050% Decay 6.6 High Current Damage Tolerance Test Results

For both the 0.5” and the 1” diameter CSST products from all four manufacturers,

the 8 µs x 20 µs current waveform applied at peak currents of 1 kA, 5 kA, and 10 kA produced no visible punctures of the CSST metal sidewalls. Damage from the 8 µs x 20 µs current waveforms to the CSST was observed as discolorations and localized melting and re-solidification of the sidewall at the arc entry locations. The average total charge delivered to each CSST sample for the 8 µs x 20 µs transient current waveform was approximately 0.026 Coulombs, 0.11 Coulombs and 0.23 Coulombs for the 1 kA, 5 kA and 10 kA peak currents, respectively. Only visual inspections were performed; none of the CSST samples were checked after testing to see if there existed any pin hole punctures in the metal sidewall at the arc entry areas. Figures 28-33 show typical high current, damage tolerance test results from the 8 µs x 20 µs at 1, 5 and 10 kA peak currents for the 0.5” and 1” diameter CSSTs.

The tests using the 10 µs x 350 µs current waveform applied at 1 kA peak

current also resulted in no punctures of the CSST metal sidewall for each of the samples. The total charge for the 10 µs x 350 µs waveform at 1 kA peak current was approximately 0.5 Coulombs. Visible damage to the CSST sidewall using this current level was similar to that observed for the 10 kA level, 8 µs x 20 µs waveform tests, with localized surface melting and re-solidification of the CSST sidewall at the arc entry locations.

Tests with the 10 µs x 350 µs waveform applied at peak currents of 5 kA and

10 kA consistently produced jacket punctures and burning of the CSST metal sidewall. This result occurred on all four manufacturers’ CSST products for both 0.5” and 1”



diameters. The CSST pipe size seemed not to affect the resulting diameters of the jacket melt-through holes. The sizes of the resulting holes in the CSST sidewalls were observed to be 0.19”-0.25" wide for both 5 kA and 10 kA level tests. At these peak currents, this waveform delivered approximately 2.5 Coulombs and 5.0 Coulombs total charge, respectively. Figures 34-39 show typical examples of the high current, damage tolerance test results from the 10 µs x 350 µs at 1 kA, 5 kA and 10 kA peak currents for the 0.5” and 1” diameter CSSTs. Table 9 summarizes all of the high current, damage tolerance test results performed on all four manufacturers’ 0.5” and 1” diameter gas pipe samples.

Nearly all of the steel sidewall burn-through holes occurred on the raised portions

of the pipe corrugations. This can be partly explained by the smaller radius of curvature of the raised sections producing more pronounced electric field intensities than those of the larger-radius, indented portions of the corrugations. The size of the pre-puncture of the dielectric jacket also had some influence on the resulting burn-through hole diameter of the CSST sidewall. It was observed that smaller pre-punctures of the CSST jacket created slightly larger burn-through hole diameters than larger pre-punctures of the jacket. This effect was only qualitatively observed, as neither the diameters of the burn-through holes nor the sizes of the pre-puncture holes in the jackets were recorded. A likely explanation for this effect is that with a smaller hole through the dielectric jacket, the potential gradient is “funneled” to a smaller footprint of metal on the sidewall, due to the difference in dielectric constants of the polyethylene jacket and the air above the punctured area (roughly 3:1). With the arc being more concentrated at a smaller surface area of the sidewall, localized higher current densities result in higher Joule heating of the pipe metal during current conduction. Similar effects have been observed with lightning tests performed on over-painted metals and structures for the aerospace industry. All recorded oscillographs for the High Current Damage Tolerance Tests are shown in Appendix C.

A sanity check of the skin depth of a 0.010” thick stainless steel conductor shows

that for the 10 µs x 350 µs waveform, the maximum skin depth would be approximately 0.17 inches, greater than the thickness of the conductor itself. The formulation used for the skin depth of a conductor is:

[meters]

This indicates that the entire cross sectional area of the CSST sidewall is indeed

involved in the transient current conduction of both the 10 µs x 350 µs and 8 µs x 20 µs waveforms. For the standardized lightning test current waveforms used in these tests, both have the approximate highest energies at low frequencies of around several kilohertz. For example, the 10 µs x 350 µs waveform is at 3.3% of its peak current at 10 kHz during the exponential rise of the current, with the 8 µs x 20 µs waveform at a similar percentage of peak at around 20 kHz.



Figure 28 – Typical 1” CSST Sidewall Damage from 8 µs x 20 µs Waveform, 1 kA/ 0.026 Coulombs





Figure 31 – Typical 0.5” CSST Sidewall Damage from 8 µs x 20 µs Waveform, 1 kA/ 0.026 Coulombs



















Table 9 – High Current, Damage Tolerance Test Summary

Test No.

CSST Mfr./

Sample No.

CSST Dia.

(inches)


(kV)

Peak Current

(kA)

Total Charge

(C)

Action Integral (kJ/Ω)

Polarity (+/-) Notes/Results

19 September 2012

Calibrations of 8 µs x20 µs Current Generator Using 2 m Long, ½” Dia. Copper Pipe as Calibration Fixture

691 Calibration, 5 kA Level 11.7 5.44 0.12 0.04 + 8 µs x20 µs, 5 kA;

tpk = 8.1 µs, t50% decay = 24.1 µs


tpk = 9 µs, t50% decay = 25.4 µs


tpk = 8.7 µs, t50% decay = 24.3 µs Arc Entry Tests on 2 m Sections of CSST; Insulating Jackets Pre-Punctured at Arc Entry

Point Using Soldering Iron; 38 AWG Wire Fed Through Jacket to CSST Outer Wall

694 A/6 0.5 23.5 10.5 0.23 1.83 + No Puncture of CSST Wall








(CSST Sample A/1-0.5” was a HV Test Article)


(CSST Sample A/2-0.5” was a HV Test Article)

703 A/6 1 2.64 1.03 0.03 0.019 + No Puncture of CSST Wall











Table 9 – High Current, Damage Tolerance Test Summary (Continued)

Test No.

CSST Mfr./

Sample No.

CSST Dia.

(inches)


(kV)

Ipeak (kA)

Total Charge

(C)



20 September 2012

712 C/8 1 22.8 10.5 0.23 1.78 + No Puncture of CSST Wall







(CSST Sample C/6-1” was a HV Test Article)



721 C/6 0.5 2.75 1.12 0.03 0.02 + No Puncture of CSST Wall



724 C/9 0.5 11.4 No Data + No Puncture of CSST Wall





729 C/14 0.5 22.8 10 0.22 1.7 + No Puncture of CSST Wall

730 B/6 0.5 22.8 9.9 0.22 1.66 + No Puncture of CSST Wall









Test No.

CSST Mfr./

Sample No.

CSST Dia.

(inches)


(kV)

Ipeak (kA)

Total Charge

(C)



20 September 2012




739 B/8 1 2.82 1.08 0.03 0.02 + No Puncture of CSST Wall









(CSST sample B/6-1” was a HV Test Article)

21 September 2012

748 D/6 1 23.2 10.6 0.23 1.85 + No Puncture of CSST Wall









757 D/6 0.5 2.71 1.05 0.03 0.02 + No Puncture of CSST Wall





Test No.

CSST Mfr./

Sample No.

CSST Dia.

(inches)


(kV)

Ipeak (kA)

Total Charge

(C)



21 September 2012








25 September 2012

Calibrations of 10 µs x350 µs Current Generator, Using 2 m Long, ½” Dia. Copper Pipe as Calibration Fixture

766 Calibration, 1 kA Level 3.6 1.1 0.53 0.3 – 10 µs x 350 µs, 1.1 kA;

tpk = 10.4 µs, t50% decay = 344 µs

767 Calibration, 5 kA Level 16.8 5.15 2.49 6.48 – 10 µs x 350 µs, 5.2 kA;

tpk = 10.3 µs, t50% decay = 342 µs

768 Calibration, 10 kA Level 34 10.3 5 26.1 – 10 µs x 350 µs, 10.3 kA;

tpk = 9.8 µs, t50% decay = 344 µs

26 September 2012

769 D/15 0.5 3.6 1.04 0.43 0.27 – No Puncture of CSST Wall

770 D/16 0.5 3.6 1.02 0.5 0.27 – No Puncture of CSST Wall

771 D/17 0.5 3.6 1 0.49 0.27 – No Puncture of CSST Wall

772 D/18 0.5 11.5 3.3 1.59 2.83 – Puncture of CSST Wall




(CSST sample D/2-0.5” was a HV Test Article)

776 D/5 0.5 17 4.98 2.37 6.34 – Puncture of CSST Wall

(CSST sample D/5-0.5” was a HV Test Article)






Test No.

CSST Mfr./

Sample No.

CSST Dia.

(inches)


(kV)

Ipeak (kA)

Total Charge

(C)



26 September 2012

779 D/23 0.5 34.5 10 4.87 26.3 – Puncture of CSST Wall

780 D/15 1 3.6 1.04 0.49 0.28 – No Puncture of CSST Wall



783 D/18 1 16.9 5.02 2.37 6.38 – Puncture of CSST Wall

784 D/19 1 17 5.02 2.37 6.42 – Puncture of CSST Wall

785 D/20 1 17 5 2.37 6.39 – Puncture of CSST Wall


787 D/22 1 34.5 10.3 4.89 27 – Puncture of CSST Wall


789 B/15 0.5 3.6 1.04 0.51 0.28 – No Puncture of CSST Wall



792 B/18 0.5 17 4.94 2.38 6.26 – Puncture of CSST Wall

793 B/19 0.5 17.1 5 2.44 6.6 – Puncture of CSST Wall

794 B/20 0.5 17.4 5.02 2.45 6.67 – Puncture of CSST Wall



797 B/23 0.5 35 10.2 5 27.5 – Puncture of CSST Wall

798 B/16 1 35 10.5 4.98 28.2 – Puncture of CSST Wall

27 September 2012

799 B/17 1 34.9 10.4 4.97 27.9 – Puncture of CSST Wall







Test No.

CSST Mfr./

Sample No.

CSST Dia.

(inches)


(kV)

Ipeak (kA)

Total Charge

(C)



27 September 2012


804 B/22 1 3.6 1.02 0.49 0.28 – No Puncture of CSST Wall



807 C/15 0.5 3.6 1.01 0.5 0.27 – No Puncture of CSST Wall



810 C/18 0.5 17.1 5 2.4 6.42 – Puncture of CSST Wall

811 C/19 0.5 17.2 5.03 2.42 6.53 – Puncture of CSST Wall

812 C/20 0.5 17.2 5 2.42 6.55 – Puncture of CSST Wall



815 C/23 0.5 34.7 10.2 4.96 27.1 – No Puncture of CSST Wall; Possible Re-Solidification of

Insulating Jacket

816 C/16 1 34.8 10.3 4.98 27.7 – Puncture of CSST Wall






822 C/22 1 3.6 1.05 0.5 0.28 – No Puncture of CSST Wall



825 A/13 0.5 3.65 1.03 0.5 0.28 – No Puncture of CSST Wall

826 A/14 0.5 3.65 1.01 0.5 0.27 – No Puncture of CSST Wall




Test No.

CSST Mfr./

Sample No.

CSST Dia.

(inches)


(kV)

Ipeak (kA)

Total Charge

(C)



27 September 2012

827 A/15 0.5 3.65 1 0.5 0.27 – No Puncture of CSST Wall

828 A/16 0.5 17.2 4.99 2.41 6.48 – Puncture of CSST Wall



831 A/19 0.5 35 10.3 4.98 27.4 – Puncture of CSST Wall



834 A/15 1 34.8 10.3 4.97 27.6 – Puncture of CSST Wall


28 September 2012

836 A/17 1 34.7 10.3 5 27.7 – Puncture of CSST Wall




840 A/21 1 3.69 1.08 0.52 0.3 – No Puncture of CSST Wall



7.0 CSST HIGH CURRENT MODEL VALIDATION TESTING The final series of lightning tests on the CSST gas piping in support of GTI

Proposal Validation of Installation Methods for CSST Gas Piping to Mitigate Lightning Related Damage (Ref. 1) were empirical validation tests for a circuit model of different CSST installations using various bonding conductor lengths. These tests were performed with a single manufacturer’s CSST product (both 0.5” and 1” diameters) of a fixed length, connecting different lengths of 6 AWG bond wire in parallel, and measuring the resulting current divisions and waveforms on both the CSST and the bond conductor. These tests were witnessed by Dr. M. Stringfellow of PowerCET and C. Ziolkowski of GTI.



7.1 High Current Model Validation Test Setup The model validation tests used the 10 µs x 350 µs standardized current test

waveform, applied to a common electrical node with two parallel conductive paths back to ground: the CSST and a 6 AWG bonding conductor. The common “input” node was a black iron pipe (BIP) manifold, electrically isolated from the ground plane, fitted with reducers/adapters appropriate for the end fittings of the CSST test samples, and a standard bronze bonding clamp for the parallel 6 AWG bond conductor. The only current return paths from the transient generator back to the common ground was via the CSST and the bonding conductor. Three current transformer (CT) probes were used to measure the total applied currents and the division of current between the CSST and the bond wire.

The test setup was arranged so that the CSST test sample and the bond

conductor were routed horizontally over a wooden table, parallel to each other, and separated by a distance of 30 cm. At the far end of the table was a copper ground plane connected to the negative side of the 10 µs x 350 µs transient current generator’s capacitor bank. The bonding conductor was terminated with a standard crimp-on type terminal ring bolted to the ground plane. The remote end of the CSST test sample was bonded to this ground plane either directly via a short braided strap clamped around the end fitting, or oriented such that a sharp-ended rod electrode with an initiating wire could create an arc-over return path at a pre-punctured location in the CSST’s dielectric jacket, as detailed in Section 6.5. Electrical bonding resistances were measured using a 4-wire bond meter; none of the interconnecting joints in the test setup exceeded 1 mΩ. Table 10 lists all relevant electrical bonding resistances measured before model validation testing.

Prior to testing, Dr. Stringfellow measured the self inductance of a 4.5-meter

section of 0.5” diameter CSST using his own handheld RLC meter. The measurement was 9.9 µH (at 10 kHz), or 2.2 µH/meter, very close to the per unit length inductance values reported in Section 6.2. Similar measurements of the test setup with a 1” diameter copper pipe in place of the 4.5 m CSST test article measured 8.9 µH for the total inductance of the pipe, ground plane and generator elements. Inductance measurements of the 6 AWG bonding conductor were 1.75 µH/m (at 10 kHz), which compares well to handbook values of 1.2 µH/m for 6 AWG stranded conductors.

Figure 40 shows a plan view diagram of the general test setup. Figures 41-43

show various photographs of the model validation test setup. Figures 44-45 show sample 10 µs x 350 µs transient current generator calibration oscillographs using the stand-in 4.5-meter copper pipe.



Table 10 – Electrical Bonding Resistances of Model Validation Test Setup

Electrical Bond Location Resistance (mΩ)

CSST Reducer/Adapter to BIP Manifold 0.05

Transient Generator Output to BIP Manifold 0.18

Bonding Conductor Bronzed Clamp to BIP Manifold 0.19

Bonding Conductor Termination to Remote Ground Plane 0.04

Remote Ground Plane to Negative Side of Generator Capacitance 0.42

Figure 40 – CSST High Current Model Validation Test Setup, Plan View

Common Ground Plane

CSST remote end ground reference

Insulated Table

Bond Conductor ground reference

6 AWG Bond Wire

CSST BIP Manifold



Figure 41 – CSST High Current Model Validation Calibration Setup

Figure 42 – CSST High Current Model Validation Test Setup

Copper Calibration

Pipe

10µs x 350µs High Current

Generator

Return Ground Plane

4.5m CSST Section

10µs x 350µs high current generator

Return Ground Plane

6 AWG Bonding

Conductor

CT Probe



Figure 43 – Transient Generator Output Node at BIP Manifold (High Current Model Validation Tests)

CSST

10µs x 350µs high current generator

Output

BIP Manifold

6 AWG Bonding

Conductor

CT Probe

Bronze Bond Clamp

CSST



Figure 44 – Typical 10 µs x 350 µs Waveform Calibration, t50% Decay

Figure 45 – Typical 10 µs x 350 µs Waveform Calibration, t10-90%



7.2 High Current Model Validation Test Results

Dr. Stringfellow’s lumped parameter, circuit models compared well to the recorded empirical data resulting from the high-current model validation tests. Several CSST/bond conductor scenarios were tested, which are listed in Table 11. The resulting waveforms were recorded on a digital storage oscilloscope, and the resulting oscillographs are reported in Appendix D. Bond conductors of 1, 3, 6, 10, and 16-meter lengths were connected in parallel with either 4.5 m of the 0.5” diameter CSST or 4.5 m of the 1” diameter CSST. The tests were performed at various peak currents for the applied 10 µs x 350 µs waveform, using levels of 3.1 kA, 8.0 kA, 10.7 kA, and 12 kA.

Resulting current divisions between the bonding conductor and the CSST were

comparable with both the 0.5” and 1” diameter CSSTs, which is reasonable considering both sizes have similar per unit length inductances. In general, with a bonding conductor in parallel with the CSST test sample, the applied 10 µs x 350 µs waveform resulted in longer duration current waveforms in the bond conductor and shorter duration waveforms in the CSST, with a corresponding lower charge delivered to the CSST. As the length of the bonding conductor was increased, a higher proportion of current divided to the 4.5 m CSST section, with longer duration waveforms, and higher amounts of electrical charge transferred to the CSST. No damage was observed for any tests when any of the above listed lengths of bonding conductor were connected in parallel with a 4.5-meter section of CSST. However, these results are misleading because not all of the above tests were performed in the same fashion, with differences in both the magnitude of the applied currents as well as the method of current return from of the CSST (arcing versus conduction).

One of these differences is with the termination method of the CSST metal pipe. For all but six of the CSST/bond conductor tests performed, the remote ends of the 4.5-meter long CSSTs were directly grounded to the return of the transient generator using a short braided strap. Only Test Nos. 7-9 and 37-40 used the pre-punctured jacket in conjunction with the sharp-ended rod electrode and initiating wire arrangement to develop an arc at the return side of the CSST. Test No. 3 was performed on a 4.5 m section of CSST without any parallel bonding conductor, using the arc entry/exit method, with the result that a hole was burned through the CSST sidewall.

Tests where the current passing through the CSST enters or exits with an arc are much more severe than tests where the current conducts end to end through the CSST. This is because of the large difference in current densities between the small surface area of the CSST sidewall at a pre-punctured area through which an arc would be concentrated, versus the entire CSST’s cross-sectional area conducting the transient currents in the end-to-end tests. The intent of this series of high current tests was not to ascertain what length of parallel bonding conductor prevents a CSST wall burn-through, but to collect empirical data with which to compare results from PowerCET’s analytical circuit models. These models are the tools for determining CSST bonding provisions for various installation scenarios. It is important to understand that the high current test setup utilized for the above described tests is not meant to be representative of typical CSST installations, and that the high current transient generator provides only one aspect of the lightning environment (the current impulse) without producing the levels of



voltages that could be expected with a lightning strike event. Similar considerations should be exercised with regards to determining the maximum length of an effective bonding conductor for various CSST installations. The inductance of a bonding conductor produces a voltage transient which is dependent on the rate of change of the lightning transient current, and this voltage must be withstood by both the bonding conductor’s insulation and the insulation of the CSST for its associated L·di/dt voltages. This insulation withstand level is dependent on many factors such as the mechanical condition of the insulation, the proximity to nearby conductors, etc., in addition to the CSST gas pipe and bond conductors’ respective lengths. These factors could result in less than the ideal withstand voltage levels of the CSST jacket dielectrics.

Figure 46 shows a typical high current transient 10 µs x 350 µs waveform applied to the CSST/bond conductor BIP manifold node for the model validation tests. Figure 47 shows the recorded waveforms from the model validation tests for the 4.5 m long CSST (1” diameter) in parallel with a 1.0-meter long, 6 AWG bond strap (Test No. 16). Figure 48 shows the recorded waveforms from the model validation tests for the 4.5 m long CSST (1” diameter) in parallel with a 3.0-meter long, 6 AWG bond strap (Test No. 18). Figure 49 shows the recorded waveforms from the model validation tests for the 4.5 m long CSST (1” diameter) in parallel with a 6.0-meter long, 6 AWG bond strap (Test No. 14). All three of these tests had a peak applied 10 µs x 350 µs current of 10.2 kA. Figure 50 shows the recorded waveforms from the model validation tests for the 4.5 m long CSST (1” diameter) in parallel with a 10.0-meter long, 6 AWG bond strap (Test No. 38). Figure 51 shows the recorded waveforms from the model validation tests for the 4.5 m long CSST (1” diameter) in parallel with a 16.0-meter long, 6 AWG bond strap (Test No. 39). Note that this last Figure does not show the applied 10 µs x 350 µs current, however the charge voltage of the transient generator was set to produce 10.7 kA peak, as was measured from Test No. 38.



Figure 46 – Typical Applied 10 µs x 350 µs Current Impulse, 10.2 kA peak (Test Nos. 11-18)

Figure 47 – CSST and Bond Wire Currents, 4.5 m, 1” CSST in Parallel with 1.0 m, 6 AWG Bonding Conductor (Test No. 16)

CSST Current Bond Wire Current

Applied Total Current



Figure 48 – Transient Currents in 4.5 m, 1” CSST in Parallel with 1.0 m, 6 AWG Bonding Conductor, 10.2 kA peak (Test No. 18)

Figure 49 – Transient Currents in 4.5 m, 1” CSST in Parallel with 6.0 m, 6 AWG Bonding Conductor, 10.2 kA peak (Test No. 14)

Bond Wire current

CSST current


Bond Wire current

CSST current




Figure 50 – CSST and Bond Wire Currents, 4.5 m, 1” CSST in Parallel with 10.0 m, 6 AWG Bonding Conductor, 10.2 kA peak (Test No. 38)

Figure 51 – CSST and Bond Wire Currents, 4.5 m, 1” CSST in Parallel with 16.0 m, 6 AWG Bonding Conductor, 10.2 kA peak (Test No. 39)

Bond Wire current

CSST current


Bond Wire current

CSST current



At the request of Dr. Stringfellow, additional high current damage tolerance tests were performed on the 4.5 m sections of CSST samples, using the arc entry method with a pre-punctured location in the insulating jacket. Both the 0.5” and 1” diameter CSST were tested (with no parallel bonding conductors), using the 10 µs x 350 µs current waveform. Progressively lower current impulses were applied to find the minimum charge transfer which resulted in sidewall burn-through. A range of 1.5 kA to 600 A of peak current was applied to the different CSST samples, with total charge delivered ranging from 0.77 Coulombs to 0.33 Coulombs.

Initially, a test performed at 950 A peak current with 0.476 Coulombs delivered did not result in a puncture of a 1” diameter CSST. Tests at 865 A peak current with 0.507 Coulombs of charge resulted in burn-through of the 1” diameter CSST sidewalls, with holes approximately 0.031” in diameter. Similar tests on the 0.5” diameter CSSTs had burn-through holes in the sidewall of the metal pipes when a peak current of 600 A/ 0.327 C was delivered. This type of damage correlates with similar observations made during the high current, damage tolerance tests described in Section 6.6 of this report, except at lower applied currents. For example, the 10 µs x 350 µs high current, damage tolerance tests performed at 1 kA/ 0.5 C did not produce any burn-through holes in the CSST sidewalls, for either the 0.5” size or the 1” diameter samples. The highest 8 µs x 20 µs current damage tolerance tests performed at 10 kA/ 0.23 C delivered did not result in any punctures of the CSST metal sidewalls.

The results of these additional tests suggest that the maximum electrical charge transfer from the CSST stainless steel pipe without resulting in wall burn-through is between 0.33 and 0.23 Coulombs, with nonlinear factors other than the driving current influencing this maximum. To help quantify this level of electric charge, two cycles of 115 Vrms, 60 Hz line voltage at 15 Arms would provide a source charge of 0.45 Coulombs.

Some of the CSST test articles with a bond conductor in parallel had no visible evidence of burn-through in the CSST pipe sidewall, even though the total charge delivered through an arc was higher than 0.23 Coulombs. Specific examples of this are Test No. 38, with a 10 m bond wire, which had a resulting 5.5 µs x 57 µs current waveform in the CSST with 0.516 Coulombs of charge transferred; Test No. 39 (16 m bond wire), with a resulting 5.7 µs x 75 µs current waveform in the CSST with 0.815 Coulombs of charged transferred; and Test No. 8, with a 6 m bond wire in parallel, resulting in a 6.4 µs x 30 µs current waveform in the CSST with 0.195 Coulombs of charged transferred.

A typical arc entry test arrangement using a sharp-ended rod electrode with a 38 AWG initiating wire centered over a pre-puncture in the CSST jacket is shown in Figure 52. This setup is the same as that used for the earlier high current, damage tolerance tests. Figure 53 shows test damage to the 1” diameter CSST sidewall at 947 A/ 0.476 C delivered, with no pipe sidewall punctures. Figure 54 shows the typical damage to a 1” diameter CSST when 864 A/ 0.507 C was delivered, resulting in burn-through in the pipe sidewall. This test was repeated twice with the same results (Test Nos. 28-29). Figure 55 shows the results of 598 A/ 0.327 C applied to a 4.5 m long, 0.5” diameter CSST, with a hole melted in the pipe sidewall. Figure 56 shows the



results of 8.5 kA/ 0.815 C applied to a 4.5 m long, 1” diameter CSST, with a 16 m long bond strap in parallel.

The results of these additional high current, damage tolerance tests appear in Table 12. Appendix D contains all of the relevant oscillographs of the applied 10 µs x 350 µs test currents and the resulting divisions in the CSST and parallel bond straps.

Figure 52 – Typical Arc Entry to CSST Test Setup

Figure 53 – 1” CSST Arc Entry Damage, 947 A/ 0.467 C (Test No. 19)

CSST Section under Test

Rod Electrode

Initiating Wire

Pre-Punctured Jacket







Figure 56 – 1” CSST Arc Entry Damage, in Parallel with 16 m Bond Conductor, 8.3 kA/ 0.815C (Test No. 39)



Table 11 – Model Validation Test Summary

Test No.


(kV)

Applied Current

(kA)

Applied t10-90%/

t50% decay (µs)

CSST Peak

Current (kA)

CSST t10-90%/

t50% decay (µs)

Bond Wire Peak

Current (kA)

Bond Wire t10-90%/

t50% decay (µs)

CSST Length

(m)

Bond Strap

Length (m)

Notes/Results

4 December 2012

Calibrations of 10 µs x350 µs Current Generator, Using 4.25 m Long, 3.8 cm Dia. Cu Pipe as Calibration Fixture

1 12.5 3.5 6.4/350 -- -- -- -- -- -- Calibrations on 4.25 m Copper Pipe

Switching to a Higher, 10 mA HV Power Supply. Recalibrating 10 µs x350 µs Current Generator

2 30 8.77 6.65/351 -- -- -- -- -- -- Calibrations on 4.25 m Copper Pipe

Installing 0.5” Dia. CSST, 4.5 m Length. Conducting Transient Current at One End and Arcing at Grounded Electrode at Opposite End

3 30 7.99 7.29/384 30 7.99 7.29/384 -- 4.5 -- Puncture of 0.5” Dia. CSST;

Approx. 1/8” Dia. Hole; Time to Current Peak of 15.7 µs

5 December 2012

Installing Only 6 AWG Bond Wire, 6 m Length. Conducting Transient Current at One End with Current Return at Grounded Opposite End

4 36.0 12.0 10.2/310 -- -- 11.38 10.2/310 -- 6.0 Discrepancy Between CT1

(Applied) and CT3 (Bond Wire); Swapping CT1 for CT3

5 36.0 11.3 10.2/310 -- -- 12.02 10.2/310 -- 6.0 Repeat of Test No. 4 with CT3 at Applied, CT1 at Bond Wire

Adjustments Made to 10 µs x350 µs Current Generator to Enable Higher Wattage Outputs. Re-Checking Applied Waveform on 6 m, 6 AWG Bond Conductor

6 36 10.3 8.3/345 -- -- 10.9 8.3/345 -- 6 Verification of Generator Output



Table 11 – Model Validation Test Summary (Continued)

Test No.


(kV)

Applied Current

(kA)

Applied t10-90%/

t50% decay (µs)

CSST Peak

Current (kA)

CSST t10-90%/

t50% decay (µs)

Bond Wire Peak

Current (kA)

Bond Wire t10-90%/

t50% decay (µs)

CSST Length

(m)

Bond Strap

Length (m)

Notes/Results

5 December 2012

At request of Dr. Stringfellow, “peak” measurements will be recorded instead of “t10-90%” measurements for the applied waveform. Installing New 0.5” Dia. CSST Run. Conducting Transient Current at One End of CSST and Bond Wire (Same Node).

Current Return Via Both Arc-Over at CSST to Grounded Electrode and at Opposite End of Bond Strap (Separate Nodes)

7 36 10.8 6.44/342 Not Recorded 9.56 34.9/405 4.5 6 No Puncture of 0.5” Dia. CSST;

tpeak Applied = 18.65 µs, tpeak Bond = 60 µs

Now Recording Times to Peak in Lieu of Front Times of the Current Measurements

8 36 10.7 18.1/345 4.72 6.35/30.4 8.84 67.3/414 4.5 6 No Puncture of 0.5” Dia. CSST;

Applied Charge = 4.5 C, CSST Charge = 0.195 C

9 36 No Data Recorded 4.5 1 No Puncture of 0.5” Dia. CSST; Arc Over at Generator Output to

Measurement CT Current Transformer (CT) 1 and Channel 1 of Oscilloscope Damaged From Test No. 9, Channels 2-4 OK. New Configuration:

Channel 2, Applied: New CT Model 3525 with A10 Attenuator; Channel 3, CSST: CT Model 4160; Channel 4, Bond: CT Model 4160

10 10.3 3.1 19.9/337 0.41 4.3/11.8 2.82 27/339 4.5 1 No Puncture of 0.5” Dia. CSST; Applied Current Appears to Be

Saturating CT

Attaching the remote end of CSST directly to current return ground plane (no longer arcing over at a pre-punctured location)

11 10.2 3.07 15.7/343 0.39 5.2/22.8 2.69 27/369 4.5 1 No Puncture of 0.5” Dia. CSST; Applied Current Appears to Be

Saturating CT1

12 10.2 3.03 17.9/351 1.31 6.8/31.6 2.35 67.1/413 4.5 6 No Puncture of 0.5” Dia. CSST; Applied Current Appears to Be

Saturating CT1




Test No.


(kV)

Applied Current

(kA)

Applied t10-90%/

t50% decay (µs)

CSST Peak

Current (kA)

CSST t10-90%/

t50% decay (µs)

Bond Wire Peak

Current (kA)

Bond Wire t10-90%/

t50% decay (µs)

CSST Length

(m)

Bond Strap

Length (m)

Notes/Results

5 December 2012

CT1, Model 3525, has Reached its Ampere-Second Limit (0.5 As) and Is Saturating; Change to a 1” Dia. CSST Sample

13 10.2 3.04 15.1/334 1.51 7.1/41.3 2.22 101.3/451 4.5 6 No Puncture of 1” Dia. CSST

14 10.2 3.03 14.5/350 1.5 7.6/41.2 2.26 103.3/476 4.5 6 Repeat of Test No. 13; No Puncture of 1” Dia. CSST

15 10.2 3.1 10.3/356 0.4 9.5/34.1 2.54 27/368 4.5 1 No Puncture of 1” Dia. CSST


17 10.2 3.1 13.1/347 1.08 6.5/32.9 2.4 69.5/425 4.5 3 No Puncture of 1” Dia. CSST


6 December 2012

Installing 4.5 m Long CSST (0.5” dia.) In Parallel with Bond Wire. Conducting Transient Current at One End with Current Return at Grounded Opposite End

37 37 10.7 15.6/368 7.07 8.3/42.2 8.22 122/459 4.5 10

No Puncture of CSST; Total Charge = 4.66 C

CSST Path Charge = 0.38 C CT Probe on CSST at Amp-Sec Saturation

Installing 4.5 m Long CSST (1” dia.) In Parallel with Bond Wire. Conducting Transient Current at One End with Current Return at Grounded Opposite End

38 37 10.7 16.3/363 7.49 9/57 7.34 155/510 4.5 10


CSST path Charge = 0.52 C CT Probe on CSST at Amp-Sec Saturation




Test No.


(kV)

Applied Current

(kA)

Applied t10-90%/

t50% decay (µs)

CSST Peak

Current (kA)

CSST t10-90%/

t50% decay (µs)

Bond Wire Peak

Current (kA)

Bond Wire t10-90%/

t50% decay (µs)

CSST Length

(m)

Bond Strap

Length (m)

Notes/Results

6 December 2012

39 37 10.7 Not Recorded 8.3 11/75 6.53 181/567 4.5 10


CSST Path Charge = 0.82 C

40 37 10.7 Not Recorded 7.71 9.2/54.9 7.48 161/510 4.5 16


CSST Path Charge = 0.56 C



Table 12 – Additional Damage Tolerance Tests on CSST Wall Burn-through vs. Total Charge Delivered

Test No.

CSST Dia.

(inches)


(kV)

Peak Current

(kA)

Total Charge

(C) Notes/Results

6 December 2012

Using Transient Waveform with Soldering Iron Pre-Puncture at Arc Entry Location

19

1

3.4 0.95 0.48 No Puncture of CSST Wall

20 5.4 1.5 0.76 Puncture of CSST Wall, Approx. 0.063” Dia. Hole










30

0.5

3.4 0.86 0.52 Puncture of CSST Wall, Approx. 0.031” Dia. Hole



33 3 0.77 0.42 Puncture of CSST Wall, Approx. 0.031” Dia. Hole

34 3 0.77 0.42 Puncture of CSST Wall, Approx. 0.031” Dia. Hole




OTD & GTI CONFIDENTIAL1

Simulation Scenario 1

U10kA

10 350

CSST 30m

#14 AWG15m (50’) 10x350µs

#6 CSST Bond max 30m (100’)

Gas Meter

Electricity Meter

(100’)Grounded Equipment

Ground Rod 25 Ohms

(50’)

Scenario 1 Simulation Model

• Conductors modeled by lumped resistance and

L12uH

L22uH

R6.05

L62uH

R7.05

L72uH

+ --CSSTIns

Romex2mR.05

Romex2mL2uH

CSST2mL2uH

R16.05

L162uH

+ V c

CSSTSheathRon: 0.01Roff: 10000

R17.132

R2.132

R1.132

CSST2mR.132

R31L262 H

R32lumped resistance and inductance

• Flashover modeled by voltage‐controlled switch

• 10kA 10x350µsTitle GTI Simulation

Number Scenario 1

PowerCET Corpo

Learth1uH

G

EarthCurr

Rearth25

Bond2mR.003

Bond2mL2uH

R3.003

L32uH

R4.003

L42uH

L82uH

R9.003

L92uH

R10.003

R12.003

L122uH

R13.003

L132uH

R14.003

L142uH

R15.003

V c

-- V c

Roff: 10000Vthres: 25000Vhyst: 25000Init Cond: OFF

L242uH

L252uH

R5.003

L52uH

R8.003

L112uH

R11.003

L102uH

L152uH

R25.1

R26100K

C11uF

L182uH

R38.003

L322uH

R39.003

L332uH

R40.003

L342uH

R41.05

L352uH

R42.05

L362uH

R43.05

R44.05

L382uH

L372uH

+ V c

-- V c

CSSTSHeath2Ron: 0.01Roff: 10000Vthres: 25000Vhyst: 25000Init Cond: OFF

BondCurr

+--ArcQ

CSSTArcCurrR46.0001

R4710K

G

R19.132

Lightning10x350DC: Amps undefinedAC: Amps 1.0 AC: Phase Tran: Exp 0 to 10.3e3Distort: Sine undefined

G

R29100000

L192uH

R18.132

L172uH

.1322uH.132

L272uH

R33.132

L282uH

R34.132

L292uH

R35.132

R24.164

L232uH

R23.132

L222uH

L302uH

R22.132

• 10kA, 10x350µs lightning current injected at gas meter

Date October 10 2012 Size ASheet 1 of 1 Rev 1File: MFS Scenario 1.Sch

G

Validation of CSST Installation Methods Appendix D


Scenario 1 Results ‐ No Bond

• No bonding conductor

Fl h t CSST

Transient - New, MFS Scenario 1.Sch + MFS Scenario 1.Anl, 10 October 2012

BondCurr (left) ArcQ (right)

EarthCurr (left) CSSTArcCurr (left)

11.0K 5.00

y 3.59839E+0x 6.95249E-6

• Flashover to CSST

• Arc current 10kA 10x350µs

• Arc charge 4.75C

• CSST rupture 4.00K

5.00K

6.00K

7.00K

8.00K

9.00K

10.0K

1.50

2.00

2.50

3.00

3.50

4.00

4.50

Time0 150u 300u 450u 600u 750u 900u 1.05m 1.20m 1.35m 1.50m

-1.00K

0

1.00K

2.00K

3.00K

-500m

0

500m

1.00

Scenario 1 Results – 30m Bond

• Bonding conductor length 30m (100 feet)




11.0K

110m

120m

y 5.87029E+3x 5.11628E-6

length 30m (100 feet)

• Flashover to CSST

• Arc current 3kA ~6x30µs

• Arc charge 0.15C 4.00K

5.00K

6.00K

7.00K

8.00K

9.00K

10.0K

40.0m

50.0m

60.0m

70.0m

80.0m

90.0m

100m

110m

• No CSST rupture

Time0 20.0u 40.0u 60.0u 80.0u 100u 120u 140u 160u 180u 200u

-1.00K

0

1.00K

2.00K

3.00K

-10.0m

0

10.0m

20.0m

30.0m



Scenario 1 Results– 8m BondTransient - New, MFS Scenario 1.Sch + MFS Scenario 1.Anl, 10 October 2012


EarthCurr (left) CSSTArcCurr (lef t)

10 0K

11.0K 18.0u

y 9.77139E+3x 2.59861E-5

• Bonding conductor length 8m (26 feet)

3.00K

4.00K

5.00K

6.00K

7.00K

8.00K

9.00K

10.0K

6.00u

8.00u

10.0u

12.0u

14.0u

16.0ulength 8m (26 feet)

• No flashover to CSST

Time0 20.0u 40.0u 60.0u 80.0u 100u 120u 140u 160u 180u 200u

-1.00K

0

1.00K

2.00K

-2.00u

0

2.00u

4.00u

Scenario 1 Results

Bond Length in

t

Flashover Voltage in

kV

CSST Arc Peak

C t

CSST Arc Waveform

i

CSST Arc Charge in C l b

CSST Rupture

meters kV Current in µs Coulombs

None >50 10 10x350 4.75 Yes

30 >50 3.27 6.5x30 0.12 No

24 >50 2.81 6.5x28 0.095 No

16 >50 2.1 6.5x28 0.065 No

12 >50 1.75 6x25 0.048 No

8 35 0 0 0 No8 35 0 0 0 No

4 17.5 0 0 0 No



Scenario 1 – Results Summary

• Bonding conductor up to 8m (26 feet)– Prevents flashover to CSST

• Bonding conductor from 8m ‐ 30m (100 feet) – Reduces flashover arc magnitude and duration

– Prevents CSST rupture

• No bonding conductor• No bonding conductor– Flashover arc high current, high energy

– CSST rupture results

Scenario 1 – Validation Tests

• Validation will use the 10x350 10kA waveform since lower energies do not cause damagesince lower energies do not cause damage

• One test with no bonding conductor• Bonding conductors of 8m, 16m, 30m, and 60m(?) of #6 AWG wire

• Perform tests with 30m of ½” and 1” diameter product

• Three shots per product; total 30 tests if 60m is included, 24 tests without



Scenario 2

10kA 10 350

ManifoldCSST 30m (100’)Manifold

Bond

10x350µs

Gas Meter

Electricity Meter

Ground Rod 25 Ohms

Grounded Object U


(100’)

Scenario 2 Results

Bond Length in

t


kV

CSST Arc Peak

C t

CSST Arc Waveform

i


CSST Rupture


None >50 6.9 10x350 3.5 Yes

30 >50 3.1 7.7x42 0.160 No

30* >50 4.2 7.2x39 0.204 No

24 48.3 0 0 0 No

24* >50 3.7 7x37 0.165 No

12 28 9 0 0 0 No12 28.9 0 0 0 No

8 20.6 0 0 0 No

4 11.1 0 0 0 No

* No Manifold Bond



Scenario 2 – Results Summary

• Bonding conductor up to 24m (79 feet)– Prevents flashover to CSSTPrevents flashover to CSST

• Bonding conductor from 26m ‐ 30m (100 feet) – Reduces flashover arc magnitude and duration– Prevents CSST rupture

• Manifold bond reduces arc currents & charge slightly

• No bonding conductor– Flashover arc high current, high energy– CSST rupture results

Scenario 2 – Validation Testing

• Perform test with 10x350 10kA waveform• Use 30m of ½” and 1” CSST with manifold at far endUse 30m of ½ and 1 CSST with manifold at far end• Test with 30m of #6 bond and 30m of #14 manifold ground wire – Both in place– Bond wire only– Manifold ground only– Neither in place

• Repeat these scenarios with 24m of #6 bond wire– Verify that manifold ground suppresses arc

• This is a total of 48 tests with 3 shots per sample



Scenario 3

U

10kA 10x350µs

#14 AWG15m (50’)

CSST 30m

Gas MeterElectricity Meter

Grounded Equipment

Ground Rod 25 Ohms Grounded Pipe 10 Ohms

(50’)


(100’)

Scenario 3 Results

Bond Length in

t


kV

CSST Arc Peak

C t

CSST Arc Waveform

i


CSST Rupture


None >50 6.7 10x350 3.2 Yes

30 >50 2.0 7.8x32 0.075 No

28 >50 1.8 7.8x32 0.066 No

26 47.7 0 0 0 No

12 27.1 0 0 0 No

8 19 5 0 0 0 No8 19.5 0 0 0 No

4 10.7 0 0 0 No



Scenario 3 Results Summary

• Bonding conductor up to 26m (85 feet)– Prevents flashover to CSST

• Bonding conductor from 28m ‐ 30m (100 feet) – Reduces flashover arc magnitude and duration

– Prevents CSST rupture

• No bonding conductor• No bonding conductor– Flashover arc high current, high energy

– CSST rupture results

Scenario 3 – Validation Test

• Perform tests with 10x350 10kA waveform• Need to determine what to use as “grounded• Need to determine what to use as “grounded equipment”

• Perform tests with ½” and 1” CSST• Perform tests with

– No bond wire30 f #6 b d i if b th h– 30m of #6 bond wire; verify no burn through

– 24m of #6 bond wire; verify no arcing

• This is a total of 18 tests with 3 shots per sample



Verify Inductance

• Verify the CSST self‐inductance by forming a l f k l th d di itloop of known length and suspending it remote from local ground

• Measure with precision RLC meter

• Repeat this for ½” and 1” CSST

Other Considerations

• Each test calls for 30m of CSST– Can this physically be accomplished at the LTI labs– Is it necessary to replace the sample between shots if there is

no arcing– Start with scenarios least likely to arc and work up

• This will require large quantities of CSST, connectors, and manifolds– Can manufacturers prep samples to save time

• Should we include 60m case of Scenario 1 given simulation gsuggests this would prevent burn through?

• Should we do repetitions (10 shots?) on the same sample at “sub‐critical” arcing to see if cumulative damage leads to burn through?


1

VALIDATION OF INSTALLATION METHODS FOR CSST GAS PIPING TO

MITIGATE LIGHTNING RELATED DAMAGE

LABORATORY TESTS TO VALIDATE SIMULATION DATA

M I C H A E L F . S T R I N G F E L L O W

C H I E F S C I E N T I S T P O W E R C E T C O R P O R A T I O N

Validation of CSST Installation Methods Appendix E

CSST Laboratory Model Validation Tests 1 April 2013

2

INTRODUCTION

The testing described in this report is part of the project originally proposed to the NFPA Research Council by SEFTIM to validate the effectiveness of electrical bonding for protection against the effects of indirect lightning strikes of corrugated stainless steel tubing (CSST) used for the delivery of fuel gas in buildings. The project is now being conducted under the auspices of the Gas Technology Institute.

The testing is being carried out to provide the following information:

1. To characterize the electrical self-inductance and resistance of CSST.

2. To determine the arc current and charge magnitudes necessary to damage CSST for various surge current waveforms.

3. To measure the current distribution in sample lengths of CSST in parallel with bonding conductors of various lengths.

4. To provide comparison between laboratory tests results and computer simulations in order to validate the use of simulations for predictions of bonding effectiveness.

The validation testing is required because of claims that CSST has anomalous electrical characteristics when conducting lightning surge currents and also that simulations may be in error because they do not account for these anomalies. These claims include the following:

1. CSST has unusually high inductance because of its corrugated structure.

2. CSST has unusually high inductance because of the magnetic permeability of the steel.

3. Flashover between adjacent corrugations can occur because of the surge impedance of the CSST material.

SIMULATIONS

As part of the original study, simulations were developed to calculate the distribution of surge current between the various paths described in a number of scenarios. These scenarios are based on those proposed by SEFTIM [1, 2], with minor modifications to account for the typical power system grounding practice in the USA.

For example, Scenario 1 envisaged the injection of indirect lightning current into the building via a metallic gas pipe, with current dividing between a connected CSST line of 30 m length and its bonding wire. The bonding wire between the CSST and the connection to the ground electrode at the electricity service entrance was considered to be a number 6 AWG wire of between 1 and 30 m long. Flashover between the CSST at its end was modeled by assuming a 50 kV breakdown voltage to grounded equipment, connected to the service entrance via 15 meters of #14 AWG insulated wiring.

The lightning current injected was a rather severe 10x350 µs wave with a peak current of 10 kA.

A sketch of this scenario is shown in Fig. 1.



3

Fig.1: Simulation 1 Schematic

Simulations of this scenario were run on a personal computer using a commercial version of SPICE (Simulation Program with Integrated Circuit Emphasis), originally developed by the University of California at Berkeley. The simulation was done to calculate the current distribution between the gas pipe, electrical power system and a bonding conductor connected between the incoming gas pipe at the meter and the electrical ground located at the panel.

Simulations were run with bonding conductors having lengths from 1 meter to 30 meters, using values of self-inductance calculated from standard formulae and resistance from direct measurement of CSST samples. The initial run of these simulations predicted that the current in the CSST in this scenario would have a much shorter waveform than the injected indirect lightning current waveform. It was also calculated that the electrical current and charge in the arc to the CSST was below the level where rupture would occur. This level was assumed to be a magnitude of approximately three times that required for longer duration waveforms (typically greater than 1 coulomb) because of the short duration of the calculated arc waveform,

For an injected current of 10kA with a 10x350 µs waveform and a bonding conductor of 30 m length, this simulation predicts that flashover occurs to the CSST and that the resultant arc has a peak current of just over 3kA with a waveform of approximately 6x30 µs. This results in a charge transfer of 0.15 coulombs through the arc to the CSST, considered insufficient to rupture it. The predicted waveforms for Simulation 1 with 30 meters of bonding conductor are shown in Fig. 2.

The simulations show that for shorter lengths of bonding conductor no flashover took place, while for longer bonding conductors flashover occurred but the current waveform was very short, resulting in insufficient charge transfer to damage the CSST.



4

.

Fig. 2: Predicted distribution of current from simulated Scenario 1 with 30 m bonding conductor

VALIDATION TESTING

Validation testing was proposed to be carried out in a high-voltage laboratory with a test set-up as close as possible to Scenario 1. It was considered impossible to conduct these tests with lengths of conductor similar to those used in Scenario 1 simulations (30 m of CSST, up to 30 m of bonding conductor, 15 m of power conductor and a 25 ohm service entrance ground) because of the constraints of the available surge generators. It was decided instead to use a simplified version of Scenario 1, eliminating the service entrance ground and limiting the length of CSST to 5 meters. The goal of these tests was to measure the waveforms in the CSST and bonding conductor and to compare these with the waveforms obtained from the computer simulation. A further goal was to measure the required current and charge necessary to breach the wall of CSST, especially to confirm the larger charges required for shorter duration currents.

SURGE GENERATOR SET-UP

The high-current surge generator for these tests was configured by Lightning Technologies Inc. in their laboratory facility in Pittsfield, Massachusetts. The generator consists of a bank of 29 capacitors, rated at 5 µF each at a maximum rated voltage of 50 kV, a non-inductive surge resistor of approximately 3 ohms and a high-current pneumatic switch. Capacitor charging was provided by a 50 kV 25 mA Hipotronics power supply. The test set-up also included copper ground planes for termination of the CSST test samples and the bonding wires as well as a custom electrode with fusible wire for establishing an arc to the CSST test samples.




Time0 20.0u 40.0u 60.0u 80.0u 100u 120u 140u 160u 180u 200u

-1.00K

0

1.00K

2.00K

3.00K

4.00K

5.00K

6.00K

7.00K

8.00K

9.00K

10.0K

11.0K

-10.0m

0

10.0m

20.0m

30.0m

40.0m

50.0m

60.0m

70.0m

80.0m

90.0m

100m

110m

120m

y 5.87029E+3x 5.11628E-6



5

The wave shaping was partially provided by the inductance of the test samples, which limited the length of CSST used to about 5 meters (16 feet). Current waveform measurements were made using wideband current transformers coupled to a four-channel digitizing oscilloscope (Fig. 3).

Fig. 3: Wideband current transformer around CSST

GENERATOR CALIBRATION

The surge generator was set up and calibrated using a 4 meter length of 1-1/2 inch copper pipe in place of the CSST. This provided a configuration with known characteristics. The resistance and total inductance in the generator loop and test sample was additionally measured by an Agilent LCR meter. This meter applies low amplitude ac signals at frequencies from 120 Hz up to 100 kHz to measure the impedance of circuit elements. The measured value of self-inductance was 8.38 µH and resistance 3.28 ohms. Calculated inductance of the copper pipe is about 3.93 µH, which suggests that the contribution to inductance from the rest of the generator circuit and test set-up is about 4.45 µH.

Test shots showed that this circuit provided surge currents with an amplitude of 8.8 kA for a charging voltage of 30 kV (Fig 4). The SPICE simulation for this test set-up with the measured values of resistance and inductance gave values within 1% of those measured (Fig. 5). The generated waveform had a risetime (1.25 times the 10% to 90% wavefront levels) of 6.5 µs and a wavetail (time to half peak magnitude) of 351 µs. The risetime was outside the 20% tolerance for the nominal 10x350 µs waveform, but since the test was concerned mainly with energy transfer, this was considered acceptable for the validation tests.

Close agreement between the measured and simulated results also confirmed that the inductance and resistance of the test circuit measured with the LCR meter at low signal levels are valid for surge conditions.



6

Fig. 4: Calibration waveform of surge generator 30kV charge (8.8 kA 6.5x351 µs)

Fig. 5: SPICE simulation of generator calibration set-up (8.8 kA 6.4x352 µs)

Transient - New, LTICalTest.Sch + UnTitled, 08 December 2012

TPi1 (left)

Time0 200u 400u 600u 800u 1.00m 1.20m 1.40m 1.60m 1.80m 2.00m

-1.00K

0

1.00K

2.00K

3.00K

4.00K

5.00K

6.00K

7.00K

8.00K

9.00K

y 4.42551E+3x 5.51268E-4



7

CSST TESTS

A number of tests were run with only lengths of CSST to verify the waveform as well as the inductance of the CSST. The ground termination of the CSST in these initial tests included a test electrode and arc-initiation wire that was placed about 1/8”away from the metallic tube through a small hole melted in the protective jacket (Fig. 6). A series of these tests was run with both high and low current magnitudes to determine the threshold or lowest current needed to penetrate the CSST. Where a hole was created by the arc, the insulating jacket was peeled back and the hole photographed.

Fig. 6: Arc-initiation surge tests with CSST

One example from Test #3 is shown in Fig. 7, in which an arc current of peak current 8.0 kA was applied, giving a discharge waveform of about 6.7x385 µs and resulting in a charge transfer through the arc of approximately 4.3 coulombs. The hole has dimensions of about 3 mm x 2 mm.



8

Fig. 7: Arc hole in CSST sample

CSST AND BONDING WIRE TESTS

The test set-up was then reconfigured to measure the current between the test sample of CSST and a parallel bonding conductor connected between the manifold and ground plane (Fig 8). CSST test lengths used were 4.44 m (175 inches) and bonding conductor lengths of 1, 3, 6, 10 and 16 meters. Inductance of the generator circuit including the bonding wire was made using the LCR meter. The largest value of inductance measured of 26.3 µH was obtained as expected with the 16 meter bonding wire.

Fig. 8: Test set-up with CSST and parallel bonding conductor

The tests confirmed that the waveform of current in the CSST when shunted by the lower resistance bonding wire had a shorter waveform than the test wave. The actual waveform varied, depending on the relative length of the bonding wire as well as the diameter of the CSST test sample. Measured waveforms for



9

the 10 m bonding wire configuration with 4.445 m of 1” diameter CSST (Test #38) are shown in Fig 9 and the simulation results for this in Fig 10.

Fig. 9: Measured current distribution between ½” CSST and 10 meter bonding wire (red is total current, blue is CSST current and green is bond wire current)

Fig. 10: Simulated current distribution between ½” CSST and 10m bonding wire

Transient - New, LTICalTest.Sch + LTICalTest.Anl, 17 December 2012

ITOT (left) ICSST (left)

IBOND (left) ArcQ (right)

Time0 50.0u 100u 150u 200u 250u 300u 350u 400u 450u 500u

-1.00K

0

1.00K

2.00K

3.00K

4.00K

5.00K

6.00K

7.00K

8.00K

9.00K

10.0K

11.0K

-50.0m

0

50.0m

100m

150m

200m

250m

300m

350m

400m

450m

500m

y 6.76317E-2x 6.16279E-5



10

TEST RESULTS

CURRENT AND CHARGE NECESSARY TO CREATE HOLE IN CSST

Original testing with a variety of CSST products from different manufacturers failed to produce any rupture or hole in a CSST test sample at 10x350 µs waves with peak magnitudes of 1kA or below. This is equivalent to a total transferred charge of about 0.5 coulombs. The first tests of the present series also failed to consistently produce holes at this current level. Later testing, however, was able to create holes with currents as low as 700 A peak (0.3 C) in ½ inch CSST and 900 A (0.4 C) in 1 inch CSST. In these latter tests, the arc was tightly controlled by the initiating wire and small hole in the CSST insulating jacket, which might explain this result.

The largest arc currents measured with a bonding wire present (16 m) were 7.2 kA and 7.7 kA in ½ inch and 1 inch samples of CSST respectively, both with a waveshape of 6x50 µs. The transferred charge in these two tests was measured at 0.56 C and 0.82 C respectively. In both cases, no hole was created in the CSST wall. These data suggest that the arc charge/time to rupture curve derived from previous testing on thin metals used for aircraft skins [4] (Fig 11) is likely applicable. When comparing the recent CSST test data with the earlier results from aircraft skins, likely limits for the arc current necessary to damage CSST as a function of arc duration can be estimated. These estimates are shown in Fig. 12, with the green curve showing a limit below which arc damage is unlikely, and the red line a limit above which damage is likely. Approximate damage threshold levels for CSST are 0.12 C to 0.25 C minimum (for arcs of 5 ms to 10 ms duration), 0.25 C to 0.5 C for 10x350 µs waves and greater than 3 C for the shorter waves (<100 µs) found on conductors in buildings in bonded scenarios.

Fig. 11: Time for arc puncture of aluminum aircraft skins (after Plumer & Robb, 1982)



11

Fig. 12: Estimated arc charge required to rupture CSST versus arc duration

SELF-INDUCTANCE AND RESISTANCE OF CSST

Measurements of the inductance of both CSST and copper conductors using low amplitude signal currents at low frequency (typically 10 kHz) using the Agilent LCR meter are consistent with those at higher currents derived from the surge generator. CSST appears to exhibit no unusual characteristics and its self-inductance may be estimated using standard formulae using the actual length of the conductor with effective diameters equivalent to the outside diameter of the material (3/4 inches for ½ inch CSST and 1-1/4 inches for 1 inch CSST). Detailed measurements are given in Appendix C. Similarly, the electrical resistance of CSST at high currents appears to be identical to its resistance measured using low frequency low-voltage instruments.

No flashover was observed between adjacent corrugations under any test conditions. Based on measured voltages developed along the CSST samples in laboratory testing and theoretical calculations, such flashover is considered highly unlikely.

SIMULATION ACCURACY

All simulations produced predicted waveforms in the various paths that were close to those measured in the laboratory. Simulated peak current magnitudes differed by between 1% and 4% of those measured. Predicted waveforms were within about 5% of those measured, with the largest discrepancy being on the wave tail. These minor discrepancies occurred in cases where it was difficult to measure the inductance of the circuit in the presence of parallel conductors. In all cases, the simulated arc waves resulted in calculated charge transfer within 10% of that measured in the laboratory tests. These results are well within experimental error and quite sufficient to validate the simulation models for their intended purpose. Details of some measured and simulated waveforms are given in Appendix A.



12

CONCLUSIONS

The laboratory tests successfully replicated a simplified version of the SEFTIM Scenario 1 with 10x350 µs surge currents of up to 10 kA magnitude injected into a parallel length of CSST and bonding conductor. Computer simulations of these tests closely modeled the results and confirmed that CSST has no unusual electrical characteristics. Tests were also able to characterize the likely charge necessary to damage CSST in the event of electrical arcs terminating on the metallic tubing.

The information gathered validated the use of computer simulations as well as providing better numerical data to refine the model.

Michael F. Stringfellow, Ph.D., P.E. Chief Scientist PowerCET Corporation



13

A P P E N D I X A MEASURED TEST WAVEFORMS

Fig, A1: Test 2 - 30 kV surge generator charge with 4 meter copper pipe

Fig A2: Test 3 - 30 kV generator charge with 4.45 meters of 1/2” CSST only



14

Fig A3: Test 4 - 36 kV generator charge 6 meters #6 wire only check of three CTs and scope channels

Fig A4: Test 7 – 36 kV generator charge with 4.5 meters CSST and 6 meter #6 parallel bonding wire



15

Fig. A5: Test 8 – 36 kV generator charge with 4.4 meters CSST and 6 meter # 6 parallel bonding wire

Fig A6: Test 11 – 10.2 kV generator charge with 4.4 meters ½” CSST and 1 meter #6 parallel bonding wire (Channel 2 CT saturation)



16

Fig A7: Test 12 - 10.2 kV generator charge with 4.5 meters 1” CSST and 6 meter # 6 parallel bonding wire

Fig A8: Test 14 – Repeat of test 13



17

Fig A9: Test 17 – 10.2 kV generator charge with 4.4 meters of 1” CSST and 3 meter #6 parallel bonding wire

Fig A10: Test 18 – Repeat of test 17



18

Fig A11: Test 37 – 37 kV generator charge with 4.4 meters ½” CSST with 10 meters of #6 parallel bonding wire

Fig A12: Test 38 – 37 kV generator charge with 4.4 meters 1” CSST with 10 meters of #6 parallel bonding wire



19

Fig A13: Test 39 – 37 kV generator charge with 4.4 meters of 1” CSST with 16 meters of #6 parallel bonding wire

Fig A14: Test 40 – 37 kV generator charge with 4.4 meters of ½” CSST with 16 meters of #6 parallel bonding wire


20

A P P E N D I X B COMPUTER MODELS AND SIMULATED WAVEFORMS

Fig B1: Surge generator model of calibration set-up with 1-1/2” copper pipe

Title LTICaltest1

Number

PowerCET Corp

Date December 17 2012 Size SSheet 1 of 1 Rev 2File: LTICalTest.Sch

C1145.7u

G

R20.1

G

R4100k

C21uF

+ --ArcQ

ICSSTR63.18

R110

G

G

+B130kV

R3.001

L13.93u

fxy

W

X

Y

Z

Vfunc2

+ Vc

-- Vc

Vsw2

+ Vc

-- Vc

Vsw3

+ Vc

-- Vc

Vsw1

ITOTL24.45uH

fxy

W X Y Z

Vfunc1



21

Fig B2: Simulated current waveform for calibration set-up

Transient - New, LTICalTest.Sch + LTICalTest.Anl, 20 December 2012

ITOT (left)

Time0 50.0u 100u 150u 200u 250u 300u 350u 400u 450u 500u

-1.00K

0

1.00K

2.00K

3.00K

4.00K

5.00K

6.00K

7.00K

8.00K

9.00K

y 8.89471E+3x 6.44983E-5



22

Fig B3: Model of surge generator with 4.45 meters of 1/2” CSST and with 1m #6 parallel bonding wire

Title 1/2" CSST 1m Bond

Number

PowerCET Corp

Date December 20 2012 Size SSheet 1 of 1 Rev 2File: LTIHalfInch1mBond.Sch

C1145.7u

G

R20.1

G

R4100k

C21uF

+ --ArcQ

ICSST

R110

G

G

R63.2

R3.203

IBOND

L15.43u

R71.5m

L31.1uH

+B110.2kV L2

4.45u

+ Vc

-- Vc

Vsw2

+ Vc

-- Vc

Vsw3

+ Vc

-- Vc

Vsw1

fxy

W X Y Z

Vfunc1

fxy

W

X

Y

Z

Vfunc2

ITOT



23

Fig

Fig B4: Simulated waveforms of current and arc charge for 1/2” CSST and 1 meter bonding conductor

Transient - New, LTIHalfInch1mBond.Sch + LTIHalfInch1mBond.Anl, 20 December 2012



Time0 20.0u 40.0u 60.0u 80.0u 100u 120u 140u 160u 180u 200u

-500

0

500

1.00K

1.50K

2.00K

2.50K

3.00K

3.50K

-1.00m

0

1.00m

2.00m

3.00m

4.00m

5.00m

6.00m

7.00m

8.00m

9.00m

10.0m

11.0m

12.0m

y 3.11237E+3x 9.02778E-6



24

Fig B5: Model of surge generator with 4.45 meters of 1” CSST and with 16m #6 parallel bonding wire

Title 1" CSST 16m Bond

Number

PowerCET Corp

Date December 20 2012 Size SSheet of Rev File: LTIOneinch10mbond.Sch

C1145.7u

G

R20.1

G

C21uF

+ --ArcQ

ICSST

R110

G

G

+B137kV

R3.080

L14.5u

ITOT

L24.45u

+ Vc

-- Vc

Vsw2

fxy

W X Y Z

Vfunc2

fxy

W X Y Z

Vfunc1

+ Vc

-- Vc

Vsw1

+ Vc

-- Vc

Vsw3

IBONDR5.025

L319uH

R63.18

R4100k



25

Fig B6: Simulated waveforms of current and arc charge for 1” CSST and 16 meter bonding conductor

Transient - New, LTIOneinch10mbond.Sch + LTIOneinch10mbond.Anl, 20 December 2012



Time0 50.0u 100u 150u 200u 250u 300u 350u 400u 450u 500u

-1.00K

0

1.00K

2.00K

3.00K

4.00K

5.00K

6.00K

7.00K

8.00K

9.00K

10.0K

11.0K

-100m

0

100m

200m

300m

400m

500m

600m

700m

800m

900m

y 8.34139E+3x 9.84936E-6



26

Fig B7: Model of surge generator with 4.45 meters of 1/2” CSST and with 16m #6 parallel bonding wire

Title 1/2" CSST 16m Bond

Number

PowerCET Corp

Date December 17 2012 Size SSheet 1 of 1 Rev 2File: LTIHalfInch16mBond.Sch

C1145.7u

G

R20.1

G

R4100k

C21uF

+ --ArcQ

ICSST

R110

G

G

R3.203

IBOND

R63.2

+B137kV

L319.0uH

R7.03

L14.8u

+ Vc

-- Vc

Vsw2

fxy

W

X

Y

Z

Vfunc2

+ Vc

-- Vc

Vsw3

+ Vc

-- Vc

Vsw1

fxy

W X Y Z

Vfunc1ITOT

L24.45u



27

Fig B8: Simulated waveforms of current and arc charge for ½” CSST and 16 meter bonding conductor

Transient - New, LTIHalfInch16mBond.Sch + LTIHalfInch16mBond.Anl, 20 December 2012



Time0 50.0u 100u 150u 200u 250u 300u 350u 400u 450u 500u

-1.00K

0

1.00K

2.00K

3.00K

4.00K

5.00K

6.00K

7.00K

8.00K

9.00K

10.0K

11.0K

-100m

0

100m

200m

300m

400m

500m

600m

y 7.71439E+3x 9.26999E-6



28

A P P E N D I X C MEASURED & CALCULATED SELF-INDUCTANCE OF CSST

MEASURED & CALCULATED INDUCTANCE OF CSST LOOPS

The measurement of the self-inductance of linear conductors is non-trivial, since any laboratory measurement requires the injection of a current into the conductor as well as its return to the source through a second conductor. The self-inductance of the return conductor as well as mutual inductance between circuit elements makes this approach subject to variation resulting from the geometry of the test set-up.

PowerCET’s approach is to measure the self-inductance of loops of conductor, which ensures that any test current injected flows only on the CSST and not on any other required return circuits. The measured value of inductance measured in this way is then compared with the calculated value of inductance using the standard formula given below:

μ R(ln(8R/r)‐2)

Where L is the self-inductance of the circular loop, R is the radius of the loop, n the number of turns and r the radius of the CSST. This formula ignores skin effects, considered a valid assumption for the frequencies involved.

Three samples were measured in this way, a 6.32 m length of ½” CSST formed into a two-turn circular loop of inside diameter 1.003m, a 4.445m length of ½” CSST formed into a two-turn loop of 0.346m inside diameter and a 4.445m length of 1” CSST formed into a two-turn loop of 0.338m inside diameter. The self-inductance of each of these three samples was measured using an Agilent LCR meter. The calculated self-inductance was derived from the above formula, taking the effective radius of each CSST sample as half its maximum diameter (0.095m for ½” CSST and 0.016m for 1” CSST). The results of these measurements and calculations are shown in Table C1 below:

Sample Measured Self-Inductance @ 10kHz

Calculated Self-Inductance

Two turns, 6.32m of ½” CSST 9.61µH 10.15µH

Two turns, 4.445m of ½” CSST 6.35µH 6.39µH

Two turns, 4.445m of 1” CSST 5.38µH 5.32µH

Table C1: Measured and calculated self-inductance of CSST loops



29

MEASURED & CALCULATED INDUCTANCE OF LINEAR CSST

Measurements were made of the self-inductance of the LTI Scenario 1 test circuit in a number of configurations in order to estimate the contribution of the various circuit elements (Fig. C1).

Fig C1: Measurement of self-inductance of test circuit

The self-inductance of the various linear conductors were then calculated using the formula attributed to Rosa:

μ /2 l(ln(2l/r)‐1) Where L is the self-inductance of the linear conductor, l is its length and r its radius. Like the earlier formula for circular loops, this also ignores skin effects.

Using the measurements and test sample calculated self-inductance enabled an estimate of the generator circuit contribution, as follows:



30

Test Circuit Total Inductance Calculated Test Sample Inductance

Deduced Generator Inductance

Surge generator with 1m #6 shorting conductor:

5.95µH 1.64µH 4.31µH

Surge generator circuit with 4.5m copper pipe:

8.38µH 4.50µH 3.88µH

Surge generator circuit with 4.445m length of ½”of CSST

9.88µH 5.19µH 4.69µH

Surge Generator with 10m #6 bonding conductor

19.20µH 14.40µH 4.80µH

Surge Generator with 16m of #6 bonding conductor

26.3µH 21.20µH 5.10µH

Table C2: Measurements of generator circuit self-inductance

These data suggest a mean generator self-inductance of 4.56µH, which would indicate a self-inductance of the 4.445m of linear ½” CSST of 5.32µH, about 3% higher than that calculated from the Rosa formula.

CONCLUSIONS

The measured data for self-inductance of CSST show that it may be calculated to an accuracy of better than 10% using standard formulae. The Rosa formula for linear conductors may be applied, taking the effective radius of the CSST as one half of its largest diameter. The estimated self-inductance of linear ½” and 1” CSST conductors for various lengths is given below in table C2.

Table C2: Calculated self-inductance of CSST



END OF REPORT

Documents

Validation of Installation Methods for CSST Gas Piping to · PDF fileValidation of Installation Methods for CSST Gas Piping to Mitigate Indirect Lightning Related Damage Page 1 Executive