This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved.

Manual of Petroleum Measurement Standards (API MPMS) Chapter 5 – Metering, Section 3 – Measurement of Liquid Hydrocarbons by Turbine Meter Manual,

Ad Hoc Flow Conditioning Task Force

Technical Report on Flow Conditioner Installation

and Effects on Turbine Meters 31JAN17

Introduction Turbine Meters are used extensively for the custody transfer of liquid hydrocarbons worldwide. The potential impact of inaccurate volumetric measurement due to improper turbine meter flow conditioning and installation effects could be significant. Flow conditioners and flow conditioning are designed to substantially reduce installation and operational effects. Phase II testing concentrated on operational effects. Phase I results on installation effects (for example 3 elbows out of plane) were covered first in an addenda to API MPMS Chapter 5.3 and then included in Section 5.3.5.3. The purpose of this Technical Report (TR) is to provide a summary of flow conditioning testing performed on turbine meters in liquid hydrocarbons. Initial testing was conducted in water and those findings were included as an addendum to API MPMS Chapter 5.3 in 2009, subsequent testing in hydrocarbon liquids was carried out through July 2016. Phase II testing focused on operational effects, specifically the relationship of strainer design, strainer basket disturbances, flow conditioning, and their effects on the flow meter deviations in hydrocarbon applications (viscosities, densities, and Reynolds number). Phase II testing focused on five flow conditioners: 1 tube bundle and 4 high performance flow conditioners. Their performance in various distorted flow profiles was measured by determining meter factor deviations. Installation and operational effects were created with piping geometry, various strainer designs and various blockages. Obstructions were placed in the strainer basket, as illustrated in Figure 2 Strainer blockage replication obstructions. Multiple turbine meters were used in the testing that included flat bladed, (un-rimmed and rimmed) and helical. The testing was also not intended to establish whether a specific meter and conditioner combination worked better than a different meter and conditioner combination. Dr. Mattingly’s NIST studies of flow profile in the 1980’s were an instigating force in the initiation of the Ad Hoc Flow Conditioning Task Force (TF). The purpose of both Phase I (beginning about 2005 on water as a liquid) and Phase II (2010-2016) was to validate the reports that obstructions in strainers caused meter factor to shift of 0.25% or more,

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved. unpredictably, and to evaluate 20D, tube bundles and flow conditioning performance on piping disturbances and random, erratically occurring effects such as strainer basket obstructions. This could likely reduce proving frequency. More importantly high performance flow conditioners would give more consistent meter reproducibility and repeatability. These random events with flowing stream debris can’t be “proved out” as effectively or rapidly as they can be “tuned out” by high pressure flow conditioning. The conditions for the testing had process variables of flow, pressure, temperature, and viscosity that varied negligibly compared to typical field operation. The Ad Hoc Flow Conditioning TF had discussions in 2011 and 2012 regarding what might be considered acceptable variation. The result was that the TF felt meter factor variations of < 0.03%± were the limits of the artifact (turbine meter, flow conditioning, piping, and strainer) in the laboratory performance. The TF concurred that meter factor variations of 0.07% or more under the test conditions were an indication of a real shift. In discussion by the users and the Ad Hoc Task Force, it was noted that in field conditions, variances larger than above may be considered acceptable. Acceptable meter factor variations are defined by the operations and the appropriate risk managements (such as loss/gain and line balance for leak detection). 1. Scope API MPMS Chapter 5.3 and parts of API MPMS Chapter 6 cover the installation requirements and performance characteristics of turbine meters in liquid hydrocarbon service. This research work provides data that should be considered for future incorporation into these standards. Phase I of this flow conditioning task force was performed on water prior to 2009. As part of Phase I, an Addendum was included in API MPMS Chapter 5, Section 5.3.5.3 and Appendix A.1 recommending the need for flow conditioning rather than straight pipe of any length. Phase II was intended to prove or disprove whether the results on water would translate to light hydrocarbons, higher viscosities, larger line and strainer sizes and different Reynolds numbers. Phase II of the flow conditioning project tested several sizes and types turbine meters, strainers and piping arrangements with various types and arrangements of commercially available flow conditioners. This was carried out on a range of petroleum liquids, to try to determine which flow conditioner arrangements provide adequate turbine meter accuracy for liquid hydrocarbon applications. Previous work by the Ad Hoc Flow Conditioning TF, determined that meter performance as reflected by meter factor deviation, repeatability and reproducibility, was sensitive to

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved. flow profile effects caused by obstructions in strainer baskets as well as strainer basket movement. Phase II testing continued with obstructions similar to those in Phase I (water testing) but obstructions “finger left” (D) and “finger right” (E) were eliminated part way through Phase II testing, as they were found to cause minimal change to meter factor deviation. 1.1 Configuration of Test Meter Runs (see photos and sketches below)

Figure 1A - Typical Test Meter Run Configuration

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved.

Figure 1B – Alternate Test Meter Run Configuration

On either configuration (above) there were 5 different strainer obstructions:

A. half-moon left B. half-moon right C. full moon D. “finger” left E. “finger” right

Figure 2, shows the devices utilized to replicate installation effects, or varying strainer blockages. Obstructions A and B were constructed having approximately half the strainer discharge bore. Obstruction C was constructed having approximately equal the strainer discharge bore. Obstructions D and E dimensions are approximately 1”x4” regardless of strainer discharge bore. Strainers used in testing had nozzle diameters of 4”, 6”, and 8”.

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved.

Figure 2 Strainer Blockage Replication Obstructions

Each separate test consisted of runs 1-7 as listed below with six meter provings of five runs each. Each test was conducted at 53% and 80% of the maximum BPH rate of each size meter. The strainer, flow conditioning section and meter remained constant while five different obstructions were sequentially installed in the strainer basket, test runs as follows:

1. no obstruction 2. obstruction A (see A-type below showing position) 3. obstruction B (see B-type below showing position) 4. obstruction C (see C-type below showing position) 5. obstruction D 6. obstruction E 7. no obstruction

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved.

Approximately 50 total tests were completed to provide the data that is used in this Technical Report. 2. Discussion of Results

Phase II testing on liquids similar to light hydrocarbons (approximately 2 cSt and 15 cSt), replicated the results of Phase I testing on water.

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved.

2.1 Strainers

Figure 5 Strainer without Positive Positioning

If the basket moved position, during operation this also demonstrated a negative effect on the readings. We overcame both issues by bolting the basket to the basket ring in the body.

2.1.1 Strainer effects on flow profile and meter performance (performances and

effects) are caused by:

a) Nozzle to body ratio, size (nozzle/body), exit nozzle velocity, b) Basket movement during operation and consistent relocation on reinstallation

after inspection, c) Basket location in the strainer body (annular width between basket and body

wall) with respect to outlet nozzle, d) Debris on the basket wall, or e) Exit nozzle fluid velocity.

2.1.2 The data from our testing shows that the larger the strainer inlet/outlet and

body, the lower the variation in Meter Factor (MF) on proving runs.

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved.

a) For the same nozzle size and configuration, larger body/basket yielded slightly less variance in MF on the test runs

b) For the same body size larger outlet nozzle size gave noticeably less variance in MF on the test runs

c) It appears that the MF stability improves noticeably at strainer outlet nozzle velocities ≤ 15 fps and is greatly improved ≤ 10 fps

d) Only conical or concentric reducers were used at the strainer nozzle outlet. Eccentric reducers would not exhibit the same behavior.

e) The strainer basket being secured/fixed inside the body reduces the varying perturbations that the basket adds to the flow profile.

2.2 Flow Conditioners 2.2.1 Flow conditioner installation location has considerable effect on performance.

Users should always follow the OEM installation recommendations.

2.2.2 Tube bundles provided erratic flow conditioning performance. Water testing first demonstrated that tube bundles exhibited unacceptable results based on meter factor stability.

2.2.3 Similar meter deviations when testing with tube bundles were observed on

the initial Phase II petroleum tests. Therefore, tube bundles were eliminated from all remaining petroleum testing that constituted the majority of the tests.

2.2.4 For all the remaining tests referenced in 2.2.3 just above, only high

performance flow conditioners were used.

Note: High performance flow conditioners are defined as those that provide pseudo-fully developed flow in laboratory and field piping configurations.

Figure 6 Meter Factor Variance at 53% of Max Flow

‐0.060%

‐0.040%

‐0.020%

0.000%

0.020%

0.040%

1 2 3 4 5 6

MF Variance with Obstruction 

Change

Obstruction A through E + No Obstruction

Meter Factor Variance at 53% of Meter Max Flow

MF Variance with TB

MF with HP FC

Linear (MF with HP FC)

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved.

Figure 7 Meter Factor Variance at 80% of Max Flow

3.0 Recommendations In all turbine meter applications, high performance flow conditioners should be used to optimize meter performance.

3.1 Strainer baskets should be as rigid as possible: 3.1.1 the strainer basket flange and perforated plate need to resist movement

during operation 3.1.2 have a method to ensure exact placement in the strainer body when

removed for cleaning and replacement.

3.2 For custody transfer measurement systems with turbine meters, straight pipe of 20D alone (or for that matter any length) did not demonstrate acceptable performance to be considered “flow conditioning”.

3.3 For custody transfer measurement systems with turbine meters, tube bundles did not demonstrate acceptable performance and should not be considered high performance flow conditioning.

3.4 Testing has indicated that additional API flow conditioning testing for other non-positive displacement meters should be considered.

3.5 Correct size gaskets, flange alignment and pipe bore size should be used to ensure no obstruction/restriction in the pipeline.

4. Observations for other Chapters

‐0.100%

‐0.080%

‐0.060%

‐0.040%

‐0.020%

0.000%

0.020%

0.040%

0.060%

0.080%

1 2 3 4 5 6

MF Variance with Obstruction 

Change

Obstruction A through E + No Obstruction

Meter Factor Variance at 80% of Meter Max Flow

MF Variance with TB

MF Variance with HP FC

Linear (MF Variance with HP FC)

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and staff of the API Standards Dept. Copyright API. All rights reserved.

4.1 Users should always follow the manufacturer’s recommendations for storage of spare turbine meters.

Note: During Phase II (as well as Phase I) testing, when meters were removed from the piping and then reinserted several days or weeks later, the meter often took several hours of operation before becoming repeatable to the extent of the typical Ad Hoc data. 4.2 Users should be aware that following periods of storage in atmospheric

conditions, all flow meters require a “break-in” period in the flowing stream to stabilize.

NOTE: This “break-in” period could eliminate the need to prove several times shortly after meter installation.

4.3 If no OEM instructions are provided, the meter should be cleaned upon removal. Storage with the shaft vertical and meter ends sealed is recommended.

4.4 Slip-on flanges are discouraged because of issues of bore alignment of the

high-performance flow conditioner or turbine meter alignment (steps, gaps, offsets). Butt weld flanges with the bore machined smooth should be used. Socket weld flanges may be acceptable.

4.5 Users should always follow manufacturer’s recommendations for metering

system equipment installation, operation, and maintenance.