1

The Summer Undergraduate Research Fellowship (SURF) Symposium

6 August 2015

Purdue University, West Lafayette, Indiana, USA

MEASUREMENT AND CALIBRATION OF CENTRIFUGAL COMPRESSOR PRESSURE SCANNING INSTRUMENTATION

José R. Rivas-Padilla University of Puerto Rico

Mayagüez, Puerto Rico, US

Fangyuan Lou Purdue University

West Lafayette, Indiana, US

Herbert “Trey” Harrison Purdue University

West Lafayette, Indiana, US

Nicole L. Key Purdue University

West Lafayette, Indiana, US

ABSTRACT

The compressor is a key component of a jet engine necessary to compress air for the combustion process. Research to optimize compressor efficiency through the understanding of air flow behavior has led to increased efforts in creating modern compressor test facilities. In collaboration with Honeywell, the High Speed Compressor facility at Zucrow Laboratories has built a centrifugal compressor test cell with instrumentation to measure the temperatures and pressures of the air flow. This facility conducts experiments to characterize the compressor performance at varying conditions. During prolonged experimentation cycles, it is difficult to acquire reliable pressure measurements of the flow, as pressure sensor “drift” increases the uncertainty of measurements over time. Pressure scanners must be calibrated periodically to maintain reliability and calibration services are costly. To reduce cost, a pressure calibration process was automated with the use of a CPC6000 pressure calibrator. LabVIEW visual instrument (VI) software was developed to interface with the calibrator and the Digital Sensor Array (DSA) pressure scanners. The CPC6000 determines the pressure standard, and the VI software uses the calibrated pressure to overwrite the calibration coefficients of the DSA. The calibration coefficients that were outside of the calibration range were successfully overwritten by the VI software. The DSA was calibrated to measure pressures within 0.05% of full-scale (FS) measurement. Automating the calibration process will reduce the cost of periodic calibration by an external resource. It will also maintain repeatability in the uncertainty of the measurements through testing cycles that last beyond the 6 month guaranteed calibration period of the manufacturer.

INTRODUCTION

The jet engine is composed of 5 main components: inlet, compressor, combustion chamber, turbine and exhaust. The compressor is a key component in this combustion process, as it increases the pressure energy of the airflow to condition it for the combustion process. The High-Speed Compressor Laboratories in the Maurice J. Zucrow Laboratories at Purdue University have built an experimental test-cell for the study of a Honeywell experimental centrifugal compressor. This test-cell uses highly sensitive instrumentation to measure temperature, pressure, air-flow and impeller velocity. The pressure measurements are used to determine the performance and health of the compressor at any given time during experimentation. To obtain high fidelity pressure measurements, the use of piezoresistive pressure sensors is implemented to measure the static and total pressure at different points in the compressor. Piezoresistive sensors have a piezo material whose resistivity changes as stresses are applied and it deforms. This change in resistivity provides a change in voltage, and with an analog to digital conversion, the voltage values are converted to engineering units (psi). As time progresses the material begins to react differently to varying temperature and pressure cycles. This will in time create an offset in the pressure readings, and the accuracy of the sensor degrades [4]. To maintain the accuracy of the sensor without discarding it, it must be calibrated following a recommended calibration schedule from the manufacturer. From Figure 1 we observe that calibration costs are high as each module costs $400 to calibrate. There are currently 38 DSA modules in the High-Speed Compressor Laboratory and yearly calibration costs can go beyond $30,000. The Scanivalve DSA modules are required to be calibrated every 6 months.

2

The calibration operating procedures for the Scanivalve DSA pressure scanners were studied to develop a LabVIEW visual instrument (VI) capable of calibrating the pressure sensors in the High-Speed Compressor Laboratory from the Maurice J. Zucrow Laboratories at Purdue University. A Scanivalve digital sensor array (DSA) pressure scanner module and the Mensor CPC6000 pressure calibrator were integrated into a VI to create a fully automated calibration process. The DSA3217 module (Figure 2) was used as the model for the calibration tests and preliminary development of the VI. All the understanding of Scanivalve calibration operations and procedure was based on the study of operating procedures available in the [1] and [3]. It is important to note, that the High-Speed Compressor Laboratory also uses the DSA3016 modules in an enclosure configuration (Figure 3), which may have varying operating procedures and may be incompatible with the information presented hereafter. This report provides the user with a deep understanding of the calibration procedures of the DSA3217 module and the operation of the FieldCAL VI. The purpose is to continue upgrading and customizing the FieldCAL VI with features that will allow the user to: calibrate other DSA modules (such as the DSA3016), calibrate modules with varying pressure ranges (2.5psid, 5.0psid, and 100psid), and even calibrate multiple modules at once.

NOMENCLATURE DSA: Digital Sensor Array LabVIEW: Dataflow Visual Programming Language VI: Visual Instrument FS: Full Scale Range % FS: Percentage of Full Scale FieldCAL VI: LabVIEW pressure calibration software HARDWARE AND EQUIPMENT The Digital Sensor Array (DSA) is a stand-alone electronic pressure scanner which can accept up to 16 pneumatic inputs. Each DSA incorporates 16 individual, temperature compensated, piezoresistive pressure sensors, an analog to digital converter, and a microprocessor. Each pressure sensor is characterized over pressure and temperature. The basic DSA3217/16Px module provides 16 channels of ‘gauge’ pressure range sensors. These 16 pressure transducers are configured into two groups of eight each. Single range modules tie two reference manifolds together and the two calibration valves to provide the user with a single reference and a single calibration port to connect to [1]. The Mensor CPC6000 Modular Precision Pressure Controller is a multi-channel/multi-range pressure system which will calibrate the DSA3217 module. It can be equipped with up to four (two per channel) highly stable, temperature compensated, pressure transducers. Two precision regulators (pump regulator and solenoid valve regulator) provide precise control and highly stable pressures [2]. The CPC6000 will be used as the pressure standard to calculate the FS% value of the DSA3217 module to determine if it is calibrated. TYPES OF CALIBRATION

1. Zero Offset Calibration (Quick Zero)

The DSA’s internal calibration valve incorporates a zero offset correction feature (‘Quick Zero’). This feature provides a quick means to correct the minor zero drift problems of piezoresistive sensors. The ‘Quick Zero’ calibration mode pneumatically shorts the positive and negative side of the pressure sensor to ambient. This means that both sides of the sensor should be measuring the same reference pressure and is considered as the natural zero state of the sensor. The difference between the two pressure readings is factored into engineering units and mathematically removed [1].

Figure 2: Pressure Scanner DSA 3217 Module

(Taken from Ref [1])

Figure 3: Scanivalve DSA3016 8 module enclosure

(Taken from Ref [1])

Figure 1: Calibration Costs per Year

3

2. Full Calibration

A full calibration deletes all calibration coefficients and re-applies known pressures (a pressure standard from a certified calibrator) over a series of temperatures. This duplicates exactly what is done at the factory during the module’s original manufacture and calibration. This calibration requires an environmental chamber capable of

reaching the entire 0℃ to 69℃ temperature range and a suitable pressure standard [1]. This type of calibration is beyond the current capabilities of the High-Speed Compressor Laboratory since we lack an adequate environmental chamber. However, a user with such a facility could add ‘Full Calibration’ features to the ‘FieldCal VI’ and perform an automated ‘Full Calibration’.

3. Field Calibration

As an alternative to a ‘Full Calibration’, a Field Calibration can be performed with a suitable pressure standard, such as a Scanivalve Mensor CPC6000 calibrator. During a ‘Field Calibration’, a series of pressures are applied to the module at a single temperature. The temperature of the module does not have to be controlled to a specific temperature [1]. Scanivalve provides a free program called “PressCal” which operates with an executable program that calculates a new set of coefficients for that given temperature, then applies a correction to the pressure coefficients across all temperatures. This technique is suitable for maintaining modules long term and has proven to be just as good as a new full calibration. The FieldCal VI was developed to perform an automated ‘Field Calibration’ using a Mensor CPC6000 pressure calibrator. CALIBRATION SETUP AND PARAMETERS

The calibrator and the pressure scanner module will be both connected with pneumatic tubing to a 90-120 psi supply tank which will supply pressure to both components. The supply pressure from the tank will allow the calibrator to quickly respond when calibrating high pressure modules (greater than 15 psi). For this particular calibration setup, the pressure supplied by the highly stable internal pump of the CPC6000 will be enough to calibrate a 5.0 psid DSA3217 pressure scanner module. An Ethernet switchbox was used to connect the computer and module to the network. This arrangement will allow the connection of multiple modules to the same network for future development of the calibration software. A GPIB connection with an IEEE-488 connector was used to connect the GPIB port of the CPC6000 calibrator to a USB

port in the computer. This arrangement would allow the LabVIEW software to interface with both components and automate the calibration process. LabVIEW will connect with the Scanivalve executable program ‘FieldCal.exe’ to apply the correction of the calibration coefficients, and finally upload all the new calibration coefficients to the DSA3217 module. To prepare the DSA3217 for calibration it is important to understand a very important calibration requirement which is the full scale pressure value of the module. Each module has its own particular pressure range. The modules available in the High-Speed Compressor Laboratory include the 2.5 psid, 5.0 psid, and 100 psid modules. This is the full scale (FS) pressure value of each module. The accuracy of the calibration performed on each module has to meet a percentage of full scale (%FS) requirement. The full scale percentage calculation is shown below:

% 𝑭𝑺 = 𝑷𝒓𝒆𝒔𝒔𝒖𝒓𝒆 𝑺𝒕𝒂𝒏𝒅𝒂𝒓𝒅 − 𝑫𝒆𝒗𝒊𝒄𝒆 𝑷𝒓𝒆𝒔𝒔𝒖𝒓𝒆

𝑴𝒐𝒅𝒖𝒍𝒆 𝑭𝒖𝒍𝒍 𝑺𝒄𝒂𝒍𝒆 𝑽𝒂𝒍𝒖𝒆 (𝑭𝑺)∗ 𝟏𝟎𝟎

When the difference between the pressure standard (calibrator control pressure) and device pressure deviates from a maximum allowed % FS value, the module is in need of calibration. Table 1 shows the allowed %FS deviation for several differential pressure modules.

Differential Module Pressure Range (FS)

Allowed Deviation of %FS

10 in of 𝐻2𝑂 ±0.20% 𝐹𝑆

20 in of 𝐻2𝑂 ±0.20% 𝐹𝑆

1.0 psid ±0.12% 𝐹𝑆

2.5 psid ±0.08% 𝐹𝑆

5 to 500 psid ±0.05% 𝐹𝑆

501 to 750 psid ±0.08% 𝐹𝑆

For each pressure scanner module, five calibration pressures are provided to be supplied by a pressure standard. However, if more calibration pressures are provided, this will further discretize the calibration curve and should provide more accurate calibration results. Table 2 presents the values recommended by Scanivalve for a 5.0 psid module.

Equation 1: Full Scale Percentage

Table 1: Allowed % FS for Varying Differential Pressure Modules (Taken from Ref [1])

4

Pressure Range

Press. 1

Press. 2

Press. 3

Press. 4

Pres. 5

10 in of 𝐻2𝑂 -0.39 -0.19 0.00 0.19 0.39

20 in of 𝐻2𝑂 -0.79 -0.39 0.00 0.39 0.79

1 PSID -1.09 -0.55 0.00 0.55 1.09

2.5 PSID -2.73 -1.37 0.00 1.37 2.73

5 PSID -5.45 -2.73 0.00 2.73 5.45

15 PSID -16.35 -8.18 0.00 8.18 16.35

30 PSID -16.35 0.00 10.9 21.8 32.7

50 PSID -16.35 0.00 18.2 36.4 54.5

100 PSID -16.35 0.00 36.3 72.7 109.0

The recommended %FS value of 0.05 will be used to determine if the module is calibrated, and the 5 pressures for a 5.0 psid module presented in Table 2 will be used to create a five point calibration curve for each channel of the DSA3217 module. The ‘FieldCal.exe’ program uses the least squares regression to fit all new points taken during calibration to produce a straight line. Another line is constructed using the DSA's existing master points. The difference between these two lines represents the amount of correction applied to all of the existing DSA's master points.

Figure 3 illustrates how the calibration setup is connected to each component to allow pneumatic and digital interactions between the equipment and the LabVIEW FieldCal VI respectively.

CALIBRATION SEQUENCE AND LOGIC The calibration sequence consists of three major calibration steps: pre validation sampling sequence, calibration sequence, and post validation sampling sequence.

1. Pre Validation Sequence

The pre validation sampling sequence measures 9 pressures supplied by the calibrator to the module. Each channel (16 channels for the DSA3217) will be evaluated to see if each pressure point measured is below 0.05 %FS when compared to the pressure supplied by the calibrator. If the absolute value of the difference between the CPC6000 reading and the DSA3217 reading is above 0.0025 psi (0.05 %FS of a 5.0 psid module), then the module requires calibration.

Table 2: Scanivalve recommended calibration pressures (Taken from Ref [3])

Figure 3: Hardware and communications diagram Figure 4: Calibration sequence logic flowchart

5

2. Calibration Sequence

After the pre validation sequence is complete, the FieldCal VI proceeds to the calibration sequence, where 5 point calibration is performed to create a new calibration curve. The calibration sequence is the most complicated and it is represented in the logic flowchart in Figure 4.

The calibration sequence generates three text files from the module. The first text file is an ‘insert.txt’ file. The ‘insert.txt’ file is generated after a CAL command string has been sent to the module followed by the supplied pressure standard. The module will compare this input with the pressure reading from the module. After the command is sent, the module will quickly respond with new INSERT commands with 16 (one per channel) new analog to digital conversion coefficient (pressure counts)inputs.

Example: Input: CAL -5.45 Output: INSERT 31 0 -5.450000 -19543 M The number after the INSERT command is the temperature of the module in degrees Celsius (31℃), followed by the calibrated channel (Channel 0), then the pressure supplied by the calibrator (-5.45000), and finally the new pressure count conversion factor (-19543). After all the INSERT commands have been received from the module, two more commands are sent to generate a text file with a list of the calibration configuration variables (‘listc.text’), and a list with the old calibration master coefficients (‘master.txt’). Table 3 shows the three required files with their command syntax and name.

Action DSA Command File name that

holds results

Get DSA's Calibration

Configuration Variables

LIST C LISTC.TXT

Get DSA's Master points LIST M 0 69 MASTER.TXT

Get the new calibration

INSERT commands

CAL <applied

pressure>

INSERT.TXT

The ‘FieldCal.exe’ program accepts information about a DSA with these three text files. Once these files have been created, the ‘FieldCal.exe’ program is run to generate a new master point file named ‘NewMast.txt’. This file is of proper format and ready for upload to the DSA.

3. Post Validation Sequence

The calibration sequence is completed after the ‘NewMaster.txt’ is generated and uploaded to the module. Then the post validation sequence begins. This sequence

is identical to the pre validation sequence and will validate that the calibration sequence has adjusted all the coefficients properly. If the max %FS value in all the channels is under 0.05, then the calibration is successful.

CALIBRATION RESULTS AND DISCUSSION Figure 5 is the designed user interface which contains all the necessary information to monitor the calibration process through the three sequences. The lights in the front panel of the visual instrument indicate the current process being performed in the back panel.

The two matrix arrays on the right indicate all the values in the pre validation sequence (top right) and post validation sequence (bottom right) for each channel. Two lights will monitor the calibration status (greater or less than 0.05 %FS) for each validation sequence. The front panel’s purpose is purely to indicate current calibration status to the user and no interaction is required. Figure 6 and Figure 7 are the graphs which represent the distribution of %FS values for Channel 0 at the pre validation sequence and post validation sequence respectively. The 9 pressure values are equally distributed across the -5 to 5 psi pressure range of the module.

Figure 6: Graph of pressure measurements prior to calibration with 0.05 FS limits

Table 2: Scanivalve recommended calibration pressures

Figure 5: FieldCal VI front panel and user interface

6

In Figure 6, 4 of the sampled pressures were out of the allowed 0.05 %FS calibration limits. After the calibration sequence, Figure 7 shows that the 9 pressures sampled were within the allowed calibration limits. The max %FS value was reduced from 0.106 to 0.029 after calibration. The “new master points” consist of 9 calibration pressures over 9 temperatures (0℃, 5℃, 15℃, 25℃, 35℃, 45℃, 59℃,

64℃, and 69℃) across all the channels on the module (up to 32). The “FieldCal.exe” calibrates the module across the entire temperature range. This allows the module to provide high fidelity measurements, even if the temperature would vary several degrees. ‘FieldCal.exe’ uses the INSERT commands generated from the CAL commands at one temperature to extrapolate calibrated pressure counts to all the other temperatures. The entire calibration sequence takes approximately 16 minutes to perform all three sequences. This calibration cycle is fully automated and only requires that the user sets up the instruments and runs the FieldCal VI. The file directory has to be manually setup to indicate where in the computer the files will be generated to archive. A calibration log with all previously performed calibrations will be maintained as a record to return the module to previous configurations if needed. CONCLUSION AND FUTURE SOFTWARE FEATURES The FieldCal VI is capable of successfully calibrating the Scanivalve DSA3217 modules by performing a fully automated Field Calibration. The Field Calibration has been identified by Scanivalve to provide the same results as a Full Calibration if performed correctly. The FieldCal VI was developed following the same calibration procedures as the manual calibration procedures available in [3]. The results show a reduction from 0.106 to 0.029 in %FS value which is less than the 0.05 %FS recommended calibration standard for DSA modules of pressure ranges between 5 and 500 psid. With the help of Scanivalve the

‘FieldCal.exe’ executable program was integrated into the VI to generate new master coefficients for the operating temperature range of 0℃ to 69℃. Further work will be done to add features to the FieldCal VI that would allow it to calibrate multiple modules at once. The ‘FieldCal.exe’ allows the calibration of up to 8 modules at once, and the FieldCal VI can be conditioned to allow this with a proper setup of pneumatic tubing manifolds that allow the calibrator to supply the same pressure standard to all modules. The software could also calibrate modules of several pressure ranges by adding the 5 point calibration points appropriate for the modules calibration range: 2.5, 5.0, 100, and 150 psid. This software is expected to be used in all the test-cells in the High-Speed Compressor Laboratories. Each test-cell will be able to customize the FieldCal VI to their necessities and calibration requirements as the software’s flexibility allows this potential. Each test cell will also have a calibration log with the record of previous calibration files generated after performing a ‘Field Calibration’ sequence with the FieldCal VI.

ACKNOWLEDGMENTS

Thanks to Honeywell for its sponsoring and funding in the development of the experimental centrifugal compressor test-cell of the High-Speed Compressor Laboratories. Also, thanks to the Scanivalve Corporation for its technical support in providing valuable information to integrate its ‘FieldCal.exe’ program into the FieldCal VI. This project was also possible thanks to the technical support provided by John C. Fabian as senior research assistant of the laboratory, and project development feedback provided by Kyle Hughes as one of the SURF program graduate assistants.

REFERENCES

[1] DSA3217/3218 Series Pressure Scanner: Operating and

Service Manual. (2014, March). Liberty Lake,

Washington, United States of America: Scanivalve

Corporation.

[2] Operating Instructions: Modular Precision Pressure

Controller CPC6000. (2015, June). San Marcos, Texas, United States of America: Mensor.

[3] PressCal: Field Calibration Software for DSA3000/

3200/ 3300 Series Modules. (2012, March). Liberty Lake, Washington, United States of America: Scanivalve Corporation.

Figure 7: Graph of pressure measurements post calibration with 0.05 FS limits

7

[4] Solinst Canada. (2015). Solinst Technical Bulletin: Understanding Pressure Sensor Drift. Retrieved from Solinst: http://www.solinst.com/products/dataloggers-and-telemetry/3001-levelogger-series/technical-bulletins/understanding-pressure-sensor-drift.php

8