Repair & Metrology Services

Transcription

1

2 Repair & Metrology Services Presents SETTING UP A 100kV MEASUREMENT SYSTEM Author: Les Wesson Co-author: Stephan Mare

3 Content 1. Introduction 2. Setting up a suitable area 3. Purchasing equipment 4. DC voltage measurements 5. AC voltage measurements 6. Problems 7. Corona 8. ILC results 9. Conclusion

4 Introduction There is a growing requirement in industry to measure both ac and dc high voltage up to at least 100 kv. With only a limited amount of high voltage facilities in South Africa, RMS decided to set up a SANAS accredited high voltage laboratory.

5 Introduction cont. Most DC/Low Frequency laboratories have some high voltage capabilities, the norm is to use a 1000:1 high voltage probe that is limited to 40 kv dc and 28 kv ac for voltage measurements.

6

7 Setting up a suitable area The first requirement was to set out a safe area to measure and generate the high voltage. A safety cage was designed and built. The cage was designed to only allow the high voltage to be generated when the area was safe and secure. A door switch was installed onto the door frame which will only allow the high voltage to be generated when the door is closed.

8

9 Purchasing of the equipment The first item on the shopping list was a high voltage generator, capable of supplying 100 kv ac and 100 kv dc. An Adret 10010PT high voltage source was purchased. This was followed by A CPS HVP-505 high voltage probe. We misunderstood the specifications and when the HV probe arrived we found that the probe was capable of measuring 100 kv dc, but only 75 kv ac.

10

11 Purchasing of the equipment Cont. According to Bram Stoker We learn from failure not from success. This proved to be correct, More careful research was needed. Performance versus cost was the main criteria. Finally a decision was made to purchase a Vitrek 4700 and HVL-150 high voltage divider capable of measuring 150 kv ac and dc.

12

13 Traceability to international standards The Vitrek voltage divider and display unit were purchased with an ISO calibration certificate for traceability. This was a critical requirement as the accepted method of calibrating high voltage dividers is to source a known 1000 V dc and use the certified ratio to calculate the higher voltage being measured.

14 DC high voltage measurements Comparisons were conducted on the newly purchased CPS HV probe and Vitrek HV divider against a 40 kv probe and matching DMM. The new CPS HV Probe does not need to be matched to a measurement instrument as it uses a digital output to a computer via a USB cable.

15 DC high voltage measurements Cont. The new Vitrek HV divider is a stand alone system using a HV tower and a matching readout unit. The Vitrek HV divider arrived with a Calibration Certificate up to 90 kv. We compared the new divider and new HV probe and were able to measure dc voltage up to 100kV with results that were better than we expected.

16 DC high voltage measurements Cont Again? The following results were obtained from a comparison with a Fluke 80k-40 probe and newly acquired CPS divider. Nominal Value (kv) 80k-40 Results (kv) CPS Divider Results (kv) Difference (%)

17 AC high voltage measurements A Fluke 80k-40 probe was again used to prove that the new CPS HV probe was operating correctly. The 80k-40 could only be used up to 28 kv. The results were disappointing. It was found that holding the probe and supporting the probe on a work surface provided very different values. Shielding the probe using metal foil improved the measurement results, but left us wondering about the best method to obtain repeatability.

18 High ac voltage measurements After receiving the new Vitrek HV divider we connected both this and the CPS HV probe and tried to perform some comparison measurements. It was found that at around 30 kv a hissing noise could be heard and the readings became unstable, at 50 kv the hissing noise had changed to a crackling noise, a flash over was expected at any moment.

19 Problems The hissing and crackling noises that were heard were as a result of corona. Any small change in the ac mains input to HV source made a significant change to the HV measurements. Poor stability resulted in a wide spread of results. Positioning of the probes and connections made a difference to the measurement results.

20 Corona Description What is Corona? Corona is a luminous discharge due to ionization of the air surrounding a conductor caused by a voltage gradient exceeding a certain critical value, the dielectric strength in air at atmospheric pressure (sea level) is roughly 30 kilovolts per centimeter, but the distance decreases at higher altitudes.

21 Corona It s a cause of EMI interference The presence of Corona in a measurement system results in electromagnetic interference. This electromagnetic interference will have a erroneous effect on the measurement instrument or device, causing unstable readings.

22 Corona discharge Corona discharge usually forms at highly curved regions on electrodes, such as sharp corners, projecting points, edges of metal surfaces, or small diameter wires. Sharp curvatures causes a high potential gradient at these locations, so that the air breaks down and forms plasma there first.

23 Corona discharge cont. On sharp points in air corona can start at potentials of 2-6 kv. In order to suppress corona formation, terminals on high voltage equipment are frequently designed with smooth large diameter rounded shapes like balls, rings or domes. Corona rings are often added to conductors to extend the high voltage measurement capabilities of a measurement system.

24 List of problems 1 Mains voltage instability. 2 Corona effect on the high voltage system. 3 Selection of the correct wiring. 4 Removing all sharp edges and points from the system. 5 Improve the earth system. 6 Improve measurement results on the CPS HV probe.

25 Problem 1 Mains voltage instability. An Elgar ac power source was installed, this had two effects, firstly the input voltage into the HV source was more stable and secondly the output adjustment of the Elgar was much finer than the variac control of the HV source. The combination of the Elgar and the Adret variac allowed for finer adjustment and repeatable input voltages into the HV unit.

26

27 Problem 2 Corona effect Initially home made corona rings were made to test the improvement in the measurement system. Researching corona rings proved to be difficult as there are no set standards which mention corona ring dimensions in the industry. 345 mm corona rings were purchased locally. These were not the best units, they were the only units available locally.

28 Problem 3 Connections The cables connecting the source to the measurement dividers were discarded and copper tubing was used. The copper tubing reduced the corona but was difficult and time consuming to work with. The copper tubing was replaced with high voltage cable. The cable also worked well and is easy to change the test set up.

29

30

31 Problems 4 Sharp Edges All sharp points and corners in the system were rounded. Rounded fasteners and bolts were used where possible.

32 Problem 5 Earth system The earth system was improved by running a copper bar around the cage and inserting an earth pin through the floor of the enclosure.

33 Problem 6 The CPS probe repeatability A frame was built to house the probe and a dome was attached to the top of the probe. A display unit was added to the base of the probe together with a corona ring, this also serves as protection for the display unit. By placing the probe on this stand and having the input attached to a fixed corona ring the repeatability problems are now solved.

34

35

36 Results The measurement results were more stable. Good repeatable measurements were obtained on the Vitrek divider and the CPS probe. This enabled us to increase our measurement system up to 100 kv both ac and dc.

37 ILC Results The HV laboratory has taken two years to reach a point where we are satisfied that we are able to make both dc and ac voltage measurements up to 100 kv. During the process of improving the HV system we conducted a number of comparisons and ILC s. The final results show that we can produce repeatable results.

38 DC High Voltage ILC Results These results were obtained from an ILC in July Nominal Value (kv) Lab 1 Results (kv) Uncertainty (%) Lab 2 Results (kv) Uncertainty (%) En Value

39 AC High Voltage ILC Results These results were obtained from an ILC in July Nominal Value (kv) Lab 1 Results (kv) Uncertainty (%) Lab 2 Results (kv) Uncertainty (%) En Value

40 Conclusion The final ILC results were better than some of the previous results which were measured with home made copper rings instead of built for purpose domes. This and inconsistent results from measurements recorded on the earlier audit sample that was used were sorted out. The laboratory was finally ready to apply for SANAS accreditation. We would like NMISA to be able to provide traceability for High Voltage measurements in South Africa.

41 Conclusion Cont. The results of the ILC show that the RMS laboratory is able to measure and generate high voltage both ac at 50 Hz and dc from 1 kv to 100 kv with uncertainties of less than 0.1%.

42 Thank you Repair & Metrology Services (Pty) Ltd No. 10 Enterprise Close, Linbro Office Park, Sandton