Selfadaptive Evaluation Software for High-Voltage Impulse Tests and Calibration of Digital Recorders and Impulse Calibrators

W. Strauss

Table of Contents:

Abstract

1 Selfadaptive Evaluation Software for High-Voltage Impulse Tests

2 Non-Standard L.I. recorded during High-Voltage Impulse Tests

3 The IEC 1083-2 Testdata-Generator
3.1 Performance Test of TR-AS Evaluation Software
3.2 The IEC 1083-2 TDG reference waveforms

4 Calibration Laboratory for Impulse Measurement
4.1 Calibration system to establish traceability
4.2 Calibration facilities

5 Conclusion

6 References

Abstract

The final draft of IEC 1083-2: Digital recorders for measurements in high-voltage impulse tests, Part 2 /7/ specifies the evaluation of software used for the determination of the parameters of impulse waveforms.

The actual discussion deals with definition and software methods for evaluating the impulse parameters of non-standard lightning impulses which have not an easy double-exponential shape but additional oscillations, overshoot and non-linearities.

The paper describes the selfadaptive evaluation software TRAS developed during the last ten years considering the actual definitions of impulse shapes and a big number of non-standard shapes measured in many high-voltage laboratories around the world including power transformer and coil tests.

The TRAS evaluating software was validated with the IEC 1083-2 testdata generator (TDG) for all groups of reference waveforms and passed all tests without objection, the results of this performance test are shown.

The user in the test laboratory is obliged to use measuring equipment according to these standards /1/. To document the compliance this equipment must be calibrated accordingly with traceability to national standards. The calibration of the digital recorder must be performed with the same evaluation software used for measurements in the high-voltage laboratory, so a uniform software for standard and non-standard waveshapes must be available /9/. For this task the first calibration laboratory in Europe for electrical impulse parameter was founded in Germany in 1994, the accreditation was performed by the Physikalisch-Technische Bundesanstalt (PTB). The calibration system to establish traceability to national standards and the calibration facilities of the laboratory are described /5/.

1 Selfadaptive Evaluation Software for High-Voltage Impulse Tests

This can be solved for standard wave shapes normally without problems using smoothing and interpolation methods for evaluating voltage and time parameters with good stability and nearly independent from the resolution of the digital recorder. Automatic detection of polarity and full or chopped waves are standard /3/4/6/.

For non-standard waves additionally the problem to generate a suitable meancurve arises which is in accordance to the respective standard IEC 60-2 /2/.

At HV Laboratories of Universitys etc. similar impulse tests with evaluating of impulse parameter may be performed in several ways trying different approximation methods to find at least the perfect approximation of this meancurve - if exists.

In a h.v. testfield of industry this is not a practicable way.

One idea for good evaluation software for non-standard waves is to get a result in short time as good as possible with option for subsequent modification by hand - with optical controlled helping facility of results on screen.

In most cases it is not possible to generate e.g. for the front or tail elapse an exponential or double exponential approximation automatically, because the necessary curve points for calculating the approximation parameters are errornous caused e.g. by noise, ripple, ringing, overshoot, multi-frequent oscillations, non-linear effects and the limited resolution of the digital recorder.

Development of evaluating software should not follow only by the classical methods of analysis based on standard functions linear, sin, exp, splines, xn and so on. The classical method of the good old test engineer in the h.v. test field who took a curve lineal and drawed a nearly perfect approximated meancurve into the oscillogram should be taken under consideration, too.

In many cases a meancurve similar to that hand-drawn curve can be generated automatically by a computer program. The TRAS-Evaluation-Software has such a meancurve feature shown in the following examples.

2 Non-Standard L.I. recorded during High-Voltage Impulse Tests

Picture 2.1.1: 00833 Transformer test of a low-impendant winding
	automatic evaluation calculates an acceptable meancurve for the peak the front time will be detected too small because of the high initial overshoot modification must be done subsequently by hand-evaluation, see next picture
Picture 2.1.2: 0833A Transformer test of a low-impendant winding
	with subsequent hand-evaluation of the front by moving the 90%-point horizontally, until both helping lines suggest a successfull approximation after each movement the actual time parameters are calculated because the 30%-point is fixed, moving of the 90%-point influences the virtual origin and therefore the time-to-chopping, too
Picture 2.1.3: 0833B Transformer test of a low-impendant winding
	additionally with subsequent hand-evaluation of the peak value by modifying the horizontal 100% raster line, e.g. to a higher value for demonstration and in a second step adapting the front time by moving the 90%-point horizontally, until the both helping lines suggest a successful approximation
Picture 2.1.4: 01385 Transformer Test
	automatic calculation of the meancurve in the peak area works successful any automatic calculation of approximation parameters for a single or double-exponential approximation will probably not work satisfying for this shape
Picture 2.1.5: 0024
	oscillations with two frequencies are superimposed in the front- and peak-area automatic calculation of the meancurve in the peak area works successful beta=4.4% was calculated for evaluating parameters refer to the right window
Picture 2.1.6: 0023
	the same test setup as before, but chopped wave automatic calculation of the meancurve in the peak area works successful also for chopped waves with oscillations beta=3.1% was calculated with respect to the picture before with a beta of 4.4% the peak voltage calculation differs approx. 1.3%
Picture 2.1.7: 0527 Transformer Test
	automatic calculation of the meancurve in the peak area works successful any automatic calculation of approximation parameters for a single or double-exponential approximation will probably not work satisfying for this shape
Picture 2.1.8: 00743 Transformer Test
	oscillations in front and peak area on a chopped wave automatic calculation of the meancurve in the peak area works successful also for chopped waves with oscillations
Picture 2.1.9: 0577 Insulator test with steep impulse voltage
	automatic meancurve calculation is not possible for such short impulses because any smoothing would affect the front time and therefore the steepness, the highest sample is the peak modification if required must be done subsequently by hand-evaluation as described before Note: T1=61ns, Tc=103ns

3 The IEC 1083-2 Testdata-Generator

3.1 Performance Test of TR-AS Evaluation Software

Picture 3.1.1: Test Record of IEC-TDG reference waveforms

Test Result:
The TR-AS Software passed all tests for all groups of impulses and for all parameters specified according to final draft of IEC 1083-2 Part 2, table 2 - specified limits of the parameters of the reference waveforms (1996-05-15).

Remark:
All test results from the standard and non-standard waves listed were evaluated automatically regarding overshoot and meancurve calculation. The calculated meancurves are shown in the respective pictures below. The applied meancurve of the IEC 1083-2 TDG for the calculation of the nominal values should be included in the standard.

3.2 The IEC 1083-2 TDG reference waveforms

IEC-TDG No. 1	IEC-TDG No. 6
IEC-TDG No. 2	IEC-TDG No. 7
IEC-TDG No. 11	IEC-TDG No. 12
IEC-TDG No. 13	IEC-TDG No. 14
IEC-TDG No. 3	IEC-TDG No. 8
IEC-TDG No. 4	IEC-TDG No. 9
IEC-TDG No. 5	IEC-TDG No. 10
IEC-TDG No. 15

4 Calibration Laboratory for Impulse Measurement

The calibration laboratory was supplied with the necessary standard reference measuring equipment by the holder DR. STRAUSS SYSTEM-ELEKTRONIK GMBH, comprising digital recorder, impulse calibration generator and digital voltmeter calibrated by the Physikalisch-Technische Bundesanstalt (PTB) in Braunschweig for the required comparison with the national standard.

The calibration laboratory was integrated into the Quality Management System of the holder company according to DIN-ISO 9001 resp. EN 29001.

After the examination of these documents the assessment of the calibration laboratory and several comparison tests were performed by the PTB.

Finally the accreditation was performed on 06/06/1994 by the German Calibration Service (DKD), supervised by the PTB and represented in the Deutschen AkkreditierungsRat (DAR) as calibration laboratory for electrical impulse parameter with DAR registration number DKD-K-11701.

4.1 Calibration system to establish traceability

The PTB Calibration Certificate is the supposition for the DKD calibration laboratory to issue the DKD Calibration certificate. The DKD Calibration Certificate is the supposition for the calibration laboratory of organisation to issue the Works Calibration Certificate. The Works Calibration Certificate is the supposition for the test laboratory of organisation to issue the Works Test Record.

4.2 Calibration facilities

• impulse calibration generators
• unit step voltage generators
• digital recorders
• impulse voltmeters

The DKD calibration laboratory started in June 1994 and performed during the first two years more than 70 calibrations with tracebility to national standards.

These calibrated digital recorders and impulse calibration generators are the basis for many users around the world to etablish an accredited test laboratory in their company and to document the quality measures inside the quality management system of the organisation according to DIN/ISO 9001.

The IEC-Standard 1083-1 is an applicable instruction for the tests and procedures to find the relevant errors of digital recorders and impulse calibration generators used for measurements during high voltage and high current impulse tests. Only digital recorders with a special design for h.v. application with calibrated and stable input stage, low internal noise and suitable software will meet all accuracy requirements.

The calibrated digital recorders TRAS 25-8, TRAS 100-8 and TRAS 100-10 fulfil all requirements for measurements during high voltage and high current tests and for reference measurements.

The calibrated impulse calibration generators KAL 1000 and KAL 1000 RIG fulfill the requirements according to IEC 1083-1 for reference impulse calibrators.

5 Conclusion

The development of evaluation software should not be restricted in future standards regarding the methods to determine the required impulse parameters. That means, everybody should feel free to develope his own approximations for this task and only the evaluation results must fulfill the limits of errors and other requirements of the standard.

Methods and instructions for evaluating non-standard waves measured e.g. during transformer tests should be defined in additional standards with respect to the special requirements and physical conditions of such tests.

The IEC-Standard 1083-1 is an applicable instruction for the tests and procedures to find the relevant errors of digital recorders and impulse calibration generators used for measurements during high voltage and high current impulse tests. Only digital recorders with a special design for h.v. application with calibrated and stable input stage, low internal noise and suitable software will meet all accuracy requirements.

6 References