The aim of this work package is to research and recommend methods for lightning impulse voltage calibration for testing UHV equipment. The target is to provide new input to IEC 60060-2 for time parameters and voltage measurement on ultra-high voltage above 2500 kV, with uncertainty for peak voltage better than 1 %. It is further to resolve unexplained effects on measurements from front oscillations, corona, proximity and signal cables. This work package will build on results of the EMPIR JRP 14IND08 ElPow. The work will be extended to higher voltage level, from 1500 kV studied in EMPIR JRP 14IND08 ElPow up to 3500 kV targeted in this project.

Typically, NMIs can provide traceability for lighting impulse voltage measurements systems up to 500-700 kV, with the highest calibration service level in the world being 1500 kV. Due to the lack of traceable calibration services, the related IEC standard 60060-2 allows a linear extrapolation by a factor of 5. Theoretically, this allows approved measurements up to 7500 kV, but there is not yet scientific proof that this extrapolation method is valid above 2500 kV [9,11]. Due to the large size of the ultra-high voltage LI measurement systems in industrial laboratories, the only alternative is to perform the calibration on-site using a transportable lower voltage reference. Such transportable references are typically limited to 700 kV, allowing approval for measurement systems up to 3500 kV, provided that the linearity extension method is validated up to that voltage.

The lowest expanded uncertainty for LI reference measurement systems up to 1000 kV is 0.5 % for the test voltage value Ut, and 1% for the front time T1 and the time to half value T2. Using state-of-the-art methods for a linear extension, typical errors of 1 % for Ut have been reported at 2700 kV [11] and more than 3 % errors above 3000 kV [9]. The errors in front times were significantly larger, and the reason(s) for that are not understood. At the moment, there is insufficient metrology support for LI measurement systems above 2000 kV.

•    Task 2.1 will evaluate the effects that impact LI measurement accuracy and propose best practices for extrapolation to ultra-high voltage level up to 3500 kV.

•    Task 2.2 will modify and characterise transportable LI measurement systems, using the results of Task 2.1.

•    Task 2.3 will validate the approach of Task 2.1, and the LI measurement system improvement and characterisation in Task 2.2, by a joint measurement campaign at the 4000 kV LI facility at TU Delft.

The aim of this work package is to develop new methods for the determination of the voltage dependence (linearity) of HV capacitors. New techniques for linearity extension to high voltages will be validated. The techniques will be applied for onsite calibration of industrial (e.g. polypropylene) capacitors and also to precision gas capacitors. The target is to expand voltage scale beyond the limitations of existing methods while maintaining uncertainties for HVAC better than 80 V/V up to 800 kV. Low uncertainty requires the accurate determination of the voltage dependence of reference HV capacitors and the validation of industrial techniques used for linearity extension to high voltages. Presently, traceability in HVAC voltage measurement is provided by using inductive voltage transformers or capacitive voltage dividers with a typical uncertainty of around 50 µV/V at 200 kV, and 500 µV/V at 800 kV. The higher uncertainty at these high voltage levels is caused by the uncertainty in the extrapolation of this instrumentation from values at lower voltages.

This work package will develop the equipment and measurement techniques required to determine the voltage dependence of inductive voltage transformers and capacitive voltage dividers and compare the developed approach with Latzel’s mechanical oscillation method [14] developed in the 1980s for gas capacitors.

•    In Task 3.1 existing methods for the determination of the voltage dependence of HV capacitors and for extrapolation techniques to higher voltages via linearity extension will be reviewed.

•    Task 3.2 will develop new reference HV capacitors with ultra-low voltage dependence, and measurement techniques to traceably determine the actual voltage dependence of these HV capacitors and improved extrapolation techniques.

•    Task 3.3 will validate the performance of the new instrumentation and measurement techniques developed in Task 3.2

The aim of this work package is to validate the PD calibration procedure of PD analysers used for HVDC insulation diagnosis by means of an adjustable synthetic reference PD generator for HVDC cable systems and GIS setups.

The relation between PD pulse patterns and defects in internal insulation is not yet well known for HVDC stress. A qualification procedure using a PD waveform generator has been developed to study insulation under d.c. stress [12]. However, the procedure needs to be extended to different PD patterns associated with each defect. Due to the vast variability of the phenomenon and acquisition techniques, large amounts of data are needed to identify PD patterns associated with the real defect. Specific methods to recognise PD patterns, as well as PD calibrators to reproduce PD patterns applicable to measurements under d.c. stress, do not exist. Nowadays, there is no traceability of HVDC GIS partial discharge measurements due to the limited bandwidth of the sensors. This work package will address these identified shortcomings by developing novel calibration methodologies as well as appropriate high bandwidth sensors. Means for condition monitoring in HVDC grids and converter stations by PD detection and classification will be provided.

•    Task 4.1 will extend and validate the PD procedure for qualifying PD analysers, operating between 1 MHz and 30 MHz, with respect to their ability to locate and characterise the type of PD source encountered in d.c. power grids.

•    Task 4.2 will develop a standardised procedure using a characterisation setup to test the performance of partial discharge sensors for HVDC GIS, operating between 30 MHz and 300 MHz, and to develop a suitable metrology for PD measurements to locate and characterise the type of PD source encountered in HVDC GIS.