Unambiguous real‑world interference waveforms to test static electricity meters
Testing the effect of interference on the accuracy of electricity meters requires test waveforms that represent what is measured at real meter connection points. Digitised waveforms contain 100’s of 1000’s of data points and these cannot easily be written in a normative standard so that industry can use the standard test waveforms. So, signal processing methods need to be developed to unambiguously specify the nature of waveforms so that they can be synthesised and generated with sufficient fidelity by testing laboratories.
The current waveforms from electrical appliances, grid metered supply points and from laboratory measurements are generated by non‑linear power electronic loads and are rich in harmonics; furthermore, they are modulated in time, in some cases chopped or impulsive, in other cases amplitude modulated over a duty cycle, but with sudden changes. The ‘anatomy’ of the waveforms must therefore be decomposed in both the frequency and time domains if the features of the signal are to be isolated, understood and reproduced accurately.
Fast Fourier Transforms (FFTs) are widely used in industry to decompose harmonic information. However, they have no resolution to identify timing-changing events such as non‑stationary pulses or modulations which are nearly always present in real world waveforms. So, in practice, time resolution is introduced using windowing methods as used in standards such as IEC61000‑4‑7, but unfortunately windowing reduces the frequency distribution and the ability to discriminate interharmonics. The trade‑off between time and frequency resolution is fixed and cannot be overcome, it can be optimised and adapted by using advance transforms such as Wavelets as used, for example, in image processing.
Two methods have been selected, namely Wavelets and a piece-wise linear approximation of the time domain waveform. These methods offer efficient and accurately representations of the selected test waveforms so that they can be normatively published in an unambiguous form suitable for synthesis by meter testbeds.
For more information, see the project webpage >>
Short Name:MeterEMI,Project Number:17NRM02
Universitat Politècnica de Catalunya (Spain)
Universiteit Twente (Netherlands)