Detection Methods for Current Signals Causing Errors in Static Electricity Meters

Abstract

In recent years, the shift to Distributed Generation (DG) and the use of smarter domestic appliances has led to an increasing integration of power electronics (active infeed converters, power drive systems etc.) at the household level. However, the use of more power electronics results in the generation of highly distorted currents entering the distribution grid. Previous research shows that such current waveforms can cause large errors in static electricity meters. Thus, there is an imperative need to study the characteristics of these current waveforms and their impact on meter readings by performing extended measurements in households. Since it is not practical to store all the high granularity waveform data of such measurements, suitable detection methods and trigger levels need to be defined to only capture the potentially problematic current waveforms. In this paper, signal processing techniques (differentiation, Short Time Fourier Transform and Wavelet Transform) are applied to current signals in order to extract features suitable for use as a trigger. Results show that the Discrete Wavelet Transform and the filter with derivative method give the most promising results and work reliably even for very noisy signals.