Theory and Experimental Validation of Two Techniques for Compensating VT Nonlinearities

Inductive instrument transformers (ITs) are still the most used voltage and current sensors in power systems. Among the numerous applications that require their use, one of the most important is surely represented by harmonics measurement. In this case, the recent literature shows that, since they suffer from both a filtering behavior due to their dynamics and from nonlinear effects produced by their iron core, they can introduce errors up to some percent. This article wants to deeply investigate, in the very same experimental conditions, about the performance of two digital signal processing techniques, recently introduced for the improvement of harmonics measurements performed through ITs, namely, SINusoidal characterization for DIstortion COMPensation (SINDICOMP) and compensation of harmonic distortion through polynomial modeling in the frequency domain (PHD). These methods have been applied to two different voltage transformers, having different specifications, by using two measurement setups based on different architectures. The impact of the voltage generator employed during the identification on the achieved accuracy is theoretically and experimentally evaluated. Modified versions of SINDICOMP and PHD compensation, which are more robust against nonidealities of the measurement setup, are presented. The performances of the techniques are evaluated by adopting voltage waveforms similar to those that can be encountered during the normal operation in a real distribution grid.