Identifying Harmonic Sources by Profiling Discrete Harmonic Cycle Time

Harmonic emissions cause power quality issues in electrical systems, making source identification necessary to determine the best way of remedying them. Existing methods rely on measuring variations in harmonic amplitude, angle, and phase which require multiple measurement points or extensive system knowledge. We identified a methodological research gap because Discrete Harmonic Cycle Time (DHCT) has not been evaluated as a measuring principle for harmonic source identification. We propose a method using DHCT to enable single point measurements to profile harmonic sources. To determine the cycle length, we used a combination of sifting, a filter bank of cascaded high-pass and notch filters, and zero-crossing detection. For comparing the devices, we extracted motifs from the time series of discrete cycle lengths and applied principal component analysis. While it has previously been known in other contexts that harmonics can have varying cycle lengths, our laboratory results show that this is also true for the emissions of power electronic devices and that the differentiation can be used to identify the devices, thus bridging this knowledge gap. This method is independent of system impedance and topology. While further validation in more complex environments is needed, our results suggest that devices can be identified using this measurement principle. Since the measuring principle is orthogonal to other methods, it has potential as a complementary tool in harmonic source identification.