Summary
One of the greatest challenges found in power quality has been, and still is, the establishing of a reliable noninvasive method for quantifying the harmonic contributions of an electric utility and a consumer at the point of common coupling (PCC). This difficulty arises from the fact that the harmonic distortions measured on the PCC are unidentifiable mixtures from both parties. Currently, some work has focused on exploring the use of blind source separation (BSS) techniques to promote the sharing of harmonic responsibilities in a noninvasive way. However, the usage of this methodology is very recent and still requires thorough evaluations, not only from the point of view of the BSS techniques available in the literature, but also from the perspective of the electrical power systems and the nature of their loads. Thus, seeking to contribute to the consolidation of the use of BSS techniques in the sharing of harmonic responsibilities, this work aimed to investigate the impact of the variation of harmonic current sources on the accuracy of BSS methods. In addition, three algorithms suitable for complex‐valued signals were evaluated computationally and experimentally; the results obtained computationally were compared with the results of the method based on the covariance characteristic of the random vectors, and the experimental results was compared to that obtained through an invasive procedure.