Identifying Microgrid Disturbances Using Independent Component Analysis

This paper focuses on the identification of islanding and power quality (PQ) disturbances in microgrids using signal processing techniques such as wavelet transform (WT) and independent component analysis (ICA). WT is known to be a suitable approach for detecting various disturbances such as PQ events, islanding, transients and harmonics due to better multi-resolution analysis. However, it is observed that WT is susceptible to increased noise levels or transients in the voltage signal extracted at the point of common coupling (PCC). Therefore, ICA is proposed to find the significant components of the voltage signal which will help in identifying the disturbances more accurately. A comparative analysis, both qualitative and quantitative, is presented in this paper to demonstrate the superior detection capabilities of the proposed ICA approach when compared with the WT approach under noisy scenarios. The proposed technique is validated in real time using OPAL-RT's OP5600 to demonstrate its feasibility for practical applications.