Signal Decomposition With Reduced Complexity for Classification of Isolated and Multiple Disturbances in Electric Signals

In a previous work, the authors discussed and introduced a technique for the classification of isolated and multiple disturbances in electric signals. However, in that work, the decomposition of the electric signal into the fundamental, harmonic, and error component can be a very difficult task to be accomplished in real time. In this regards, this contribution proposes and analyzes the decomposition of electric signals into fundamental and error components. For each component, higher order statistics (HOS)-based features are selected and extracted and feed Bayesian classifiers that are designed for each class of disturbance. Comparison results with a standard HOS-based classification technique indicate that the proposed technique can offer improved performance not only for isolated disturbance, but also for multiples ones.