Real-time harmonic extractor based on optimization of fuzzy neural network

The real-time harmonic monitoring has become a very important concern in so many applications, for instance in electrical power systems, which are suffering from harmonic distortion due to nonlinear loads. The limitation in terms of computing time and complexity of data-processing related to the conventional techniques makes it appealing to investigate alternative techniques. From another hand, the majority of applications, such as filter design, need fast extraction of various harmonic components. Consequently, a technique based on optimized fuzzy neural networks (FNNs) by genetic algorithms (GAs) for the fast extraction of various harmonic components is presented here. It employs a type of fusion between artificial neural networks (ANNs) and fuzzy logic whose membership functions are adjusted suitably, due to the training capability of ANNs to accurately estimate the individual harmonic components of a distorted signal. The suggested technique is carried out, simulated, implemented on a real-time hardware platform and tested. Results clearly show that the proposed technique is appreciably faster, more accurate and less computationally complex than other techniques.