Application of echo state network for harmonic detection in distribution networks

摘要

Harmonic pollution is not a new concept but that is an important phenomenon on power systems and one of the major power quality meaning in electrical power. The high usage of non-linear loads and power electronic devices caused power system harmonic problems raises. Harmonic detection becomes important issue for prevention of the power quality problems. Also, exact detection of harmonics amplitude in power system voltage or current is necessary to design filters for eliminating harmonics. In this study, a method based on an echo state network (ESN) is presented to estimate the harmonic components amplitude. The proposed ESN works with only a quarter cycle data point inputs, but other estimation techniques use one-complete cycle or at least half-cycle data. So, the proposed method is suitable for real-time applications. Moreover, penguins search optimisation algorithm is used in the proposed method for optimising the best values of ESNs parameters causing the output of the ESN enhanced. The method is tested with a number of simulated signals and the test results indicate that the proposed method detects accurately the harmonic components amplitude compared to latest similar methods.