Harmonic Detection Based On Extreme Learning Machine

With the progress and development of society,more and more new power equipment and power electronic products continue to appear in daily life.While improving the quality of life,it also brings more harmonic pollution,making people pay more attention to the detection and control of power system harmonics.Many scholars have done a lot of research on how to carry out high-efficiency and high-precision power system harmonic detection,and achieved effective research results.However,with the continuous change of power system structure,and the power system harmonic is an unstable nonlinear signal,it is difficult for traditional processing methods to achieve better results in its detection and control.In this paper,the harmonic detection model based on Extreme Learning Machine(ELM)and its improved model is studied.The details are as follows:This paper studies the harmonic detection model based on ELM combined with dimension reduction and analyzes the influence of the number of neurons and dimension reduction algorithm on the harmonic detection model.Selects the optimal number of neurons and the best dimension reduction algorithm to establish the harmonic detection model.Then,aiming at the randomness of the weight of ELM,the harmonic detection model based on the optimization algorithm of the Kernel ELM is studied.The optimal parameters C and g are selected by the optimization algorithm to establish the harmonic detection model.Secondly,due to the limitation of online harmonic detection,the harmonic detection model based on Online Sequence ELM is studied,and the influence of initial sample number and each batch of learning samples on harmonic detection is analyzed.The optimal number is selected to establish harmonic detection model.Finally,the Online Sequence ELM is selected to establish the harmonic detection model.In order to improve the detection effect of the harmonic,the multi-dimensional scaling algorithm and Online Sequence ELM are selected to establish the harmonic detection model.Through the experimental simulation,it can be seen that the accuracy of multi-dimensional scaling algorithm combined with the harmonic detection model of Online Sequence ELM is above 0.9 for each harmonic.The experimental results show that the predicted value is almost consistent with the actual value,which can prove that the harmonic detection model established in this paper has good effect on the detection of each harmonic.In this paper,the ELM algorithm is applied to harmonic detection,and a harmonic detection model is established by combining with dimension reduction algorithm.The experiment results show that the ELM algorithm combined with dimension reduction algorithm is worth further research on harmonic detection model.This paper constructs a fast and accurate harmonic detection model to realize the accurate detection of the harmonics,and provides some reference for the establishment of harmonic detection system and its application in real life.Figure[35]table[9]reference[91]...