Harmonic Detection And Analysis Of Power System Based On Adaptive Algorithm

Harmonic pollution is an enduring topic,harmonic pollution is a long-standing problem,with a large number of industrial non-linear load put into used,such as a large number of power electronic equipment put into use,the problem have seriously increased.And the harmonic make the power quality decline which bring problems to all walks of life.So the problem caused by harmonic should be need to consider seriously.Active Power Filter(APF)take harmonic suppression and reactive power compensation as the main functions by using the latest power electronics technology and control theory.However,the performance of APF depends on whether the harmonic current can be quickly and accurately detected.Firstly,this dissertation introduces the producing and harmful of harmonics,briefly explains some methods of eliminating harmonics,summarizes several existing mature harmonic detection techniques,and indicates that the harmonic current detection is a crucial part of the APF.Secondly,this dissertation briefly introduces the principle and structure of adaptive filter.On the basis of the theory of steepest descent method,makes formula derivation of two basic types of adaptive algorithms.Based on the adaptive interference cancellation principle,the interference cancellation model is applied to the adaptive harmonic current detection,and the LMS and RLS algorithms are used to detect the harmonic and analyze the advantages and disadvantages.At the same time,it explores the influence of simulation step size on convergence speed and steady-state error of LMS algorithm under the different step size,and leads to the thinking about variable steps.Then,by summarizing the shortcomings of the harmonic current detection method with the step-length LMS algorithm,the idea of variable step size is introduced.Based on the existing variable step size algorithm and summarize its advantages to propose a new variable step size algorithm.By using the time-averaged estimate of the system error as the dynamic factor,and combining the advantages of the class Sigmoid function,the fixed parameter value in the original algorithm is set as the dynamic parameter,so as to reduce the interference of the high-order harmonic current to the step.Also makes the parameters in the improved algorithm less set,so that the algorithm has a certain versatility and good stability,this method not only improves the convergence rate of the algorithm,but also ensures the steady-state error of the algorithm,and the algorithm can track mutations system very well.Finally,a new set member of affine projection algorithm(SM-APA)is proposed to reduce the computational complexity of the affne projection algorithm(AfA).The set member of affine projection algorithm reduces the computation complexity under the rapid convergence.The new algorithm used the method of weights updates selectively to reduce the corresponding calculation.At the same time,in order to improve the steady-state accuracy of the detection results,the variable step size idea is added to the SM-APA algorithm,the steady-state error is reduced while speeding up convergence,and the tracking ability of the algorithm is improved.