Adaptive Filtering Algorithm And Its Application In Power Harmonic Detection

Adaptive algorithms are filters that change their parameters over time and adjust their performance according to certain predetermined criteria.It is widely used in echo cancellation,channel equalization,signal prediction,interference suppression,and beamforming.Its performance is generally evaluated based on one or more of the following factors: 1.convergence rate,2.mean square error(MSE),3.robustness of value,4.computational complexity,5.tracking performance,and other indicators.Ideally,one would like to have an adaptive filter that is computationally efficient,robust,and convergent at the fastest rate,and that produces the lowest steady-state MSE possible.In addition,people also hope to design an easily implementable filter for low-cost,low-precision VLSI chips.However,in adaptive filter design,as with any engineering design problem,it is not possible to achieve all of the desired features at the same time,and there is actually a trade-off.For example,a Least Mean Square(LMS)adaptive filter is computationally simple and robust,but convergence is very slow,especially when a colored signal is input.Recursive Least Square(RLS)adaptive algorithm has fast convergence and reduces the steady-state MSE,but has high computational complexity and poor robustness.Therefore,researchers have been trying to improve the performance of different adaptive measurement filters.This article aims to study how to improve the performance of adaptive filtering algorithms and reduce the computational complexity of the algorithm.The adaptive filter algorithm proposed in this paper is applied to harmonic current detection.First of all,this paper briefly describes the basic structure and principle of the adaptive filter.For different situations,a series of algorithms based on adaptive filter performance improvement are proposed.It mainly includes the Convex Combine Proportionate Affine Projection Algorithm Based on Coefficient Difference(CDPAPA)algorithm and the variable Kernel Width Maximum Correntropy Criteria Subband Adaptive Filter(VKW).-MCC-SAF algorithm and Novel Affine Projection Sign Subband Adaptive Filter(NAPSSAF)algorithm.Then,APA(Affine Projection Algorithm Based on Dichotomous Coordinate Descent,DCD-APA)based on the bipartite coordinate descending method is introduced to solve the shortcomings of high complexity of Affine Projection Algorithm(APA).The algorithm maintains APA performance while the computational complexity is low and it is easier to implement hardware.However,constant step size DCD-APA contradicts between fast convergence and tracking performance.Therefore,this paper proposes a new variable step size DCD-AP(VSS-DCD-AP).The algorithm can not only obtain faster convergence speed,but also have lower steady-state error.Second,although the recursive least squares(RLS)algorithm has a faster convergence rate,the calculation is complicated.For this reason,the DCD method is introduced into the RLS algorithm,and a new variable forget factor(VFF)method is proposed at the same time,thereby constructing a new variable forgetting factor DCD-RDS(VFF-DCD-RLS)algorithm..The algorithm has strong tracking ability and low computational complexity.Finally,the harmonic current detection problem is discussed,and a common model of harmonic current detection in active power filter is introduced.We use the low-complexity adaptive algorithms proposed in this paper to solve harmonic detection problems.