基于粒子群优化自适应最小二乘法的电网动态谐波估计

发布时间：2018-01-06 08:41

本文关键词：基于粒子群优化自适应最小二乘法的电网动态谐波估计　出处：《深圳大学》2017年硕士论文　论文类型：学位论文

【摘要】：电力技术的快速发展,越来越多的非线性电子元件和设备应用到电力系统中,所产生的谐波问题成为主要的电能质量问题。为了解决电网谐波问题,国内外学者对此已经做了许多深入的研究,并取得许多斐然的成果。本文对国内外现有的电力系统的检测方法如:有傅里叶变换(FT)、快速傅里叶变换(FFT)、小波变换(wavelet transform)、人工神经网络、卡尔曼滤波(Kalman filter,KF)等方法进行了深入细致的研究,对它们的检测性能和适用性进行了探讨。本文首先给出了几种常用的电能质量检测方法,有效值法(root mean square,RMS)、傅里叶变换法、小波变换法、卡尔曼滤波算法和自适应最小二乘法(recursive least square,RLS),对各种方法进行了仿真分析并对各个检测方法的优缺点给出了分析和比较。然后,本文了研究基于粒子群优化自适应最小二乘法(particle swarm optimized recursive least square,PSO-RLS)的电网谐波估计方法。在研究和分析了RLS算法后,针对RLS算法的不足,本文提出利用粒子群(particle swarm optimization,PSO)算法求解自适应最小二乘法(RLS)所需的最优化的电网谐波参数,即状态向量的权重的初始值,在得到优化后的初始权重参数后再利用RLS算法对电网谐波参数进行参数估计。本文所提出的方法克服了自适应最小二乘法(RLS)对初始参数敏感的问题,优化了RLS算法的谐波估计效果。最后,本文应用PSO-RLS算法对静态和动态的电压信号进行仿真分析,并比较了不同的噪声环境下参数估计效果,此外,还应用本文所提方法对电网动态子谐波和间谐波进行了仿真分析。仿真结果表明,与可变约束最小二乘方法(Variable Constraint based Least Mean Square,VCLMS),遗传算法(Genetic Algorithm,GA)优化参数估计相比,本文所提方法估计效果更优。
[Abstract]:With the rapid development of power technology, more and more nonlinear electronic components and devices are applied to power system. The harmonic problem becomes the main power quality problem in order to solve the harmonic problem. Scholars at home and abroad have done a lot of in-depth research, and achieved a lot of remarkable results. In this paper, the existing detection methods of power systems at home and abroad, such as Fourier transform Fourier transform (FTFT). Fast Fourier transform (FFT), wavelet transform (WT), wavelet transform (WT), artificial neural network (Ann), Kalman filter (Kalman). KF) and other methods are studied in detail, and their detection performance and applicability are discussed. Firstly, several commonly used power quality detection methods are given in this paper. The effective value method is root mean square, Fourier transform and wavelet transform. Kalman filter algorithm and adaptive least square method recursive least squared RLS). All kinds of methods are simulated and analyzed, and the advantages and disadvantages of each detection method are analyzed and compared. In this paper, the particle swarm optimized recursive least square based on particle swarm optimization is studied. After studying and analyzing the RLS algorithm, the deficiency of the RLS algorithm is pointed out. In this paper, a particle swarm optimization is proposed. PSO) algorithm is used to solve the optimal harmonic parameters of the power network, that is, the initial value of the weight of the state vector, which is required by the adaptive least square method (RLSs). The RLS algorithm is used to estimate the harmonic parameters of the power network after the optimized initial weight parameters are obtained. The method proposed in this paper overcomes the adaptive least square method (RLS). Sensitive to initial parameters. The harmonic estimation effect of RLS algorithm is optimized. Finally, the static and dynamic voltage signals are simulated and analyzed by using PSO-RLS algorithm, and the effect of parameter estimation in different noise environment is compared. In addition, the dynamic subharmonics and interharmonics of the power network are simulated and analyzed by using the method proposed in this paper. The simulation results show that. And variable Constraint based Least Mean squared VCLMSs. Genetic algorithm (GA) is more effective than genetic algorithm in parameter estimation.
【学位授予单位】：深圳大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM714;TP18

【参考文献】