基于遗传规划的符号回归方法在电能质量分析中的新应用  

New application of symbolic regression method based on genetic programming in power quality analysis

作　　者：于鹏[1] 

机构地区：[1]渤海大学,辽宁锦州121013

出　　处：《电子设计工程》2013年第7期20-23,共4页Electronic Design Engineering

摘　　要：谐波分析是电能质量分析的一个重要问题。Fourier变换与小波变换各有其局限性。文中提出一种采用基于遗传规划的符号回归方法进行电量分析的新方法。文章通过对传统研究方法的分析提出新方法。结合一个具体实例证明这种方法的可行性。设计了全桥逆变仿真实验用于产生分析数据,运用Eurqa软件采用符号回归的方法进行遗传选择计算得到了正弦分量与畸变分量的表达式,从而验证了符号回归应用于电能质量分析的实用性。与传统方法相比,该方法时间信息损失少,函数关系明确,在电能质量分析领域具有广泛的适用性和较好的应用潜力。Harmonic analysis is an important question in power quality analysis. Fourier transform and small wave transformation both have limitation. Through the analysis of traditional method, a new method of electrical variable analysis is proposed in this paper. The new method uses the way of symbolic regression based on genetic programming. An example is used to illustrate the practical use of the method. A H-bridge simulation model experiment is designed to generate the data for analysis. The expression of sinusoidal wave form and odd wave form is got through genetic regression computing using the software of Eurqa. The practical application of symbolic regression on power quality analysis is proved. To be compared with traditional method, the new method is low information loss and high clearance of expression. The new method will have widely usage and large potential in the field of power quality analysis.

关 键 词：符号回归 谐波 遗传规划 电能质量 

分 类 号：TM712[电气工程—电力系统及自动化]