基于混沌量子蜂群优化SVR的多峰MPPT算法研究  

Research on Multi-peak MPPT Algorithm Based on SVR Optimized by Chaotic Quantum-inspired Artificial Bee Colony

作　　者：高相铭[1] 杨世凤[2] 潘三博[1] 

机构地区：[1]安阳师范学院物理与电气工程学院,河南安阳455000 [2]天津科技大学自动化学院,天津300222

出　　处：《电气传动》2015年第12期55-60,共6页Electric Drive

基　　金：国家自然科学基金项目(U1204515);上海市教委科研创新项目(14YZ160)

摘　　要：针对局部阴影条件下光伏阵列P—V曲线具有多峰特性,而传统的MPPT算法无法有效跟踪其最大功率点的问题,研究了一种多峰MPPT算法。采用混沌搜索和量子计算对传统蜂群算法进行改进,并将改进后的算法应用于支持向量回归机(SVR)的参数寻优,最终实现了基于混沌量子蜂群优化SVR的多峰MPPT算法。通过仿真和实验对算法的有效性进行了验证。实验结果表明:该算法在均匀光照或局部阴影条件下均能够准确跟踪全局最大功率点,有效克服了传统MPPT算法容易陷入局部极值的问题。Owing to the P—V curve of photovoltaic arrays having the multi-peak characteristic that makes theconventional MPPT method failed under partially shaded conditions,a multi-peak MPPT algorithm was researched inorder to solve the maximum power point tracking problem effectively. The chaos searching and quantum computingmethod was used to improve the performance of artificial bee colony and the improved algorithm was adopted tooptimize the parameters of SVR. Finally,a multi-peak MPPT algorithm was established based on SVR（support vectorregression）optimized by chaos quantum-inspired artificial bee colony. The validity of algorithm was demonstrated bysimulation and testing. The experiment results show that the new MPPT algorithm could track the global maximumpower point accurately under uniform illumination conditions and partially shaded conditions,and sololve the problemseffectively that conventional MPPT algorithm was liable to trap in local extremum.

关 键 词：混沌量子蜂群算法 多峰特性 最大功率点跟踪 参数寻优 

分 类 号：TM615[电气工程—电力系统及自动化]