混合有源滤波器参数的多目标优化设计

发布时间：2018-12-17 00:38

【摘要】：智能电网的高速发展以及非线性装置的使用给日常生产、生活带来便利的同时也造成了严重的谐波危害,使得电能质量下降,因此谐波的治理尤为重要。目前治理谐波的主要手段是投入滤波装置,而混合有源滤波器(Hybrid active power filter,HAPF)能防止波形严重畸变并且提供适量的无功补偿,是其中最具前瞻性的滤波系统。分析山西电网的谐波污染情况,并考虑到HAPF参数优化时存在多目标、非线性的特点,本文将适用于工程优化的多目标布谷鸟搜索算法(Cuckoo Search algorithm for Multi-objective optimization,MOCS)应用在变电站挂网运行的500kVA的HAPF中,以实现其参数优化设计。掌握MOCS算法的工作原理及具体流程的基础上,分析了它在多目标参数优化问题中的不足。针对该算法中大量聚集在边界上的解会影响迭代后期收敛速度的缺陷,引入高斯扰动算子对布谷鸟巢位置进行改进,并对超出边界范围的鸟巢进行越界处理,提出了高斯扰动边界策略MOCS算法。针对标准MOCS算法易陷入局部最优的问题,采用适于全局优化的混沌理论及云模型来进行改进,提出了混沌云模型MOCS算法。然后利用典型的多目标基准函数对两种改进算法分别进行性能测试,实验结果显示出改进后各方面的优势。对并联型注入式基波谐振HAPF参数进行设计并说明每个模块的参数对HAPF滤波性能的影响。以经济性指标、满意度函数以及无功补偿为HAPF的多目标函数,直流侧电压及开关频率为约束条件,用两种改进的MOCS算法对其进行参数优化。利用Matlab/Simulink对HAPF进行仿真,与传统的优化设计方法对比,验证本文两种改进算法与HAPF优化配置模型的实用性与经济性。
[Abstract]:The rapid development of smart grid and the use of nonlinear devices bring convenience to daily production and daily life, but also cause serious harmonic harm, which makes the power quality decline, so the harmonic management is particularly important. At present, the main method of harmonic control is to put in the filter device, and the hybrid active power filter (Hybrid active power filter,HAPF) can prevent the serious distortion of the waveform and provide appropriate reactive power compensation, so it is the most prospective filtering system. The harmonic pollution of Shanxi power network is analyzed. Considering the multi-objective and nonlinear characteristics of HAPF parameter optimization, this paper presents a multi-objective Cuckoo search algorithm, (Cuckoo Search algorithm for Multi-objective optimization, which is suitable for engineering optimization. MOCS) is applied to the HAPF of 500kVA in substation to realize the parameter optimization design. On the basis of mastering the working principle and concrete flow of MOCS algorithm, the deficiency of MOCS algorithm in multi-objective parameter optimization problem is analyzed. In order to overcome the defect that a large number of solutions gathered on the boundary in the algorithm will affect the convergence rate in the late iteration period, Gao Si perturbation operator is introduced to improve the location of the cuckoo nest, and the nest beyond the boundary range is treated across the boundary. The MOCS algorithm of Gao Si perturbation boundary strategy is proposed. In order to solve the problem that standard MOCS algorithm is prone to local optimization, chaotic cloud model MOCS algorithm is proposed by using chaos theory and cloud model, which is suitable for global optimization. Then the performance of the two improved algorithms is tested by using the typical multi-objective benchmark function. The experimental results show the advantages of the improved algorithm. The HAPF parameters of the parallel injection fundamental resonance are designed and the influence of the parameters of each module on the HAPF filtering performance is explained. Taking the economic index, satisfaction function and reactive power compensation as the multi-objective function of HAPF, DC side voltage and switching frequency as constraints, two improved MOCS algorithms are used to optimize the parameters. The simulation of HAPF using Matlab/Simulink is compared with the traditional optimization design method, which verifies the practicability and economy of the two improved algorithms and the optimal configuration model of HAPF.
【学位授予单位】：太原科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM761

【参考文献】