基于鱼群算法优化ELM的电力电容器组电容量预测研究  

Study on Capacitance Prediction of Power Capacitor Bank Based on Artificial Fish Swarm Algorithm

作　　者：葛文红 徐佳 汪悦生 王明刚 颜平丽 吴静 GE Wenhong;XU Jia;WANG Yuesheng;WANG Mingang;YAN Pingli;WU Jing(State Grid Huaibei Power Supply Company,Anhui Huaibei 235000,China;State Grid Anqing Power Supply Company,Anhui Anqing 246003,China)

机构地区：[1]国网淮北供电公司,安徽淮北235000 [2]国网安庆供电公司,安徽安庆246003

出　　处：《电力电容器与无功补偿》2018年第3期11-15,32,共6页Power Capacitor & Reactive Power Compensation

摘　　要：为实现10 kV变电站电容器组电容量的准确预测,提高电容器的使用寿命和输变电系统的稳定性,文章提出一种基于AFSA优化ELM的权值和偏置的10 kV变电站电容器组电容量预测模型。该模型将AFSA算法应用于优化ELM的输入权值和隐含层偏置,实现ELM最优化预测。通过对PSO-ELM、GA-ELM和DE-ELM 3种方法的对比分析可知,AFSA-ELM进行10 kV变电站电容器组电容量预测具有预测精度高的优点,该方法可为电容器组电容预警提供决策依据和参考。For achieving accurate capacitance prediction of capacitor bank for 10 kV substation,improving the operation life of the capacitor and stability of the power transmission and transformation system,in this paper,a kind of capacitance prediction model of capacitor for 10 kV substation based on AFSA optimized ELM weight and offset is proposed.In this model,the AFSA algorithm is used to optimize input weight and hidden layer offset of ELM so to achieve ELM’s most optimization prediction.It is known through comparison and analysis of such three methods as PSO-ELM,GA-ELM and DE-ELM that capacitance prediction for capacitor of 10 kV substation with AFSA-ELM has the advantages of high prediction accuracy and the method and the method can provide decision base and reference for the capacitance prediction of capacitor bank.

关 键 词：鱼群算法(AFSA) 极限学习机(ELM) 遗传算法 电容器 差分进化算法 

分 类 号：TM53[电气工程—电器] TP18[自动化与计算机技术—控制理论与控制工程]