基于神经网络PID控制的聚合温控负荷优化调度  

Optimal scheduling of aggregated thermostatically controlled loads based on neural network PID control

作　　者：马巨海[1] 焦宗旭 徐程琳 杨婕[1] 马锴[1] MA Juhai;JIAO Zongxu;XU Chenglin;YANG Jie;MA Kai(School of Electrical Engineering,Yanshan University,Qinhuangdao,Hebei 066004,China)

机构地区：[1]燕山大学电气工程学院

出　　处：《燕山大学学报》2019年第5期417-422,共6页Journal of Yanshan University

基　　金：国家重点研发计划资助项目(2018YFB1702300);国家自然科学基金资助项目(61973264,61573303);河北省自然科学基金资助项目(E2017203284);中国博士后科学基金资助项目(2016M601282)

摘　　要：本文设计了基于BP神经网络的温控负荷控制系统,实现了智能电网中的需求侧频率调节辅助服务。针对控制系统中的控制器部分,设计了基于神经网络的PID控制器,通过BP算法对控制器的3个参数进行实时在线优化调整,并利用温控负荷的温度设定值调节策略,减小了聚合温控负荷的跟踪误差,改善了系统的性能,提升了频率调节服务质量。仿真结果表明,所设计的神经网络PID控制器可以有效地提高频率调节信号的跟踪精度。In this paper,a control system of thermostatically controlled loads based on back propagation algorithm (BP) neural network was designed,which can realize the frequency adjust service in the smart grid.For the controller part of control system,the proportion-integral-differential (PID) controller based on neural network is designed,and the BP algorithm is used to adjust and optimize the parameters of the controller.The objective is to reduce the tracking error,improve the performance of the system,and improve the quality of frequency regulation service by using the temperature setting adjustment strategy.The simulation results show the effectiveness of the designed neural network PID controller.

关 键 词：温控负荷 频率调节 BP网络 PID控制器 智能电网 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]