摘要
为提高风光储系统的可调度性,建立了风光储发电系统储能电站控制模型。结合风/光电站的运行实际,将风力、光伏发电的功率预测结果表示为模糊变量,给定实际功率数据曲线,以其短期预测曲线作为计划出力曲线,以实际出力曲线和计划出力曲线的均方根误差最小、预测合格率最大、充放电功率平缓及储能电量维持合理范围为目标,以风/光电站及储能电站发电限值为约束条件,通过相关机会目标规划方法对多目标任务进行追踪,利用模糊模拟及动态粒子群算法进行求解,依据超短期预测曲线进行滚动计算,得出控制时段内实际出力曲线。算例仿真结果表明,该储能控制策略符合风光储电站的运行实际,可使风光储发电系统很好地跟踪计划出力曲线。
To improve the schedulability of wind-photovoltaic-storage hybrid system,the control model of the windphotovoltaic-storage hybrid system is established.Combined with the actual operation of the wind-photovoltaic power station,the power prediction results of the wind-photovoltaic-storage power station are expressed as fuzzy variables.It’s actual power data curve is given,its short-term prediction curve is taken as the planned output curve.The objective is to minimize the root mean square error between the actual and the planned output curve,maximize the predicted qualification rate,slow charging and discharging rate,and maintain a reasonable range of energy storage,and generation limit of the wind-photovoltaic-storage power station and energy storage station is the constraint condition.The multi-objective task is tracked by the relevant opportunity target planning method.The fuzzy simulation and dynamic particle swarm optimization algorithm are used to solve the problem.The rolling calculation is carried out according to the ultra short-term prediction data,then the actual output curve in the control period is obtained.The simulation results show that,the energy storage control strategy is practical and the windphotovoltaic-storage hybrid system tracks the planned output curve well.
作者
牛瑞杰
郭俊文
李晓博
舒进
马骁骅
刘琳
NIU Ruijie;GUO Junwen;LI Xiaobo;SHU Jin;MA Xiaohua;LIU Lin(Xi’an Thermal Power Research Institute Co..Ltd.,Xi’an 710054,China;State Grid Henan Yanshi Electric Power Supply Branch,Yanshi 471900,China)
出处
《热力发电》
CAS
北大核心
2020年第8期150-155,共6页
Thermal Power Generation
基金
中国华能集团有限公司总部科技项目(HNKJ18-H34)。
关键词
风光储系统
储能
优化控制
模糊模拟
动态粒子群
相关机会目标规划
多目标
wind-photovoltaic-storage hybrid system
energy storage
optimized control
fuzzy simulation
dynamic particle swarm
relevant opportunity target planning method
multi-object
