基于预测控制的储能系统多时间尺度动态响应优化研究

Multi-time Scale Dynamic Response Optimization of Energy Storage System Based on Predictive Control

风电固有的随机性使其难以在长时间尺度上被准确预测,且预测误差随预测时间的延长而增大,导致用于抑制风电随机性的储能控制策略往往与实际需求偏差较大。为了有效应对风电的动态随机性,应用由不同响应能力及容量的储能单元组成的混合储能系统,提出一种基于多时间级分层思想与模型预测控制(model predictive control,MPC)理论相结合的混合储能分层MPC优化策略,实现对风电和储能的精准控制。在对风电进行周期预测的基础上构建储能分层动态响应优化模型,上层优化利用大功率慢速响应储能单元在长时间尺度上对风电进行能量转移,下层优化利用小功率快速响应储能单元在短时间尺度上平滑风电波动,且上层优化结果作为下层优化的状态变量。通过分层MPC,不同时间级的风电周期预测信息不断滚动更新,各层间的储能优化控制也可得到有效协调。针对微网中的风储联合发电系统,通过改变分层MPC的滚动时域及步长,得到适用于算例的最优参数。仿真结果验证了该策略可以动态响应风电变化,并有效改善风储系统输出的峰谷特性和波动性。

The inherent randomness of wind power makes it difficult to predict the amount of wind that will be stored in a turbine over a long span of time.The prediction error increases if the prediction time is extended,resulting in a large deviation between the power output after the application of the control strategy used to deal with wind power changes and the actual demand.In order to effectively cope with the dynamic randomness of wind power,this paper applies a hybrid energy storage system composed of energy storage units with different response capabilities and capacities.It proposes a hybrid energy storage hierarchical model predictive control (MPC) optimization strategy based on multi-time hierarchy and MPC,to achieve precise control of wind power and storage.Based on the periodic prediction of wind power,a layered dynamic response optimization model of energy storage is constructed.The wind power is transferred over a long time scale by an upper layer by using an energy storage unit with high power and slow response.The lower layer uses an energy storage unit with low power and fast response to smoothen the wind power fluctuations over a short time scale,and the results of the upper layer optimization are taken as the state variables of the lower layer optimization.Through the layered model predictive control,the predictive information of the wind power cycle at different time levels is updated continuously,and the optimal control of energy storage among layers can also be effectively coordinated.For the wind-storage-combined power generation system in the micro-grid,the optimal parameters are obtained by changing the rolling time domain and step size of the layered MPC.The simulation results confirm that the strategy can dynamically respond to wind power changes and effectively reduce the peak-to-valley characteristics and volatility of the output.