基于随机模型预测控制的能源局域网优化调度研究

Stochastic Model Predictive Control for Energy Management Optimization of an Energy Local Network

针对高可再生能源渗透率的能源局域网中可再生能源输出不确定性、负载需求波动性和实时电价随机性对能源局域网优化调度的影响,提出了基于随机模型预测控制的能源局域网优化调度方法。通过分析能源局域网系统架构及其各类设备的特性,构建了混合整数二次规划模型的能源局域网能量优化调度目标函数,并运用模型预测控制框架实现该能源局域网的在线优化调度。提出了能够快速实现场景缩减,选出典型场景的两阶段场景缩减方法。通过与传统开环调度方法以及标准模型预测控制调度方法的比较,证明文中所提方法的可行性与鲁棒性。

This paper proposed a stochastic model predictive control（SMPC） based energy local network（ELN） optimization and scheduling method for reducing the impacts introduced by the intermittent output of renewable energy resources, fluctuant of load demand and random real-time electricity price of a high renewable energy penetration level ELN. We constructed a mixed integer quadratic programming model as the objective function of the ELN operation by analyzing the features of all the elements in the ELN, and this optimization model can be online operated in a stochastic model predictive control（SMPC） framework. Forecast uncertainty for power production of renewable energy resources, real-time electricity price and load demand were described by scenarios, and a two-stage scenario cutting method was proposed to choose the typical scenarios. Simulation results show that the method proposed in this paper is effective and feasible by comparing with the open-looped operation method and MPC based operation method.