采用改进粒子群算法的微电网短期调控模型

Short-Term Scheduling Model for Micro-Grids With Improved Discrete Particle Swarm Optimization

现有的大部分微电网调控模型以预测发电功率作为调控目标,且在进行成本核算时没有考虑电池的状态变化带来的影响,致使调控周期长、调控策略经济性差且易受功率预测误差影响。针对上述问题,基于对风力发电、光伏发电和储能设备的成本分析,提出了一种以机组的启停为策略、以成本最小为目标的微电网短期调控模型。为寻求最优的调控策略,在保留传统粒子群速度更新方法的基础上,修改了位置更新方法并引入惩罚函数,提出一种改进的离散粒子群算法(discrete particle swarm optimization-Ⅱ,DPSO-Ⅱ)。仿真和实际算例结果表明,所提模型能够达到预期的调控目标,且具有较好的经济性和鲁棒性;所提DPSO-Ⅱ算法的寻优性能较传统基于遗传算法和离散粒子群算法有较大提高,因而具有潜在的应用价值。

While most existing scheduling models for micro-grid target to matching predicted power generation and neglect influence of battery capacity state, leading to long regulation cycle, big regulation error and difficult implementation, this paper proposes a short-term scheduling model for micro-grids consisting of wind and PV generations and batteries, aiming at minimizing total cost of operation, maintenance and start-stop. To find optimal strategy, a discrete particle swarm optimization algorithm with penalty function（DPSO-Ⅱ） is proposed. Simulation results show that DPSO-Ⅱ out-performs genetic algorithms and DPSO algorithms in optimization ability and speed. The proposed scheduling scheme meets balance requirements between supply and demand with better economical efficiency and robustness, and thus has potential value in practice.