基于SVM-MOPSO混合智能算法的配电网分布式电源规划

摘要

针对分布式电源(Distributed Generation，DG)并网给电力系统带来的随机扰动，综合考虑配电网运行效益，计及风光时序特性，以经济性、电能质量及环保性为目标，搭建了机会约束规划模型。采用混合智能算法求解，即基于支持向量机(Support Vector Machine，SVM)算法模拟优化变量到目标函数以及约束条件映射的不确定性函数，运用多目标粒子群算法(Multi-Objective Particle Swarm Optimization，MOPSO)求解模型，得出 Pareto 非劣决策集并给出典型解及理想解。算例结果表明，该规划方法考虑到 DG 的随机性特征、时序特性和并网概率分布，能提高算法执行效率，证明了所提方法的合理性和有效性，且Pareto前沿的引入，给决策者充分选择空间，更具有工程性。%Regarding stochastic disturbance in power system brought by grid-connected distributed generation (DG), generally considering operational effectiveness, along with timing characteristics of wind speed and sunlight intensity, taking economy, power quality and environmental efficiency as goals, the optimization model of stochastic chance-constrained programming is built. The hybrid intelligent algorithm is used, which simulates the uncertainty functions based on support vector machine (SVM) and solves the model by multi-objective particle swarm optimization (MOPSO), and then the Pareto non-inferior decision set is obtained. Simulation results show that the planning model can fully take into account randomness, timing characteristics and grid-connected probability distribution of DG, and improve the efficiency of the algorithm, then verify the rationality and validity of the proposed approach. Moreover, the introduction of Pareto front gives fully choices to policymakers and possesses more engineering value.