梯级水电站优化调度的改进粒子群算法

Improved Particle Swarm Optimization Algorithm for Optimal Scheduling of Cascade Hydropower Stations

针对粒子群算法易陷入局部最优的缺点,提出了一种双适应度方法、动态邻域算子和随机动态调整惯性权重机制有机结合的混合改进策略。算例计算表明,该改进策略能增强粒子的局部收敛能力,加快算法的收敛速度,便于处理复杂约束条件,为求解具有复杂约束条件的非线性规划问题提供了一种简单有效的方法。文中探讨了梯级水电站优化调度的相关问题,考虑了丰枯分时电价因素,建立了梯级水电站长期优化调度数学模型,并应用改进粒子群算法进行求解。实际梯级水电站计算表明,该模型使枯水期大部分时间出力均匀平稳,丰水期能兼顾防洪和蓄水的不同要求,有利于电力系统的稳定运行。改进粒子群算法计算速度快、收敛精度高,为梯级水电站长期优化调度提供了一种简单实用的求解方法。

To overcome the defect of local optimization of PSO, this paper presents a hybrid-advanced strategy, which combines dual adaptation degree method, dynamic neighborhood field operator and randomly dynamically adjusting inertia weight method. The simulation results show that the proposed strategy can increase the local convergence ability and accelerate convergence of particles, and thus, it is a simple effective approach to solve nonlinear programming problems with complex constraint conditions. This paper discusses the correlative problems in optimal scheduling of cascade hydroelectric power plants and establishes a long-term optimal scheduling model based on flood and dry time-of-use power prices. The computation results for actual cascade hydroelectric power plants show that this model can help coordinate power generation and water consumption, and decrease profitless spill water of cascade hydroelectric power plants. This model can keep up balanced power output in the dry season, and is propitious to stable operation of a power system. The advanced PSO algorithm is a simple effective approach to long-term optimal scheduling of cascade hydroelectric power plants.