基于改进蝙蝠算法的梯级水电站经济调度

Improved Bat Algorithm for Economic Dispatch in Cascade Hydropower Stations

梯级水电站中长期经济调度是一个典型非线性优化问题,通常要求在满足复杂的水力、电力约束条件,兼顾求解效率的同时实现梯级发电量最大。为有效解决这一问题,通过改进标准蝙蝠算法(bat algorithm,BA)更新策略和引入差分进化算法(differential evolution,DE)变异、选择操作,提出一种改进的蝙蝠算法(improved bat algorithm,IBA)。对标准蝙蝠算法更新策略进行以下改进:1)蝙蝠个体脉冲频率不随种群迭代而更新;2)蝙蝠个体脉冲发射率和脉冲音量随种群迭代而更新;3)无条件接受全局搜索产生的新解,有条件接受局部搜索产生的新解;4)改进飞行速度公式,缩小新个体与当前种群最优个体的偏离值。同时,针对蝙蝠算法种群多样性差、易陷入局部最优的缺点,引入差分进化算法中的变异、选择操作,实现动态控制变异概率。建立兼顾梯级最小出力最大化的梯级总发电量最大模型,利用大渡河流域瀑布沟、深溪沟、枕头坝一级梯级水电站经济调度问题实例,从流域长系列径流资料中选取典型年,对IBA的主要控制参数(缩放因子、最大迭代次数)进行测试与分析。采用IBA、BA、逐步优化算法(POA)对同一典型年进行模拟调度。从枯期出力特征、梯级发电量、算法运行时间3项指标综合来看,对于复杂的梯级水电站经济调度问题,改进的蝙蝠算法能够在枯水期给电网提供尽可能大而稳定的出力,同时缩短计算时间,获得更高精度解。

The mid-long term economic scheduling of cascade hydropower stations is a typical nonlinear optimization problem. It is usually re- quired to maximize the total power generation giving consideration to efficiency of solution while simultaneously meeting complicated hydraulic and electrical constraints. To solve this problem effectively, an improved bat algorithm （IBA） was proposed by improving the updating strategies in Bat Algorithm and introducing the mutation and selection operation in Differential Evolution algorithm. The following improvements were made to the updating strategies in BA： 1） The pulse frequency of each bat was not updated with the population iteration. 2） The impulse emission rate and the pulse volume of each bat were updated with the population iteration. 3） New individuals generated by global search were accepted un- conditionally in IBA, but individuals generated by local search were accepted conditionally. 4） The flight velocity formula was improved to re- duce the deviation between the new individual and the optimal individual of the current population. In order to increase the diversity of the population and to avoid local optimum, the mutation and selection operation in DE was introduced, and the mutation probability was thus dynamic. A model was established to maximize the annual energy output and maximize the minimal output of cascaded hydropower stations. Based on the economic dispatch problem of cascade hydropower stations of Dadu River, the main control parameters （zoom factor and maximum iterations） of IBA were tested and analyzed. IBA, BA and POA were used to simulate the scheduling at the same year. Taking account to the characteristics of output in the dry season, the total power generation and the running time, we can conclude that compared with the other two algorithms, the simulation results obtained by using IBA verified its superiority in both efficiency and precision.