基于BP神经网络水质预测,引入"稀释比",建立了耦合水质目标的三峡水库非汛期多目标优化调度模型。基于加权和形式的凸模糊决策,将多目标转化为单目标,再应用遗传算法求解优化调度模型。优化计算结果给出了既能满足发电量需求又有改善支流水质作用的综合效益最优的三峡水库非汛期调度运行方式,从而有效地实现了多目标耦合的水库优化调度策略。
A multi-objective optimal dispatching model combined with water quality objectives are developed for the non-flood season operation of the Three Gorges reservoir by introducing a dilution ratio and adopting a back-propagation neural network water quality forecast model.This dispatching model is solved using a genetic algorithm,a transformation of multi-objectives into single objectives and a technique of weighted convex fuzzy decision-making.The optimization results with the model provide optimal schemes that not only meet the demand for power generation scheduling but also improve the water quality of the tributaries,so that the overall efficiency of the reservoir's non-flood season operation is optimal.Thus,this new model realizes the scheduling optimization strategy in effective multi-objectives coupling.