摘要
针对一维泥沙数学模型维数高、求解耗时长以及水沙联合调度模型多目标难以求解的问题,结合遗传算法与神经网络的特性,以发电量最大和有效库容最大为基本目标,构建了水库水沙联合优化调度模型。利用约束法和权重法,将多目标模型转化为单目标模型,采用加速遗传算法进行求解,其中泥沙冲淤量使用自适应BP神经网络进行拟合预测。三峡水库实例计算结果表明:运行20 a,与原设计运行方式相比,采用该优化调度模型优化运行年均发电量增加7.732%,泥沙淤积量增加0.044%,在淤积量增加很小的情况下能大幅度增加发电量,模型能较好地解决水库水沙联合调度问题,在工程实际中是有效可行的。
In order to solve problems including multi-dimensions and long solution times of a one-dimensional sediment model, multi-objectives and the solution difficulty of a water-sediment coordinative dispatch model, a water-sediment coordinated optimized dispatch model was established on the basis of characteristics of genetic algorithms and neural networks. The model took the maximum power generation and effective storage as the elementary objects. The multi- objective model can be transformed into a single-objective model with the method of restricting and weighting. The single- objective model is solved using the method of accelerating genetic algorithm. The adaptive BP neural network was used to fit the prediction for the sedimentation and scour. Three Gorges Reservoir case results show that the average annual electric energy production and the sedimentation and scour increase by 7. 732% and 0. 044% , respectively, compared to those of the original design after the project operated for 20 years. The power generation can be markedly increased in the case of little sedimentation and scour. The model can solve water-sediment coordinative optimized dispatch well and is also effective in engineering practice.
出处
《水利水电科技进展》
CSCD
北大核心
2013年第2期9-13,共5页
Advances in Science and Technology of Water Resources
基金
国家自然科学基金(51179069)
关键词
水沙联合优化调度
加速遗传算法
自适应BP神经网络
三峡水库
water-sediment coordinative optimized dispatch
accelerating genetic algorithm
adaptive BP neural network
Three Gorges Reservoir