三峡梯级枢纽多目标生态优化调度模型及其求解方法  

Multi-objective optimization model for ecological operation in Three Gorges cascade hydropower stations and its algorithms

作　　者：卢有麟[1] 周建中[1] 王浩[1,2] 张勇传[1] 

机构地区：[1]华中科技大学水电与数字化工程学院,湖北武汉430074 [2]中国水利水电科学研究院水资源研究所,北京100038

出　　处：《水科学进展》2011年第6期780-788,共9页Advances in Water Science

基　　金："十一五"国家科技支撑计划资助项目(2008BAB29B08);水利部公益性行业科研专项经费资助项目(200701008)~~

摘　　要：针对三峡梯级枢纽综合效益的充分发挥及其对长江流域典型生态系统修复及持续改善的科学需求,通过分析发电效益与生态效益之间的制约竞争关系,以发电量最大和生态缺水量最小为目标建立了梯级电站多目标生态优化调度模型,对三峡梯级枢纽多目标生态优化调度进行了研究。同时,针对传统优化方法难以同时处理多个调度目标的固有缺陷,提出一种改进多目标差分进化算法对所构建模型进行高效求解。该方法针对差分进化算法在多目标协同优化和全局寻优能力等方面的不足,依据问题的特点重新设计了差分进化算法的进化算子,同时设计了一种多目标混沌搜索策略以加强算法的局部搜索能力。最后,依据多目标生态优化调度问题的特点设计了一种不需要设置惩罚因子的约束处理方法。通过三峡梯级枢纽多目标生态优化调度的实例应用,验证了本文所构建模型的合理性以及所提出算法的有效性和工程实用性。A multi-objective optimization model for ecological operation in Three Gorges cascade hydropower stations is proposed. The model considers the restrictive relationship between power generation and ecological restoration in Three Gorges, and balances the needs of competing interests of power generation and ecology. The objective functions are designed to maximize power generation while rrtinimizing ecological water requirements, for which a multi-objective differential evolution （DE） algorithm is proposed. The evolutionary operators in the DE algorithm are modified according to the inherent nature of multi-objective optimization problems. A local chaotic search strategy is designed to perform multi-objective optimization, thus solving the problem of premature convergence in the DE algorithm. A non- penalty-based constraint handling algorithm is proposed to handle complicated problems in multi-objective optimization for ecological operation. The model is applied to the optimized ecological operation in Three Gorges cascade hydropower stations. The result demonstrates the effectiveness and practicality of the proposed model in engineering applica- tions.

关 键 词：生态调度 多目标 差分进化算法 混沌序列 约束处理 三峡梯级 

分 类 号：TV697[水利工程—水利水电工程]