水库多目标调度模型及算法研究

发布时间：2018-05-03 05:22

本文选题：水库调度 + 多目标模型　；参考：《水力发电学报》2017年12期

【摘要】：协调水库发电、灌溉和生态环境效益是众多水库决策者面临的棘手问题。本文以考虑发电保证率要求的总发电量最大、下游灌溉和生态流量需水满足度最大为目标建立水库多目标调度模型,以动态规划和离散微分动态规划为计算核心,在分别对发电保证率、灌溉需水满足度和生态环境需水满足度建立惩罚函数的基础上,提出求解多目标模型的变惩罚系数法(VPC),并应用VPC法进行模型求解,通过筛选得到多目标非劣解的Pareto前沿。通过与非支配排序遗传算法II(NSGA-II)计算结果进行对比,表明VPC算法可以得到更高质量的非劣解集。实例研究表明,该模型和求解方法具有较好的适应性和较强的实用价值,可以为水库调度管理者提供重要的决策参考。
[Abstract]:Coordination of reservoir power generation, irrigation and ecological environmental benefits is a thorny problem faced by many reservoir decision makers. In this paper, a multi-objective model of reservoir operation is established with the goal of maximum total electricity generation considering the requirement of power generation guarantee rate and the maximum water requirement of downstream irrigation and ecological flow, with dynamic programming and discrete differential dynamic programming as the core of calculation. On the basis of establishing penalty function of power generation guarantee rate, irrigation water demand satisfaction and ecological environment water demand satisfaction, the variable penalty coefficient method for solving multi-objective model is put forward, and the model is solved by VPC method. The Pareto frontier is obtained by screening the multiobjective noninferior solution. By comparing the results with the non-dominated sorting genetic algorithm II NSGA-II, it is shown that the VPC algorithm can obtain a higher quality non-inferior solution set. The case study shows that the model and the solution method have good adaptability and strong practical value, and can provide important decision reference for reservoir dispatching managers.
【作者单位】：武汉大学水资源与水电工程科学国家重点实验室;
【基金】：国家重点研发计划(2016YFC0401309;2016YFC0400907) 国家自然科学基金(51479141)
【分类号】：TV697.1

【相似文献】