基于可行域搜索映射的并行动态规划

为解决传统动态规划在处理水库群联合优化调度时面临的约束处理机制选择和计算时间长的问题,引入映射思想,基于映射和集合论知识构建可行域搜索映射模型,并结合动态规划的并行性,提出基于可行域搜索映射的并行动态规划。该算法通过构建时段可行搜索空间和并行模式,以规避无效状态组合计算并充分发挥计算机多核优势,提高计算效率。以李仙江流域三库联合调度为实例进行研究,从年发电量、计算耗时等方面将改进算法与传统动态规划以及逐步优化算法(POA)进行详细的对比分析。结果表明,该算法能在保证解全局收敛性的前提下减少计算耗时,制定梯级水库最优调度策略。     

基于动态规划的紫坪铺水库发电优化调度模型研究

紫坪铺水库是我国特大型都江堰灌区的水源工程,担负着成都、德阳、绵阳等7市37县(区)综合供水和成都地区的防洪任务,兼有发电、生态环境保护等综合效益,直接受益人口2600多万人,是我国西部大开发首批十大标志性工程之一。针对紫坪铺水库发电优化调度问题,按照实际调度情况,运用动态规划算法,建立了以防洪、供水、工程安全运行为约束条件,以发电效益最优为目标函数的紫坪铺水库发电优化调度模型,模型可为紫坪铺水库发电优化调度提供决策依据。     

基于天然水文情势的水库调度规则研究

为了在水库生态调度中模拟天然水文情势,基于水文情势的主要特征设计了15个水文情势评价指标,并将某应用于水库生态调度模型的目标函数中;基于水文分期特征将水文年划分为不同时段,提出了参考各时段实时可利用水量的水库生态供水规则;通过分层分解水库不同供水任务,降低了水库生态调度模型求解的复杂程度,并将模型应用于万家寨水库。结果显示:①本文设计的水文情势评价指标及其在水库生态调度模型中的应用减少了生态供水与天然水文情势间的差异;②基于水文分期和实时水情的水库生态供水规则有效挖掘了生态供水潜力;③为了维护河流健康,万家寨水库上游水资源开发利用率不宜超过40%,而水资源开发利用率长期高于60%则会造成严重的生态破坏。     

综合利用水库水量水质统一调度模型

水库水量水质统一调度是水资源可持续管理的一项重要内容。本文以河北省洋河水库为例，建立了引水、供水、灌溉、防洪等综合利用的水库水量水质统一调度大系统多目标分解聚合模型，给出了变动惩罚系数多维多目标离散微分动态规划与自优化模拟技术相结合的兴利调度方法、变罚系数离散微分动态规划防洪调度模型和水库富营养化模拟算法。     

混合粒子群算法在水库中长期发电优化调度中的应用

针对传统粒子群算法(PSO)早熟收敛、局部搜索能力不足等问题,提出一种混合粒子群算法(HPSO)求解水库优化调度问题。该算法引入混沌思想的遍历性特点生成初始解,提高初始种群质量;采用自适应惯性权重法平衡个体搜索行为,避免陷入局部最优。百色水库中长期发电优化调度实例研究表明,HPSO比PSO有更好的收敛性和优化结果,同时计算时间较传统动态规划方法大幅缩减,且优化结果相近,是一种水库优化调度可供选择的计算方法。     

沭水东调工程跨流域水库群联合供水研究

在分析日照市沭水东调工程特点的基础上,以城市供水量最大、水库弃水量最小为目标建立了多目标跨流域水库群联合供水调度模型。同时,将遗传算法嵌套于长系列变动时历法中,对工程涉及的沭河流域青峰岭、小仕阳、峤山水库以及傅疃河流域日照水库进行联合供水调算,探讨沭河流域各水库的可调限制库容和日照水库需水动态限制库容等调水指标。结果表明,水库群联合供水较各水库单独供水可增加利用沭河水量939万m3,反映了水库群联合调度有助于充分挖掘沭河流域各水库的供水潜力,对解决日照市区供水缺口十分有利。     

基于改进粒子群算法的多水库复杂联合供水优化调度研究

丹凤县县城不断增长的用水需求与工程供水能力之间的矛盾日益突出,为了充分挖掘工程供水的潜力,解决现有供水工程运行中弃水多、供水保证率低的问题,以弃水量最小为目标建立模型,采用改进粒子群算法对青峰水库和龙潭水库的联合供水进行优化求解。以月为计算时段,在优先考虑河道生态流量、蒸散发和渗漏损失的情况下,针对青峰水库200×10⁴ m³和220×10⁴ m³两种拟定兴利库容方案情景进行模拟。结果表明：迭代次数达36次时目标函数趋于稳定;与常规调度相比,两库的年平均弃水量明显减少,供水保证率有所提高;采用线性递减惯性权重改进的粒子群算法在青峰和龙潭两水库联合优化供水计算中具有优越性和可行性。     

多水库联合供水的优化调度方法

针对大中型水库增加城市供水项目的实际情况,运用运筹学的理论与方法,建立了多水库联合运用情况下,系统供水优化调度的线性规划模型.该模型以多水库年总供水效益最大为目标函数,以月为计算时段,以各月灌溉和城市供水量为决策变量,以水库水量、供水需求等为约束条件.对主要优化参数各月灌溉和城市供水的单方效益等进行了研究.利用该方法,可避免用动态规划法形成的"维数灾".     

基于引水限制调度线的并联水库系统供水优化算法及应用

以实现梯级水库群或者水源连通工程水量补偿式调节为目标,针对并联水库的联合供水特征,采用引水限制调度线控制引水水库可引水量,引入粒子群算法构建了并联水库优化调度模型。通过优化引水调度线提出可以改进并联水库系统的调度规则和调度方法,将优化和常规调度有机结合,并根据并联水库系统实际运行条件设置模型的目标函数、约束条件和参数设置方法。以X市S水库和L水库组成的并联水库系统为实例,采用该模型进行优化模拟。结果表明:并联水库系统经优化年均弃水量比两库单独供水的弃水量减少5 496万m~3,末库容比两库单独供水时提高203万m~3。在95%供水保证率时,并联水库系统供水系统的年供水能力比双库单独供水能力之和提高了5 556万m~3。优化形成的引水限制调度线分布形态合理,并联供水系统经过优化大幅度提高了水库系统的供水能力,证明了基于引水限制调度线的并联水库系统供水优化算法的可行性,该方法对并联水库联合调度具有重要参考价值。     

基于改进粒子群算法的水库优化调度研究

在分析以往水库优化调度模型优缺点的基础上,提出了基于动态调节惯性权重的粒子群优化方法的水库优化调度模型,对基本粒子群算法进行了改进。改进的算法通过时变权重的设置来实现,从而克服了PSO搜索精度不高,易陷入局部最优的缺点,并通过引入罚函数解决强约束问题。以某综合利用水库优化调度为实例进行研究,并与动态规划模型计算结果进行对比分析,实例计算表明:改进PSO算法原理简单,易于编程实现,而且占用计算机内存小,收敛速度快,搜索效率高,能以较快的速度收敛到全局最优解,是一种有效的搜索算法。     

寺坪水电站优化调度模型

随着社会经济的快速发展,水资源供需矛盾日趋尖锐,如何合理调度有限的水资源已成为水资源管理中的现实而紧迫的任务。通过进一步研究水库调度的机理,采用非线性规划作为优化求解方法构建了水库优化调度模型,非线性规划结合了线性规划和动态规划各自的优点,并将水库调度中的各种因素融入数学模型中,较准确地计算模拟期间各种优化变量数值条件下的目标函数值,并比较得出最优值。之后,基于多目标思想,给出了一组Pareto前沿解集,通过寺坪水电站的应用,构建不同的目标函数,得到一系列决策方案,以便决策者选择偏好的决策方案。     

跨流域供水水库群联合调度规则研究

针对跨流域供水水库群联合调度存在的主从递阶结构,提出了调水规则和供水规则相结合的跨流域供水水库群联合调度规则。其中,调水规则由一组基于各水库蓄水量的调水控制线表示,根据其间的相对位置关系,决定是否调水,调水量如何分配等;供水规则由各库供水调度图表示,对应于不同用水户的限制供水线将水库的兴利库容分为若干调度区。建立了适合于主从递阶结构的水库群联合调度二层规划模型,采用并行种群混合进化的粒子群算法对模型进行求解。中国北方某大型跨流域调水工程的实例研究证明了模型的合理性和有效性。     

长江上游控制性水库群联合调度及水资源影响分析

针对长江上游控制性水库群联合调度问题,建立了大规模混联水库群联合优化调度模型,并提出离散微分动态规划(DDDP)和逐步优化算法(POA)相结合的混合方法,实现大规模混联梯级水库群联合优化调度问题的高效求解。在此基础上,结合流域长系列历史径流资料,进行了长江上游控制性梯级水库群调度模拟,分析了联合调度的发电效益;并在此基础上,结合相关研究成果,探究并分析了梯级水库群建成投运后,联合调度对流域水资源的影响。成果表明,梯级水库群的建成及联合调蓄对于长江中下游枯水期的流量补偿效益十分明显,供水、航运以及压咸补淡等综合效益十分显著。     

水库供水发电多目标优化调度模型及应用研究

提出了一种研究确定水库最优供水量的多目标水库优化调度数学模型,此模型以供水量最大和发电量最大为目标函数,考虑水量平衡、防洪、发电、航运及水库综合利用要求约束条件.提出一种交互式的求解方法对模型进行解算,此方法首先采用约束法,通过松驰供水量最大目标将多目标模型转换成多个单目标模型,在单目标模型中引入2个参数,用于调整模型的计算结果.对单目标模型,采用动态规划法求解,求得多目标模型的不劣解集.提出一个用于选择多目标最优解的决策偏好系数,对多目标模型的不劣解集进行对比分析,从而确定多目标模型的最优解.应用所提出的方法研究了广东省白盆珠水库的调度方案,提出了白盆珠水库的最优供水方案.     

Reservoir Operation Optimization: A Nonstructural Solution for Control of Seepage from Lar Reservoir in Iran

Abstract

In this paper, the results of utilizing a deterministic dynamic programming model for operation of Lar Reservoir in Iran are discussed. This reservoir has experienced extensive seepage from the start of its operation. The optimization model consists of a three-step cycle, which began with the optimization of reservoir operation for a given set of streamflows. The optimal policies are then analyzed in a regression procedure to obtain a set of operating rules. After the first run, operating rules from the previous run were placed as a new constraint on the water releases with some pre-assigned tolerance and the cycle continues. The model also consisted of mathematical functions for modeling the seepage from Lar Reservoir as a function of storage head in the reservoir. The loss function in the model was also modified in order to incorporate parameters that reduce the seepage. Results of different scenarios showed the significant effect of optimal policies on reduction of seepage and increasing the reliability of water supply to Tehran Metropolitan Area. A pumping station was also proposed to utilize the inactive part of the reservoir, in access of over 100 MCM, in order to reduce the seepage. The effectiveness of different pumping capacities to reduce the seepage was also investigated.     

A Novel Optimization Method for Multi-Reservoir Operation Policy Derivation in Complex Inter-Basin Water Transfer System

Joint operation of multiple reservoir system in inter-basin water transfer-supply project is a complex problem because of the complicated structure and cooperated operation policy. The combination of high-dimensional, multi-peak and multiple constraints makes it incredibly difficult to obtain the optimal rule curves for multi-reservoir operation. In view of this, we constructed a joint optimization operation model, considering both water supply and transfer, and proposed the concept of “shape constraints”. To obtain the solution of this high-dimensional optimization model, a novel progressive optimum seeking method, namely Progressive Reservoir Algorithm-Particle Swarm Optimization (PRA-PSO), is presented based on the nature of progressive optimization algorithm (POA) and standard particle swarm optimization (PSO). The water transfer project in northeast China, consisting of three routes eight reservoirs, is selected as a case study. The results show that (1) PRA-PSO is yielding much more promising results when compared with other optimization techniques; (2) shape constraints would narrow the scope of feasible solution area but increase the convergence of algorithm; (3) because of the strong interaction between water transfer and water supply action, the progressive setting of PRA-PSO should be in accordance with the order of reservoir water transfer. The case study indicates the novel optimization method could effectively increase the chance of jumping out of local optimal points, thereby searching for better solutions.     

Water Transfer Triggering Mechanism for Multi-Reservoir Operation in Inter-Basin Water Transfer-Supply Project

This paper proposes a new water transfer triggering mechanism for multi-reservoir system to divert water from abundant to scarce regions with a constant diversion flow in an inter-basin water transfer-supply project. Taking into account of the uncertain nature of inflow, the storage of reservoir is taken as a signal for decision-making to indicate water abundance or water scarcity. In this study, a set of rule curves based on storage of donor reservoir and storage of recipient reservoir are used together to determine when to start water transfer. To initiate water diversion to each recipient reservoir effectively, several water transfer rule curves of the donor reservoir are set for each recipient reservoir respectively in the multi-reservoir system with one donor reservoir and several recipient reservoirs, which is the main difference in comparison with other water transfer triggering mechanisms. In addition, a systematic framework is developed to integrate the water transfer rule curves with hedging rule curves to simultaneously solve the water transfer and water supply problems, since they interact with each other during the operation process. In order to verify the utility of the new water transfer triggering mechanism, an inter-basin water transfer-supply project in China is used as a case study and an improved particle swarm optimization algorithm (IPSO) with a simulation model is adopted for optimizing the decision variables. The results show that the proposed water transfer triggering mechanism can improve the operation performances of the inter-basin system.     

A Factorial-based Dynamic Analysis Method for Reservoir Operation Under Fuzzy-stochastic Uncertainties

In this study, a factorial-based fuzzy-stochastic dynamic programming (FFS-DP) method is developed for tackling multiple uncertainties including fuzziness, randomness and their interaction in reservoir operation management (ROM). FFS-DP is framed on the integration of stochastic dynamic programming, fuzzy-Markov chain, vertex analysis and factorial analysis techniques. It can not only deal with the conventional optimization problem for reflecting dynamic and uncertain features in ROM, but also obtain detailed effects of uncertain parameters and their interactions on the system performance. The developed method is applied to a case study of a reservoir operation system, where the local authority is in charge of allocating relative scant water to the downstream municipality. The results obtained can help the local authority identify desired water release policies under uncertain system conditions. Besides, the results simultaneously indicate that significant factors and their interactions can be identified in ROM. Moreover, the results can be further analyzed for generating optimal parameter inputs to obtain maximized system benefits.     

考虑调水和供水规则的水库优化调度

针对用于调节当地地表水资源的水库受多水源影响需要制订新调度规则的问题，考虑非常规水调蓄能力低的特点，在优先配置非常规水的基础上对受水区水库设置调水和供水规则，建立了考虑调水和供水效率的水库优化调度模型，并采用理想点法和遗传算法对调度规则寻优。将模型应用于威海市区米山水库，结果表明：与标准调度策略相比，优化后的水库调度规则使多年平均调水量减少2 692.4万m³,多年平均弃水量减少417.5万m³,提高了天然来水和外调水的利用率；枯水年对外调水的依赖性高于丰水年和平水年，平均调水量仅减少398万m³;在该调度规则下，减少外调水量的同时增加非常规水的配置可将农业供水保证率从75%提高到88.4%。