基于PSO的梯级水库联合防洪调度

在分析梯级水库联合防洪调度、上下游水库及防洪保护对象之间复杂的水文水力联系和水库群对防洪保护对象防洪补偿调度方式的基础上,建立了梯级水库联合防洪补偿调度模型,并提出了一种基于粒子群算法的模型求解方法。以湖南省资水流域上柘溪与金塘冲水库为例,根据不同典型年的地区洪水组合,通过模型求解拟定了梯级水库联合补偿调度原则。分析研究结果表明,这种基于PSO的梯级水库联合防洪调度原则是合理的,该模型能在一定程度上简化梯级水库联合防洪补偿调度的复杂性,为解决梯级水库联合防洪补偿调度提供了一条简洁有效的途径。     

Flood Control Operation of Reservoir Group Using Yin-Yang Firefly Algorithm

Flood control operation (FCO) of a reservoir is a complex optimization problem with a large number of constraints. With the rapid development of optimization techniques in recent years, more and more research efforts have been devoted to optimizing FCO problems. However, for solving large-scale reservoir group optimization problem, this is still a challenging task. In this work, a reservoir group FCO model is established with minimum flood volume stored in each reservoir and minimum peak flow of downstream control point during the dispatch process. At the same time, a flood forecast model for FCO of a reservoir group is developed by coupling Yin-Yang firefly algorithm (YYFA) with ε constrained method. As a case study, the proposed model is applied to a three-reservoir flood control system in Luanhe River Basin consisting of reservoirs, river channels, and downstream control points. Results show that optimal operation of three reservoirs systems can efficiently reduce the occupied storage capacity for flood control and flood peaks at downstream control point of the basin. The proposed method can be extended to FCO of other reservoir groups with similar conditions.

     

Deriving Reservoir Refill Operating Rules by Using the Proposed DPNS Model

The dynamic programming neural-network simplex (DPNS) model, which is aimed at making some improvements to the dynamic programming neural-network (DPN) model, is proposed and used to derive refill operating rules in reservoir planning and management. The DPNS model consists of three stages. First, the training data set (reservoir optimal sequences of releases) is searched by using the dynamic programming (DP) model to solve the deterministic refill operation problem. Second, with the training data set obtained, the artificial neural network (ANN) model representing the operating rules is trained through back-propagation (BP) algorithm. These two stages construct the standard DPN model. The third stage of DPNS is proposed to refine the operating rules through simulation-based optimization. By choosing maximum the hydropower generation as objective function, a nonlinear programming technique, Simplex method, is used to refine the final output of the DPN model. Both the DPNS and DPN models are used to derive operating rules for the real time refill operation of the Three Gorges Reservoir (TGR) for the year of 2007. It is shown that the DPNS model can improve not only the probability of refill but also the mean hydropower generation when compare with that of the DPN model. It's recommended that the objective function of ANN approach for deriving refill operating rules should maximize the yield or minimize the loss, which can be computed from reservoir simulation during the refill period, rather than to fit the optimal data set as well as possible. And the derivation of optimal or near-optimal operating rules can be carried out effectively and efficiently using the proposed DPNS model.     

A Multi-Objective Optimization Model for Coordinated Regulation of Flow and Sediment in Cascade Reservoirs

This paper presents a multi-objective optimization model for the coordinated regulation of flow and sediment in cascade reservoirs. The model was developed to address two contradicting issues: sediment trapping and flow regulation. The benefits of flood control, hydropower generation and navigation, and sedimentation in cascade reservoirs were considered as the target functions; then the corresponding submodels for reservoir operation and sediment computation were established. The model was implemented by reducing it to a single objective nonlinear model using the constraint method. Non-inferior solutions were obtained by solving the model with catfish effect particle swarm optimization algorithm. The model was applied to the cascade system of Xiluodu and Xiangjiaba reservoirs in the lower Jinsha River in the flood recession period. Under the safety of flood control and navigation, a non-inferior set for impounding time, power generation, and siltation-loss rate of capacity was obtained and optimal solutions with different weights were derived. The results demonstrate that the model is a useful tool in coordinated operations of cascade reservoirs.     

水电站水库群防洪补偿联合调度模型研究及应用

针对具有下游防洪任务的水电站水库群，提出基于预报及库容补偿的水库群防洪补偿联合调度逐次渐进协调模型，推求水库汛期防洪库容动态控制方案。该模型运用了大系统分解协调理论及贝尔曼的逐步逼近思想，以各水库为独立的子系统建立3层递阶结构，针对库间水力、电力联系及防洪库容限制等复杂的约束条件进行不同层次的协调。模型经过多次迭代计算，得到最佳的水库群防洪库容协调方案。该模型应用于清江流域梯级水库，计算结果表明，在不降低水库及梯级原有的防洪标准前提下，能有效利用上游水布垭水库的防洪库容，分担隔河岩水库部分防洪任务，并显著提高梯级水库发电量。     

溪洛渡-向家坝-三峡梯级水库联合蓄水方案与多目标决策研究

为协调好梯级水库联合蓄水调度过程中防洪、发电、蓄水和航运等目标之间的矛盾,实现各水库蓄水时机与蓄水进程的协同优化,以溪洛渡-向家坝-三峡梯级水库为例,推求了可权衡防洪与兴利之间矛盾的梯级水库联合蓄水方案,主要研究内容主要包括以下3个部分:(1)风险分析,推求了汛末各分期内坝前最高安全水位约束和联合蓄水方案的防洪风险;(2)兴利效益分析,分析了联合蓄水方案的发电和蓄水等兴利效益;(3)多目标决策,评价了联合蓄水方案的防洪风险、发电和蓄水效益,得出了最优非劣质蓄水方案。研究结果表明:溪洛渡-向家坝-三峡梯级水库的较优联合蓄水方案,分别为9月5日起蓄的同步起蓄方案和9月1日-9月5日-9月10日起蓄的异步起蓄方案,较原设计蓄水方案,年均发电量可分别增加23.76和20.43亿k W·h,增幅分别为3.25%和2.78%;蓄水率可由96.73%分别提高至97.51%和97.57%,两者均可在不降低原防洪标准的前提下,提高梯级水库的综合效益,前者的发电效益较优,而后者的蓄水效益较优。     

A Triple-stage Operation Method for Deriving Operation Rules for Cascade Reservoirs during Catastrophic Flood Events

Alleviating flood stress and reducing flood disaster losses are vital issues when operating flood control systems. To address catastrophic floods, this study establishes an optimal flood control operation model for minimizing the excess flood volume of downstream flood control points. A triple-stage operation method (TSOM) is proposed to solve the model by dividing flood control operation processes into three stages. The constructed model is applied to the operation of a flood control system located in the middle and lower reaches of the Ganjiang River, China. The results demonstrate that the TSOM has superiority in solving efficiency, reducing excess flood volume by more than 30–230 million cubic meters compared with the progressive optimality algorithm (POA) for most designed floods with 500-year return periods. Furthermore, the discharge and water level processes created by the TSOM fluctuate less than those created by the POA, which meets the requirements of practical flood control operations.

     

A Simulation-Based Optimization Model for Flood Management on a Watershed Scale

Using structural and nonstructural measures for flood damage reduction is a long-standing problem in water resources planning and management. In the present study, an algorithm is presented for the optimal design of structural and nonstructural flood mitigation measures based on simulation-based optimization approach. For this purpose, the MIKE-11 simulation model, a one-dimensional hydrodynamic model, was used to calculate the potential damages of different flood scenarios under the various combinations of structural and nonstructural measures and this model was coupled with the NSGA-II multi-objective optimization model to provide the optimal Pareto solutions between two conflict objectives of minimizing the investment costs of flood mitigation measures and the potential damages of the floodplain. The proposed model was then applied to a small watershed in central part of Iran as a case study and the optimal trade-off solutions were calculated for different flood scenarios. Using these trade-offs, for each level of funding, decision makers can assign the optimal design of flood mitigation measures considering decision criteria.     

Optimal Design of Barrage Profile on Anisotropic Soil Using Multi‐Objective Optimization Approach

Weirs and barrages are costly hydraulic diversion structures; therefore, any attempt to improve their design is a worthy contribution. Diversion structures, such as weir or barrage, may be designed on permeable formations considering homogenous soil properties. But in reality, soil properties are hardly homogeneous. In this paper, an approach is described to determine an economically efficient barrage profile by considering soil’s anisotropic behaviour. Hydraulic conductivity is considered to be an anisotropic soil property. An optimization-based methodology is developed to obtain the optimal barrage profile. The minimization of the material cost and minimization of the exit gradient is considered for multi-objective formulation. The multi-objective formulation is solved using NSGA-II, and a Pareto optimal front is obtained for different degrees of anisotropy. The flow interaction under a diversion structure in anisotropic soil is incorporated using the Modified Lane theory and is embedded in optimization formulation. The developed methodology is illustrated with a barrage profile as a hydraulic structure. A parametric study is carried out to study the effects of varying barrage design elements on the barrage’s optimum material cost.

     

Derivation of Aggregation-Based Joint Operating Rule Curves for Cascade Hydropower Reservoirs

Operating rule curves have been widely applied to reservoir operation, due to their ease of implementation. However, these curves are generally used for single reservoirs and have rarely been applied to cascade reservoirs. This study was conducted to derive joint operating rule curves for cascade hydropower reservoirs. Steps in the proposed methodology include: (1) determining the optimal release schedule using dynamic programming to solve a deterministic long-term operation model, (2) identifying the forms of operating rule curves suitable for cascade hydropower reservoirs based on the optimal release schedule, (3) constructing a simulation-based optimization model and then using the non-dominated sorting genetic algorithm-II (NSGA-II) to identify the key points of the operating rule curves, (4) testing and verifying the efficiency of the generated joint operating rule curves using synthetic inflow series. China’s Qing River cascade hydropower reservoirs (the Shuibuya, Geheyan and Gaobazhou reservoirs) were selected for a case study. When compared with the conventional operating rule curves, the annual power generation can be increased by 2.62% (from 7.27 to 7.46 billion kWh) using the observed inflow from 1951 to 2005, as well as by about 1.77% and 2.52% using the synthetic inflows generated from two alternative hydrologic simulation methods. Linear operating rules were also implemented to simulate coordinated operation of the Qing River cascade hydropower reservoirs. The joint operating rule curves were more efficient and reliable than conventional operating rule curves and linear operating rules, indicating that the proposed method can greatly improve hydropower generation and work stability.     

复杂防洪系统优化调度的三层并行逐步优化算法

逐步优化算法（POA）在求解水库群防洪优化调度问题方面,应用广泛且较为有效。但是随着水库数量的增加和计及洪水演进、干支流汇入等因素,该算法求解复杂防洪系统调度的效率降低。本文在POA算法基础上,引入莱维飞行更新策略、模式搜索法以及并行技术,提出了三层并行POA算法（TPPOA）,并应用于赣江中下游复杂防洪系统的防洪优化调度中。结果表明,相较常规防洪调度规则和POA算法,TPPOA获得的调度策略可以有效降低下游防洪控制点洪峰流量,减轻下游防洪压力。此外,在6核心CPU计算机下,TPPOA计算速度约为POA算法的3~7倍,计算速度更快。本文为复杂防洪系统的优化调度提供了一种新的方法。     

Simulation with RBF Neural Network Model for Reservoir Operation Rules

Reservoirs usually have multipurpose, such as flood control, water supply, hydropower and recreation. Deriving reservoirs operation rules are very important because it could help guide operators determine the release. For fulfilling such work, the use of neural network has presented to be a cost-effective technique superior to traditional statistical methods. But their training, usually with back-propagation (BP) algorithm or other gradient algorithms, is often with certain drawbacks. In this paper, a newly developed method, simulation with radial basis function neural network (RBFNN) model is adopted. Exemplars are obtained through a simulation model, and RBF neural network is trained to derive reservoirs operation rules by using particle swarm optimization (PSO) algorithm. The Yellow River upstream multi-reservoir system is demonstrated for this study.     

考虑分级防洪目标的梯级水库汛控水位调度模型及应用

围绕中小洪水减压、大洪水保安和特大洪水降损等防洪目标，构建了面向分级防洪目标的梯级水库汛控水位优化调度模型，针对不同频率设计洪水，基于“模拟-优化”框架，采用仿生进化算法对调度模型进行了高效求解，并利用熵权法对调度方案进行了多目标评价，比选了梯级水库汛控水位方案。以金沙江中下游6个梯级水库和三峡水库组成的梯级水库为研究对象，研究结果表明：相比原汛限水位方案，在不降低原防洪标准的前提下，选定的方案可使梯级水库水位抬高0.36～6.22 m,发电量增加6.28亿～19.26亿kW·h,增幅为1.18%～4.27%,经济效益显著，可为梯级水库汛控水位的规划设计提供技术支撑。     

Optimal Design of Check Dams in Mountainous Watersheds for Flood Mitigation

One of the measures for flood control is to construct a series of small barriers, known also as check dams, on tributaries of watershed stream network. Check dams are generally used in mountainous areas in order to control sediment transport and attenuate flood peak. In this paper, a simulation-based optimization model is developed to determine size, shape and the number of check dams for flood mitigation. HEC-HMS model is used to simulate watershed rainfall-runoff process considering various check dam designs. The model is coupled with a multi-objective evolutionary algorithm, called non-dominated sorting differential evolution (NSDE), to find the trade-off solutions considering three objective functions: 1) minimizing the investment cost, 2) minimizing the flood peak discharge and 3) maximizing the time to peak discharge. The proposed model is applied to a mountainous watershed in Iran and (near) optimal strategies, including the suitable number of check dams in each sub-watershed, and optimal dam size (e.g. optimal height, bottom width and side angles) in each sub-watershed are obtained. The results show that cost-effective designs can decrease peak discharge up to 53%, 54 and 54% corresponding to 2-yr, 5-yr and 10-yr flood return period scenarios, respectively. In addition, the check dams can also increase the time to peak for up to 88%, 81 and 77%, corresponding to 2-yr, 5-yr and 10-yr flood scenarios, respectively.     

雅砻江下游梯级水库综合调度规则优化方法

以综合运行水位控制线、防弃水线、防破坏线、防洪限制水位和死水位作为水库调度指导线,并构成水库调度图,围绕该调度图制定水库综合调度规则,进而建立了调度图模拟-优化模型。选取雅砻江下游梯级水库作为研究对象,采用模拟-优化技术对调度图进行优化。采用优化后的调度图指导水库群模拟运行,模拟运行结果验证了该调度图具有良好的可操作性和实用价值。     

长江流域梯级枢纽泥沙调控关键技术

通过将水沙动力学与水库调度优化理论相结合,围绕长江流域梯级枢纽泥沙调控关键技术开展研究。在改进恢复饱和系数、非均匀沙挟沙力等关键计算模式的基础上,完善并集成长江上游梯级水库群水沙数学模型、长江中下游复杂江湖河网水沙数学模型和典型河段平面二维水沙数学模型,构建了江河湖库水沙输移模拟模块。基于泥沙对防洪、发电、航运和长期使用4个主要目标的影响,构建了长江泥沙优化调控目标函数,进而集成水沙输移模拟模块构建了长江泥沙优化调度模型,并提出了基于预构泥沙信息库、结合BP神经网络的模型求解方法和非劣解集的改进灰靶评价方法。在此基础上,以三峡水库为例,提出了汛期“蓄清排浑”动态运用方案以及长期分阶段泥沙调控策略。     

三峡水库动态汛限水位与蓄水时机选定的优化设计

阐述了三峡水库运行初期的调度规则，建立了实时调度模型。提出了防洪、发电和航运等指标体系，在此基础上，建立了用于动态汛限水位和蓄水时机优化的混合规划数学模型，并设计了一种混合编码方式，运用遗传算法对该模型进行优化求解。利用宜昌站1882～2001年实测日流量资料进行模拟优化，生成了一系列非劣解；然后运用模糊决策的方法，得到了相对合理的动态汛限水位与蓄水时机方案。结果表明，优化设计能权衡防洪与兴利之间的关系，充分挖掘汛初与汛末的潜力，提高三峡水库运行初期的综合效益。     

基于GA-Elman的河流水位预测方法研究

河流水位的变化过程是一个复杂的非线性过程,传统的神经网络预测存在误差较大、收敛速度慢、稳定性差等问题。为了实现对河流水位的有效预测,提出基于遗传算法(GA)优化Elman神经网络的河流水位预测模型。将GA与Elman网络进行有效结合,解决了单一Elman网络存在的不足。选取永定河的监测站点水文数据对河流水位进行预测与检验,并分别将其与Elman网络与BP网络预测结果进行对比。对比结果表明:GA-Elman水位预测模型的收敛速度快、精度高,可根据预测结果实现对水库、拦河闸合理调用,实现对河流水资源的有效配置,以满足灌溉、发电、防洪等工作的需求。     

基于长期与中长期嵌套的水库优化调度

为解决水库优化调度过程中时间尺度不同而导致的水库综合效益不高的问题,以防洪与兴利为目标,将水库长期优化调度与中长期优化调度进行嵌套,建立水库多目标优化调度嵌套模型。其中,多目标问题通过约束法将防洪目标转换为硬性约束后,再转化为求解发电量最大值的单目标问题。算法方面,综合对比后确定长期优化调度采用动态规划算法求解、中长期优化调度采用遗传算法求解。为验证嵌套模型的优化效果,以澄碧河水库为例,进行水库优化模拟调度。结果表明:在满足防洪目标的前提下,嵌套模型优化调度方案相较于长期优化调度方案和实际调度方案效益均更高,验证了嵌套模型在水库优化调度问题中的优越性。     

黄河上游梯级水库防凌优化调度方案研究

针对黄河上游凌汛期龙羊峡水库出库流量调度问题,在分析上游梯级水库防凌调度原则及调度运行方式的基础上,考虑青海电网电量平衡的需求,建立了黄河上游梯级水库防凌优化调度模型,并提出了改进的微粒群算法用于求解该模型。结果表明:凌汛期在保证龙羊峡水库出库总水量不变的情况下,通过调整龙羊峡水库在各时段的出库水量,既满足了青海电网电量平衡的要求,优化了青海电网电量结构,同时也增加了刘家峡水库防凌库容;通过优化调度,增加了凌汛期梯级水电站的可发电量,提高了梯级电站效益。