跨流域水库群引水与供水联合优化调度研究

发布时间：2018-08-04 09:02

【摘要】：我国的水资源时空分布不均,与社会经济发展水平及战略布局不相匹配,水资源供需矛盾突出,仅靠挖掘本流域的水资源潜力无法完全解决当地的水资源短缺问题,跨流域调水工程已成为水资源重新分配和缓解缺水地供需矛盾的重要途径。目前,针对跨流域水库群联合优化调度模型、求解方法的研究取得了一系列研究成果。然而,随着跨流域调水规模的不断扩大,水库群的拓扑结构也越来越复杂,调度规则形式与求解方法适应性方面出现了一系列新的问题。例如,如何解决大规模复杂水库群优化调度的“维数灾”问题,联合供水任务在成员水库间的供水分配问题和多受水水库的引水分配问题等。目前,针对跨流域水库群引水规则的研究较少,而引水规则制定时需要依据不同水库群的拓扑结构,同时需要不断完善规则以提高引水效率。此外,鉴于跨流域水库群调度规模扩大加剧了调度问题求解的复杂性,对优化计算效率和求解精度也提出了越来越高的要求。随着计算机技术不断发展和多核并行计算平台日益多元化,多核并行计算逐渐成为提高水库群优化调度计算效率和求解精度的一个重要途径。因此,本文从调度模型和求解方法两个方面展开研究,主要研究成果如下： (1)首先介绍了研究区的区域概况和社会经济发展布局,识别研究区域水资源开发利用存在的问题,说明实施跨流域调水工程措施的必要性。然后,介绍了辽宁省东水西调北线工程——辽西北供水工程的特点,并对系统结构做出合理的概化,明确水库群调度目标及所需要的特征资料。最后,基于历史长系列的入流量过程,分析研究区域的用水和径流规律,为调度模型的资料输入、调度规则的合理性分析等做基础的工程背景及数据准备。 (2)以辽西北供水工程为背景,根据跨流域水库群拓扑结构预设了相应的调度规则。①引(调)水与供水规则基本形式及其使用规则。其中,限制供水与引(调)水的启动标准分别为供水与引(调)水调度线,根据时段初蓄水量所处的调度区确定是否限制供水或引(调)水决策。此外,受水水库的引水决策是调水水库的调水决策和受水水库的引水决策的一个组合决策过程。②联合供水任务的供水规则,采用主控水库法来制定联合供水任务的供水决策,然后构建供水分配模型实现供水任务的分解。③构建引水分配模型实现多受水水库的初始引水量分配问题。实例结果分析表明了所制定规则的合理性和有效性。 (3)为提高整个系统的供水和引水效益,确定供水量最大和引水量最小两个目标函数以及相应的约束条件,构建基于调度规则的跨流域水库群联合调度优化模型。采用模拟-优化方法求解所制定的调度规则,其中,模拟模块嵌入优化模块以寻求联合调度图中调度线的关键点位置和引水供水分配系数。实例结果也表明了模型的科学性与合理性以及模拟-优化方法的实用性。 (4)针对跨流域水库群联合优化调度具有高维非线性和动态性的特点,提出一种改进微粒群算法,用于求解跨流域水库群引水与供水联合优化调度模型。该算法针对基本微粒群算法易陷入局部最优,并且进化后期收敛速度慢的缺陷,引入粒子交叉变异遗传更新策略、模拟退火策略和反射边界策略,以增加种群的多样性并提高算法的进化速度和全局搜索能力。数值实验显示该算法适用于求解高维复杂问题,实例研究也表明该算法的收敛性能得到了改善,提高了整个系统的供水与引水效益,是一种求解大规模复杂水库群优化调度的实用方法。 (5)为进一步提高跨流域大规模复杂水库群优化调度的计算效率和求解精度,采用并行PSO算法进行联合调度模型的多核并行求解。该算法充分利用PSO搜索速度快、天然并行性等特点,引入多种群思想保证种群的多样性,提高算法的全局收敛能力；采用基于分治策略的Fork/Join框架实现将子种群分配到不同CPU内核进行独立求解；利用java并发过程中的同步和通信机制实现子种群间的信息交流,避免陷入局部最优。最后通过实例分析表明,多核并行PSO算法能够充分利用多核资源,有利于提高联合优化调度模型的计算速度和求解精度,是解决大规模复杂水库群优化调度的一种高效实用的方法。 (6)针对已有引水规则的引水效率有待进一步提高的问题,本文提出一种引水规则的有效改进形式,以及改进规则下受水水库时段引水量的确定方法。该改进规则同时考虑了调水水库和受水水库的蓄水状态,并且在受水水库上采用双引水控制线形式。当调水水库调水且受水水库蓄水状态处于双引水控制线之间时,引水量在满引(初始引水量与引水能力的较小值)与不引之间线性插值。最后采用基于并行PSO算法(PPSO)的模型迭代方法求解,结果表明改进引水规则对供水影响较小,但极大提高了丰、平水年的引水效率。
[Abstract]:The uneven spatial and temporal distribution of water resources in China is not matched with the social and economic development level and strategic layout, and the contradiction between supply and demand of water resources is prominent. It is impossible to fully solve the local water shortage problem by digging the potential of water resources in this basin. The trans basin water diversion project has become an important part of the redistribution of water resources and the relief of the contradiction between the supply and demand of water shortage. At present, a series of research achievements have been achieved in the study of the joint optimal scheduling model of cross basin reservoirs, and a series of research results have been obtained. However, with the continuous expansion of the scale of water diversion in the cross basin, the topology of the reservoir group is becoming more and more complex, and a series of new problems have appeared in the form of scheduling rules and the suitability of the solution methods. For example, such as How to solve the "dimensionality disaster" problem of the large-scale complex reservoir group optimization scheduling, the distribution of water supply between the members of the reservoir and the diversion of water reservoir, and so on. At present, the study on the diversion rules for the cross basin reservoirs is less, and the timing of the diversion should be based on the topology of different reservoir groups. In addition, in view of the complexity of the scheduling problem and the increasing complexity of the scheduling problem, the efficiency and accuracy of the optimization are also raised. With the continuous development of computer technology and the increasing diversity of multi core parallel computing platform, the multi core parallel meter has been developed. The calculation has gradually become an important way to improve the efficiency and accuracy of the optimal operation of the reservoir group. Therefore, this paper studies the two aspects of the scheduling model and the solution method, and the main research results are as follows:
(1) first, it introduces the general situation of the study area and the layout of the social and economic development, identifies the problems existing in the development and utilization of the water resources in the study area, and illustrates the necessity of implementing the measures for the implementation of the water diversion project in the cross basin. Then, it introduces the characteristics of the East Water Diversion Project of the east of Liaoning Province, the west north water supply project, and makes a reasonable estimate of the system structure. In the end, the law of water and runoff in the study area is analyzed based on the history of long series of inflow, and the engineering background and data are prepared for the basis of the data input of the scheduling model and the reasonableness analysis of the scheduling rules.
(2) taking the water supply project in the northwest of Liaoning as the background, the corresponding scheduling rules are preset according to the topology of the cross basin reservoir group. (1) the basic forms of water supply and water supply rules and their use rules. Among them, the starting standard of water supply and diversion (adjustment) is the water supply and (adjustment) water dispatching line, which is determined according to the dispatching area of the initial water storage in the period of time. Whether the water supply or the water diversion decision is restricted. In addition, the water diversion decision of the water reservoir is a combination decision process of the water diversion decision of the water reservoir and the decision of the water diversion of the water reservoir. The water supply assignment model is constructed to realize the initial water diversion distribution of the water reservoir. The case results analysis shows that the rules are reasonable and effective.
(3) in order to improve the water supply and water diversion benefits of the whole system, to determine the maximum water supply and the minimum two target functions of the water diversion and the corresponding constraints, the optimization model of the joint dispatching of the reservoir group based on the scheduling rules is constructed. The simulation optimization method is used to solve the scheduling rules, in which the simulation module is embedded in the optimization module. In order to find the key point position of the scheduling line and the water supply distribution coefficient in the joint schedule, the example results also show the scientificalness and rationality of the model and the practicability of the simulation optimization method.
(4) an improved particle swarm optimization (PSO) algorithm is proposed to solve the joint optimal scheduling model of water diversion and water supply in a cross basin reservoir. The algorithm is introduced to the problem that the basic particle swarm optimization is easy to fall into the best local and slow convergence in the later stage of evolution. Particle cross mutation genetic updating strategy, simulated annealing strategy and reflection boundary strategy are used to increase the diversity of the population and improve the evolutionary speed and global searching ability of the algorithm. Numerical experiments show that the algorithm is suitable for solving high dimensional complex problems. The example study also shows that the convergence performance of the algorithm is improved and the whole system is improved. The efficiency of water supply and diversion is a practical method for solving the large-scale and complex reservoir group optimal operation.
(5) in order to further improve the computational efficiency and accuracy of the optimal scheduling of large and complex reservoirs in the cross basin, the parallel PSO algorithm is used to solve the multi core parallel scheduling model. The algorithm makes full use of PSO search speed and natural parallelism to ensure the diversity of the population and improve the global convergence of the algorithm. By using the Fork/Join framework based on divide and conquer strategy, the subpopulation is allocated to different CPU kernel for independent solution, and the information exchange between subpopulations is realized by the synchronization and communication mechanism in the concurrent process of Java, and the local optimum is avoided. Finally, the multi core parallel PSO algorithm can make full use of the multi core parallel algorithm. Nuclear resources are beneficial to improve the calculation speed and solving precision of the combined optimal scheduling model. It is an efficient and practical method to solve the optimal scheduling of large and complex reservoirs.
(6) in view of the problem that the water diversion efficiency of the existing water diversion rules needs to be further improved, this paper proposes an effective improvement form for the regulation of water diversion, as well as the method for improving the water diversion amount of the water reservoir under the regulation under the rules. The improvement rule also takes into account the water storage state of the water reservoir and the water reservoir, and uses double quotes on the water receiving reservoir. Water control line form. When water diversion reservoir is water reservoir and water reservoir is stored in double water diversion control line, the quantity of diversion is linear interpolation between the full citation (initial water intake and water diversion capacity) and the non citation. Finally, the model iterative method based on parallel PSO algorithm (PPSO) is used to solve the problem. The result shows that the water supply rules are improved to water supply. The effect is small, but it greatly improves the water diversion efficiency of the year.
【学位授予单位】：大连理工大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TV697.11

【参考文献】