芜湖市三山水厂泵站与清水池联合优化调度研究

发布时间：2018-03-27 19:07

本文选题：优化调度　切入点：BP神经网络　出处：《合肥工业大学》2017年硕士论文

【摘要】：随着我国经济的迅猛发展,人民生活质量的稳步提升,社会节水、节能意识的日益增强,人们对供水系统的期望也越来越高。如何在稳定供水的前提下,尽可能的节约成本已经成为社会关注的焦点。泵站和清水池作为供水系统中重要的组成部分,寻求更加科学合理的运行模式是我国供水系统进一步发展的必然要求。目前我国泵站管理水平相对比较落后,传统的依赖经验管理的调度模式已经无法满足人们对日益增长的供水质量的要求,泵站运行的优化调度面临着新的挑战。本文以芜湖市三山区供水系统为依托,运用BP神经网络、遗传算法、动态规划等算法,建立时用水量预测模型、供水系统管网分析模型、泵站优化调度模型以及一、二级泵站与清水池联合优化调度模型,对泵站优化调度中所涉及的各种内容进行了系统研究,主要内容如下:(1)时用水量预测:根据芜湖市三山区时用水量的变化规律及各种影响因素,选择BP神经网络建立芜湖市三山区时用水量预测模型。模型采用3层网络结构,以过去24h的时用水量作为输入变量,将下一小时的时用水量作为输出变量,对未来用水量进行科学预测。通过芜湖市三山区实际数据检验表明,所建模型具有较高的预测精度。(2)供水系统管网分析模型:根据芜湖市三山区压力监测点的布置位置,选择BP神经网络建立芜湖市三山区供水系统管网分析模型。其中网络结构为3层,模型以7个压力监测点的时平均压力和三山水厂的时供水流量作为输入变量,将三山水厂的时平均供水压力作为输出变量。经过芜湖市三山区供水管网的实例检验,所建模型基本能够模拟管网的运行状态,可以为优化调度服务。(3)泵站优化调度:研究一、二级泵站优化调度模型的建模方法,以流量、水压、水泵性能等为约束,建立三山水厂一、二级泵站优化调度模型,并运用遗传算法进行求解。(4)一、二级泵站与清水池联合优化调度:针对一、二级泵站的工作特性,充分利用峰谷电价差和清水池的调蓄容积,以一、二级泵站日运行总电费最小和水泵机组启停次数最少为双目标,建立一、二级泵站与清水池分级优化调度模型,并分别运用动态规划方法、遗传算法进行求解,制定了一、二级泵站的优化调度运行方案。实例计算结果显示,优化后的运行方案经济效益大幅提升。
[Abstract]:With the rapid development of our country's economy, the steady improvement of people's life quality, the increasing consciousness of saving water and energy saving, people's expectation of water supply system is higher and higher. Cost saving as much as possible has become the focus of social attention. Pumping stations and clear water pools are important components of the water supply system. It is necessary to seek more scientific and reasonable operation mode for the further development of water supply system in China. At present, the management level of pump stations in China is relatively backward. The traditional dispatching mode relying on experience management can no longer meet the demand of increasing water supply quality, and the optimal operation of pump station is facing new challenges. This paper relies on the water supply system in Sanshan District of Wuhu City. BP neural network, genetic algorithm, dynamic programming and other algorithms are used to establish water consumption prediction model, water supply network analysis model, pump station optimal dispatching model and the first, second stage pump station and clear water pool joint optimal dispatching model. This paper makes a systematic study on the various contents involved in the optimal dispatching of pumping stations. The main contents are as follows: (1) Prediction of hourly water consumption: according to the changing law of hourly water consumption and various influencing factors in Sanshan District of Wuhu City, BP neural network is selected to establish the forecasting model of hourly water consumption in Sanshan District of Wuhu City. The model adopts a three-layer network structure, takes the hourly water consumption of the past 24 hours as input variable, and takes the hourly water consumption of the next hour as the output variable. Scientific prediction of water consumption in the future is carried out. The results of practical data test in Sanshan District of Wuhu City show that the model has a higher prediction precision and a pipe network analysis model of water supply system: according to the location of pressure monitoring points in Sanshan District of Wuhu City, BP neural network is selected to establish the analysis model of water supply system in Sanshan District of Wuhu City, in which the network structure is three layers. The model takes the hourly average pressure of seven pressure monitoring points and the hourly water supply flow of Sanshan Water Plant as input variables. Taking the average hourly water supply pressure of Sanshan Water Plant as the output variable, the model can basically simulate the running state of the pipe network, and can serve for the optimal dispatching of the pumping station by the example of the water supply network in Sanshan District of Wuhu City. The modeling method of optimal dispatching model of two-stage pumping station is to set up the optimal dispatching model of the first and second stage pumping stations of Sanshan Water Plant with the constraints of flow rate, water pressure and pump performance, and use genetic algorithm to solve the problem. Combined optimal dispatching of two stage pumping stations and clear water pools: according to the working characteristics of the first and second stage pumping stations, the difference between peak and valley electricity prices and the storage capacity of the clear water pool are fully utilized to make use of one, The minimum daily total electricity charge of secondary pumping station and the least number of start-up and stopping times of pump unit are the double objectives. A hierarchical optimal dispatching model of the first, second stage pump station and clear water tank is established, and the dynamic programming method and genetic algorithm are used to solve the problem, and a new model is developed. The calculation results show that the economic benefit of the optimized operation scheme is greatly improved.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TV675

【参考文献】