城市明渠引排水系统的智能控制技术研究

发布时间：2018-02-28 14:40

本文关键词： 城市明渠远程监控动态规划自适应动态规划　出处：《浙江大学》2017年硕士论文　论文类型：学位论文

【摘要】：水资源是重要的战略性资源,而明渠是城市引排水系统中的重要组成部分,对水资源的科学管理、调度和利用有着深远的意义。所以,对明渠引排水系统进行合理的建设和控制,以保证城市的供排水需求就显得尤为重要。在硬件设施方面,我国的明渠引排水系统建设比较落后,大多数明渠泵站仍以人工值守的方式进行控制。在研究方面,国内外对明渠的相关研究多集中于水位、流量和容积上的控制,而忽略了与能耗相关的控制代价。另外,由于难以获取精确的数据,且存在降水等不确定性因素,大多数基于模型的控制算法难以在实际工程当中应用。本文针对这些问题进行了相关研究。1.针对控制方式落后的人工值守泵站,本文设计了一个明渠引排水系统的远程监控自动化改造方案,阐述了设计方案中监控系统的组织架构,具体介绍了中央监控服务器中组态平台、中转数据库平台和ODBC数据链接的设计和建立流程。该方案最终在实际工程项目当中得到了成功应用,并达到了预期的效果。2.针对已知明渠模型和降水预测的情况,本文提出了一种改进的单明渠在线动态规划控制算法,控制目标是跟踪目标水位和减小与能耗相关的控制代价。改进动态规划算法通过引入状态反馈和滚动更新等机制,减小降水预测误差对系统带来的影响。本文对改进动态规划算法进行了仿真,并设置了传统动态规划方法作为对照,结果显示改进的动态规划算法在水位跟踪和节能降耗方面有着更佳的表现。3.针对明渠模型未知和存在降水、城市排水等干扰的的情况,本文提出了一种在线运行的多明渠自适应动态规划控制算法,目的同样是在跟踪明渠目标水位的同时减小与能耗相关的控制代价。本文设计的自适应动态规划算法为一个闭环结构,包含执行神经网络和评判神经网络两个RBF神经网络。本文对自适应动态规划算法进行了仿真,并与传统的PID算法进行对比,结果表明自适应动态规划在水位跟踪和减小控制代价方面优于传统PID控制算法。
[Abstract]:Water resource is an important strategic resource, and open channel is an important part of urban drainage system, which has far-reaching significance for the scientific management, regulation and utilization of water resources. It is very important to construct and control the open channel drainage system reasonably in order to ensure the water supply and drainage demand of the city. In terms of hardware facilities, the construction of the open channel drainage system in our country is relatively backward. Most open channel pumping stations are still controlled by manual monitoring. In the research aspect, the research on open channel mostly focuses on the control of water level, flow rate and volume, while neglecting the control cost related to energy consumption. Because it is difficult to obtain accurate data, and there are uncertain factors such as precipitation, Most model-based control algorithms are difficult to be applied in practical projects. 1. In this paper, a remote monitoring and control automation reconstruction scheme of open channel drainage system is designed, and the organization structure of the monitoring system in the design scheme is expounded, and the configuration platform in the central monitoring server is introduced in detail. The design and establishment process of the transit database platform and ODBC data link. Finally, the scheme has been successfully applied in the actual project, and achieved the expected effect. 2. Aiming at the known open channel model and precipitation forecast, In this paper, an improved on-line dynamic programming control algorithm for single open channel is proposed. The control target is to track the target water level and reduce the control cost related to energy consumption. The improved dynamic programming algorithm introduces the mechanism of state feedback and rolling update. In this paper, the improved dynamic programming algorithm is simulated, and the traditional dynamic programming method is used as a comparison. The results show that the improved dynamic programming algorithm has better performance in water level tracking, energy saving and consumption reduction. In this paper, a multi-channel adaptive dynamic programming control algorithm for on-line operation is proposed. The aim is to reduce the control cost associated with energy consumption while tracking the target water level in the open channel. The adaptive dynamic programming algorithm designed in this paper is a closed loop structure. There are two RBF neural networks including executive neural network and critical neural network. This paper simulates the adaptive dynamic programming algorithm and compares it with the traditional PID algorithm. The results show that the adaptive dynamic programming is superior to the traditional PID control algorithm in water level tracking and reducing control cost.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP273;TP277;TU99

【参考文献】