供水管网水质监测系统优化构建关键技术研究

Over the past few years, there have been many water-quality contamination events in water distribution systems (WDSs), posing significant impacts on resident life and social economy. Therefore, contamination events within WDS have become a serious problem that needs to be urgently addressed to ensure the stability and sustainable development of the society. Water-quality monitoring has been one of the most key methods to ensure the safe delivery of water within WDSs. This consequently motivates an important research problem, that is, how to optimally design the water-quality monitoring system? To this end, this study (i) develops a comprehensive water-contamination indicator, with which the contaminated water amount, contaminated time, contaminated range and the consumed residents are accounted for. Additionally, the probability density and uncertainty analysis approaches are proposed to improve the understanding of the indicator. The comprehensive water-contamination indicator as well as its probability density and uncertainty analysis is expected to reveal the mechanisms of contamination events in affecting the water quality within WDSs; (ii) develops a theory framework and mathematical model to optimally design water-quality monitoring system based on the marginal benefit of the comprehensive water-contamination indicator. Such a framework would reveal the relationship between the structure properties of the water-quality monitoring system and the value of the comprehensive water-contamination indicator, and hence it assists to simultaneously optimally design the scale of the water-quality monitoring system as well as its deployment; (iii) develops a hybrid method that combines an artificial neural network and an evolutionary algorithm to enhance the efficiency of the optimization problem. The outcome of this project is an offer of the key technology support for water-quality protection within WDSs, and hence would possess important social and scientific significance.

近年来，供水管网水质污染事故时有发生，对居民生活和社会经济都产生严重影响，已成为社会稳定与发展迫切需要解决的重大问题。水质监测是实现供水管网输水安全的重要手段，而如何优化构建水质监测系统是该研究领域的重要科学问题。为突破解决该科学难题，本研究拟提出基于污染水量、污染时间、污染范围和受害群体规模的综合水质污染指标，并建立污染指标的全概率密度及不确定性分析技术方法，以全面揭示水质污染对管网水体的危害机制；提出基于综合水质污染指标边际效益的监测系统优化构建理论和技术方法，以实现监测系统规模（监测点数目）与布局（监测点位置）的同步优化，并阐释监测系统结构特征（监测点数目与位置）与水质污染指标之间的关联机制；建立人工神经网络与智能算法耦合的优化方法，以突破解决水质监测系统优化模型求解效率低下的技术难题。本项目将为城市供水管网水质安全防控与预警提供关键技术支撑，研究成果具有重大社会和科学意义。

供水管网水质安全直接关系到广大人民群众身体健康与生命安全，是城市发展与社会稳定的重要前提。如何构建经济有效的水质监测系统以最大程度降低水质污染事故危害是目前该研究领域面临的一个难题。本项目以供水管网水质监测系统优化构建为研究对象，以突破解决以往研究中关键技术难题为目标，建立了基于污染水量、污染时间、污染空间范围和受害群体规模的综合水质污染指标，并形成了污染指标的全概率密度及不确定性分析技术；提出了基于综合水质污染指标边际效益的监测系统优化构建理论和技术方法，以实现监测系统规模与布局的同步优化；研发了人工神经网络与智能算法耦合的优化技术，以突破解决水质监测系统优化模型求解效率低下的技术瓶颈。研究的重要结果包括：①传统的污染指标期望值在表征水质监测系统有效性上存在明显不足，忽略了极端污染事故的危害性，而本研究提出的污染指标全概率密度分析方法可有效的解决该问题，为供水管网水质物联网的构建提供了理论基础（科学意义）；②建立了基于多目标优化的监测系统规模与布置方案同步优化方法，基于帕累托前沿上的优化解，通过分析不同监测系统结构特征（规模与布局）与水质污染指标体系之间的关联机制，确定了边际效益最高的监测系统优化布置方案（工程价值）；③建立了水质监测物联网与人工取样检测相结合的污染物高效定位方法，形成了水质监测预警-人工检测定位的协同技术。研究成果应用于实际供水管网，结果表明基于污染指标全概率密度的水质监测方案的监测效率比传统方法提高了约15%，水质监测预警-人工检测定位效率比传统污染物定位方法提高了约3倍。以该基金资助的研究成果发表学术论文5篇，其中SCI论文3篇（2篇为第一作者且该基金项目为第一资助）。