基于伴随方程的河渠污染源反演模型及水污染优化控制研究

Study on prediction and control of river pollution based on water quality numerical model has always been a research focus in the water environment research field. With the further study of the water quality model, emphasis has been transited from model exploitation to improving model reliability and evaluation ability. Currently, the urban water environment monitoring has been extensively conducted and large amount of real-time monitoring data provide the possibilities of water quality model optimizing and water pollution control. The mechanism model of water pollution, with the capability of parameter optimization and key index sensitivity analysis, will be undoubtedly more promising and practical in the future. This research will focus on the normal problems of the water quality model in water pollution early-warning and control, which are the main odds of model correction and delay of control plan. The adjoint equation method will be combined to the channel water quality numerical model for carrying out research on the model parameter calibration, pollution source identification and pollution control scheme optimization. Based on these models, we propose to build a laboratory scale computer-controlled prototype to study the pollution source identification and control scheme optimizing model. The executive ability and operation efficiency of the whole model will be tested in dealing with prediction, inversion and control function. The theoretical and fundamental study is aimed to develop the control algorithm and find the efficient pollution solution method, which will be the theoretical support and technical basis on building a real-time water pollution control system assembled with pollution monitoring, predicting and control scheme optimizing.

基于水质数学模型的水污染预报和控制研究一直是水环境领域的研究热点。随着水质模型研究的深化，研究的重心逐渐从模型开发转移到改善模型的可靠性和评价能力方面。当前我国的河渠水环境监测建设逐渐完备，大量的实时监测数据为开展水质模型优化和水污染控制研究提供了可能，具备参数优化和关键指标敏感性分析能力的机理性水质模型在水环境领域无疑具有更大的发展潜力和实用价值。本课题针对当前河渠水污染预警和控制系统中普遍存在的模型校正困难、控制方案滞后等问题，将伴随方程方法应用到河渠水质数学模型中，开展对模型参数率定、污染源识别和污染控制方案优化等问题的研究。在此基础上，通过建立一套实验室污染源识别和控制模型系统，研究系统在集合预报、反演和控制功能下的执行能力和运行效率。通过前期的机理性、基础性研究，建立控制算法，寻找高效解决方案，为建立河渠水污染实时监测、预报和控制系统提供理论支撑和技术依据。

基于水质数学模型的水污染预报和控制研究一直是水环境领域的研究热点。随着水质模型研究的深化，研究的重心逐渐从模型开发转移到改善模型的可靠性和评价能力方面。当前我国的河渠水环境监测建设逐渐完备，大量的实时监测数据为开展水质模型优化和水污染控制研究提供了可能，具备参数优化和关键指标敏感性分析能力的机理性水质模型在水环境领域无疑具有更大的发展潜力和实用价值。本项目针对当前河渠水污染预警和控制系统中普遍存在的模型校正困难、控制方案滞后等问题，将伴随方程方法应用到河渠水质数学模型中，开展了对模型参数率定、污染源识别和污染控制方案优化等问题的研究。开发了基于复杂地形的浅水方程Well-balanced浅水模型及其算法，对浅水水流模拟中TVD坡度限制器的性能进行了对比分析，改进了水深平均的二维浅水模型中植被拖曳力的计算方法。将伴随方程方法分别应用于解析解模型和一维、二维浅水动力学模型中，开发了基于伴随同化方法的河湖水动力水质数学模型，成功的反演了河道水质模型的纵向离散系数和瞬时多点源的源强。成果应用于白沙洲水厂水源地水质安全预测和污染控制、泵站控制条件约束下的南水北调东线河渠水力计算、洪泽湖区污染物突发事件影响预测和优化控制、以及三峡水库糙率系数和入库流量反演中。通过开展的机理性、基础性研究，建立的控制算法为建立河渠水污染实时监测、预报和控制系统提供理论支撑和技术依据。成果在水污染控制领域具有一定的推广的应用价值。