【作者】 王鸿翔；

【导师】 张土乔； David Z.Zhu；

【作者基本信息】 浙江大学 ， 市政工程， 2009， 博士

【摘要】 本文围绕给水管网水质模型校正问题进行了相关基础性研究,并通过对二次加氯站布置研究讨论了水质监控的相关问题。首先,基于多组分物质混合理论,利用EPANET2.0和EPANET-MSX工具包编程模拟计算,采用蒙特卡罗模拟的方法对影响水质模型准确性的主要参数进行了不确定性分析,结果表明,各水源不同水质主体余氯衰减系数K_b和管壁余氯衰减系数K_w较之其它参数对于水质模型节点余氯浓度不确定性影响程度较大;在此基础上,对两个水源不同水质的水流在管网中混合运行后的各管段主体水流余氯衰减系数进行了确定估计,并基于该方法对多个水源不同水质供水管网管道的主体水流余氯衰减系数的确定进行了扩展,经过算例验证表明,该研究为管网水质模型中主体水流余氯衰减系数的确定提出了一种更为快捷的方法;基于管壁余氯一级衰减模型,建立了多工况下管壁余氯衰减系数校正的优化模型,采用Max-Min蚁群算法求解给水管网管壁余氯衰减系数校正数学模型,算例证明了该校正模型方法的正确性和可行性,为更加准确地描述管网水质模型和更好地解决水质问题,提供了新的研究途径。其次,基于D-优化方法和一次二阶矩(FOSM)理论,定量研究了给水管网水质模型管壁余氯衰减系数和模型预测值的不确定性问题,研究表明,在节点流量较大情况下,采集校正数据得出的校正参数的不确定性比较小,模型预测值不确定性相对较大,供水管网水质模型不确定性研究为水质模型系统全面校正提供了一定的依据和方法;基于水质模型参数及模型预测值不确定性研究,提出了以模型参数不确定性最小及采样点设计花费最小为目标函数,建立了水质模型校正的采样点优化布置多目标数学模型,采用NSGA-Ⅱ多目标遗传算法对该模型进行了求解研究。通过算例分析,结果表明该方法有效可行,对于水质模型校正的采样设计问题具有一定的参考价值。再次,以管网供出的达标水量最大为目标函数,结合加氯站布置的部分影响原理,建立了二次加氯选址的数学模型,采用粒子群优化算法对一级二次加氯点布置优化进行了优化选址,取得了较好效果;引入管网余氯与来自水源有机物反应衰减的水质模拟模型,以管网中包括水源地在内的各加氯站总加氯量最小为目标函数,考虑余氯与有机物反应后各节点余氯浓度及所对应副产物浓度符合要求为约束,建立了二次加氯投加量的优化模型,采用遗传算法对该模型进行了优化求解。基于以余氯浓度值为基础的管网水质服务水平的理论,对二次加氯前后管网水质服务水平情况进行了评价比较,结果表明,采用二次加氯站优化结果相对未设置二次加氯站之前,不仅提高了服务水平,有效地减小了投加氯总量,而且避免了余氯量过大及由此可能产生的负面作用,为管网水质模型和水质监控管理运行提供了参考依据。最后,在实例研究中实现了管网的水质模型主体余氯衰减系数和管壁余氯衰减系数确定和校正,对基于水质模型校正的采样点优化布置问题进行了求解,并在准确校正水质模型之后,对管网二次加氯站进行了优化选址和投加量方案的优化,分析计算结果表明,本文提出的相关模型和方法具有较好的实用性。更多还原

【Abstract】 Calibration can be defined as a process in which a number of WDS model parameters are adjusted until the model mimics behavior of the real WDS as closely as possible. The focus of this dissertation is the relative research about water quality model estimates, calibration and monitoring, besides booster chlorination setup.Firstly, chlorine bulk and wall decay coefficients have been estimated and calibrated. Based on EPANET2 and EPANET-MSX toolkits, the effect of individual parameters uncertainty, as well as the combined effect of the parameters uncertainty, on water quality in a distribution system was analyzed by Monte Carlo simulation (MCS). It concludes that chlorine bulk and wall decay coefficients have larger influence on model prediction comparing with other parameters. Then chlorine bulk coefficients in two-source water distribution system networks, two sources of which have different source bulk decay coefficients, have been estimated. The method can be used in multi-sources networks, which can provide a novel means to get the bulk chlorine coefficients instead of measuring in the lab. Besides, a general calibration model under multi-mode for identifying wall decay parameters is formulated, based on the simple first-order reaction of chlorine and the Lagrange time-based approach of dynamic water quality model. Multi-mode model was analyzed to collect more node residual chlorine for calibration. Max-min ant colony system algorithm was proposed to solve the calibration model that is coupled with hydraulic and water quality simulation models using EPANET2 Toolkits. It concludes that the methods can obtain the solution quickly and effectively.Secondly, based on D-optimization and FOSM theory, the uncertainties involved with estimated parameters and the model prediction uncertainties for critical demand conditions due to the parameter uncertainties are calculated. Then location optimization of sampling and measurement for water quality calibration has been proposed based on estimated parameters and the model prediction uncertainties, and has been resolved by multi-objective genetic algorithm NSGA-II. The number of measurement can be determined from the obtained pareto front solution. NSGA-II is very befitting for resolving such two-objective optimization problem.Thirdly, booster chlorination added to water distribution to maintain disinfectant residuals and avoid high dosages at water sources has been researched. The objective function which maximizes the certified water quantity to consumer, considering partial influence on nodes by booster chlorination, has been proposed. The mathematical model has been solved by particle swarm optimization algorithm; given a set number of boosters’ stations. By considering the natural organic matter reacting with the chlorine and the disinfectant by-products concentration under allowed maximization as one of the restrictions, the chlorine injection rates of chlorine boosters have been optimized by minimizing the total chlorine mass to be injected to maintain required residuals, using genetic algorithms. Then though introducing performance evaluation system of networks service level, the level of service has been compared between no booster chlorination and re-chlorination. It concludes that booster chlorination not only increase the service level of networks, but can allow lower average chlorine and decrease the variability of chlorine residuals through a water distribution network, which lead to a lower dosage and may reduce the formation of disinfectant by-products.Finally, the methods of calibrating chlorine bulk and wall decay coefficients have been used in a realistic network. Then sampling design based on water quality model calibration has been optimized and booster chlorination stations have been displayed including the optimal location and chlorine injection rates in different time. It concludes that the model and method which have been proposed in this research can be applied to solve relative problems.更多还原