URBAN DRAINAGE NETWORK REHABILITATION CONSIDERING STORM TANK INSTALLATION AND PIPE SUBSTITUTION
URBAN DRAINAGE NETWORK REHABILITATION CONSIDERING STORM TANK INSTALLATION AND PIPE SUBSTITUTION
复制标题
考虑安装雨水池和更换管道的城市排水管网修复
DOI：
10.4995/thesis/10251/129869
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Ulrich Aurele
中科院分区：
文献类型：
--
作者：
Ngamalieu Nengoue;Ulrich Aurele
排水网络修复是管理者和责任人需要实施的基本过程之一，以使有缺陷的网络适应气候变化和城市化的不利影响。在文献中，作者提出的两种方案是管道替代或安装暴风罐（STs）。在本文中，第三种方案建议将管道替代和STs安装结合起来进行排水网络修复。对不同管网的模拟结果表明，在排水管网修复中联合使用换管和安装化粪池系统比单独使用两种方案效果更好。不幸的是，这种复原方法需要大量的计算时间。他们需要很多时间来提供可接受的解决方案，而这些解决方案往往陷入局部最小值。本文的目的是提出一种基于管道替代和STs安装结合使用的排水网络修复方法。该方法考虑了搜索空间约简（SSR）技术。所采用的策略在结构化方法中结合了四个关键选项，旨在减少问题的搜索空间（SS）：减少可能安装STs的节点数量。减少可能发生直径变化的线的数量，减少由每个STs的截面构成的离散化。减少管道中候选直径的数量。在搜索空间减小的情况下，本文所采用的单目标优化伪遗传算法（PGA）可以在更短的时间内探索搜索空间，从而获得更好的结果。对于MO优化，NSGA-II可以在优化过程后快速提供不同考虑场景的Pareto front。总体目标分为具体目标，具体如下：第一个具体目标包括制定一个优化问题，验证考虑安装STs和更换管道的修复比单独实施的两种策略中的任何一种都能提供更好的效果。充分评估用于形成目标函数的成本函数构成第二个具体目标。考虑的不同成本有：投资成本和洪水损失成本。第三个具体目标是在PGA和雨水管理模型的基础上，建立一个考虑安装化粪池系统和更换管道的修复模型。由于投资成本和洪水损失成本的类型不同，无法进行汇总。投资成本是真实存在的，而洪涝灾害成本取决于降雨概率。因此，本文的第四个具体目标是提出考虑STs安装和管道替代的排水网络修复MOEA。对于群优化和多目标优化，计算时间增加。也有人怀疑客观解陷入局部极小值。第五个目标是提出一种SSR方法来解决这一问题。第六个具体目标是进行灵敏度分析，验证SSR对优化过程最终结果的影响。因此，选择了不同的种群大小和停止准则值，并对不同配置进行了模拟。本文的第七个具体目标是提出一种新的基于SSR技术的MO优化康复方法。对于本文提出的每一个具体目标，本文都对一个排水网络进行了应用，并取得了令人满意的结果。在一个简单网络中应用基于PGA算法的简单优化方法，在一个中等规模网络中应用SO优化、MO优化和SSR优化方法。最后，在一个大型网格网络中应用考虑SSR的MO优化方法。
Drainage networks rehabilitation is one of the fundamental process that managers and responsible need to implement to adapt defective networks to climate change and urbanization adverse effects. In the literature, pipes substitution or storm tanks (STs) installation are the two scenarios presented by authors. In this thesis, a third scenario proposed combine pipes substitution and STs installation for drainage networks rehabilitation. Results of several simulations on various networks showed that the combine use of pipes substitution and STs installation in drainage networks rehabilitation provides better results than separation of the two rehabilitation scenarios. Unfortunately, such rehabilitation methodologies are computationally time consumers. They need much time to provide acceptable solutions which are often caught up in local minima. The aim of this thesis is to propose a drainage networks rehabilitation methodology based on the combine use of pipes substitution and STs installation. The methodology considers search space reduction (SSR) technique. The adopted strategy combines in a structured methodology four key options aiming at reducing the search space (SS) of the problem: Reduce the number of nodes in which STs could potentially be installed. Reduce the number of lines in which there could potentially be a change in diameter Reduce the discretization that is made of the section of each of the STs. Reduce the number of candidate diameters in the pipes. Once the search space is reduced, the pseudo genetic algorithm (PGA) used in this thesis for single objective (SO) optimization can easily explore the search space in less time resulting in the obtention of better results. For the MO optimization, the NSGA-II can provide rapidly Pareto fronts for the different considered scenarios after the optimization process. The general objective was divided in specific objectives detailed as follow: The first specific objective consists of formulate an optimization problem that verifies that rehabilitation considering STs installation and pipes substitution provides better results than any of the two strategies implemented separately. Adequately assess the cost functions used to form the objective functions constitutes the second specific objective. The different costs considered are: Investments costs and flood damage costs. The third specific objective is to develop a rehabilitation model considering STs installation and pipes substitution, based on PGA and the Storm Water management Model. Investment costs and flood damage costs could not be summed due to their types. Investment costs are reals while, flood damage costs depend on the rainfall probability. So, the fourth specific objective of this thesis is to propose a MOEA for drainage networks rehabilitation considering STs installation and pipes substitution. For SO and Multi-objective (MO) optimization, the computation time is elevated. It was also suspected that the objective solutions were caught up in local minima. The fifth objective is to propose an SSR methodology to solve this issue The sixth specific objective consist of carry out a sensitivity analysis to verify the effects of the SSR on the final result of the optimization process. So, different population sizes and stop criteria values were selected and simulation for different configurations were performed. The seventh specific objective of this thesis is to propose a new rehabilitation methodology considering SSR technique for MO optimization. For each specific objective presented in this thesis, an application to a drainage network has been made and the obtained results were satisfactory. A simple network was used to apply the simple optimization methodology based on PGA algorithm A medium size network was used to apply the SO optimization, the MO optimization and the SSR methodology. Finally, a large and mesh network was used to apply the MO optimization methodology considering SSR.