SCEUA算法在地下水污染溯源中的应用研究

发布时间：2018-01-15 01:04

本文关键词：SCEUA算法在地下水污染溯源中的应用研究　出处：《济南大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着经济的快速发展,全国各地的地下水都受到了不同程度的污染。但我国地下水污染预防、监测和修复系统还不完善,因此导致人类赖以生存的地下水环境遭受到了严重的威胁。近年来,人们越来越关注地下水的污染问题。由于地下水污染具有隐蔽性,因此对地下水污染预防和治理的关键是及时有效地确定污染源。然而,目前我国对地下水污染源的确定方法还不够成熟,存在较多不足。本文针对这一问题探索了地下水污染溯源的方法。优化算法是进行地下水污染溯源研究的一种有效的方法。由于优化算法在不断地更新发展,不同优化算法有其各自优点,将其应用到许多领域后体现了很好的优越性,解决了一些原本传统方法不能解决的问题。本文将SCE-UA优化算法与地下水溶质运移程序(MT3DMS)相结合的基础上,建立地下水污染溯源智能搜索模型(SCEUA-MT3DMS)。首先通过优化算法模型来生成污染种群样本点,然后再利用种群样本点,通过地下水污染物运移数值模型来模拟污染物在含水介质中的渗流与迁移过程,同时将模拟预测的结果进一步作为优化算法模型选择的依据。优化算法模型经过种群变异、反射和进化等过程,从众多潜在污染源位置中自动选择出可能性最大的污染源位置及污染物的排放强度。建成地下水污染溯源智能搜索模型后,通过对单污染源和双污染源、已知位置和未知位置及稳定流和非稳定流等不同情况下模拟实验,一般在进化代数T=40,复型个数p=40的情况下,均能够准确定位污染源真实位置,所反演的污染浓度与真实值间的相对偏差D2均小于2%,可以满足精度的需要,验证了该方法的可行性,反映了智能搜索模型能够高效地收敛到全局最优解,对污染源位置和浓度的反演定位准确,能够进一步应用于水文地质条件更为复杂的实例污染源反演问题。宁阳工业园区地下水污染相对集中且污染比较严重,水文地质资料等相对全面,将该模型应用到宁阳工业园区进行区域大模型的实际案例,考虑到区域大模型存在的问题,采用了网格嵌套数值模拟方法,子模型实现了对研究区的局部细化,减少了不同化工企业之间的相互影响,提高智能搜索模型地下水污染溯源反演效率和反演精度,有效的缩短了模型的运行时间。经分析选定位置合适的某化工厂、监测井及两种特征污染物多氯联苯(PCB)和硝酸根离子,运用该模型进行有关污染源位置和污染物排放浓度的反演定位。反演结果基本可以满足精度需要,也比较符合污染源的实际位置及浓度情况,通过实例体现了该地下水污染溯源智能搜索模型较好的准确性和可靠性。研究表明,通过将SCE-UA算法与地下水溶质运移程序(MT3DMS)耦合来进行地下水污染反演是一种有效的方法,可以快速准确的定位污染源,达到了溯源的目的,体现了方法的正确性和可行性。
[Abstract]:With the rapid development of economy, groundwater in various parts of the country has been polluted to varying degrees. However, the system of prevention, monitoring and restoration of groundwater pollution in China is not perfect. In recent years, people pay more and more attention to the problem of groundwater pollution, because of the hidden nature of groundwater pollution. Therefore, the key to the prevention and control of groundwater pollution is to determine the source of pollution in time and effectively. However, at present, the method of determining the source of groundwater pollution in China is not mature enough. There are many shortcomings. This paper explores the method of groundwater pollution traceability. Optimization algorithm is an effective method for groundwater pollution traceability research. Because the optimization algorithm is constantly updated and developed. Different optimization algorithms have their own advantages, and their application in many fields shows a good advantage. Some problems that could not be solved by the traditional method were solved. In this paper, the SCE-UA optimization algorithm is combined with the groundwater solute transport program MT3DMS. An intelligent searching model for traceability of groundwater pollution was established. Firstly, the pollution population sample points were generated by optimizing the algorithm model, and then the population sample points were used. The seepage and migration process of pollutants in aqueous medium is simulated by means of a numerical model of groundwater pollutant transport. At the same time, the results of simulation prediction are further used as the basis for the selection of the optimization algorithm model, which goes through the process of population mutation, reflection and evolution. The most likely pollution source location and pollutant emission intensity are automatically selected from many potential pollution sources. After the establishment of the intelligent searching model of groundwater pollution traceability, the single and double sources of pollution are analyzed. Under different conditions, such as known and unknown position, stable flow and unstable flow, the simulation experiments can accurately locate the true location of the source of pollution in the case of evolutionary algebra T _ (40) and the number of complex types (p _ (40)). The inverse relative deviation between the pollution concentration and the real value is less than 2, which can meet the need of accuracy. The feasibility of this method is verified, which reflects that the intelligent search model can converge to the global optimal solution efficiently. The exact location of the location and concentration of pollution sources can be further applied to the source inversion of more complicated hydrogeological conditions. The pollution of groundwater in Ningyang Industrial Park is relatively concentrated and serious. The hydrogeological data are relatively comprehensive. This model is applied to the practical case of the regional large-scale model in Ningyang Industrial Park. Considering the problems of the regional large-scale model, the numerical simulation method of grid nesting is adopted. The sub-model realizes the local refinement of the study area, reduces the interaction between different chemical enterprises, and improves the retrieval efficiency and accuracy of the intelligent search model for groundwater pollution traceability. The operation time of the model was effectively shortened. A chemical plant with suitable location was selected, monitoring well and two characteristic pollutants, PCB) and nitrate ion. The model is used to locate the pollution source location and pollutant emission concentration. The inversion results can meet the need of accuracy, and also accord with the actual location and concentration of the pollution source. The example shows that the intelligent searching model of groundwater pollution traceability is accurate and reliable. It is an effective method to retrieve groundwater pollution by coupling SCE-UA algorithm with MT3DMS. it can locate pollution sources quickly and accurately. The purpose of traceability is achieved, and the correctness and feasibility of the method are demonstrated.
【学位授予单位】：济南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：X523

【参考文献】