【作者】 徐华；

【导师】 薛恒新；

【作者基本信息】 南京理工大学 ， 系统工程， 2009， 博士

【摘要】 空气和水与人们的生活息息相关,随着环境保护工作力度的加强,传统的对污染源进行定期检查和应付污染事故的手段已不能有效的、及时的满足社会的需要,利用现代技术来武装环保执法队伍,改进环保手段是提高环保能力的重要途径。为了确保环境最大安全,有必要对监测布点、预警应急、污染处理等管理的理论和方法进行深入系统地研究。本文的具体内容如下：介绍了论文工作背景,对后面几章中重点讨论的优化布点、水质评价、预警与应急、环境污染预测、污水处理控制和再利用等目前国内外的研究现状进行了分析和研究,为后面几章打好基础,并给出了论文的研究目的和将要解决的主要问题。研究了一种新颖的蒙特卡罗估计选择办法(MCES)。采用多宇宙并行量子遗传算法对MCES方法进行优化,使其具有更快速的收敛能力和更好的搜索效率。该方法用于环境监测优化布点,具有简便、快速、结果合理稳定、易于推广等优点。另外用一种新的TSK模糊系统CTSK (Central TSK Fuzzy System)进行地下水质的评价取得了很好的效果。对于大样本数据集,设计出了CCMEB-CGBR(CCMEB-based Constrained GBR)算法。通过建模来实现环境污染预测系统。通过引入进化学习方法提出了视觉TSK模糊系统新的进化学习训练算法。新的基于进化学习的算法不但具有更好的全局收敛能力,同时依然能继承视觉模糊建模方法的原有优点。另外,大气污染物浓度的变化具有较强的非线性特性,本章研究了自适应神经模糊推理系统(ANFIS)方法基于Sugeno模糊模型,在此基础上给出了网络权值的修正算法。用共轭梯度法提高其前提参数的学习速度。结合国内外应急管理的现状,制定了预警应急体系的工作原则,系统设计了预警应急管理体系的总体框图。并且探讨了预警与应急管理体制的未来发展趋势。采用粒子群优化算法对一种新的级联MLP神经网络CATSMLP进行优化使其鲁棒性增强,具有更快速的收敛能力和更好的寻优能力。将其用于藻类污染预警取得了很好的效果。针对具有凸多面体参数不确定性的污水处理控制系统,研究其鲁棒H_∞保成本控制器的设计问题。应用所设计的鲁棒H_∞保成本控制器进行研究,结果表明了本章方法的可行性及有效性,并且系统最终的稳态误差较小。并且提出污水等可作为可再生能源的有效建议和方法。对论文作了总结,提出了本文的主要创新点,针对论文研究工作中的不足提出了对未来研究发展前景的展望。更多还原

【Abstract】 Air and water is closely related to people’s daily life. With the strengthening of environmental protection efforts, the traditional means of to carry out regular inspection of pollution sources and to cope with pollution accidents can no longer meet the needs of the society effectively and timely. Using of modern technology to arm the environmental law enforcement team and improving environmental protection methods are important approaches to improve the environmental capacity. In order to ensure maximum security environment, it is necessary to carry out a deep and systematic research on management theories and methods such as monitoring points, emergency warning, pollution controlling and so on.The main contents of this papers are as follows:Firstly, introduces the background of this paper, mainly analysis the present research optimal point, water quality assessment, early warning and contingency, environmental pollution forecast, control of wastewater and reuse etc. will be focused on in later section at home and abroad, lying a good foundation for later chapters, and it also list out the purpose of the research paper and the main problems to be solved.Secondly, studies in a novel Monte Carlo Estimate Methods (MCES). Multi-universe parallel quantum genetic algorithm method for optimizing the MCES, thus it will have a faster convergence and better search efficiency. This method used for optimization of environmental monitoring points has the following advantages:simple, rapid, reasonably and stable results, easy to promote etc. Then evaluating the quality of underground water with a new derivation of the TSK fuzzy system CTSK (Central TSK Fuzzy System) achieves pretty good results. This article has designed a novel algorithm named CCMEB-CGBR(CCMEB-based Constrained GBR) for large data sets.Thirdly, is about to achieve environment pollution prediction system by modeling. The evolutional learning method is introduced to train the parameters of the Visual TSK fuzzy system. The proposed new method not only has the better global learning abilities but also can inherit the virtues of the original visual TSK fuzzy system modeling method. In addition, changes in the concentration of air pollutants has a strong non-linear characteristics, this chapter research on Adaptive Neuron-Fuzzy Inference System (ANFIS) method which is based on the Sugeno fuzzy model, on this basis it gives out the correction algorithm of network weights. We use conjugate gradient method to improve the learning speed of the premise parameters.Fourthly, combines the status of emergency management at home and abroad, establishes principles of early warning emergency system, systemically designs the overall block diagram of early warning system for emergency management. And it discusses the future development trend of early warning emergency management system. By using particle swarm optimization algorithm to optimize a new cascade of MLP neural network CATSMLP, it can enhance its robustness, enables itself with faster convergence and better capacity optimization. Be used for algae pollution warning has a good effect.Fifthly, mainly talks about a control system for sewage treatment with parameter uncertainty convex polyhedron, studies the design problem about its robust Hoo guaranteed cost controller. With the application of the design of robust H∞guaranteed cost controller simulation study, simulation results show that the method is the feasibility and effectiveness, and finally the steady-state system smaller error. And presents methods and effective proposals about such as the sewage and so on can be used as renewable energy sources.Finally, is a summation of the essay; it lists out the main innovation in this article and proposes research and development for the future outlook for the lack of essay’s research work.更多还原