面向工业园区的污染气体排放监测系统设计

发布时间：2018-03-27 16:09

本文选题：工业园区　切入点：气体排放监测　出处：《北京交通大学》2016年硕士论文

【摘要】：随着我国工业的快速发展,污染源的种类日益增多,特别是工业园区、工业集中区及周边环境,环境污染负荷逐渐增加,环境污染事故时有发生。工业园区是污染气体排放的重点分布区域,园区企业因工艺废气排放、仓储装卸、物料转移、烟气排放等引起气体排放及污染扩散问题,引起的居民投诉日益增加。工业园区的企业种类较多,产生的大气污染物也较多,需要监测的大气污染物种类较多,成分复杂,监测的范围广,因此研究并设计面向工业园区的气体排放监控系统已经成为工业园区污染排放管理的当务之急。本文首先系统地介绍了监测系统总体框架,阐述了无线传感器网络的基本概念,随后重点设计了无线传感器网络的三大模块：感知模块、处理器模块和无线通信模块。从监测系统的性能要求和应用需求角度出发,设计了园区监测设备的硬件平台。在无线传感器监测系统的核心算法中,本文的气体扩散模型采用高斯烟羽模型。文中首先讨论了园区监测传感器布点方案,并通过传感器相关系数去除传感器冗余,从而减少不必要的设备成本和通信开销。针对园区多个排放源的情况,构建源强反算模型,最终将多源反算问题归结为最优化问题的求解。本文基于粒子群算法进行改进,提出了一种自适应迭代次数随机粒子群(AIT-SPSO)算法来增强算法的全局搜索能力和减少算法复杂度,能够精确快速的计算排放点的源强值,从而满足系统对园区进行实时监测的要求。此外,针对传感器零点漂移问题,建立传感器网络邻接拓扑图,采用BP神经网络训练传感器及其邻接传感器测量的历史数据,将训练结果与实测结果进行比较,通过软件实现对传感器漂移的判定。工业园区监测系统应用软件是使用JAVA编程,基于Eclipse平台开发的轻量级系统。采用SpringMVC+Hibernate作为底层开发框架。数据库采用的是SQLServer 2008数据库,按需设计数据库表单,并结合Hibernate技术实现对象数据持久化功能。在模块实现介绍中还给出了软件开发后期系统测试的优化,主要是针对系统后台性能的改善进行展开。
[Abstract]:With the rapid development of industry in China, the types of pollution sources are increasing day by day, especially in industrial parks, industrial concentration areas and surrounding environment, the environmental pollution load is gradually increasing. Environmental pollution accidents occur from time to time. Industrial parks are the key areas for the distribution of polluting gas emissions. The enterprises in the parks cause gas emissions and pollution diffusion problems due to process exhaust gas emissions, storage, handling, material transfer, flue gas emissions, etc. There are more types of enterprises and more air pollutants in industrial parks. The types of air pollutants that need to be monitored are more, the composition is complex, and the scope of monitoring is wide. Therefore, the research and design of gas emission monitoring system for industrial park has become the urgent task of pollution emission management in industrial park. Firstly, the general framework of monitoring system is introduced systematically, and the basic concept of wireless sensor network is expounded. Then, three modules of wireless sensor network are designed: sensing module, processor module and wireless communication module. In the core algorithm of wireless sensor monitoring system, the gas diffusion model of this paper adopts Gao Si smoke plume model. The sensor redundancy is removed by sensor correlation coefficient, which reduces the unnecessary equipment cost and communication overhead. In view of the situation of multiple emission sources in the campus, the inverse calculation model of source strength is constructed. In this paper, an adaptive iterative times random particle swarm optimization (AIT-SPSO) algorithm is proposed to enhance the global search ability and reduce the complexity of the algorithm. It can calculate the source strength of the emission point accurately and quickly, so as to meet the requirement of real-time monitoring of the park. In addition, to solve the problem of sensor zero drift, the adjacent topology graph of sensor network is established. BP neural network is used to train the historical data of sensors and their adjacent sensors. The training results are compared with the measured results, and the drift of sensors is judged by software. The application software of industrial park monitoring system is programmed with JAVA. A lightweight system based on Eclipse platform is developed. SpringMVC Hibernate is used as the underlying development framework. The database adopts SQLServer 2008 database, and the database form is designed on demand. In the introduction of the module, the optimization of system testing in the later stage of software development is given, which is mainly aimed at improving the background performance of the system.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：X84

【相似文献】