基于数据融合的工厂污水无线监测系统研究

发布时间：2018-03-04 14:33

本文选题：无线传感网络　切入点：水质监测和评估　出处：《宁夏大学》2017年硕士论文　论文类型：学位论文

【摘要】：水是生命之源,对人类的生存和发展起着至关重要的作用。随着工业的进步,工厂污水排放不断增加,致使水污染情势日益严峻,实现水环境保护与管理的重要措施之一就是对水环境进行有效的监测和评估。本文分析了当前国内外水环境监测系统的研究状况,针对企业排污监测的特点,将数据融合算法和无线传感器相结合应用于工厂污水监测系统中,提出了基于数据融合的工厂污水无线监测系统。本文的主要工作如下:(1)介绍了目前国内外水环境监测系统的技术水平和研究现状,综合无线传感网络技术和数据融合算法,提出将数据融合算法和无线传感网络相结合应用于工厂污水监测系统中。(2)本文以淮安某一氨氮化肥厂为研究对象,设计多层次的数据融合算法,将温度、PH、氨氮、溶解氧、浊度五个污水参数进行数据融合,评估出污水等级。本文采用分布式检测结构,分别在数据层、特征层、决策层进行数据融合。数据层采用格拉布斯准则和中位值平均法对单个传感器多次测量结果进行融合,剔除失真数据,提高测量精度;特征层采用自适应加权算法对监测区域内多个同类传感器的数据进行融合,以求得该区域各个参数的一个整体特征值;决策层采用GA-BP神经网络对五个特征值进行融合得出污水等级。(3)系统硬件电路设计。包括检测节点、网关、监测中心。各个部分采用模块化的思想进行设计,检测节点包括传感器模块和组网通信模块,传感器模块负责对各参数数据的采集,组网通信模块采用CC2530负责数据的融合和无线收发;网关采用CC2530+STM32+GPRS,CC2530负责组网通信,STM32负责数据融合处理,GPRS负责数据的远程发送;监测中心采用电脑进行操作,不需要硬件设计。(4)软件设计。对系统硬件的各个部分进行相应的软件设计,在检测节点处首先采用格拉布斯准则剔除可疑数据,然后利用中位值平均法滤波,提高传感器采集数据的可靠性和精度;网关处使用自适应加权算法对多个同类传感器数据进行融合,以得出一个最优值;监测中心:采用GA-BP神经网络,首先用遗传算法去优化BP神经网络,然后将训练好的GA-BP神经网络用于对水质的等级评估,得出水环境的一个整体指标。最后使用LabView进行上位机友好界面的编写和Web对外发布。(5)对设计的整个污水监测系统进行测试,评估系统性能。测试结果表明本文将数据融合算法和无线传感网络相结合设计出的监测系统,不仅通信稳定、可扩展性好而且能准确对污水进行等级评估,具有可靠性高、通用性强、准确性高等特点,具有良好的应用前景。
[Abstract]:Water is the source of life and plays a vital role in the survival and development of human beings. With the progress of industry, the discharge of sewage from factories is increasing, and the situation of water pollution is becoming more and more serious. One of the important measures to realize the protection and management of water environment is to effectively monitor and evaluate the water environment. This paper analyzes the current research situation of water environment monitoring system at home and abroad, and aims at the characteristics of sewage monitoring in enterprises. The data fusion algorithm and wireless sensor are combined in the plant sewage monitoring system. This paper presents a wireless monitoring system for factory sewage based on data fusion. The main work of this paper is as follows: 1) the technical level and research status of water environment monitoring system at home and abroad are introduced, and the wireless sensor network technology and data fusion algorithm are integrated. Data fusion algorithm and wireless sensor network are applied to plant sewage monitoring system. In this paper, a certain ammonia nitrogen fertilizer plant in Huai'an is taken as the research object, and a multi-level data fusion algorithm is designed. The five parameters of turbidity are fused to evaluate the grade of sewage. In this paper, the distributed detection structure is used in the data layer, the characteristic layer, The data layer uses Grubbs criterion and median average method to fuse the multiple measurement results of a single sensor to eliminate the distorted data and improve the measurement accuracy. The feature layer uses adaptive weighting algorithm to fuse the data of several similar sensors in the monitoring area to obtain a global characteristic value of each parameter in the region. The decision layer uses the GA-BP neural network to fuse the five eigenvalues to get the design of the hardware circuit of the system, including the detection node, the gateway and the monitoring center. Each part is designed with the idea of modularization. The detection node includes sensor module and network communication module. The sensor module is responsible for the collection of each parameter data, and the network communication module uses CC2530 for data fusion and wireless transceiver. The gateway adopts CC2530 STM32 GPRSN CC2530 to be responsible for data fusion processing and remote data transmission, and the monitoring center uses computer to operate, and STM32 is responsible for data fusion and data transmission. There is no need for hardware design. (4) Software design for each part of the system hardware is carried out. At the detection node, the Grubbs criterion is used to eliminate the suspicious data, and then the median average method is used to filter the suspicious data. Improve the reliability and accuracy of sensor data collection; Gateway uses adaptive weighting algorithm to fuse multiple similar sensor data to obtain an optimal value; Monitoring Center: using GA-BP neural network, First, the BP neural network is optimized by genetic algorithm, and then the trained GA-BP neural network is used to evaluate the water quality. Finally, we use LabView to compile the friendly interface of the host computer and Web to release the whole sewage monitoring system. The test results show that the monitoring system designed by combining the data fusion algorithm with the wireless sensor network is not only stable in communication, good in expansibility, but also accurate in evaluating the level of sewage, and has high reliability. It has the characteristics of high generality and high accuracy, so it has a good application prospect.
【学位授予单位】：宁夏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP274

【参考文献】