Study Of Data Fusion And Data Mining In Chlorine Gas Monitoring WSN

The need of chemical raw materials is increasing with the rapiddevelopment of our economy. Chlorine is yellow and poisonous underatmospheric pressure. Chlorine is also a kind of Strong oxidizer and chlorinatingagent. Chlorine is widely used in disinfection、industrial bleaching、productionof pesticides and so on. If the chlorine gas leak，the harmfulness will be verystrong, affected extensively and will be very hard to disposal. Therefore，measures should be taken both in the process of production and transportation, incase of serious consequences caused by chlorine leak.The chlorine monitoring network has been designed in this paper. In order tomonitor the chlorine gas which be in the transportation or application. Thewireless sensor network (WSN) has been designed considered of thecharacteristics of chlorine gas. Thus the monitoring of remote and large-scalewill be very convenient. In this paper，the Chlorine concentration monitoringterminal has been designed according to the characteristics of the chlorine gasWSN, which including sensor module, wireless communication module, displaywarning module, power supply module and data fusion processing module. Andthe sensor module respectively included temperature, humidity and concentrationsensor. Wi-Fi is the main communication way of communication module.The chlorine gas concentration and surrounding environmental index beentransferred by Wi-Fi from monitoring terminal to intermediate nodes. Theintermediate nodes have been sited near the monitoring station so that the energycould be supplied. The data will be processed by intermediate nodes beforetransfer to monitoring center. And the monitoring curve will be intuitivedisplayed through PC. Neural network algorithm could be used to predict thefuture trend of WSN according to the history data. So the monitoring people willlearn the states of the WSN earlier, and the warning model is improved. The RBF Neural network algorithm was used for data fusion in the paper, RBF algorithmhas been embedded in intermediate nodes, and the data fusion process wasrealized by intermediate nodes. And the fusion results were transmitted tomonitoring terminal. Thus the energy consumption has been reduced whichcoursed by data transformation. Also the BP neural network algorithm was usedto forecast the network trend.