Research Of Sewage Water Qualitative Prediction Based On Wireless Sensor Network

With the rapid development of industry, the amount and type of industrial sewage increased rapidly, and the pollution become widespread and serious, which are threatening the human health and safety, detection and treatment of sewage have been researched and attention from scholars. With the development of Internet technology, Wireless Sensor Network(WSN) as the Internet of things the underlying technology is applied more and more. In recent years, Wireless Sensor Network(WSN) is widely used in industrial environment monitoring, including in the sewage treatment industry. In addition in order to realize the intelligent control and predict warning for sewage treatment, in view of the important water quality parameters in wastewater treatment, we through the forecast evaluation to meet the requirements of practical application.First of all, in this paper, we design a ZigBee wireless sensor network model of wastewater quality data collection. For the issues of how to optimize routing for the network and how to reduce network energy, in this paper analyzes ZigBee Routing algorithm, for ZBR algorithm more routing hops in the cluster, route discovery broadcast storm, unnecessary consumption of energy and other issues suggest improvements. In this paper we propose a new routing algorithm ZBR- L. Within the cluster routing, it through the establishment of routing table neighbors optimized routing path to reduces routing overhead; Secondly, when routing discovery it limiting the RREQ request packet send, to save energy, make the network nodes in longevity. At last, the improved algorithm has carried on the simulation analysis, comparing with the ZBR algorithm, which proved that ZBR-L in packet delivery of network traffic indicators has a better result.Second, we through the wireless sensor network to transmit data to the database, and study the water quality data. A sewage effluent COD value forecast model was established by the neural network, and the use of measured data simulation. But because of BP neural network to forecast has some errors and other problems, and artificial bee colony algorithm(ABC) with global searching capability characteristics. Artificial bee colony algorithm to neural network weights and threshold optimization has been used, and establish ABC-BP model to reduce errors. Compared result with before, the experiments show that artificial colony optimization results are the most close to the expected output, error smaller and more effective.