Investigation Of SBR Sewage Control Technology Based On Back-propagation Neural Network

With the development of computer and automatic control technology, theapplication of SBR process is more widely used in the treatment of urban sewage andindustrial wastewater. It has become the best choice of the treatment of small-mediumurban sewage and industrial wastewater. SBR process is still a developing sewagetreatment technology. The development of control strategy is the key of the SBRenergy consumption under the conditions that the reasonable hardware equipment isprovided. Intelligent Control determines the development direction for SBR controlstrategy. Moreover, sewage treatment of highly complexity, multi variable, andnonlinear, neural network has created the conditions for real-time control, also itmanifests high potential in energy saving system. Today, the wastewater treatmentsystem is mostly adopting Computer Control Technique of the DCS and PLC. Manyscholars set up to studying sewage treatment control system based on embeddedsystem.Embedded Linux has some shortcoming—development complexity, lack ofstrong technical support. Therefore the paper presents Windows7componentizedembedded operating system which is similar to desktop Windows systems--WindowsEmbedded Standard7instead. The paper researches the customization of WES7andanalyzes the function modules in detail.The paper introduced the back-propagation neural network, focused on trainingalgorithms. Based on the Windows Embedded Standard7operating system, Matlab2013a neural network toolbox is used to establish a three layers structure (3-8-2) BPneural network. The network could predict aeration quantity and aeration time of theaeration section, through the water inlet COD, initial DO, initial MLSS in SBRprocess. In Dev C++IDE, the C language program is analyzed and rewrited to realizethe prediction of the value of pH time series of SBR process with BP neural network.