| There are strong coupling, uncertainty, time-varying parameters, strong nonlinear and so on in the process of wastewater treatment. At present, activated sludge wastewater treatment is widely used in domestic sewage treatment plant, which mainly using microbial growth,reproduction and biological reaction to remove organic pollutants. In wastewater treatment process, an important indicator is the dissolved oxygen content of the aeration tank, the accurate prediction and control of dissolved oxygen concentration is a prerequisite to ensure the ideal effect of the sewage treatment system. At present, the traditional measurement algorithms such as iodometric method, oxidation electrode method, spectrophotometric method and fluorescence quenching method have been unable to meet the accurate measurement of dissolved oxygen in sewage treatment. At the same time, soft sensing technique well solves the problem in sewage processing.Aiming at the problems existing in the prediction of dissolved oxygen in the current wastewater treatment process, this paper has done the following research work on the basis of analyzing and summarizing the existing research results.(1)This topic research significance was analyzed, sewage treatment method of three steps and activated sludge process were stated simply, and the soft sensing technology of domestic and foreign research status are reviewed.(2)Soft sensing technology for the basic principle of the system is described, and the soft measurement modeling process and modeling method are explained one by one, which modeling steps is divided for the selection of process variables, data acquisition and processing, establishment of soft sensing model and correction of the soft sensing model. The modeling methods are described around the process mechanism, regression analysis, state estimation and other seven kinds of methods. Four kinds of intelligent optimization algorithm basic principle are introduced, such as BP neural network, SVM algorithm, particle swarm algorithm and genetic algorithm. For the sewage treatment process combined with soft measurement and intelligent optimization algorithm to do the groundwork.(3)On the foundation of the benchmark simulation model BSM1, using Simulink to build the adaptive dynamic simulation model of the concentration of dissolved oxygen(do),and analyzes the two main factors’ influence of the aeration tank volume of oxygen and oxygen necessary consumption on the concentration of dissolved oxygen(do). On the steady-state and dynamic model of dissolved oxygen concentration in a comparative study,the dynamic model of dissolved oxygen concentration is more accurate and stable, and the mechanism of dissolved oxygen concentration has a further understanding of the mechanism.( 4) Aiming at the problem that the concentration of dissolved oxygen cannot be measured online, using genetic algorithm to optimize the weight and threshold value of BPneural network, a model of dissolved oxygen soft sensing model based on genetic algorithm to optimize BP neural network is established. The prediction of dissolved oxygen concentration was successfully achieved. Through the contrast experiment before and after optimization, it was found that the prediction accuracy of the optimized model was greatly improved.( 5) Based on the research of particle swarm optimization algorithm, the model of dissolved oxygen concentration is established by using the particle swarm optimization support vector regression machine. According to the soft sensing theory, the dissolved oxygen concentration of the support vector regression machine was predicted by MATLAB software.At the same time, the RBF algorithm is added to forecast the contrast experiment. Through the comparison of the results, the practicability and accuracy of the particle swarm optimization support vector regression model is verified, which provides a guarantee for further research on the prediction and control of dissolved oxygen concentration in wastewater treatment. |
PDF Full Text Request