Wastewater Treatment Modeling And Control Of Dissolved Oxygen Concentration Based On Artificial Intelligence

The system of Sewage treatment is a complex non-linear and time-varying systems. It is very difficult to establish its precise mathematical model throughout the process control in wastewater treatment. For the specific application, we can only use the traditional control methods, and the accuracy is not high, adaptive capacity is weak. Therefore, relying solely on the traditional control methods have been unable to achieve the control requirements. At present, artificial intelligence systems are constantly developing, and in many nonlinear systems they has been widely used. In which adaptive neural network has a strong ability to any degree of accuracy that it can approach any nonlinear function; The fuzzy control in industrial applications has also been the rapid development, the process of fuzzy control has superior dynamic response and it has a strong adaptability in the changes in process parameters.The papers analyzes the dissolved oxygen concentration control requirements of sewage treatment systems'process first, established the idealized simplified model of the activated sludge treatment system. According to this simplified model proposed an adaptive fuzzy neural network control program and its application to the wastewater treatment system. This control method combines the fuzzy logic systems and neural network advantages, it's not only has a strong self-learning ability, but also can automatically generate fuzzy rules, and adjust the membership function through practical requirements of control object. In the controller training process to get the optimal parameters of the controller, so that the network used in wastewater treatment system can be able to quickly achieve a better control requirements.The papers established the sewage treatment systems based on fuzzy neural network PID control by researching the above adaptive fuzzy neural network design of wastewater treatment system. This new method does not depend on the precise model of sewage treatment systems, and this makes the whole control in wastewater treatment applications, a more universal significance. Through Dynamic Recurrent Neural Network(Elman) which identified the control object first, and the control object is also activated sludge wastewater treatment system of dissolved oxygen concentration in aeration tank, then combined the fuzzy control and neural networks, through the neural network to realize fuzzy logic, using the advantages of the neural networks and fuzzy control-line to adjust the parameters of PID at the same time. So the controller of a fuzzy neural network that is not only has the self-learning ability, but also made full use of advantages of PID control, achieved good simulation results.