Tap Water Residual Chlorine Control Process Based On Fuzzy Neural Network

Tap water is essential material resource of people’s daily life and production. It has alarge impact on the national health, safety and economic operation. The water residualchlorine is an important indicator of water quality. It’s too high or too low has a direct impacton the normal use of tap water, so chlorine control is a key procedure in the water treatmentprocess. But there are some difficulties with the chlorine control objective’s characteristics oflarge time-varying, large time lag and large inertia, especially when the disturbance occurred,chlorine control would adjust in time and keep the error in range. Therefore, a rational designof residual chlorine control is essential.This paper is dedicated to solve the chlorine control problems in water treatment processunder the existence of large time-varying and large time lag. The paper adopts fuzzy neuralnetwork controller.The content of this paper has researched mainly included the following aspects:Firstly, this paper elaborates each part of water treatment process, to illustrate thenecessity and importance of residual chlorine control, combined with China’s Standards ofDrinking Water Quality. According to the analysis of chlorine disinfection, residual chlorine,chlorination control methods and residual chlorine side effects, we get to know the chlorinecontrol method and factor preliminary.Secondly, this paper introduces the basic theory of fuzzy control and neural network, andwith comparison of both, we get the overview of fusion and classification of fuzzy control andneural network, and study the design ideas and procedures of fuzzy neural network controllerbriefly.Thirdly, we focus on the application of fuzzy neural network controller, and analyze anddesign its structure and algorithm. According to the simulation in MATLAB, we can find outthat fuzzy neural network in the residual chlorine control system can acquire very goodcontrol quality against to traditional PID controller.Finally, according to the defect of the referred fuzzy neural network controller, we adopts ANFIS controller based on T-S model in the residual chlorine control process, takingflow quantity, turbidity, pH and feedback of residual chlorine as four input variables, thechlorinator opening as output variable. Under the circumstance of the fixed processing time ofthe clean water tank, the residual chlorine of supply water is controlled well and fastresponded. According to simulation of the both fuzzy neural network controller, theapplication of residual chlorine control based on fuzzy neural network is indeed feasible.