Modeling water treatment process using fuzzy neural network based on subtractive clustering

Because of nonlinear, time-varying and time-delaying property, itpsilas difficult to model water treatment process by traditional method, so a Takagi-Sugeno fuzzy model based on subtractive clustering algorithm is proposed in this paper. Firstly, subtractive clustering is used to partition the input space and to determine the initial values of premise parameters and fuzzy rules. Moreover, an improved hybrid study algorithm consisting of a back propagation algorithm and least square algorithm is implemented to optimize the parameters. Finally, this proposed method is used to model the water treatment process, and the simulation results show that it offers the advantages of high precision, fast convergence and fast computing speed.