| The concentration of ferrous ion is the most important technical index in iron precipitation process based on goethite method. In practice, the concentration of ferrous ion cannot be measured online and the iron precipitation process is characterized with strong nonlinearity, multi-variable and time delay. Therefore, it is very difficult to stabilize the concentration of ferrous ion. For the characteristics of iron precipitation process, nonlinear predictive control for concentration of ferrous ion based on LSSVM and PSO algorithm is developed.(1) The mechanism of the iron precipitation process based on goethite method is introduced, and the principal factors influencing the concentration of ferrous ion are analyzed. Then, the predictive model for the concentration of ferrous ion is established based on LSSVM. Based on the industrial data collected from the practical process, the key parameters in LSSVM model are optimized using a self-adaption search technique. Simulation results show the high precision of the LSSVM model.(2) It is a nonlinear constrained optimization problem to obtain the control law for nonlinear prediction control strategy. It is very difficult to solve the optimization problem using the traditional partial derivative-based optimization method. Therefore, a PSO algorithm is proposed for rolling optimization in nonlinear predictive control. To avoid the local minimum, a dynamic weighted chaos PSO algorithm is developed based on Tent mapping and the search ability of PSO is improved. Several test functions verify the effectiveness of the proposed algorithm.(3) Nonlinear predictive control strategy for the concentration of ferrous ion in iron precipitation process is proposed based LSSVM and PSO. Simulation results show that the proposed nonlinear predictive control method can stabilize the concentration of ferrous ion and, which is very useful to practical production. This paper contains28figures,6tables,6equations and64references. |
PDF Full Text Request