Wastewater treatment prediction based on chaos-GA optimized LS-SVM

Wastewater treatment is a complicated biochemical process with nonlinearity and time-delay. Its mathematical model is difficult to establish. In order to optimize the process control and reduce the power energy consumption, parameters such as Chemical Oxygen Demand (COD) should be accurately predicted. A novel least square support vector machine model is presented to predict the effluent COD in this paper. A multi-scale chaos search algorithm was proposed to optimize the model parameters, and the genetic algorithm was combined to accelerate the search speed. Simulation results show that the proposed method has higher precision, greater generalization ability and less computation. The prediction MSE was reduced from 21.43 by ANN to 6.83 by the proposed method.