Nonlinear System Modeling Using RBF Networks for Industrial Application

Radial basis function (RBF) networks, because of their universal approximation ability, have been widely applied to industrial process modeling. In this study, an Improved ErrCor (IErrCor) algorithm-an extension of error correction (ErrCor) algorithm-is proposed, in which compact structure and satisfactory generalization ability can be obtained with only one learning try. First, a second-order-based constructive mechanism guarantees the structure compactness and computational efficiency. Second, different from other algorithms that start with random or constant parameters, optimal initial parameters accelerate the convergence process and improve the convergence performance, making the IErrCor RBF network more stable. Convergence analysis is given to demonstrate and prove the reasonability and effectiveness of the proposed algorithm. Finally, the IErrCor algorithm has been evaluated and compared with several popular advanced learning algorithms such as support vector machine (SVM), extreme learning machine (ELM), and original ErrCor algorithm through a series of benchmark experiments and then been applied to effluent water quality prediction in wastewater treatment process. All the simulation results reveal the outperformance and potentiality of IErrCor RBF network in industrial applications.