Zeta Potential Modeling of Papermaking Wastewater on Neural Network

Abstract

The key of the papermaking wastewater treatment is to control the amount of added flocculants, the effect of the dosage can be clearly reflected by zeta potential's changes in wastewater, so zeta potential modeling reflects the actual flocculation process of wastewater. According to the problem of zeta potential measurement is off-line, and the flocculation process is complex and difficult to establish accurate mathematical model, this paper proposes the use of neural network to model and predicts its zeta potential. This work studies the multi-input and single-output modeling method based on BP, RBF neural network, it establishes a nonlinear function relationship between the zeta potential and the amount of flocculants, then indirectly gets the estimated value of zeta potential. The simulation results show that the neural network prediction model can achieve the prediction of the optimal zeta potential, and has good real-time and good generalization ability. Compared with the BP network, RBF network performs better with low error, less calculation, shorter cycle, so RBF neural network is an effective modeling method.