pH control of a neutralization process using neural networks

pH measurement and control has been an area of intense research and investigation over the past several years. The interest in pH control can be attributed to the nonlinear behavior of the pH system, which subsequently makes the pH difficult to control. The experimental project attempted to control the pH of a neutralization process using a neural network (NN). The neural network was trained using steady state data obtained from a mathematical model. The neural network was used both as a stand-alone feedforward controller and in conjunction with a PI controller. The performances of both these control systems were then compared with a control system using PI control only. Additionally, experiments to investigate the influence of the buffer solution on the pH process were conducted. The control system using NN-PI control performed better than the other control systems in most cases.