Study Of Fuzzy Neural Network Controller Of PH Process Based On DSP

This research is supported by "Subject of Science and Technology Development Fund of Tianjin College(20030514)".The problems of pH process control are found in many chemical processes including waste water treatment, medicine making, fermentation, paper, supply water treatment and so on. It is very important to control the pH value to a certain range in these processes. For example, in order to protect surroundings, the pH value of water supply for industry must meet some demand to avoid metal corrosion, and so on. The general method of pH measuring and pH controlling are discussed in this paper.pH process is a nonlinear characteristic process and has serious nonlinearity and delay because of its logarithm measuring relationship with the concentration of hydrogen ion. It is difficult to have an ideal control effect for traditional PID control or nonlinear PID control. An approach about fuzzy neural network controller, the FNNC control strategy, is proposed in this study. This approach makes use of the characteristics that fuzzy control does not depend on the accurate mathematical model of object, has strong robustness and arithmetic is simple and easy to understand and so on, and the capability that neural network can approach arbitrary function and study conveniently. The strategy which integrates the logic function of fuzzy control and self-learning function of neural network can deal with serious nonlinearity and delay of pH process, and has stronger robustness and ability of disturbance rejection by computer simulation.To meet the need of real time calculation in the control process and device exploiture, the TMS320VC33 high speed digital processor (DSP) is adopted as the control and process unit, and the FNNC controller's arithmetic is successfully accomplished based on ICETEK-VC33-A DSP application board.