The Study Of PH Control Based On SAE

pH control, particularly in the neutral range of 6-8, has been a difficult and challenging problem due to its time-varying and nonlinear characteristics. These characteristics cause many difficulties in controlling the pH process with conventional control technique such as the PID controller. Many researchers have exerted much of their effort in the study of modeling and controlling of pH process.Based on the pH mechanism model, two different control methods were presented in this paper and were applied to pH process control of CSTR (Continuous Stirred Tank Reaction). As for the characteristics of high nonlinearity, an equivalent pH control objective in terms of Strong Acid Equivalent (SAE) which is a weighted sum of ionic concentrations was introduced. And the new control objective SAE is linear in states. The mathematical model derived in this paper is combined the dynamic and static model of pH process.These two methods presented in the paper are the multiple models adaptive control (MMAC) and the adaptive internal model control (AIMC).In MMAC method, an approximate equivalent linear control modelwas derived then adopts multiple models to adapt to the changes of titration curves. A fixed multiple models plus adaptive models was designed in which the fixed models can accelerate the transient performance and the adaptive model eliminate the stable error. In addition, the switching method among multiple models was given and different controllers were designed to meet different sub models.The AIMC consists of the process model, its inverse model, low-pass filter and an adaptive mechanism. The inverse model of the process was serviced as the IMC controller. To obtain good performance, an adaptive control mechanism was applied to compensate for the errors between the real plant and the model.The simulation results show that the proposed methods have excellent capabilities for setpiont tracking, disturbance rejection and tolerance to model uncertainty and can solve the time variant and nonlinear problems remarkably.(Control Theory and Control Engineering)Supervised by...