| A multivariable process was designed and constructed for the purpose of testing advanced control strategies. Certain characteristics of difficult industrial control problems such as nonlinearities, interactions between controlled variables, and magnitude and velocity constraints were designed into the process.; pH neutralization process, because of its highly nonlinear characteristic and relatively low cost of operation, was chosen as the test bench. A mathematical model was developed for the system and good agreement was shown with both open loop and closed loop tests. Both simulated and experimental results of conventional multi-loop PID control strategy were compared with those of advanced Model Predictive Control (MPC) strategies.; Piecewise linearization technique was used in applying the linear MPC algorithms to this nonlinear process. A new re-initialization technique was proposed and implemented to eliminate the influence caused by multi models. By using the new technique, it is shown that linear MPC algorithms can be used to provide good control of nonlinear processes. |
