The Application Of RBF-ARX Model In Three-Tank Water Level Predictive Control System

Because controlled objects in laboratory investigation are always different from those practical industrial plants, it caused a lot of problems in the application of a new theory to the actual object and seriously hindered that. However, the Three-Tank Water Level Control System has the same characteristics with the complicated industrial process control systems. Modeling and model predictive control (MPC) to the Three-Tank Water Level Control System will be presented in this paper, which may include following aspects:First, the structure of the RBF-ARX model was introduced. The RBF-ARX model is constructed as a global model, which can describe the nonstationary nonlinear systems whose dynamic characteristics depend on time-varying working-points and may be locally linearized. It is an ARX model with time-varying Gaussian radial basis function (RBF) network-style coefficients, which can be numerated on-line easily based on the current working state. Besides, the RBF-ARX model is estimated off-line so as to avoid the possible failure of on-line parameter estimation during real-time control, and the local linear model of the system at each working-point may be easily obtained from the global RBF-ARX model. So the use of nonlinear programming techniques to solve the on-line optimization problem with constraints in RBF-ARX-MPC is also avoided.Second, the reasons causing the Three-Tank Water Level System's nonlinear characteristic were studied, and also some modeling and control strategies were compared, which shows the superiority of the RBF-ARX model in Three-Tank Water Level System modeling. Then, the RBF-ARX model of the Three-Tank Water Level control system with two inputs and two outputs was set up after sampling the water level data. At last, the parameters' identification, optimization and selection were analyzed and explained in detail.Third, the RBF-ARX model-based predictive control strategy was educed after analyzing two kinds of basic predictive control strategies, and then the realization of the control strategy, the selection of the parameters and the designing of the system were introduced in detail.Finally, the application of PID control and the RBF-ARX model-based predictive control to the real Three-Tank Water Level Control System were compared; it shows that the predictive controller based on RBF-ARX model has improved the dynamic performance and robustness of the control system. That is to say, this control method is effective, especially to those complex systems such as Three-Tank Water Level control system.