Practical Implementations of Advanced Process Control for Linear Systems

This paper describes some practical problems encountered, when implementing Advanced Process Control, APC, schemes on linear processes. The implemented APC controllers discussed will be LQR, Riccati MPC and Condensed MPC controllers illustrated by simulation of the Four Tank Process and a linearised CSTR. Advantages and disadvantages of these controllers will be discussed. All three controller types shows a set of common undesirable characteristics, which must be accounted for. At the end of the evaluation horizon the "optimal" solution has an unstable characteristics, which can be suppressed be selecting dierent control and evaluation horizon. Depending of the degrees of freedom, oset-free control of a number of the controlled variables can be achieved by integration of the innovation errors and introduction of noise models. If the measured or unmeasured disturbances increases, oset-free control cannot be achieved without violation of process constraints. A target calculation function can be used to calculate the optimal achievable target for the process. The use of hard and soft constraints for process input constraints in the MPC controllers, ensures feasible solutions. The computational load as function of controllers type, Model dimension and constraint type will be discussed. Finally the special requirements set by processes including a pure integration dynamics will be illustrated by a linearised CSTR process. The simulated results presented, will later on be implemented on and demonstrated on pilot
plant equipment on the department of Chemical Engineering DTU Lyngby.