USE OF LINEAR QUADRATIC AND QUADRATIC PROGRAMMING METHODS IN MODEL-BASED PROCESS CONTROL (LQ METHOD).

摘要

This dissertation compares two types of internal model based control methods: on-line optimization and off-line optimization methods for the application to a multi-effect evaporator system. Emphasis is on the LQ-MAC (Linear Quadratic Model Algorithmic Control) off-line optimization method and the QPMAC (Quadratic Programming Model Algorithmic Control) on-line optimization method.; A careful theoretical study of the LQ-MAC method, including stability and convergence to the input/output set-points in the face of modelling errors, is given.; The QPMAC method is a combination of the MAC, QDMC (Quadratic Dynamic Model Control), and IMC (Internal Model Control) methods. The formulations of the control problems with explicit inequality input/output constraints and input/output set points are summarized so that one can use the quadratic programming method for process control.; Both methods are used to design controllers for a multi-effect evaporator. Extensive simulation results are given. The similarities and differences of the on-line optimization and off-line optimization approaches are summarized. It is recommended to use the quadratic programming control method for this evaporation system.