The state model-based transfer function models are applied for adaptation of linear controller and disturbance compensator in a feedback/feed-forward control system of nonlinear process. An advantage of the presented adaptation method is the avoidance of artificial disturbances or iterative identification procedures for on-line estimation of process dynamic parameters. The adaptation is based on linearization of the process model at each sampling time about the current state point, independent of the process being at steady-state or transient conditions. The linear time-varying dynamics model is updated on-line using measured values of process variables and reduced to the first-order plus time delay transfer function models in order to directly apply well-developed controller tuning rules. Computational aspects of the adaptation method are discussed and computation algorithms are presented. The adaptive feedback/feed-forward control system was applied for controlling temperature in industrial methane tank, dynamic parameters of which vary in a wide range due to variations of methane-tank process load and external conditions. The heat balance-based process state model is developed and validated using observation data of real plant. Computer simulation of the proposed control system performance under extreme operating conditions demonstrates fast adaptation of controller parameters, robust behaviour and significant improvement in the controllers' performance compared to that of fixed-gain controllers.
展开▼
