Please use this identifier to cite or link to this item:
Title: New Two Degree of Freedom Control Schemes for Integrating and Unstable Processes with Time Delay
Authors: Ajmeri, M.
Keywords: Electrical Engineering
Issue Date: 2015
Publisher: IIT Patna
Abstract: In one degree of freedom control scheme, the controller is tuned either for servo or regulatory applications which results in significant degradation when servo settings are used for load disturbance rejection and vice-versa. Performance degradation can be minimized by using a two degree-of-freedom control structure. It is observed that a PD controller gives good set point tracking performance whereas a PI/PID controller is required for rejecting the load disturbance for integrating processes. However, none of the reported two degree of freedom control schemes has considered the above said controller configurations. The Parallel Control Structure (PCS) which is recently reported decouples set-point tracking and disturbance rejection responses. In this thesis, the set point tracking and load disturbance rejection controllers of the PCS are assumed as PD and PID controllers. Direct synthesis approach is used to design the controllers of the PCS for a class of integrating plus time delay process models. Furthermore, PCS is modified by introducing an additional controller which results in improved control performance for unstable plus time delay processes. Tuning rules based on the modified PCS are proposed for unstable first and second order process models. PI/PID controller in the unity feedback configuration fails to give satisfactory performance for processes with large time delays. The Smith predictor reported in literature for stable processes with large time delays yields steady state error in presence of input load disturbance for integrating processes. Furthermore, it fails to stabilize the unstable process models. Several modified Smith predictors have been therefore reported in literature for integrating and unstable plus time delay process models. However, recently reported modified Smith predictor schemes for integrating processes require tuning of large number of controller parameters. In this work, tuning rules are proposed for the modified Smith Predictor for a class of integrating process models. The novelty of the proposed approach is that improved/similar closed loop performance is achieved with less number of controller parameters as compared to the recently reported methods. It is to be noted that reported Smith predictors fail to give robust performance for balanced and delay dominated unstable first order process models and process with two unstable poles. In the present research, it is observed that introducing an additional controller in the modified Smith predictor to stabilize the delay part of the process model results in robust closed loop performance for the above said process models. Suitable values of the design parameters are recommended after studying their effect on the system performance and robustness. This is an advantage of the present work over the other reported control schemes where the authors provide suitable ranges for the tuning parameters.
Appears in Collections:03. EE

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.