Closed-loop identification with MPC for an industrial scale CD-control problem

Abstract

This paper details an approach to implement closed-loop identification in a model predictive control framework for the cross-direction control of basis weight. The closed-loop identification technique uses the concepts of Markov parameters to determine a step response model that can then be used within a model predictive controller. The technique is applied to an industrial scale paper machine simulation benchmark problem for cases of varying degrees of shrinkage. Performance of the wet and dry end full array sensors are compared. Also the performance of the closed-loop identification is compared for a nominal case and two cases in which shrinkage occurs in the drying process.

Closed-Loop Identification for Model Predictive Control of HVAC Systems: From Input Design to Controller Synthesis

“…where T s is the sampling period. After completing this transformation, we solve (21) in the continuous-time domain, and then we obtain C ac (q) by reversing the mapping (23). From the equation (19) we see that u air (t) acts as a disturbance to the temperature of the cooled air y ac (t).…”

“…The filter (44) forces the H ∞ optimization (21) to focus on the low-frequency range and to guarantee integral action in the optimal controller. The resulting controller is…”

“…A sensible solution is to employ data-driven techniques and system identification methods to identify the dynamics of the HVAC subsystems along with their local controllers, i.e., closed-loop dynamics, based on measured data, and then incorporate these closed-loop dynamics into the MPC design. By doing so, the model accuracy of the overall system to be controlled and, consequently, the control performance, can be improved [21], [22]. Closedloop identification for MPC has been studied in different settings in [23], [24].…”

Identification of low-order unstable process model from closed-loop step test

“…Some limiting conditions to form the limit cycle from relay feedback for an unstable process have been disclosed by Liu and Gao (2008b). Besides, using the pseudo-random binary sequence (PRBS) as excitation signal to the set-point, closed-loop identification methods for application of model predictive control (MPC) have been reported in the references (Sung et al, 2001;Saffer and Doyle, 2002;Bindlish, Rawlings and Young, 2003). In this paper, identification algorithms based on a closed-loop step test are proposed for obtaining low-order models of FOPDT and SOPDT for tuning unstable processes.…”

Identification of low-order process model with time delay from closed-loop step test

“…Using relay feedback to yield sustained oscillation of the closed-loop output, Padhy and Majhi [26] suggested an identification method based on the resulting limit cycle data. Besides, using the pseudo-random binary sequence (PRBS) as excitation to the process input or the set-point, closed-loop identification methods have also been reported for delay-free linear time-invariant processes [27][28][29][30][31].…”