e-journal
Robust Model Predictive Control Under Saturations and Packet Dropouts With Application to Networked Flotation Processes
This paper investigates the problem of robust model predictive control (RMPC) with saturations and packet dropouts. In this model, polytopic uncertainties are adopted to describe the inconsistency arising from the discretization process of sampling, while the occurrence probabilities of packet dropouts are time-varying and saturations are taken into account to describe input and output signals. The problem of exponential RMPC with saturations and packet dropouts is solved and characterized by a convex optimization problem. The developed results of RMPC are then applied to networked flotation processes, which are made up of three layers: direct control layer, set-point control layer, and optimization layer. The RMPC is used for compensating the output information from the optimization layer to the direct
control layer such that the desired economic objective can be achieved. Simulations are presented to show the effectiveness of the proposed method.
Note to Practitioners—With the increasing requirement from industrial processes such as chemical processes, electrolysis and physical processes, it is important to regulate the output of systems
to satisfy a desired objective. This paper addresses the problem of RMPC with saturations and packet dropouts, in which the expectations of random packet dropouts are time-varying. Such phenomena can be widely observed in networked control systems. By means of the proposed RMPC, the output of flotation processes can be operated to a specific goal. Future research topics include using parameter-dependent techniques to reduce the conservativeness and considering more practical phenomena to reflect more realistic factors.
Index Terms—Data losses, industrial processes, model predictive control, networked control systems, saturations.
Tidak ada salinan data
Tidak tersedia versi lain