Please use this identifier to cite or link to this item:
|Title:||Development of a Closed Form Nonlinear Predictive Control Law Based on a Class of Wiener Models|
|Publisher:||AMER CHEMICAL SOC|
|Citation:||INDUSTRIAL & ENGINEERING CHEMISTRY RESEARCH, 49(1), 148-165|
|Abstract:||Nonlinear model predictive control (NMPC) is increasingly being used for controlling microscale and system-on-chip devices, which exhibit complex and very fast dynamics. For effective control of such systems it is necessary to develop computationally efficient approaches for solving the NMPC problem. In this work, a Wiener type model has been used for capturing dynamics of multivariable nonlinear systems with fading memory. The resulting discrete nonlinear state space model is used to generate multistep predictions and formulate all unconstrained NMPC problem. A closed form solution to this problem is Constructed analytically using the theory Of Solutions of quadratic operator polynomials. The effectiveness of the resulting quadratic control law is demonstrated by conducting simulation studies on a proton exchange membrane fuel cell (PEMFC) system, which exhibits fast dynamics and input multiplicity behavior. The quadratic control law is expected to control the PEMFC at its optimum (Singular) operating point. The proposed laws achieve a fast and smooth transition from a Suboptimal operating point to the Optimum operating point with significantly small Computation time. The proposed law is also found to be robust in the face of moderate perturbation ill the unmeasured disturbances. The simulation results are validated by conducting experimental studies on a single cell PEMFC system and a benchmark heater-mixer setup that exhibits input multiplicity behavior. Through the experimental studies, we demonstrate that the proposed quadratic control law is able to operate the system at a singular operating point and establish the feasibility of employing the proposed control law for systems with very fast dynamics.|
|Appears in Collections:||Article|
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.