Robustness Analysis Of Nonlinear Systems Predictive Control

With the process of industrial development, its growing in scale, continuous, industrial production processes become increasingly complex, controlled object is often nonlinear, strong coupling, changes in operating point range, parameter uncertainty, large delay and other characteristics of incomplete information, and there are various constraints, but also its dynamic behavior changes with operating conditions, and system parameters changes such as environmental factors. In system performance, controller design and tuning and robustness problems in many areas, the traditional linear predictive controller has been difficult to work. Therefore, how to improve on the predictive control to solve industrial process variables that exist in many strongly nonlinear, delay, uncertainty, and a number of constraints and other issues, making it more suitable for practical application of complex industrial processes, is a serious problem. In this regard, different domestic and foreign scholars proposed a variety of nonlinear systems robust predictive control, in which H_∞robust predictive control is an effective strategy for dealing with nonlinear systems.In this paper, based on previous research work, for a class of constrained nonlinear systems, combined with the input state of stability, robust invariant set, control Lyapunov functions and other theories and methods in optimizing the feasibility, stability, robustness, computational estimation of volume and stability region of the analysis done in order to get a real sense of the theoretical results and design. With differential strategy theory, given a suboptimal robust NMPC algorithm design.For bound state perturbation, get an adjustable sub-optimal robust NMPC algorithm, and in theory, suboptimal closed-loop system verified with respect to sub-optimal bounded uncertainty is the input state stability, which in the robust NMPC strategy reduces the computation on online optimization. Using nonlinear H_∞control, for the constraints, uncertainty, nonlinear system, given the robust MPC algorithm for affine systems design. With L2 gain of nonlinear systems and input state stability, theoretically its predicted that the optimal H_∞predictive control and sub-optimal H_∞predictive control for uncertain but bounded disturbances is robust control,and finite-dimensional parametric design to further reduce The H_∞NMPC line optimization calculation. Finally, a design example shows the validity of research results.