Offset-free Nmpc with Robust Constraint Satisfaction Using Model-error Modeling

Robustness and offset-free tracking remains a challenge in the field of nonlinear model predictive control (NMPC) in the presence of plant-model mismatch. This paper focuses on offset-free tracking with robust constraint satisfaction using NMPC. In the proposed approach, the discrepancy between the plant and a fundamental model is captured using a model-error model. The fundamental model is augmented with the model-error model to predict the future evolution of the states to obtain the optimal sequence of control inputs. Whenever new information from the plant is available, the model-error model is adapted. This helps the controller to react to time varying model-errors and disturbances and to achieve offset-free tracking with robust constraint satisfaction. The advantages of the proposed approach when compared to existing approaches for offset-free tracking are demonstrated with a continuous stirred tank (CSTR) example.