An offset-free MPC formulation for nonlinear systems using adaptive integral controller.

Area of Science:

• Process Control
• Chemical Engineering
• Control Systems

Background:

• Nonlinear processes present significant challenges in control system design due to their complex dynamics.
• Existing model predictive control (MPC) strategies often struggle with offset-free performance and handling uncertainties in nonlinear systems.
• Observer-based MPC can be complex to tune for nonlinear applications.

Purpose of the Study:

• To develop a novel, easily tunable, offset-free control scheme for nonlinear processes.
• To address parametric uncertainty and external disturbances in nonlinear control systems.
• To demonstrate the effectiveness of the proposed control strategy compared to existing methods.

Main Methods:

• Utilizing a multiple model predictive controller (MMPC) to represent nonlinear processes with linear models.
• Employing a min-max approach to mitigate parametric uncertainty between linear models and the nonlinear process.
• Integrating an adaptive integral action controller in parallel with the MMPC to manage input/output disturbances.

Main Results:

• The proposed MMPC combined with adaptive integral action achieves offset-free control for nonlinear systems.
• The control scheme effectively handles setpoint changes and parametric uncertainties.
• Simulations on a pH-control system demonstrate superior performance compared to observer-based MPC.

Conclusions:

• The novel MMPC and adaptive integral action controller offer a robust and tunable solution for offset-free control of nonlinear processes.
• This approach simplifies tuning compared to traditional observer-based MPC methods.
• The technique shows significant promise for practical applications in chemical and process engineering.