Multivariable non-minimum state space model predictive control based on disturbance observer.

Area of Science:

• Control Engineering
• Chemical Engineering

Background:

• Lumped system disturbances, including external factors and internal model mismatches, significantly impact control system performance.
• Existing feedback and feedforward control methods often fail to guarantee optimal output in the presence of such disturbances.

Purpose of the Study:

• To propose a multivariable non-minimum state space predictive control method with a disturbance observer (MNMSSPC-D) for enhanced disturbance suppression.
• To improve optimal control performance and anti-disturbance capabilities in complex systems.

Main Methods:

• Development of a multivariable non-minimum state space (MNMSS) prediction model by incorporating estimated disturbance and output variables.
• Application of the rolling optimization principle within predictive control to design the MNMSSPC-D controller.
• Validation of the proposed method through simulation on a heavy oil fractionator.

Main Results:

• The proposed MNMSSPC-D method effectively integrates disturbance estimation into the state variables of the prediction model.
• The designed controller ensures state availability for the MNMSS model.
• Optimal control performance and superior anti-disturbance ability were achieved.

Conclusions:

• The MNMSSPC-D method offers a robust solution for suppressing lumped system disturbances.
• The approach guarantees optimal control performance and enhances system resilience against disturbances.
• Simulation results confirm the effectiveness of the proposed control strategy.