Nonlinear model predictive control-Cross-coupling control with deep neural network feedforward for multi-hydraulic...

Area of Science:

• Control Systems Engineering
• Hydraulic Systems
• Nonlinear Control Theory

Background:

• Multi-hydraulic systems (MHS) present complex synchronization challenges due to nonlinearities and disturbances.
• Existing control algorithms often struggle with unmodeled dynamics and external noise in MHS.

Purpose of the Study:

• To develop an advanced control algorithm for precise synchronization in nonlinear asymmetric MHS.
• To enhance disturbance rejection capabilities in MHS control.
• To validate the performance of the proposed controller through simulations.

Main Methods:

• Establishment and nonlinear feedback linearization of a general nonlinear asymmetric MHS state space model.
• Integration of nonlinear model predictive control (NMPC) with cross-coupling control (CCC).
• Introduction of a novel inverse model-based disturbance compensator and a deep neural network feedforward (DNNF) component.

Main Results:

• The proposed nonlinear model predictive control-cross-coupling control with deep neural network feedforward (NMPC-CCC-DNNF) controller demonstrated superior performance.
• Achieved a significant reduction in synchronization root mean square error (RMSE) by up to 60.8% compared to other controllers.
• Effectively handled unmodeled errors and noise, confirming robust disturbance rejection.

Conclusions:

• The NMPC-CCC-DNNF controller offers a robust and effective solution for MHS synchronization.
• The integration of NMPC, CCC, and DNNF provides enhanced control precision and disturbance attenuation.
• Simulation results validate the significant performance improvements over existing methods.