An Intelligent Model Predictive Control Framework for Low-Frequency Seismic Vibration Suppression in Active Isolation...

Area of Science:

• Engineering
• Control Systems
• Machine Learning

Background:

• Low-frequency seismic disturbances degrade precision engineering and active vibration isolation systems.
• Model predictive control (MPC) is effective but relies heavily on accurate system models.

Purpose of the Study:

• To propose a multilayer perceptron-enhanced model predictive control (MLP-MPC) framework for active vibration isolation.
• To improve vibration suppression performance under seismic disturbances.

Main Methods:

• An MLP is trained offline to estimate the free-response term in MPC predictions.
• The trained MLP replaces model-based calculations during online implementation.
• The MLP-MPC framework preserves the quadratic programming structure of conventional MPC.

Main Results:

• MLP-MPC demonstrated reduced root-mean-square (RMS) tracking error.
• Lower moving-window RMS error was observed compared to conventional MPC and Proportional-Integral-Derivative (PID) control.
• The system was evaluated on a single-degree-of-freedom active vibration isolation setup.

Conclusions:

• Integrating data-driven free-response estimation enhances predictive control for vibration suppression.
• The MLP-MPC approach offers a practical solution for low-frequency vibration control.
• The method maintains computational feasibility while improving performance.