Structural Vibrations Control with Model Predictive Algorithm Improved with Kalman Filter

Articles in Press

Document Type : Original Article

Authors

• Mortaza Ghorbani ouri ¹
• Amin Gholizad ²
• Mona Shoaei parchin ¹

¹ Department of Civil Engineering, University of Mohaghegh Ardabili

² University of Mohaghegh Ardabili

10.22067/jfcei.2025.94471.1362

Abstract

Model Predictive Control (MPC) refers to a family of controllers that utilize a predicted model of the system response in calculating control forces. The presence of uncertainties in equations and data acquisition networks is inevitable in practical problems and can impair the system’s performance in predicting future behavior and computing control forces accurately. To overcome this issue, a steady-state Kalman filter can be used to predict the response of dynamic systems. However, its effectiveness in reducing computational complexity and decreasing the prediction horizon, which is expected to improve control performance, has not yet been thoroughly investigated. In the present study, aiming to reduce the prediction steps of the MPC, the dynamic vibrations of an eleven-story shear structure equipped with active tendon dampers were controlled using a MPC algorithm combined with a steady-state linear Kalman filter. The results indicate that the improved Model Predictive Controller with the steady-state Kalman filter, with fewer computational steps (105 fewer computational steps), reduce the structural displacement response under the Kocaeli earthquake and reduces control force magnitudes by up to 57% compared with the conventional Model Predictive Control algorithm.

Keywords

• Active Tendon
• Model Predictive Control
• Prediction Horizon
• Steady-State Kalman Filter
• Vibration Control

---