NEURAL NETWORK-ENHANCED HYBRID VIBRATION CONTROL FOR BASE-ISOLATED STRUCTURES

In civil engineering, vibration control is crucial to protect structures, reduce damage and ensure comfort by reducing dynamic forces, which enhances stability. One effective approach is hybrid structural vibration control, which has made notable advancements in successfully mitigating undesired oscillations induced by external factors like wind and earthquakes. In our study, an advanced hybrid system that integrates a base isolator with an active tuned mass damper (ATMD) has been developed. The active control force for the ATMD is precisely determined through the use of an Artificial Neural Network (ANN) controller. The ANN controller is trained with a comprehensive database of forces generated by a linear quadratic regulator (LQR). To evaluate the hybrid control system's effectiveness, simulations were conducted on a rigid six-degree-of-freedom base-isolated frame structure where the ATMD is strategically placed on the lowest floor of the structure. Notably, the hybrid control system demonstrates a significant reduction in vibration amplitudes in terms of base isolator displacement and maximum drifts while not adversely affecting the response of the superstructure, where it achieves an impressive reduction of up to 87 % in displacement responses. The results highlight the ANN algorithm's efficacy in reducing structural responses, emphasizing the substantial potential of hybrid systems in advancing structural vibration control. Furthermore, the system's performance was validated through simulations using real-time historical data from three earthquakes that were not included in the ANN training (Hector, Northridge, and Imperial Valley), confirming its capability to adapt and provide reliable vibration control in practical situations beyond the scope of its initial training. In conclusion, this approach represents a significant step forward in the field of seismic response reduction, offering both efficiency and adaptability.