Optimization of structural control via a smart NEURO-FBG control system

This study improves a NEURO-FBG active control system to mature the concept of a smart structure. Originally, a system similar to the human brain is created from FBG sensors and neural networks. The system comprises three parts, namely, a structural condition surveillance system, a NEURO-FBG converter, and a NEURO-FBG controller. To solve the inherent time-consuming and reliability problem of the NEURO-FBG converter, a new technology is first proposed, and the relationship between inter-story drift and strain data is established. Global indices such as displacement and velocity of the structure are then reconstructed for searching the optimal control force of the actuator. Meanwhile, the soundness of a building with hydraulic actuators is also an important issue to be solved. To make the building sound, the characteristics of earthquakes are considered for enhancing the performance of the NEURO-FBG controller. Theoretical analysis shows satisfactory improvement to the control efficiency of both displacement and acceleration. To verify the enhanced system, a series of shaking table tests was conducted. Experimental results demonstrated that the new NEURO-FBG system can effectively manage the structure; and the controller, taking into consideration the ground acceleration effect, is more reliable and robust for practical application than a conventional controller.