Vibration mitigation of civil structures using hybrid control strategies

The greater scope of this thesis is to improve the safety and performance of civil engineering building-like structures which are exposed to external forces such as wind, earthquakes or other environmental factors. The current work looks at the use of mass damper technology as a potential solution for the dynamic response reduction of such structures. A comprehensive literature review of past studies and research advances is included, specifically focusing on mass damper applications on building-like structures. Using a systematic literature review approach, an up-to-date list of mass damper applications including more than 200 entries is reported, as well as a list of control algorithms commonly used in the engineering industry. The review suggested that, the mass damper systems suffer from various problems including robustness issues, high energy requirements for the active components and difficulties in the implementation of effective control algorithms. This thesis intends to tackle the identified challenges by conducting numerical simulations of a real high-rise tower. The tower is equipped with a hybrid mass damper (HMD) which possesses passive, semi-active and active capabilities. This work compares the performance of the Robust Model Predictive Control (RMPC) scheme to the well-established robust controller, H∞. Additionally, the thesis also investigates the performance of the Deep Deterministic Policy Gradient (DDPG) reinforcement learning controller when compared to the Linear Quadratic Regulator (LQR) controller. Methods to reduce the energy consumption of the HMD system were also investigated. This thesis proposed a mode-switching control scheme using the Deep Q-Network (DQN) algorithm. This method switches between passive, semi-active, and active configurations of the system and was found to be efficient in dissipating the dynamic responses of the tower while also considerably reducing energy requirements compared to a purely active system. The current research work showed that, the proposed control algorithms and the energy consumption reduction techniques were efficient and outperformed the well-established methods rendering them novel and innovative for the field. Overall, this thesis proposes effective and practical solutions which can be used to design safe and sustainable building-like structures in the future.