Aug 2026· Asian Journal of Civil Engineering· 0 citations· 21 references
Abstract
Torsional vibrations induced by seismic excitation can significantly increase structural demand, particularly in systems with geometric or stiffness irregularities. This proof-of-concept study proposes a hybrid adaptive control framework that combines passive gyroscopic damping with reinforcement learning (RL)-based adaptive torque control for torsional response mitigation. A simplified three-degree-of-freedom (3DOF) structural model is evaluated under bidirectional earthquake loading, while the control policy is trained using the Proximal Policy Optimization (PPO) algorithm to generate adaptive torque directly from measured system states without explicit system identification. Numerical simulations using recorded earthquake ground motions show that the proposed approach reduces root-mean-square torsional displacement to 0.012 ± 0.001 rad, representing a 65.7% reduction relative to the uncontrolled case and improved performance over passive gyroscopic control alone. Sensitivity analyses involving variations in mass, stiffness, and damping indicate stable controller performance across the structural parameter ranges considered. The trained controller achieved sub-millisecond inference time with minimal computational overhead, supporting its computational feasibility for future real-time implementation. The results demonstrate the potential of integrating physically grounded gyroscopic damping with data-driven adaptive control for intelligent torsional vibration mitigation.
This paper proposes a reinforcement-learning (RL)-based semi-active control strategy to mitigate stay-cable vibrations using magnetorheological (MR) dampers while explicitly addressing MR fluid temperature variations during damper operation. The controller is model-free and based on the Q-learning method. The controller updates its policy online using local measurements, including the displacement and velocity at the damper location. The control law avoids a fixed inverse-damper model by using a quadratic action–value approximation, which yields a closed-form current command, while a target matrix is used to stabilize the online learning. The controller performance is numerically evaluated on the A10 cable of the Dongting Lake Bridge using a temperature-dependent hyperbolic-tangent MR damper model implemented in MATLAB. A comparison with a classical LQR-based semi-active controller that relies on an inverse MR damper model is presented. The results indicate that the proposed RL controller remains effective under temperature-induced uncertainty, while still requiring lower control effort in terms of improved energy efficiency.
Y. M. Delijani, Shaohong Cheng, F. Ghrib· International Symposium on D...· 0 citations
The multi‐speed planetary hybrid drivetrain features a complex structure, where transient variations from multiple excitation sources are highly coupled with vehicle structural parameters, leading to severe torsional vibration issues, especially during engine start‐stop processes. This study proposes an unknown input‐compensated least mean square (UIC‐LMS) adaptive algorithm for active torsional vibration suppression. An engine excitation model is established through time‐ and frequency‐domain analyses, while a six‐degree‐of‐freedom dynamics model of the drivetrain is developed using the lumped mass method, with frequency‐domain analysis revealing the resonance mechanisms. To achieve precise torque estimation under uncertainties, a reduced‐order unknown input observer (UIO) is designed by reconstructing measured outputs. The least mean square (LMS) adaptive filtering then computes motor compensation torque, with real‐time parameter adjustment via a transversal filter. Hardware‐in‐the‐loop test demonstrate that the UIO improves estimation accuracy by over 78.9% compared to the Luenberger observer, showing superior robustness against unknown input disturbances and measurement noise. Under UIC‐LMS control, maximum amplitude fluctuations of transmitted torque and torsional angle in the torsional vibration damper are reduced by over 44.8%, while those in the driveshaft are reduced by more than 72.5%. Drawing parallels with advanced equivalent‐input‐disturbance rejection strategies for uncertain systems, the algorithm maintains superior effectiveness across drivetrain parameter uncertainties by integrating UIO‐based lumped disturbance observation with LMS adaptive compensation. Real‐vehicle validation confirms the algorithm's robustness and high torsional vibration suppression rates in embedded controllers.
D. Song, Yufeng Huang, Hongfei Tong et al.· International Journal of Rob...· 0 citations