Degree
Doctor of Philosophy (PhD)
Department
Department of Civil and Environmental Engineering
Document Type
Dissertation
Abstract
Semi-active vibration control is increasingly used to protect civil infrastructure and modern wind turbines. In practice, performance assessment often relies on nonlinear time-domain simulations. A fast and reliable performance assessment method is therefore important for early-stage design and evaluation of semi-active control performance.
Past research has developed energy-based equivalent-gain methods to analytically evaluate the control performance of semi-active systems. A key limitation of existing equivalent-gain evaluation methods is that their practical applicability has largely been confined to idealized stationary white-noise excitation. As a result, predictive assessment tools that remain accurate under realistic random loading conditions have been lacking.
This dissertation proposes energy-based approaches for predicting the performance of semi-active control under random loading conditions extending beyond white-noise excitation. The primary contribution is a set of closed-form theories and formulas for the dissipative fraction (deterministic dominant-frequency loading) and the dissipative probability (stochastic loading), which are developed herein. These proposed formulas quantify how often the semi-active control action is energetically dissipative for a specified excitation and controller setting, thereby establishing a direct analytical link between nonlinear clipped-control behavior and its energetic outcomes.
A further contribution is the integration of the proposed dissipative formulations with existing equivalent displacement-gain and equivalent velocity-gain formulas from the literature. These equivalent gain formulas are incorporated as analytical components of the proposed approaches to estimate selected response measures without requiring full nonlinear time-domain simulation. The proposed procedure substantially reduces computational demand and extends predictive capability to a wider class of realistic random loading conditions.
The framework is developed progressively from SDOF systems under harmonic excitation to broadband stochastic loading, and is then extended to MDOF systems through formulations that support complex structural models. Numerical studies in MATLAB demonstrate strong agreement between the proposed analytical predictions and full nonlinear simulations, and experimental validation further supports the accuracy of the proposed approach.
Overall, this dissertation proposes new approaches for the performance prediction of semi-active vibration-mitigation systems. It shifts control performance assessment from costly nonlinear simulation and white-noise-limited applicability toward an analytical framework that supports generic random loading, enabling rapid evaluation of semi-active control strategies for civil infrastructure and wind turbines.
Date
4-15-2026
Recommended Citation
Xie, Feng, "Energy-Based Probabilistic Approaches for Vibration Control Efficacy Evaluation with Applications to Wind Turbines" (2026). LSU Doctoral Dissertations. 7068.
https://repository.lsu.edu/gradschool_dissertations/7068
Committee Chair
Aly Mousaad Aly
LSU Acknowledgement
1
LSU Accessibility Acknowledgment
1