Structural dynamic load identification based on Gibbs sampling in time domain
Wang Ting 1 Wan Zhimin 2 Zhengweiguang 2
1.School of Mechanical Engineering, Nantong Vocational University Nantong 226000;
2.School of Mechanical Science and Engineering, Huazhong University of Science and Technology Wuhan 430074
Abstract:The regularization technique is always employed to deal with the ill-posed problem of load identification. However, the regularization parameters obtained by the traditional regularization techniques are constant. A novel approach of load identification in time domain based on Gibbs sampling method is proposed, which has the natural mechanism of regularization with automatic tuning as a function of time. First, the unknown load vectors and measurement noise are regarded as the random variables. Then, the hierarchical Bayesian mathematical model of load identification is built. Final, the Gibbs sampling is adopted to obtain the posterior probability density distributions of the predicted loads. Numerical studies are conducted to demonstrate the effectiveness by comparing with the Tikhonov regularization methods based on the L-curve and GCV criterion.
王婷 1,万志敏 2,郑伟光 2. 基于Gibbs抽样的结构时域载荷识别[J]. 振动与冲击, 2018, 37(2): 85-90.
Wang Ting 1 Wan Zhimin 2 Zhengweiguang 2. Structural dynamic load identification based on Gibbs sampling in time domain. JOURNAL OF VIBRATION AND SHOCK, 2018, 37(2): 85-90.