Abstract:This study advances dynamic force reconstruction in the time domain through the precise integration method (PIM), particularly effective for matrix exponential numerical solutions. Traditional PIM approaches, typically reliant on linear or step function assumptions within integration time steps, face limitations in handling large integration time step loads. Addressing this, our research introduces an innovative method employing velocity and displacement sensors in simple and complex structures, significantly enhancing load fitting accuracy, especially for nonlinear loads. We propose a refined version of PIM: the Chebyshev polynomial integration method (CPIM) and the adaptive precise integration method (APIM) based on wavelets. These methods are designed to fit force functions of varying orders across different time steps, thereby bolstering signal reconstruction robustness. The efficacy of these approaches is validated given adequate measurement response data. Furthermore, recognizing the growing emphasis on uncertainty in engineering, our work incorporates the collocation method for uncertainty quantification in load identification. To counteract the erratic results from polynomial fitting, particle swarm optimization (PSO) thresholding is employed to eliminate high-frequency noise, thus enhancing the Signal-to-Noise Ratio and minimizing signal distortion. Notably, we demonstrate that the Shannon wavelet surpasses the Chebyshev polynomial in fitting ability due to its tighter branch, rendering APIM more suitable for engineering applications compared to CPIM. APIM adeptly minimizes errors related to acquisition frequency by adaptively adjusting the number of discretization points and wavelet convergence levels to fulfill convergence criteria and matrix norm requisites. This paper proposes a multi-source uncertainty-oriented dynamic Force Reconstruction Framework, integrating adaptive fitting precise integration with PSO. Our numerical examples illustrate that APIM, as discussed herein, offers broader applications and is more time efficient than CPIM.

Multitask-Transfer-Learning Method for Random-Force Frequency Identification Considering Multisource Uncertainties

A Multi-Task Learning-Based Automatic Blind Identification Procedure for Operational Modal Analysis

Dynamic Load Identification at Natural Frequencies for Aircraft via Attention Based 1D-CNN

Comparative Analysis of Machine Learning Techniques for Identifying Multiple Force Systems from Accelerometer Measurements

Force Identification Based on the Reconstruction of Distributed Response with an Optimized Sensor Placement

Force Localization and Reconstruction Based on a Novel Sparse Kalman Filter

A hybrid model-data-driven framework for inverse load identification of interval structures based on physics-informed neural network and improved Kalman filter algorithm

Impact Force Localization and Reconstruction Via Gated Temporal Convolutional Network

A Novel Method for Force Identification Based on Filter and FEM

Time Domain Force Localization and Reconstruction Based on Hierarchical Bayesian Method

Force Reconstruction for Nonlinear Structures in Time Domain

Wheel-Rail Force Identification Method Based on CNN-BiLSTM Hybrid Model

Multi-source uncertainty-oriented dynamic force reconstruction framework based on adaptive fitting precise integration and optimized wavelet denoising

A Novel Method for Force Identification Based on the Discrete Cosine Transform

Force Identification Based on a Comprehensive Approach Combining Taylor Formula and Acceleration Transmissibility

Interval strategy-based regularization approach for force reconstruction with multi-source uncertainties

A Model-Based Deep Learning Approach to Interpretable Impact Force Localization and Reconstruction

A Uncorrelated Multi-Sources Dynamic Random Load Identification Algorithm Based on Least-Squares in Frequency

A Transfer and Deep Learning-Based Method for Online Frequency Stability Assessment and Control

Multiscale Residual Attention Convolutional Neural Network for Bearing Fault Diagnosis

Fusing LSTM neural network and expanded disturbance Kalman filter for estimating external disturbing forces of ball screw drives