Optimal Clustering on Spectral Line in Inertial Parameter Identification From Experimental Data

The accurate identification of inertial parameters is essential for studying dynamic system behaviors, where frequency-domain inertia restraint methods are widely recognized for their computational efficiency and precision. However, these conventional approaches remain insufficient in resolving uncertainties associated with mass-line frequency band selection, which may compromise inertial parameter identification accuracy. To address this limitation, the paper analyses the relationship between the frequency response function and the inertia parameters, replaces the traditional idea of determining all inertia parameters by the spectral line optimization of a single inertia parameter, establishes the optimization criterion of the inertia parameter, and improves the efficiency and accuracy of the inertia parameter by combining the method of cluster analysis. The acceleration sensor and force sensors were employed to acquire the experimental data of the modal test, enabling the frequency response function analysis and modal parameters extraction. Comparative validation against traditional methods demonstrates that the proposed approach significantly enhances the accuracy and efficiency of rigid-body inertial parameter identification. The results validate the efficacy of parameter refinement and cluster-based optimization in inertia restraint methods for rigid-body inertial parameter determination, providing an optimal clustering framework for dynamic system parameter identification.