Load identification for carbody and hanging equipment based on compressed sensing and weighted sparse regularisation

An improved methodology for identifying the dynamic coupling loads at the carbody-hanging equipment interface is proposed, aiming to enhance structural dynamic behaviour across a range of operational scenarios. The mathematical model is established to simulate the carbody and hanging equipment interactions through Euler-Bernoulli beam theory and the vehicle vertical coupled dynamics simulation approach. The load identification model is constructed through integration of compressed sensing (CS) technology with a weighted sparse regularisation strategy, while an improved Akaike Information Criterion (AIC) is employed for optimisation of regularisation parameters. It is demonstrated that the proposed method effectively accounts for both structural coupled vibrations and the influence of different excitation sources from the equipment. Furthermore, a CS framework combined with a redundant dictionary is constructed to achieve sparse signal decomposition, thereby resolving the conflict in simultaneous identification of harmonic loads and impulse components. At the core of the proposed iterative algorithm is a strategy founded on weighted sparse regularisation augmented by an improved AIC, dedicated to the optimal selection of the regularisation parameter, leading to superior accuracy and robustness against noise. Validation results from both simulation and roller rig experiments verify the proposed method and confirm its effectiveness and practical value.