Data-based deep learning for random vibration fatigue life prediction of car seat frame

Seats are important components in automobiles, susceptible to vibration and fatigue caused by random road spectra. To address this issue, we introduce three methods for predicting fatigue life using measured data. Initially, this paper compiles measured load spectra to conduct accelerated dynamic analysis and evaluate fatigue life using time domain methods. Subsequently, frequency domain analysis of random vibrations is performed based on the converted power spectral density (PSD) data. Finally, a deep learning model, fueled by experimental data, is proposed to monitor the real-time fatigue life of the car seat frame. The comparative analysis of these methodologies confirms the computational efficiency and real-time performance of the proposed data-driven deep learning model. We demonstrate that this model has substantial practical significance for evaluating the durability and reliability of structures or components in automotive, machining, aerospace, and related fields.