Fatigue Life Prediction of a Board-Level Assembly for Random Vibrations

Chen Yang,Jun Wang
DOI: https://doi.org/10.1109/icept.2017.8046639
2017-01-01
Abstract:This study combines finite element (FE) modeling and digital signal processing (DSP) to predict the fatigue life of a ball grid array (BGA) package mounted on the print circuit board (PCB) under a random vibration loading. First, the board-level assembly was modeled using ABAQUS. Then the modal analysis and a following power spectrum analysis were performed to calculate the power spectral density (PSD) response. The averaged Von-Mise stress of the critical BGA in time-domain was calculated by filtering Gaussian distribution random numbers. On the basis of the S-N curves and the Miner's rule, the rain-flow counting method was adopted to figure out the cumulative damages. The PSD and root mean square (RMS) value of the generated Von-Mises stresses agrees well with the simulation results and the predicted life also well fits the result calculated by three-band method. The methodology is effective in predicting component's life, especially in the system with a wide band response spectrum.
What problem does this paper attempt to address?