Improvement of Fatigue Life Prediction Accuracy for Electronic Components under Random Vibration

Printed circuit boards (PCBs) used in electronic devices are often subjected to random vibrations during transportation, which can cause fatigue failure at the terminals of mounted components. Accurate prediction of vibration-induced stresses is essential for reliable design and lifetime estimation. Since the main source of stress is deformation caused by the board’s vibration modes, the accuracy of deformation prediction directly affects the fatigue life evaluation. However, conventional models assuming uniform isotropic material properties cannot sufficiently capture the local stiffness variations caused by copper wiring patterns, resulting in reduced accuracy, especially for higher-order vibration modes. This study proposes a finite element modeling approach that incorporates local stiffness by dividing the PCB surface into a grid and applying binarization processing to identify copper pattern regions. Equivalent material properties reflecting copper stiffness are assigned accordingly. Furthermore, fatigue life under random vibration is predicted using Dirlik’s method for stress spectral analysis combined with the Corten-Dolan method for material fatigue correction. This approach significantly improves the accuracy of fatigue life prediction. The proposed model achieved good agreement with measured natural frequencies and mode shapes, and fatigue life predictions closely match experimental results.