Machine learning aided modelling of thermomechanical fatigue of solder joints in electronic component assemblies

Printed board assemblies, i.e. components soldered on printed circuit boards (PCBs), are exposed to thermal cycles responsible for fatigue cracking of solder joints as a result expansion mismatch between the constituting elements. Advanced finite element simulations performed using traction–separation law represent process and temperature-dependent elasto-viscoplastic model joint response. Predictions successfully assessed towards machine learning processed experimental data. In particular, high sensitivity ageing reliability geometric dimensions thickness is properly captured. Additional parameters, related PCB substrate, also studied, opening new avenues design optimization.