Hybrid Approach to Conduct Failure Prognostics of Automotive Electronic Control Unit Using Stress Sensor as In Situ Load Counter

A model/sensor hybrid approach is implemented to conduct failure prognostics of an automotive electronic control unit (ECU). A 3-D finite-element model simulating a complex ECU is built, and its predictability is calibrated and verified by an optical displacement measurement technique called moiré interferometry. On the sensor front, a silicon-based piezoresistive stress sensor is embedded into the ECU to providein situstress measurements during operations. A stress metric is defined using the stress values of 12 cells in each sensor, and it is converted toin situloading histories using the calibrated finite-element model. The modeling and verification steps that lead to the predictive finite-element model are described. The proposed hybrid approach is implemented using the data obtained from a molded ECU subjected to thermal cycling conditions, and the forces at an interface between an aluminum wire bond and metal pad are investigated.