Particle swarm optimization with extended Kalman filter for prognostication of accrued damage in electronics under temperature and vibration

Electronics in harsh environments may be subjected to extended periods of simultaneous high temperature and vibration and stored for prolonged periods of time prior to and during deployment. Damage accrued during the storage and prior usage may reduce the remaining useful life of the electronic system during deployment. Methods for assessment of accrued damage under simultaneous stresses are scarce. Test to failure data has been measured to study the effect of simultaneous thermal and vibration loadings on the reliability of BGA components. Two groups of pristine and isothermally aged components have been tested at both room temperature and 125°C while simultaneously being subjected to vibration loads. The transient response of printed circuit boards under the overlapping stresses has been characterized. Damage accrued under overlapping stresses has been investigated using physics-based leading indicators of damage. The leading indicators are state vectors based on optically measured strains and deformation, resistance spectroscopy and phase sensitive detection. An extended Kalman filter (EKF) is employed to predict remaining useful life (RUL) of the BGA components. A particle swarm optimization (PSO) technique, has been used to robustly demonstrate and quantify the repeatability of the resistance spectroscopy measurements and the accompanying prognostic algorithms.