The goal of this project is to develop a primarily comprehensive automated process chain for failure diagnostics integrating self-learning data analysis and defect detection and localization, standardized material analysis, and centralized data acquisition and correlation to electrical performance data. A consortium from four European countries Germany, France, Sweden, and the Czech Republic including the semiconductor and electronic system suppliers of Infineon, ST, and Bosch will collaborate with four real SMEs and three medium-sized companies from the failure analysis equipment and related software domain supported by three research institutes to push failure analysis in Europe to the next level of efficiency.
Responsibilities of ILH
In the project FA4.0, the multi-physical finite element field simulation of electronic assemblies will be integrated into a learning system for failure analysis. The goal is the AI-controlled execution of multi-physical field simulations for the targeted characterization and classification of typical fault patterns, the comparison with measurement data, and the efficient, AI-compatible processing of the simulation data.
The failure analysis of the silicon carbide power transistor via time‑domain reflectometry (TDR) while considering different calibration parameters to improve the accuracy of the measurement in the manufacturing process is automated. The supporting simulation model is based on a reinforcement learning approach where the simulation data is generated by co-simulation of MATLAB and ADS to efficiently train hybrid machine learning models. The proposed approach can be used in Industry4.0 processes since it can differentiate all hard and soft failures while considering the rise-time of the input step signal which is normally not considered in the conventional TDR mathematical formulae.