Deep Learning Reliability Awareness of Converters at the Edge

Deep Learning Reliability Awareness of Converters at the Edge
Electrical and Computer Engineering

Online monitoring of power electronic converters enables the user to optimize the efficiency and operation of the system over its lifetime. Sensors and efficient characterization of the system are the keys to achieve this optimization. The main challenge is the non-scalability of solutions mainly because of diverse architectures for different applications. 

This DoD-funded project sets to move beyond mainstream device modeling and traditional reliability analysis (i.e. Weibull distributions, mean-time-to-failure, etc.) and looking to more applicable analytical tools through introducing advanced sensing solutions and combining it with cutting-edge deep learning techniques. In the next three years, lead PI Dr. Babak Parkhideh will collaborate with Dr. Hamed Tabkhi and Dr. Robert Cox of the ECE Department to investigate advanced sensors and data processing units to measure and process the characteristics of emerging semiconductor devices at very high frequency/resolution empowered by cutting-edge computing technology for Just-In-Time (JIT) behavioral modeling and decision making. Through a tightly-coupled integrated research collaboration, this project will co-design and co-create an advanced real-time characterization architecture. The architecture that can be integrated into the next generation of power electronics for online reliability assessment and improving the efficiency or lifetime of the system by means of proper corrective actions.