Development and Applications of Fine-grained Digital Twins for Power Electronic Modules (22-02)

Principal Investigator: Dr. Lingfeng Wang

Digital Twin (DT) is the virtual and computerized counterpart of a physical system. It is an emerging solution for the modeling and simulation of Cyber-Physical Systems (CPS) which is regarded as one of the most promising technologies for realizing intelligent manufacturing, industry 4.0 and the smart grid. DT can benefit various industry sectors through downtime reduction, situational awareness enhancement, inventory management, fleet management, what-if simulations, etc. Thus, the concept of DT has been warmly embraced by the manufacturing industries. However, the study of DT in the evolving electrical energy area is still in the initial stage but its related DT market is expected to be extremely huge. Thus, this project aims to develop the DT modeling framework and application schemes for the power electronic modules which are in the heart of many power electronic systems and smart grids with high shares of renewable energy resources. This project aims to develop the DT of representative power electronic modules for contemporary power grids together with practical DT-based applications in critical parameter estimation and health condition assessment, which is much needed in the current energy industry and many other industries where power electronic modules are widely used. In this project, the novel DT modeling framework of the IGBT power modules will be developed based on the 3D finite element model (FEM) in the ANSYS or COMSOL environment. Applications of DT for the power electronic modules will also be developed. The main innovations of the proposed work are listed as follows:

• A fine-grained FEM-based DT modeling scheme for enabling multi-physics simulations will be developed for power electronic modules.
• An intelligent interface will be developed for the DT based on cutting-edge deep learning algorithms. The interface will be able to update the parameters and status of the DT properly and efficiently according to the inputs and measurements from the physical world.
• A novel DT-based approach will be developed for the junction temperature estimation of power modules.
• A health condition assessment method will be developed which can accurately evaluate the effect of the chip solder fatigue of power modules using DT.