Fraunhofer Institute for Integrated Systems and Device Technology IISB
Fraunhofer Institute for Integrated Systems and Device Technology IISB

Digital Twin for Applications in Power Electronics

Digital twins for power electronics provide virtual, physics-based representations of real products and their behavior during manufacturing, testing and operation. Experimentally validated electro-thermo-mechanical finite element models (FEM), lifetime analyses and physics-informed machine learning enable the investigation of loads, critical locations and damage mechanisms. This allows the The reliability and lifetime of power electronic systems can already be to be assessed during development and design, material and load parameters canto be investigated systematically.

At Fraunhofer IISB, we develop and validate physics-based simulation models and combine them with experimental measurements, reliability tests and data-driven methods. This enables allows us to support development and reliability teams in virtual reliability assessment, lifetime prediction and the optimization of power electronic products.

Virtual Reliability Assessment

Identify critical locations and failure mechanisms early and assess the reliability of power electronic systems

We use electro-thermo-mechanical finite element models (FEM) to investigate electrical, thermal and mechanical loads in power electronic systems. We cover the entire process from manufacturing to operation. This allows us to identify critical locations and potential failure mechanisms early in the virtual model. We compare the simulation models with experimental measurements to ensure robust results. To do this, we characterize electrical, thermal and mechanical material properties and measure thermal impedance, among other parameters. On this basis, we precisely calibrate material and interface properties.

Related publications

  • Zhao, Dawei, et al. (2024): “Adhesion strength of ductile thin film determined by cross-sectional nanoindentation.” International Journal of Mechanical Sciences 270: 109103. DOI: 10.1016/j.ijmecsci.2024.109103
  • Zhao, Dawei, et al. (2023): “Influence of SiC chip thickness on the power cycling capability of power electronics assemblies – A comprehensive numerical study.” Microelectronics Reliability 150: 115091. DOI: 10.1016/j.microrel.2023.115091
  • Zhao, Dawei, et al. (2020): “Combined experimental and numerical approach for investigating the mechanical degradation of the interface between thin film metallization and Si-substrate after temperature cycling test.” Microelectronics Reliability 114: 113785. DOI: 10.1016/j.microrel.2020.113785

Lifetime Prediction under Real-World Load Conditions

Reliably predict the lifetime of power electronic systems under application-specific loads

How long will a power electronic system last under the loads of its specific application? For lifetime assessment, we combine physics-based simulations with data from reliability tests and application-specific mission profiles. This includes evaluating power cycling tests (PCT), thermal shock tests (TST) and Mission Profiles. The simulation reveals electro-thermo-mechanical loads and the resulting accumulation of damage. This enables us to analyze how different load conditions affect lifetime and which failure mechanisms play a role. Experimental data from power cycling tests help us further improve the lifetime models.

Related publications

  • Bose, Abantika, et al. (2026): “An Advanced Mission Profile-based Lifetime Assessment in Power Electronics Assembly Using Physics-informed Surrogate Modeling.” 27th International Conference on Thermal, Mechanical and Multi-Physics Simulation and Experiments in Microelectronics and Microsystems (EuroSimE). IEEE. DOI: 10.1109/EuroSimE69483.2026.11511966
  • Zhao, Dawei, et al. (2025): “On the validity of rainflow counting-based lifetime assessment for power electronics assembly.” Microelectronics Reliability 168: 115651. DOI: 10.1016/j.microrel.2025.115651
  • Zhao, Dawei, et al. (2023): “Influence of SiC chip thickness on the power cycling capability of power electronics assemblies – A comprehensive numerical study.” Microelectronics Reliability 150: 115091. DOI: 10.1016/j.microrel.2023.115091

Fast Models for Monitoring and Design Optimization

Accelerate complex simulations and make them usable for design optimization and monitoring

High-resolution finite element simulations provide detailed information about the behavior of power electronic systems but require substantial computing time. We therefore develop physics-informed surrogate models that combine physical knowledge with data-driven methods. These models deliver fast, physically consistent and interpretable evaluations. Depending on the application example, the surrogate models reduce representative computing times from several hours to less than ten seconds or from ten to twelve hours to around 30 seconds. This enables us to efficiently scale FEM evaluations across different designs and load conditions and make them accessible for real-time monitoring scenarios.

Related publications

  • Bose, Abantika, et al. (2026): “An Advanced Mission Profile-based Lifetime Assessment in Power Electronics Assembly Using Physics-informed Surrogate Modeling.” 27th International Conference on Thermal, Mechanical and Multi-Physics Simulation and Experiments in Microelectronics and Microsystems (EuroSimE). IEEE. DOI: 10.1109/EuroSimE69483.2026.11511966

Design and Material Optimization

Systematically assess and optimize the impact of design, materials and load conditions on reliability

How do design decisions, material properties or load conditions affect the reliability of a power electronic system? We use our simulation models to investigate and compare different variants virtually. Sensitivity analyses reveal which parameters have a particularly strong influence on reliability. Through design space exploration and multi-objective optimization, we systematically weigh different requirements against one another. To do this, we characterize thermal, electrical and mechanical material properties and incorporate them directly into the virtual analysis. This allows us to identify critical combinations of design, materials and load conditions early and establish a robust basis for targeted design and material decisions.

Related publication

  • Zhao, Dawei, et al. (2023): “Influence of SiC chip thickness on the power cycling capability of power electronics assemblies – A comprehensive numerical study.” Microelectronics Reliability 150: 115091. DOI: 10.1016/j.microrel.2023.115091
  • Letz, S. A., Dawei Zhao and Martin März (2022): “Mesostructural impact on the macroscopic stress state and yield locus of porous polycrystalline silver.” Materials & Design 219: 110785. DOI: 10.1016/j.matdes.2022.110785