R-PODID, a KDT JU co-funded project, aims to develop an automated, cloudless, short-term fault-prediction for electric drives, power modules, and power devices, that can be integrated into power converters.
As part of the R-PODID project, Applied Materials, Inc. / Think Silicon S.A. will develop ultra-low memory AI models for predictive maintenance customized for the RISC-V based NEOXTM multicore accelerator.
The project main objectives:
- Methodology for fault-prediction model generation from sparse training sets or system simulation
- Power electronics with integrated support for embedded AI
- 24h fault-prediction for Gallium Nitride (GaN) and Silicon Carbide (SiC) based power converters
- 24h fault-prediction and fault mitigation for electric drives
- Sensors for reliability prediction in power modules
Supported by 33 partners, R-PODID innovations are implemented into the power modules and applied in the four use cases for conveyor belts, industrial lighting, automotive traction inverters, and a heavy-duty testbed.
More information can be found here: https://www.linkedin.com/company/r-podid/