ANN-based design optimization and laser structuring for EV inverters

As the demand for high-performance electric vehicles (EVs) increases, the importance of energy-efficient thermal management of inverters has become ever more important. Also, the energy density, layout, and total power output of inverter chips have been rapidly changing. To address these evolving challenges and develop optimal thermal management solutions, we have incorporated an Artificial Neural Network (ANN)-based fast optimization scheme. This approach enables us to tackle the thermal issues across various industrial platforms, both by drastically reducing the time required for design optimization and by maximizing performance. Furthermore, to physically realize these optimized cooling designs, we employed femtosecond laser processing to precisely fabricate microstructures directly on the substrate. This advanced technique confirmed that high-density microstructures can be successfully fabricated directly on temperature-sensitive, package-level devices.