Litcius/Paper detail

An Innovative and Efficient Approach for Field-Limiting Ring Design of Power Devices Based on Deep Neural Networks

Jingyu Li, Haoyue Zhao, Hao Yuan, Fengyu Du, Zixi Wang, Xiao-Yan Tang, Qingwen Song, Yuming Zhang

2024IEEE Electron Device Letters10 citationsDOI

Abstract

The design of device termination is crucial for power devices. In this letter, we present a novel approach for designing termination in Si-based power devices using deep neural networks (DNN), taking the narrowed field-limiting ring (NFLR) as an example. Our proposed method can not only achieve a prediction accuracy of over 97% for breakdown voltage (BV), but also provide the device optimization design scheme automatically and intelligently. We believe that the proposed machine learning technology can significantly reduce the cost and enhance the efficiency of power device design, establishing itself as a promising method for power device design.

Topics & Concepts

LimitingArtificial neural networkComputer sciencePower (physics)Field (mathematics)Electronic engineeringScheme (mathematics)Power semiconductor deviceDeep learningVoltageElectrical engineeringEngineeringArtificial intelligenceMechanical engineeringMathematical analysisPhysicsPure mathematicsMathematicsQuantum mechanicsSilicon Carbide Semiconductor TechnologiesAdvancements in Semiconductor Devices and Circuit DesignSemiconductor materials and devices