首页|Research from Beijing University of Technology Broadens Understanding of Machine Learning (Using Machine Learning and Finite Element Analysis to Extract Tractio n-Separation Relations at Bonding Wire Interfaces of Insulated Gate Bipolar ...)

Research from Beijing University of Technology Broadens Understanding of Machine Learning (Using Machine Learning and Finite Element Analysis to Extract Tractio n-Separation Relations at Bonding Wire Interfaces of Insulated Gate Bipolar ...)

扫码查看
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Investigators discuss new findings in artificial intelligence. According to news reporting from Beijing, People's Repu blic of China, by NewsRx journalists, research stated, "For insulated gate bipol ar transistor (IGBT) modules using wire bonding as the interconnection method, t he main failure mechanism is cracking of the bonded interface. Studying the mech anical properties of the bonded interface is crucial for assessing the reliabili ty of IGBT modules." Financial supporters for this research include National Natural Science Foundati on of China. Our news correspondents obtained a quote from the research from Beijing Universi ty of Technology: "In this paper, first, shear tests are conducted on the bonded interface to test the bonded interface's strength. Then, finite element-cohesiv e zone modeling (FE-CZM) is established to describe the mechanical behavior of t he bonded interface. A novel machine learning (ML) architecture integrating a co nvolutional neural network (CNN) and a long short-term memory (LSTM) network is used to identify the shape and parameters of the traction separation law (TSL) o f the FE-CZM model accurately and efficiently. The CNN-LSTM architecture not onl y has excellent feature extraction and sequence-data-processing abilities but ca n also effectively address the long-term dependency problem. A total of 1800 set s of datasets are obtained based on numerical computations, and the CNN-LSTM arc hitecture is trained with load-displacement (F-d) curves as input parameters and TSL shapes and parameters as output parameters. The results show that the error rate of the model for TSL shape prediction is only 0.186%."

Beijing University of TechnologyBeijin gPeople's Republic of ChinaAsiaCyborgsEmerging TechnologiesMachine Lea rning

2024

Robotics & Machine Learning Daily News

Robotics & Machine Learning Daily News

ISSN:
年,卷(期):2024.(Mar.8)