射出成形

14:00 〜 14:20

[B-107] 金属表面画像からの機械学習を用いた金属-樹脂接合の強度予測

*崔 鐘祺¹、王 鑠涵³、木村 文信³、伊藤 由華²、鈴木 幸徳²、梶原 優介³ (1. 東京大学、2. 新東工業、3. 東京大学生産技術研究所)

キーワード：

Injection Molded Direct Joining (IMDJ)、Blasting、Machine Learning

Injection Molded Direct Joining (IMDJ) is a direct joining technique that utilizes metal surface pretreatment and plastic injection molding for connection, eliminating the need for additional components. For surface treatment, blasting is employed because it uses environmentally friendly surface processing, reducing costs and achieving high productivity. In order to provide rapid and accurate predictions of joining strength, thereby optimizing design decisions and production processes while reducing the need for physical testing. This research focuses on applying machine learning technology to predict the joining strength during the metal and plastic joining process. A machine learning model has been developed to predict the joining strength based on two-dimensional images of metal surfaces obtained through a laser microscope. Additionally, by visualizing features, it is possible to identify structures beneficial to strength, thereby optimizing the blasting method.