成形加工
Online ISSN : 1883-7417
Print ISSN : 0915-4027
ISSN-L : 0915-4027
35 巻, 11 号
選択された号の論文の12件中1~12を表示しています
目次
巻頭言
解説 : 特集 押出成形
技術報告
プリンキピア ―常識を変えたモノたちの物語―
海外だより
知りたい・見たい・訪ねたい 成形加工の元気な仲間
論文
  • 佐藤 智, 染宮 聖人, 平山 紀夫, 山本 晃司, 松原 成志朗, 寺田 賢二郎
    2023 年 35 巻 11 号 p. 404-410
    発行日: 2023/10/20
    公開日: 2023/11/20
    ジャーナル フリー

    We propose a method for predicting the interfacial adhesion or bond strength of unidirectional carbon fiber reinforced thermoplastic plastics (UD-CFRTP) using a neural network (NN) and numerical material testing (NMT) that takes into account the plastic behavior of resin. In the proposed method, first, elastoplastic materials are assumed for the matrix resin, and macroscopic fracture strengths are calculated from NMTs that simulate off-axis tensile tests of UD-CFRTP. Next, a series of NMTs are performed by varying the interfacial adhesion strength between the fiber and resin, the fracture strength of the matrix resin, and the fiber volume fraction, respectively, and the relationships with the obtained macroscopic fracture strengths of UD-CFRTP are learned by the NN. Then, using the learned NNs, the microscopic interfacial adhesion strength and fracture strength of the matrix resin are predicted from the results of actual off-axis tensile tests of UDCFRTP. To verify the accuracy of the proposed method, NMTs are conducted using the predicted strengths, and the results are compared and evaluated with the results of actual off-axis tensile tests of UD-CFRTP.

feedback
Top