Scholarlink

Steel hardness, critical in automotive, construction, and military applications, can be tailored through chemical composition and heat treatment parameters. This study utilized Artificial Neural Networks (ANN) to develop Quantitative Structure-Property Relationships (QSPR) for predicting Vickers Hardness (HV) and strength of thermally treated steel grades. Using a Multi-Layer Perceptron approach in MATLAB®, ANN models were trained with datasets of 432, 703, and 1184 patterns, achieving correlation coefficients of 0.95, 0.947, and 0.90, respectively. Seven configurations, including up to 10 neurons in the hidden layer, were evaluated to optimize model accuracy. Key input parameters identified include steel composition, austenitizing time and temperature, and critical cooling time. A user-friendly Graphical User Interface (GUI) was developed to streamline model use and validate predictions against experimental and industrial data. This approach enhances steel property optimization for various applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development of Quantitative-Structure Property Relationship (QSPR) and Artificial Neural Network (ANN) Based Approaches for Estimating Mechanical Properties of Thermally Treated Steel Grades

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • R. Archa,
  • Y. Samih,
  • S. Baki Senhaji,
  • J. Jacquemin

摘要

Steel hardness, critical in automotive, construction, and military applications, can be tailored through chemical composition and heat treatment parameters. This study utilized Artificial Neural Networks (ANN) to develop Quantitative Structure-Property Relationships (QSPR) for predicting Vickers Hardness (HV) and strength of thermally treated steel grades. Using a Multi-Layer Perceptron approach in MATLAB®, ANN models were trained with datasets of 432, 703, and 1184 patterns, achieving correlation coefficients of 0.95, 0.947, and 0.90, respectively. Seven configurations, including up to 10 neurons in the hidden layer, were evaluated to optimize model accuracy. Key input parameters identified include steel composition, austenitizing time and temperature, and critical cooling time. A user-friendly Graphical User Interface (GUI) was developed to streamline model use and validate predictions against experimental and industrial data. This approach enhances steel property optimization for various applications.