Scholarlink

<p>The development of lightweight, high-strength materials is critical for next-generation aerospace and automotive applications. In this study, a comprehensive materials informatics framework is developed to predict the microhardness and tensile strength of microwave-sintered AA7075/SiC/ZrC hybrid composites. A structured experimental dataset comprising 172 samples was generated by systematically varying SiC and ZrC content, compaction pressure, sintering temperature, and sintering time, ensuring broad coverage of the processing space. Unlike conventional studies that rely solely on standalone machine learning implementations, the present work integrates advanced data visualization, rigorous model validation, and physically interpretable learning. Multiple regression algorithms—including Artificial Neural Networks (ANN), Extreme Gradient Boosting (XGBoost), Random Forest (RF), Support Vector Regression (SVR), and K-Nearest Neighbors (KNN)—were trained using optimized hyperparameters and evaluated through nested cross-validation to ensure robustness and generalizability. Among these, ANN and XGBoost demonstrated superior predictive performance, achieving coefficients of determination (R²) exceeding 0.97 for tensile strength and 0.95 for microhardness. A key novelty of this study lies in explicitly linking machine learning predictions with underlying metallurgical mechanisms. Feature importance analysis, supported by microstructural observations, reveals that tensile strength is predominantly governed by compaction pressure and reinforcement distribution, while microhardness is strongly influenced by SiC content and sintering parameters. These relationships are interpreted in terms of densification behavior, Orowan strengthening, and grain refinement mechanisms. By bridging experimental materials science with interpretable machine learning, this work provides a reliable and physically grounded predictive framework that reduces experimental effort and enables accelerated optimization of hybrid aluminium matrix composites.</p>

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

Data-driven prediction of microhardness and tensile strength in microwave-sintered ZrC reinforced AA7075/SiC hybrid composites using machine learning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月3日

查看全文
  • E Srinath,
  • K Venkateswara Reddy,
  • Guttikonda Manohar,
  • Pramod Kumar P,
  • Nagaraj Ashok

摘要

The development of lightweight, high-strength materials is critical for next-generation aerospace and automotive applications. In this study, a comprehensive materials informatics framework is developed to predict the microhardness and tensile strength of microwave-sintered AA7075/SiC/ZrC hybrid composites. A structured experimental dataset comprising 172 samples was generated by systematically varying SiC and ZrC content, compaction pressure, sintering temperature, and sintering time, ensuring broad coverage of the processing space. Unlike conventional studies that rely solely on standalone machine learning implementations, the present work integrates advanced data visualization, rigorous model validation, and physically interpretable learning. Multiple regression algorithms—including Artificial Neural Networks (ANN), Extreme Gradient Boosting (XGBoost), Random Forest (RF), Support Vector Regression (SVR), and K-Nearest Neighbors (KNN)—were trained using optimized hyperparameters and evaluated through nested cross-validation to ensure robustness and generalizability. Among these, ANN and XGBoost demonstrated superior predictive performance, achieving coefficients of determination (R²) exceeding 0.97 for tensile strength and 0.95 for microhardness. A key novelty of this study lies in explicitly linking machine learning predictions with underlying metallurgical mechanisms. Feature importance analysis, supported by microstructural observations, reveals that tensile strength is predominantly governed by compaction pressure and reinforcement distribution, while microhardness is strongly influenced by SiC content and sintering parameters. These relationships are interpreted in terms of densification behavior, Orowan strengthening, and grain refinement mechanisms. By bridging experimental materials science with interpretable machine learning, this work provides a reliable and physically grounded predictive framework that reduces experimental effort and enables accelerated optimization of hybrid aluminium matrix composites.