Scholarlink

<p>This study evaluates the mechanical behavior of monolithic and Cu-reinforced porous alumina ceramics and examines the predictive capability of an Artificial Neural Network (ANN) model for tensile and flexural strengths. Rice husk ash (RHA), yeast, and graphite waste were used as pore-forming agents in the fabrication of samples with different loadings and nano-Cu concentrations. Total porosity increased with increasing pore-former loading (24–61.8%, 21–63.8%, and 24.5–49.5% for samples based on graphite waste, yeast, and RHA, respectively), which reduced the cross-sectional area for load bearing and consequently decreased mechanical strength. For the corresponding systems, the tensile strengths were 14.3–46.4&#xa0;MPa, 5.4–37.3&#xa0;MPa, and 19–48.3&#xa0;MPa, while the flexural strengths were 30.4–123.4&#xa0;MPa, 20.7–95.7&#xa0;MPa, and 63.6–207.3&#xa0;MPa. Through pore bridging and partial filling, nano-Cu inclusion significantly improved both properties and improved stress transfer within the matrix. The ANN model developed using pore former type, Cu content, and porosity as inputs, demonstrated high predictive accuracy for tensile (R² = 0.9818; MAE = 0.0361) and flexural strength (R² = 0.9888; MAE = 0.0220). Clearly, the ANN model outperformed conventional regression techniques in terms of prediction accuracy and successfully represented the nonlinear correlations between mechanical performance and processing parameters. These findings demonstrate the strength and reliability of ANN modeling as an effective method for predicting and optimizing the mechanical behavior of porous ceramic structures.</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

Mechanical behavior evaluation of monolithic and Cu-doped porous alumina prepared with different pore formers: An ANN-based approach

  • 期刊论文

  • 发表日期：

    2026年4月21日

查看全文
  • Mohammed Sabah Ali,
  • M. A. Azmah Hanim,
  • T. T. Dele-Afolabi

摘要

This study evaluates the mechanical behavior of monolithic and Cu-reinforced porous alumina ceramics and examines the predictive capability of an Artificial Neural Network (ANN) model for tensile and flexural strengths. Rice husk ash (RHA), yeast, and graphite waste were used as pore-forming agents in the fabrication of samples with different loadings and nano-Cu concentrations. Total porosity increased with increasing pore-former loading (24–61.8%, 21–63.8%, and 24.5–49.5% for samples based on graphite waste, yeast, and RHA, respectively), which reduced the cross-sectional area for load bearing and consequently decreased mechanical strength. For the corresponding systems, the tensile strengths were 14.3–46.4 MPa, 5.4–37.3 MPa, and 19–48.3 MPa, while the flexural strengths were 30.4–123.4 MPa, 20.7–95.7 MPa, and 63.6–207.3 MPa. Through pore bridging and partial filling, nano-Cu inclusion significantly improved both properties and improved stress transfer within the matrix. The ANN model developed using pore former type, Cu content, and porosity as inputs, demonstrated high predictive accuracy for tensile (R² = 0.9818; MAE = 0.0361) and flexural strength (R² = 0.9888; MAE = 0.0220). Clearly, the ANN model outperformed conventional regression techniques in terms of prediction accuracy and successfully represented the nonlinear correlations between mechanical performance and processing parameters. These findings demonstrate the strength and reliability of ANN modeling as an effective method for predicting and optimizing the mechanical behavior of porous ceramic structures.