Scholarlink

<p>This study investigated multi-objective optimization and mechanical property prediction of FDM components with six input parameters. Tensile strength (UTS) and elongation percentage (EL) were selected as performance outputs. The GRA method was used to evaluate multi-response behaviour. The results showed that the maximum Grey Relational Grade (GRG) was 0.9714. The optimal parameters that produced the best results for elongation (2.4%) and tensile strength (37&#xa0;MPa) were 0.06&#xa0;mm layer height, 80% infill density, 220&#xa0;°C nozzle temperature, 60&#xa0;°C bed temperature, 60&#xa0;mm/s print speed and insignificant fan speed. In order to evaluate prediction capabilities, an XGBoost regression model was created using GRG as a target output. The model attained the exceptional accuracy with R2 (0.8781), MSE (0.00293), RMSE (0.05414), MAE (0.03911) and MAPE (8.16%). These outcomes indicated that XGBoost is a reliable technique for predicting multi-performance optimisation outcomes in FDM, leading to precise microstructure manipulation and therefore improving the mechanical performance of printed components.</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 forecasting of mechanical characteristics of FDM printed products using grey relational grade analysis and XGBoost

  • 期刊论文

  • 发表日期：

    2026年4月29日

查看全文
  • Ali Ahmed Alqudaihi,
  • E. M. Shaban,
  • Abdulrahman S. Sait,
  • Alaauldeen A. Duhduh,
  • Kelli DurgaPrasad

摘要

This study investigated multi-objective optimization and mechanical property prediction of FDM components with six input parameters. Tensile strength (UTS) and elongation percentage (EL) were selected as performance outputs. The GRA method was used to evaluate multi-response behaviour. The results showed that the maximum Grey Relational Grade (GRG) was 0.9714. The optimal parameters that produced the best results for elongation (2.4%) and tensile strength (37 MPa) were 0.06 mm layer height, 80% infill density, 220 °C nozzle temperature, 60 °C bed temperature, 60 mm/s print speed and insignificant fan speed. In order to evaluate prediction capabilities, an XGBoost regression model was created using GRG as a target output. The model attained the exceptional accuracy with R2 (0.8781), MSE (0.00293), RMSE (0.05414), MAE (0.03911) and MAPE (8.16%). These outcomes indicated that XGBoost is a reliable technique for predicting multi-performance optimisation outcomes in FDM, leading to precise microstructure manipulation and therefore improving the mechanical performance of printed components.