Scholarlink

<p>The high carbon and chromium content, along with heat treatability, make AISI D2 steel a great choice for strength and wear-resistant applications. However, its hardness leads to heat generation and poor surface quality. This study investigates the influence of graphene-aluminum oxide hybrid nanofluids and machining parameters on surface roughness, cutting temperature, and material removal rate (MRR) in the Computer Numerical Control (CNC) turning of AISI D2 steel. Advanced optimization approaches, including Face-Centered Central Composite Design (CCD), Desirability Function Analysis (DFA), Random Forest Regression (RFR), Linear Regression (LR), Extreme Gradient Boosting (XGBoost), Artificial Neural Network (ANN) and Genetic Algorithm (GA), are employed to optimize machining parameters. The machine learning results reveal that XGBoost surpasses RFR and LR in accurately predicting cutting temperature and surface roughness. The ANN model's optimal validation includes a&#xa0;Mean Square Error (MSE) of 0.0031603 and an R-value close to one. The Face-Centered CCD of the Response Surface Methodology (RSM) and ANN outputs validated each other, demonstrating a strong correlation between the results. The optimal machining parameters predicted by DFA and GA include a cutting speed of 80 m/min, a nanofluid concentration of 1.4%, a depth of cut of 0.3 mm, and a feed rate of 0.07 mm/rev. Under these conditions, the cutting temperature was 23.89 °C, surface roughness was 0.53 μm, and the MRR reached 2579.7 mm<sup>3</sup>/min. Confirmation experiments aligned with the GA-predicted values, with DFA predictions showing minimal percentage errors. This study confirms that integrating statistical optimization, machine learning, and hybrid nanofluid cooling significantly improves surface quality, thermal efficiency, and MRR in machining operations.</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

Improving computer numerical control (CNC) turning performance of AISI D2 steel with nanofluid composites and advanced machine learning techniques

  • 期刊论文

  • 发表日期：

    2025年4月21日

查看全文
  • Dame Alemayehu Efa,
  • Naol Dessalegn Dejene,
  • Dejene Alemayehu Ifa,
  • Sololo Kebede Nemomsa,
  • Temesgen Batu Gemechu

摘要

The high carbon and chromium content, along with heat treatability, make AISI D2 steel a great choice for strength and wear-resistant applications. However, its hardness leads to heat generation and poor surface quality. This study investigates the influence of graphene-aluminum oxide hybrid nanofluids and machining parameters on surface roughness, cutting temperature, and material removal rate (MRR) in the Computer Numerical Control (CNC) turning of AISI D2 steel. Advanced optimization approaches, including Face-Centered Central Composite Design (CCD), Desirability Function Analysis (DFA), Random Forest Regression (RFR), Linear Regression (LR), Extreme Gradient Boosting (XGBoost), Artificial Neural Network (ANN) and Genetic Algorithm (GA), are employed to optimize machining parameters. The machine learning results reveal that XGBoost surpasses RFR and LR in accurately predicting cutting temperature and surface roughness. The ANN model's optimal validation includes a Mean Square Error (MSE) of 0.0031603 and an R-value close to one. The Face-Centered CCD of the Response Surface Methodology (RSM) and ANN outputs validated each other, demonstrating a strong correlation between the results. The optimal machining parameters predicted by DFA and GA include a cutting speed of 80 m/min, a nanofluid concentration of 1.4%, a depth of cut of 0.3 mm, and a feed rate of 0.07 mm/rev. Under these conditions, the cutting temperature was 23.89 °C, surface roughness was 0.53 μm, and the MRR reached 2579.7 mm3/min. Confirmation experiments aligned with the GA-predicted values, with DFA predictions showing minimal percentage errors. This study confirms that integrating statistical optimization, machine learning, and hybrid nanofluid cooling significantly improves surface quality, thermal efficiency, and MRR in machining operations.