Scholarlink

<p>One effective technique for creating metallic parts with a near-net form is additive manufacturing (AM). Regardless of its immense potential, the inbuilt problems of high roughness and comparatively low dimensional accuracy frequently require the final stage of using subtractive processes. To guarantee that the finished components satisfy the particular needs of their intended applications, post-processing is crucial. This study investigates the machinability of selective laser-melted stainless steel 316 (SLMed-SS 316) under various cutting environments, including cryogenic CO<sub>2</sub>, dry, flood, and minimum quantity lubrication (MQL). The study assesses flank wear (Vb) and surface roughness (Ra) to determine how these environmental factors affect machined surfaces. Cryogenic CO<sub>2</sub> produced the best surface quality, reducing Ra by 49-53%, 35-41%, and 17-21% compared with dry, flood, and MQL conditions, respectively. Likewise, Vb under cryogenic CO<sub>2</sub> was 43-48% lower than dry machining, 29-33% lower than flood cooling, and 17-20% lower than MQL. Additionally, machine learning (ML) techniques random forest (RF) and bagging (BA) are used to forecast Ra and Vb. Because of their ability to handle nonlinear interactions, reduce overfitting, and produce reliable predictions, both ensemble approaches were selected. It has been demonstrated that cryogenic CO<sub>2</sub> condition significantly enhances surface quality and reduces Vb. The RF algorithm outperforms the BA algorithm by showing superior predictive performance on key metrics, R<sup>2</sup>, MAE, RMSE, RAE, and RRSE.</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

Enhancing Machinability of Additively Manufactured SS 316 through Cryogenic Cooling: Experimental and Machine Learning Perspectives

  • 期刊论文

  • 发表日期：

    2026年3月4日

查看全文
  • S. Sellakumar,
  • C. Gopinath,
  • M. Belsam Jeba Ananth,
  • B. Navin Kumar

摘要

One effective technique for creating metallic parts with a near-net form is additive manufacturing (AM). Regardless of its immense potential, the inbuilt problems of high roughness and comparatively low dimensional accuracy frequently require the final stage of using subtractive processes. To guarantee that the finished components satisfy the particular needs of their intended applications, post-processing is crucial. This study investigates the machinability of selective laser-melted stainless steel 316 (SLMed-SS 316) under various cutting environments, including cryogenic CO2, dry, flood, and minimum quantity lubrication (MQL). The study assesses flank wear (Vb) and surface roughness (Ra) to determine how these environmental factors affect machined surfaces. Cryogenic CO2 produced the best surface quality, reducing Ra by 49-53%, 35-41%, and 17-21% compared with dry, flood, and MQL conditions, respectively. Likewise, Vb under cryogenic CO2 was 43-48% lower than dry machining, 29-33% lower than flood cooling, and 17-20% lower than MQL. Additionally, machine learning (ML) techniques random forest (RF) and bagging (BA) are used to forecast Ra and Vb. Because of their ability to handle nonlinear interactions, reduce overfitting, and produce reliable predictions, both ensemble approaches were selected. It has been demonstrated that cryogenic CO2 condition significantly enhances surface quality and reduces Vb. The RF algorithm outperforms the BA algorithm by showing superior predictive performance on key metrics, R2, MAE, RMSE, RAE, and RRSE.