Scholarlink

<p>The prediction of material properties such as hardness, density, and surface roughness of additively manufactured AlSi12Mg alloy components using machine learning techniques like K-Nearest Neighbor algorithm and Artificial Neural Network, is crucial for confirming the performance and quality of components, especially in integrating process parameters and enabling real-time predictions. Current approaches do not adequately account for intricate relationships between process parameters and material properties. Traditional methods need extensive experimentation and testing which is costly, time-consuming, and inefficient. By altering process variables including laser power, scan speed, and hatch distance, the primary goal of this work is to forecast the mechanical characteristics of AlSi12Mg alloy components that are additively built, including hardness, density, and surface roughness. The experimental data suggest that the parabola is steepening as the laser power is increased, with the density and hardness being more influenced (50–60%) by the laser power. Scan speed has also been shown to be a more significant factor (40%) in surface roughness. To address the aforementioned problem, this study models the relationship between process parameters and material qualities using machine learning algorithms like K-Nearest Neighbours (KNN) and Artificial Neural Networks (ANN). This proves that material properties predictions using KNN are more accurate than ANN, due to KNN’s ability to handle the linear relationship between process parameters and material properties efficiently. This research highlights the ability of machine learning algorithms in metal additive manufacturing, which reduces experimentation time, cost, and energy. By allowing real-time process control and optimized material properties, this research improves product quality and reduces total manufacturing costs in additive manufacturing. It facilitates improved manufacturing, employing machine learning to improve operational efficiency, customization, and quality control. This work could be extended to increase the prediction accuracy and enhance the optimization of process parameters by expanding the data set and applying many other supervised machine learning algorithms.</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

Predicting material properties in AlSi12Mg alloy additive manufacturing using KNN and ANN machine learning techniques

  • 期刊论文

  • 发表日期：

    2025年3月24日

查看全文
  • M. Arunadevi,
  • L. Avinash,
  • Amit Tiwari,
  • C. Durga Prasad,
  • R. Suresh Kumar,
  • L. Raghavendra,
  • G. Veeresha,
  • Sagarkumar J. Aswar

摘要

The prediction of material properties such as hardness, density, and surface roughness of additively manufactured AlSi12Mg alloy components using machine learning techniques like K-Nearest Neighbor algorithm and Artificial Neural Network, is crucial for confirming the performance and quality of components, especially in integrating process parameters and enabling real-time predictions. Current approaches do not adequately account for intricate relationships between process parameters and material properties. Traditional methods need extensive experimentation and testing which is costly, time-consuming, and inefficient. By altering process variables including laser power, scan speed, and hatch distance, the primary goal of this work is to forecast the mechanical characteristics of AlSi12Mg alloy components that are additively built, including hardness, density, and surface roughness. The experimental data suggest that the parabola is steepening as the laser power is increased, with the density and hardness being more influenced (50–60%) by the laser power. Scan speed has also been shown to be a more significant factor (40%) in surface roughness. To address the aforementioned problem, this study models the relationship between process parameters and material qualities using machine learning algorithms like K-Nearest Neighbours (KNN) and Artificial Neural Networks (ANN). This proves that material properties predictions using KNN are more accurate than ANN, due to KNN’s ability to handle the linear relationship between process parameters and material properties efficiently. This research highlights the ability of machine learning algorithms in metal additive manufacturing, which reduces experimentation time, cost, and energy. By allowing real-time process control and optimized material properties, this research improves product quality and reduces total manufacturing costs in additive manufacturing. It facilitates improved manufacturing, employing machine learning to improve operational efficiency, customization, and quality control. This work could be extended to increase the prediction accuracy and enhance the optimization of process parameters by expanding the data set and applying many other supervised machine learning algorithms.