Scholarlink

<p>The production quality of high-pressure die casting (HPDC) products is a key indicator for evaluating the reliability of HPDC manufacturing systems. This relies on predictive techniques to estimate the production quality of HPDC products. However, the quality prediction of HPDC products faces challenges such as the complexity of the forming process and outdated or insufficient data collection. This article proposes a high-fidelity digital twin solution to overcome these difficulties. The designed HPDC process monitoring system can synchronize real-time production processes in virtual space and continuously update simulation data. At the same time, a multi-source data collection and management method has been introduced to minimize incomplete information. Additionally, a product quality prediction method using a stacked ensemble learning algorithm has been disclosed, which can quickly and reliably predict product quality under limited process parameters. Experimental results show that the stacking model achieves a coefficient of determination (<i>R</i><sup>2</sup>) of 0.925 and a root mean square error (RMSE) of 0.833, demonstrating its high accuracy in biscuit thickness prediction. The effectiveness of the proposed solution is verified through implementation in a real HPDC workshop environment.</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

Digital twin–based product quality prediction system for HPDC under partial information

  • 期刊论文

  • 发表日期：

    2025年9月23日

查看全文
  • Yu Du,
  • Xiaoxuan Wu,
  • Dong Liu,
  • Wenjie Chai,
  • Baoqi Xu,
  • Ming Cong,
  • Sunan Huang

摘要

The production quality of high-pressure die casting (HPDC) products is a key indicator for evaluating the reliability of HPDC manufacturing systems. This relies on predictive techniques to estimate the production quality of HPDC products. However, the quality prediction of HPDC products faces challenges such as the complexity of the forming process and outdated or insufficient data collection. This article proposes a high-fidelity digital twin solution to overcome these difficulties. The designed HPDC process monitoring system can synchronize real-time production processes in virtual space and continuously update simulation data. At the same time, a multi-source data collection and management method has been introduced to minimize incomplete information. Additionally, a product quality prediction method using a stacked ensemble learning algorithm has been disclosed, which can quickly and reliably predict product quality under limited process parameters. Experimental results show that the stacking model achieves a coefficient of determination (R2) of 0.925 and a root mean square error (RMSE) of 0.833, demonstrating its high accuracy in biscuit thickness prediction. The effectiveness of the proposed solution is verified through implementation in a real HPDC workshop environment.