Scholarlink

<p>In recent years, the industrial sector has increasingly embraced machine learning and data analysis. The high volume and complexity of production data require efficient data preparation processes to standardize and prepare it for analysis. This work illustrates some of the issues that may arise in different industries, using cold rolling in an aluminum factory as an example. Cold rolling, a vital process in aluminum strip manufacturing, stands to benefit significantly from these technologies. To optimize processes and improve product quality, having strip- and length-related process data is essential, particularly regarding position-dependent information. This paper outlines the implementation of a time/length transformation and its application to continuously recorded industrial production data in a continuous dimension changing environment such as rolling. Furthermore, it provides an approach to deal with strip splitting, scrap cutting, and multiple passes on a single product. In detail, an efficient, iterative method to process batches of hierarchical data using Apache Spark is introduced. The paper concludes with a brief overview of future research directions and potential applications in data analysis.&#xa0;</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

Insights into a cold rolling mill: a survey through length-related process data, strip splitting, and multiple passes

  • 期刊论文

  • 发表日期：

    2025年5月13日

查看全文
  • Alexander Haidenthaler,
  • Manuela Schreyer,
  • Patrick Pfeiffer,
  • Werner Fragner

摘要

In recent years, the industrial sector has increasingly embraced machine learning and data analysis. The high volume and complexity of production data require efficient data preparation processes to standardize and prepare it for analysis. This work illustrates some of the issues that may arise in different industries, using cold rolling in an aluminum factory as an example. Cold rolling, a vital process in aluminum strip manufacturing, stands to benefit significantly from these technologies. To optimize processes and improve product quality, having strip- and length-related process data is essential, particularly regarding position-dependent information. This paper outlines the implementation of a time/length transformation and its application to continuously recorded industrial production data in a continuous dimension changing environment such as rolling. Furthermore, it provides an approach to deal with strip splitting, scrap cutting, and multiple passes on a single product. In detail, an efficient, iterative method to process batches of hierarchical data using Apache Spark is introduced. The paper concludes with a brief overview of future research directions and potential applications in data analysis.