Scholarlink

In Germany, the injection molding industry has to deal with several challenges, such as the shortage of skilled workers. Artificial intelligence (AI) can help to overcome these challenges. In this application, an AI model is used to make predictions about part quality during injection molding cycles based on process parameters such as cavity pressure. For a robust AI, a lot of data must be made available for training. However, experimental data acquisition with injection molding tests based on a test plan and subsequent measurement is very resource-intensive. As an alternative, the AI is therefore trained with simulation data, the collection of which is more resource-efficient. The process data, such as cavity pressure, can be mapped very well with the simulation software. However, the measured dimensions, which serve as quality data, are sometimes poorly simulated. However, this can be compensated for with various experimental reference cycles. This means that AI models can also be created with simulated data that are able to provide good predictions regarding the dimensions.

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

Quality Prediction in Injection Molding Using Simulation-Based AI Training

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Elena Fischer,
  • Sascha Magerle,
  • Lukas Helmstaedt,
  • Matthias Deckert,
  • Marius Pflueger

摘要

In Germany, the injection molding industry has to deal with several challenges, such as the shortage of skilled workers. Artificial intelligence (AI) can help to overcome these challenges. In this application, an AI model is used to make predictions about part quality during injection molding cycles based on process parameters such as cavity pressure. For a robust AI, a lot of data must be made available for training. However, experimental data acquisition with injection molding tests based on a test plan and subsequent measurement is very resource-intensive. As an alternative, the AI is therefore trained with simulation data, the collection of which is more resource-efficient. The process data, such as cavity pressure, can be mapped very well with the simulation software. However, the measured dimensions, which serve as quality data, are sometimes poorly simulated. However, this can be compensated for with various experimental reference cycles. This means that AI models can also be created with simulated data that are able to provide good predictions regarding the dimensions.