Abstract Achieving production quality is a key issue faced in injection molding. Before mass production, ensuring good production quality is one of the crucial factors in injection molding. To achieve good quality, CAE technology is beneficial to assist us either to make the process window approach or to integrate with optimization strategies to improve the quality. However, there are still some questions or challenges existed. For example, for general injection molding, the difference between CAE simulation prediction and real experimental observation is very often encountered. But its mechanism of this difference is not fully understood yet. When design of experiments (DOE) procedure is performed using CAE simulation, the optimal parameters obtained from CAE prediction are expected to be forwarded into the real molding trial. However, there is no guarantee to get results with good accuracy based on those optimal parameters. In this study, we have proposed a feasible methodology to uncover the difference and its internal mechanism between CAE simulation prediction and real experimental observation. It also includes the method to reduce that difference by calibrating the machine performance. Specifically, a standard procedure to calibrate the real performance of the injection machine using CAE technology has been constructed. Moreover, to realize the integration of CAE and DOE optimization strategy, the quality difference between the virtual CAE-DOE and the physical DOE optimization has been investigated before doing the machine calibration. The result showed that the difference between the virtual CAE-DOE and the physical DOE is almost same as that of the original injection molding design. However, after the machine calibration, the quality difference between the virtual CAE-DOE and the physical DOE optimization is reduced by 67%. It is noted that the machine calibration effect is quite significant in injection molding process development.

中文翻译：

---

研究机器校准对注塑件优化实验设计 (DOE) 的影响

摘要 实现生产质量是注塑成型面临的一个关键问题。在批量生产之前，确保良好的生产质量是注塑成型的关键因素之一。为了获得良好的质量，CAE 技术有利于帮助我们制定工艺窗口方法或与优化策略相结合以提高质量。但是，仍然存在一些问题或挑战。例如，对于一般的注塑成型，经常会遇到CAE模拟预测与真实实验观察之间的差异。但其产生这种差异的机制尚未完全了解。当使用 CAE 模拟执行实验设计 (DOE) 程序时，预计从 CAE 预测中获得的最佳参数将被转发到实际成型试验中。然而，无法保证根据这些最佳参数获得具有良好准确度的结果。在这项研究中，我们提出了一种可行的方法来揭示 CAE 模拟预测与实际实验观察之间的差异及其内在机制。它还包括通过校准机器性能来减少这种差异的方法。具体而言，已经构建了使用 CAE 技术校准注射机真实性能的标准程序。此外，为了实现CAE和DOE优化策略的整合，在进行机器校准之前，已经研究了虚拟CAE-DOE和物理DOE优化之间的质量差异。结果表明，虚拟 CAE-DOE 和物理 DOE 之间的差异与原始注塑设计的差异几乎相同。但是，经过机器校准后，虚拟CAE-DOE和物理DOE优化的质量差异减少了67%。值得注意的是，机器校准效果在注塑成型工艺开发中非常显着。