ABSTRACT

In production and manufacturing processes, noise factors are often considered difficult or costly to observe. The emergence of advanced sensor technology has made it easier for some major equipment to obtain large amounts of online monitoring data during the production stage of a product. In this paper, a new Bayesian approach is proposed to extend offline RPD to online multi-response RPD by making full use of this additional information. As new observations of the noise factor are obtained gradually, the settings of the controllable factors are adjusted online to further reduce the influence of noise factor variations on production quality. This approach not only addresses the correlation among multiple responses but also considers the uncertainty of model parameters and the variability of noise factors. A case study and a simulation study demonstrate that the proposed approach is superior to existing methods.

摘要

在生产和制造过程中，噪声因素通常被认为难以观察或观察成本高昂。先进传感器技术的出现，使得一些主要设备在产品的生产阶段更容易获得大量的在线监测数据。在本文中，提出了一种新的贝叶斯方法，通过充分利用这些附加信息将离线 RPD 扩展到在线多响应 RPD。随着噪声因子的新观测值逐渐获得，在线调整可控因子的设置，以进一步降低噪声因子变化对生产质量的影响。这种方法不仅解决了多个响应之间的相关性，而且还考虑了模型参数的不确定性和噪声因素的可变性。