Application of a multivariate process control technique for set-up dominated low volume operations.

摘要

In traditional high-volume manufacturing applications, the timing of control adjustments to processes is based on parametric Statistical Process Control (SPC) methods, such as Shewhart X & R charts. In high-value, high-complexity and low-volume industries, where production runs are in the order of tens rather than thousands, traditional SPC approaches are not easily applicable. A manufactured component's complexity, with multiple critical features to monitor, increases the difficulty for a process operator to maintain all of them within their design tolerances. In response to this, this paper presents a framework of nonparametric SPC, called multivariate Set-Up Process Algorithm (mSUPA), for managing control adjustment when required. mSUPA uses a simple to interpret traffic light system for alerting process operators when an adjustment is required. mSUPA is underpinned by multivariate statistics and probability theory for validating a process set up. The case of mSUPA application to a real industry process is discussed.