A Comparison Study of Distribution‐Free Multivariate SPC Methods for Multimode Data

Abstract

The data-rich environments of industrial applications lead to large amounts of correlated quality characteristics that are monitored using Multivariate Statistical Process Control (MSPC) tools. These variables usually represent heterogeneous quantities that originate from one or multiple sensors and are acquired with different sampling parameters. In this framework, any assumptions relative to the underlying statistical distribution may not be appropriate, and conventional MSPC methods may deliver unacceptable performances. In addition, in many practical applications, the process switches from one operating mode to a different one, leading to a stream of multimode data.Various nonparametric approaches have been proposed for the design of multivariate control charts, but the monitoring of multimode processes remains a challenge for most of them. In this study, we investigate the use of distribution-free MSPC methods based on statistical learning tools. In this work, we compared the kernel distance-based control chart (K-chart) based on a one-class-classification variant of support vector machines and a fuzzy neural network method based on the adaptive resonance theory. The performances of the two methods were evaluated using both Monte Carlo simulations and real industrial data. The simulated scenarios include different types of out-of-control conditions to highlight the advantages and disadvantages of the two methods. Real data acquired during a roll grinding process provide a framework for the assessment of the practical applicability of these methods in multimode industrial applications.

On the Selection of the Bandwidth Parameter for the k‐Chart

“…The design of the k ‐chart recommended did not use the boundary found by SVDD, but the (1 − α ) t h quantile of the bootstraped kernel distances using the optimum s value. Grasso et al . used this design of the k ‐chart in comparison with fuzzy neural networks on a study of multimode process data.…”

“…used this design of the k ‐chart in comparison with fuzzy neural networks on a study of multimode process data. It should be noted that Gani et al ., Gani and Limam, and Grasso et al . assumed an in‐control reference sample and compared Phase II performance based on ARL.…”

“…Sun and Tsung noticed the similarity between SVDD and statistical process control and adapted SVDD as a control chart known as the k ‐chart. There have been several case studies published using the k ‐chart. Subsequently, there is a wide range of guidance on the implementation of this chart in practice, primarily for Phase II applications where the goal is to monitor a process for changes.…”

A phase I multi-modelling approach for profile monitoring of signal data

“…Nevertheless, they make different uses of the information included in the database. The global modelling approach consists of designing a single control chart that is globally able to monitor the process in every known state: this means that the entire database, without any distinction between different modes, is used to design the chart (Grasso et al, 2015). The multi-modelling approach, instead, consists of designing one control chart for each IC mode, such that only the information related to the IC state that matches the current observations is used for the end-use monitoring phase.…”

“…A multimode process is a process that switches from one IC operating mode to a different one, producing a stream of data from different IC distributions (or waveforms). SPC of multimode processes in discrete-part manufacturing was studied in a recent paper (Grasso et al, 2015), which focused on the comparison of nonparametric techniques where quality features of interest are simple variables and not profiles. Profile monitoring of multimode processes was recently proposed for geometric shape monitoring (Park and Shrivastava, 2014).…”