Novelty detection for practical pattern recognition in condition monitoring of multivariate processes: a case study

Abstract

Multivariate statistical process control (MSPC) can be applied for condition monitoring (CM) purposes. MSPC is implemented using a variety of techniques including neural networks (NNs). In situations when the number of process attributes is sufficiently large (e.g. 10 or more) concerns can arise with respect to training of NNs for pattern recognition. A classification method known as novelty detection (ND) can provide an effective alternative to conventional NN solutions that suffer from the above problem. Despite its great potential, ND is still unknown to the broad community of manufacturing engineers. This paper successfully demonstrates the ability of ND, using Gaussian mixture models, to notify operators of an end-milling process of the presence of faulty tool conditions. A significant achievement is that ND is used to identify abnormal time-series patterns as opposed to individual vectors of multiple simultaneous measurements related to abnormal conditions. Such patterns are found in windowed streams of signals related to 10 different process features (with an effective problem dimension of 140). The paper also investigates some of the issues related to implementation of ND for pattern recognition in condition monitoring.

SVM and ANN Application to Multivariate Pattern Recognition Using Scatter Data

“…6) Repetir los pasos (4) y (5) hasta que la ecuación de vigilancia sea alcanzada. 7) Para un nuevo patrón realizar los pasos del (2) al (6).…”

“…Patrones cíclicos se observan como una serie de máximos y mínimos ocurridos en el proceso; pueden indicar la fluctuación del voltaje de la fuente de alimentación. En los últimos años, varios autores [3,[5][6][7][8][9][10][11][12][13][14][15][16], han propuesto el uso de Redes Neuronales Artificiales (RNA) en el reconocimiento de patrones en los gráficos de control con el fin de examinar patrones, mejorar la eficiencia en comparación con los métodos usuales y obtener un diagnóstico automático de los patrones.…”

Data Mining Models of High Dimensional Data Streams, and Contemporary Concept Drift Detection Methods: a Comprehensive Review

“…Moreover, a little research has done issues related to change detection of a multivariate data streams. The referred works [24], [25], [26] propose that Multivariate approaches are assumed as extracting examples from a multivariate process where the feature is not independent. This Multivariate change detection tries to frame the multivariate process through a particular function which can check the incorporation of new data that may be an example or group of examples into the model.…”

“…The clustering or multivariate distribution modeling estimates the streaming data distribution. For modeling, a multivariate process of novel uses a most common parametric like in the works of [25]. The contribution depicted in [46] estimating the kernel density to identify the drift in concept.…”

“…[10], [11] Field-of-medicine; [12], [13] Monitoring-and-control sector; [14], [15] Management-and-strategic-planning domain; [16], [17] Personal-assistance-field; [18], [19] Ubiquitous-environmental-information-domain. [2], [7], [22] Surveys [27] Network-intrusion-detection field Unsupervised [23] Univariate Outlier-detection [24], [25], [26] Multivariate By example models, can't react to the changes occurred in the correlated components considered [27], [28] Multivariate solo monitoring of components using ensemble scalar CDT, every single component of the data stream is inspected to detect concept drift in a multivariate data, can't react to the changes occurred in the correlated components considered [29], [30] Multivariate nonparametric density models [31], [32], [26] Multivariate 'Pure' detectors, outperform with a low volume of data Supervised [33] Machine learning Uses commonly selected sequences as Hidden Markov Models, limits to define simple patterns. [34], [35], [36], [37] Machine learning Uses online algorithm, performs low on high dimensional data [34] Machine learning Bayesian Online Change Point Detection BOCPD, performs low on high dimensional data [41] Machine learning Extension of BOCPD, detects gradual changes [38], [39] Machine learning Not responsible for a gradual change.…”