Unsupervised mode detection in cyber-physical systems using variable order Markov models

Abstract

Sequential data generated from various sources in a multi-mode industrial production system provides valuable information on the current mode of the system and enables one to build a model for each individual operating mode. Using these models in a multi-mode system, one may distinguish modes of the system and, furthermore, detect whether the current mode is a (normal or faulty) mode known from historical data, or a new mode. In this work, we model each individual mode by a probabilistic suffix tree (PST) used to implement variable order Markov models (VOMMs) and propose a novel unsupervised PST matching algorithm that compares the tree models by a matching cost once they are constructed. The matching cost we define comprises of a subsequence dissimilarity cost and a probability cost. Our tree matching method enables to compare two PSTs in linear time by one concurrent top-down pass. We use this matching cost as a similarity measure for k-medoid clustering and cluster PSTs obtained from system modes according to their matching costs. The overall approach yields promising results for unsupervised identification of modes on data obtained from of a physical factory demonstrator. Notably we can distinguish modes on two levels of granularity, both corresponding to human expert labels, with a RAND score of up to 73 % compared to a baseline of at most 42 %.

Detection of operating mode changes, without a priori model and in uncertain environments

“…Some applications are to be noted in the industrial field (Hawkins et al, 2003;Paul, 1994), energy consumption (Touzani et al, 2019), turbines (Letzgus, 2020), manufacturing processes (Atashgar and Rafiee, 2020;Guo et al, 2019;Wang et al, 2014), and chemical process (Salvador et al, 2014). We can also note the recent development of this theme in current environmental fields (Awe and Adepoju, 2020; Burnett, 2019) and societal fields (Choe et al, 2016;Coughlin et al, 2021;Salmasnia et al, 2020;Su¨rmeli et al, 2017).…”

“…From a methodological point of view, the detection of regime change can be done either on the basis of models established beforehand Su¨rmeli et al (2017), each model reflecting each mode of operation, or directly from the data without having a priori model(s), the second situation being the most difficult to handle. Model-based methods can themselves be differentiated according to the nature of these models, whether they are deterministic or stochastic.…”

Generating Artificial Sensor Data for the Comparison of Unsupervised Machine Learning Methods

“…An example of a solution in which digital computing processes are analyzed in combination with analog physical processes is given in [ 12 ]. There are different ways of analyzing continuous signals, such as the continuous time Bayesian network (CTBN) [ 13 , 14 ] or hierarchical clustering [ 15 ]. Continuous signals can also be evaluated by working with time-dependent machines or neural networks [ 16 ].…”