[128] Anomaly detection-based condition monitoring
M Kas1 and F Fomi-Wamba2
1University of West Bohemia, Czech Republic
2Framatome GmbH, Germany
The impact of an anomaly is domain-dependent. In a dataset of network activities, an anomaly can imply an intrusion attack. Other objectives of anomaly detection are industrial damage detection, data leak prevention, identifying security vulnerabilities or military surveillance. Anomalies are observations or a sequence of observations in which distribution deviates remarkably from the general distribution of the whole dataset. The large majority of the dataset consists of normal (healthy) data points. The anomalies form only a very small part of the dataset. Anomaly detection is the technique to find these observations and its methods are specific to the type of data. While there is a wide spectrum of anomaly detection approaches today, it becomes more and more difficult to keep track of all the techniques. As a matter of fact, it is not clear which of the three categories of detection methods, ie statistical approaches, machine learning approaches or deep learning approaches, is more appropriate to detect anomalies on time-series data, which are mainly used in industry. A typical industrial device has multi-dimensional characteristics. It is possible to measure voltage, current, active power, vibrations, rotational speed, temperature, pressure difference, etc on such a device. Early detection of anomalous behaviour of an industrial device can help reduce or prevent serious damage leading to significant financial losses. This paper is a summary of the methods used to detect anomalies in condition monitoring applications.
