AD (Anomaly Detection)
Anomaly detection (AD) Represents an Industry 4.0 workflow for after-sales machine service where a specific set of computing actions need to be performed based on the real values of a set of sensors. The AD workflow starts with cleaning input data (f1), then executes a specific function (f2, f3 or f4) for the source sensor. Thereafter, f5 normalises the data in order to determine whether values are beyond a specific threshold to detect anomalies, and if so, which category the anomaly (e.g. critical, severe, light) belongs to. Based on the detected anomaly category, an appropriate set of functions will be invoked to deal with the anomaly (f6, f7 or f8) and to notify appropriate persons (f9-f17). Finally, anomalies are stored in a database (f18).
