Detection of Anomalies Related to the Operation of the Profinet Network Through Feature Extraction and Classification

PROFINET networks are increasingly being applied in industrial environments. Due to its expansion, it is common for new diagnostic tools to emerge dedicated to protocol. However, these tools present functions that require advanced knowledge of the end user to use them. Being more specific, it would be of the most importance if, by means of network traffic analysis, it was possible to identify the network's operating status (normal operation or with some anomaly) automatically. In view of the facts, this article aims to propose a methodology for classifying events related to the operation of the PROFINET Network Through Feature Extraction and Classification with Artificial Neural Networks. All the data are real and they were collected in an industrial plant.