Edge-Computing Diagnostic of signals on PLC with Data-Driven approach

In the new concept of Industry 4.0 one of the crucial aspects on which is based this revolution relates to the continuous monitoring and prediction of possible malfunctions and errors in the nominal behaviours of the industrial machinery. The introduction of emerging technologies such as Machine-to-Machine (M2M) communication or solutions based on Internet of Things (IoT) devices for the communication, provided the evolution of the old industrial world into a more intelligent manufacturing sector. In this context, diagnostics and prognostics of faults and their precursors has gained remarkable attention and many actual researches are being developed on the topic. Even though the industrial world is integrating more and more solutions to keep up with Industry 4.0 new trends, most of the proposed implementations do not consider standard industrial hardware and software available nowadays. This thesis focuses on developing solutions that could be directly implemented on industrial components by generating internal libraries in a PLC syntax that will work on the source of the data extraction. The arguments treated in this document starts with analysing the world introduced with the innovation of the smart factories, proceeding with the definition of the new requirements that such facilities must undergo, such as continuous monitoring and evaluation in real time. There will be then an analysis of the current architectures and general use of the PLC device. The second chapter will instead treat about the signal processing algorithms present in literature focusing on the Least Square procedure and its recursive forms. There will be then a chapter on the software implementations apported as function blocks on the PLC environment to encode an identification and diagnostic elaboration. The fourth chapter will analyse the range of application of the developed code with different systems, concluding the tests with an example of the diagnostic implementation.