Návrh a optimalizace inteligentních senzorů pro potřeby Průmyslu 4.0 a SMART technologií

Abstract

The thesis deals with the design of an intelligent system for online monitoring and quality control of production at robotic welding stations. This system uses advanced signal processing techniques, machine learning and artificial intelligence methods to determine non-standard acoustic emission waveforms of the welding process. The first iteration of the implementation of the proposed system on robotic workplaces in industry (Brose CZ spol. s.r.o.) was carried out, in which the knowledge base is extended using a combination of supervised and unsupervised machine learning.. The work was developed within the project "Platform for Industry 4.0 and robotics-oriented research in the Ostrava agglomeration". The conducted study clearly confirmed the hypotheses of using acoustic emissions for online monitoring and quality control of production on MAG robotic welding stations. The work produced a unique dataset of multi-channel acoustic recordings, which will be further expanded and offered to the wider public community. A unique multi-channel measurement system based on virtual instrumentation was developed for the work. The activities within this PhD thesis accelerated the collaboration between FEI, VSB and Brose CZ spol. s.r.o. Partial results of this thesis will be used in further research, projects and publications.