Predictive maintenance of electronics systems based on analysis with thermographic camera and fuzzy graphs

This paper presents two different ways of identifying possible defects for electronic components in order to achieve predictive maintenance. The analysis is performed on the components of a power station used in telecommunication. In the first part of the paper is performed a defect analysis based on fuzzy graphs. This leads to the components causing the most accidental stops. In the second part an analysis is carried out using the thermal imaging camera on the electronic components of the rectifier, because this components produce the most accidental stops. At the end of the paper we can see a number of possible defects that can occur if the studied components it's not working at the nominal parameters and how we can identify the respective components.