Design of a monitoring tool for the health assessment of power converters

Power electronics and in particular power converters, have a pivotal role in multiple expanding fields like: Electric mobility, renewable energy integration, microgrids, and smart grid technologies. Reliability for power electronic converters is a significant concern since, compared with other electrical modules, are characterized by a higher failure rate, as reported in different field reviews. The consequences of a converter failure are not only limited to the downtime of the system and potential damage to downstream components. It can also pose a threat to the safety of the system. Those reasons made the lifetime management of power converters an increasingly studied field. This master's thesis aims to detect the ageing of a power converter by identifying the parameters that are associated with the degradation of a component. To achieve this goal a digital twin of the circuit was built. The output voltage was simulated and then compared with the measured output voltage of the converter. Minimizing the cost function built between the two characteristics provided an estimation of the system parameter. Initially, the method was tested on a simple RLC circuit where a benchmark was made to choose the most suitable optimizer for this application. Lastly, its efficacy was tested on a buck converter. The higher complexity of the system introduced other challenges, one of which was to process the output of the simulation in order to align and normalize the data. A qualitative analysis of the estimation of the failure indicator was then performed.