Junction Temperature Prediction: A Multi-Featured Dataset Approach

The generation level from photovoltaic (PV) solar energy has grown significantly in recent years. A future trend will be using PV inverters to provide auxiliary services. However, this equipment is the most prone to failure in PV systems even before it. Thermal stress is the main trigger for aging failures. This work used the mission profile of the city of Teresina (Brazil) to perform thermal simulations on semiconductor switches through the PSIM's thermal module. No literature records were found using mission profile data for thermal analysis in this software. This paper presents a methodology for generating a dataset to predict the semiconductor's junction temperature
$T_{j}$
and its swings
$\Delta T_{j}$
. This proposal integrates operational, environmental, and component-specific features simulated through PSIM's thermal module. In this way, choosing components most suitable to withstand the expected thermal stresses and better thermal management can be allowed, assisting in developing adequate cooling strategies and service life prediction.