DEVELOPMENT OF A PYTHON TOOL FOR ANALYZING FAILURE TIME DATA WITH APPLICATIONS IN MAINTENANCE ENGINEERING

Abstract

This work proposes the development of a Python application with a Streamlit interface, capable of performing basic reliability analyses in an accessible and interactive manner. The methodology used is characterized as applied and exploratory research, with the tool being built and validated through a computational experiment. The application allows the import of CSV or Excel files containing failure data, in addition to providing a sample dataset. Classic reliability engineering functions are calculated: probability density, cumulative failure, reliability, and failure rate. Curve fitting is performed using polynomial regression, with the coefficient of determination (R²) used as the model evaluation metric. The tool was validated with real data extracted from the studies by Aguiar (2019) and Viana et al. (2018), presenting results consistent with the reference works. However, the use of polynomial regression is recognized as a limitation, which, while useful for visualization, can compromise accuracy in more rigorous probabilistic analyses. Suggestions for future work include the inclusion of distributions such as Weibull and Exponential, as well as the incorporation of statistical goodness-of-fit tests.