The article is devoted to the development of a software product that analyzes the concentration of gases (H2; CO; C2H4; C2H2) measured in transformer oil and set as a function of time. Product development consists of two stages. The first step is the task of classifying transformers. The second stage is predicting reliable operation before a possible failure based on the concentrations of dissolved gases in the transformer oil set as the initial data. During the development of the software product, the frameworks for creating web applications Flask and ReactJS were used, the models were trained on the frameworks H2O.ai, LightGBM, Sklearn.

Keywords: machine learning, regression, power transformers, gas concentration, web-service