Abstract
In the context of the digitalization of the energy industry, the task of increasing the efficiency of methods for assessing the remaining life of power transformers is becoming particularly relevant, since the aging of their insulation is a key factor limiting the service life of equipment. The aim of the work is to develop and test a software solution that provides the calculation of the coefficient of the aging rate of transformer oil-filled equipment insulation based on an algorithmic approach that integrates classical and refined methods of modeling thermal effects in the conditions of the Arctic region. The study uses the traditional Van 't Hoff-Arrhenius and Montsinger dependencies, and also suggests two clarifying methods: a modified one (taking into account the exponential characteristics of changes in six-degree intervals) and the method of averaged intermediates (linearly interpolated). These methods were combined into a complex algorithm implemented as a software package, which made it possible to automate calculations based on telemetry data of the temperature of the most heated point of the winding. The practical implementation of the calculation was performed on an array of data recovered from the transformer monitoring system by processing graphical information with neural networks. The results showed reproducibility of calculations and confirmation of the possibility of using the developed complex in real-world operating conditions. The main conclusions are that the developed algorithm allows flexible consideration of temperature conditions, reduces labor costs for processing diagnostic data and increases the validity of forecasts of the remaining resource. This opens up prospects for its integration into equipment monitoring and condition-based maintenance systems.
Keywords: transformer, Montsinger equation, Van't-Hoff-Arrhenius equation, aging of insulation, software package