Development of a machine learning based predictive maintenance algorithm for transformers in distribution substations
Abstract
Maintenance management is an important component of a well-functioning plant and it ensures that all equipment required for production is operating at 100 percent efficiency at all times. Maintenance of substation equipment is necessary to ensure maximum efficiency. Over 99 percent of the significant inputs in businesses today run with electric power. This means that an
unexpected power outage would disrupt the activities of the company. [1]. With inadequate maintenance being the most cause of transformer failures in substations, adoption of suitable maintenance strategies for plant equipment has become an essential decision [2]. The existing maintenance strategies currently employed in UMEME are not adequate to keep assets operational at all times. This affects reliability of power supply to consumers, and also results into loss of revenue. It has therefore become to introduce the concept of predictive maintenance which allows for safety compliance, preemptive actions and increased asset life. This helps in knowing what kind of faults are likely to occur hence prior investigations, maintenance scheduling adjustments and repairs are done before an asset fails. In this project, we utilized deep learning specifically Long Short-Term Memory algorithm to make the fault predictions. We were able to fore cast faults for the next one year. The model was able to do predictions with the mean absolute error of 2 and a root mean squared error of 1.414. These very low values for the errors make LSTMs very suitable for making predictions helpful for the fault analysis