Novosibirsk State Technical University has developed predictive models that allow you to plan the repair of electrical equipment in accordance with the dynamics of defect development (repair according to the actual projected operating time for a dangerous defect).
"The actual situation at the facilities of the electric power industry of the oil industry is as follows: the diagnostic service has corporate regulations, according to which thermal imaging is periodically carried out, the presence or absence of defects is recorded. There is a terminal (initial) stage of defect development, and there is an emergency stage. The most dangerous defect in electrical equipment is overheating, which can lead to the disappearance of power supply and cause losses. Based on the information that the thermal imaging control showed: if there is a defect, what temperature is at this defect, which stage of development it belongs to (developed, initial or emergency), samples are created: for objects such as a transformer substation of a certain voltage class, for power line objects, etc. We have developed predictive models and a methodology for their application for planning equipment repairs according to technical condition. Based on the samples, using models, it is possible to predict how the defect develops when repairs are needed so that this defect does not develop to equipment failure and the enterprise does not have losses due to power supply interruption," said Vladimir Levin, Associate Professor of the Department of Automated Electric Power Systems at NSTU-NETI, Doctor of Technical Sciences.
According to him, there is a way to plan repairs with clearly fixed intervals (for example, every two years), there is another way: the distance from repair to repair is floating and depends on the equipment condition. Having predicted the time before the occurrence of a dangerous defect, it is possible to plan repairs in accordance with the dynamics of the development of defects (repairs based on the actual projected operating time for a dangerous defect).
"Operating time is the average time interval after which such a defect reappears. If we have learned to reliably predict this interval, then we will prepare for repairs, anticipating the moment of the defect, and thus prevent an emergency shutdown of the facility," adds Vladimir Levin.
Predictive models are cost-effective; they allow you to save money on installing temperature sensors for detecting abnormal overheating. The plans include the development of a software and hardware complex that will generate statistics from equipment inspection protocols, make a forecast, calculate repair time, include it in plans and issue a ready—made repair plan for certain types of equipment as needed.