Power Transformer Hot-spot Temperature Inversion Based on Gappy-POD

The reliability and safety of a power transformer have always been a priority in the power system. Monitoring the hot- spot temperature in power transformers in a noncontact, realtime manner is crucial to the prevention and control of thermal energy. To meet the challenges of temperature data collection, especially in the industrial environment where arranging multiple temperature measurement points in the transformers is difficult. This study introduces Gappy orthogonal decomposition (POD) reconstruction Algorithm. This can predict the internal hot spot temperature by monitoring the surface temperature of the equipment. Based on Gappy-POD, this paper presents a method to determine the monitoring points location by greedy selection algorithm. This algorithm presents a viable approach for noncontact monitoring of the hot spot temperature of power transformers. The results show that the reconstructed temperature of the winding hot spot is comparable with the simulated temperature. Compared with method which the monitoring points location are determined randomly, the method proposed in this paper show better performance in the hot-spot inversion precision.