A GD-SVM Model for Calculating Oil-Immersed Transformer Hot Spot Temperature

With the rapid development of power grid, the equipment of power grid is more and more intensive, and the requirement of reliability is higher and higher. Oil-immersed transformer is one of the most critical equipment in power system. Its operation reliability is directly related to the safety and stability of the whole power grid. When transformer fault occurs, even under the same external operating environment, the hot spot temperature is different from the normal hot spot temperature. Generally, when the transformer overheat fault occurs, the hot spot temperature is usually higher than the normal hot spot temperature, and different fault modes will also cause different temperature variations of the transformer hot spot. By analyzing the variation law of the transformer hot spot temperature under different faults, it is helpful to evaluate and predict the load capacity of the transformer in real time and ensure the safety operation of the transformer. Hot spot temperature is a factor of measuring transformer performance, and it is also the most important limiting factor of transformer load capacity. It is also related to transformer safety and reliability, service life and manufacturing cost. Therefore, accurate calculation and prediction of the hot spot temperature has great significance for reasonable using transformer. Therefore, a support vector machine thermal calculation model based on gradient descent method is proposed in this paper. The validity and feasibility of the model are verified by the comparison of calculation and experiment results.