Transformer Overloading Capacity Research Based On Multi-parameters

Transformers overloading capacity has been calculated in order to achieve operation normally and high rate use of load capacity of transformer. Calculation method of top-oil temperature and hot-spot temperature in IEEE Standard and GB Standard were applied to compute overloading time of transformers. Moreover, the kinds of calculation methods, which were in GB Standard and IEEE Standard, were compared between all the transformers. And the factors of overloading were fully analyzed systematically. The accurate effects of each parameter were demonstrated. To help save calculation time of overloading capacity, a simplified method was presented. Results of the simplified method proved the method was accurate enough to use.Especially, the transformers overloading capacity was analyzed if the coolers in transformers were broken. Three methods were proposed to research this phenomenon. Firstly, a correction coefficient was proposed to add into the calculation method in GB Standard. The improved method was able to describe the temperature in transformers under the condition that the coolers were broken. The second way is an improved model of equivalent circuit represents the thermal heat flow equations for transformers. The parameter in the differential equations for equivalent circuit was calculated to reveal the coolers’ conditions. Thirdly, an improved BP neural network prediction model for hot-spot was proposed. The neural network could learn the temperature variation under different cooling conditions and predict hot-spot and top-oil temperature after a period of time of transformer overloading. All the three methods could be utilized to research the transformer overloading capacity if the coolers were broken.As the overloading situation could do harm to the transformers, a risk assessment of transformer overloading was under research. The risk contains the transformer life lost, the faults brought by overloading and the economy losses by faults. The temperature calculation method inside transformer in GB Standard is utilized to process the life loss of transformer brought by overloading. Furthermore, a fully new transformer risk assessment model was brought. The model was based on the deep learning which was the research focus. This model could fetch the characteristic in temperature change during transformer overloading situation and predict the top-oil and hot-spot temperature precisely. The predicted hot-spot temperatures were used to get the life loss of transformers after the overloading. Also, with the cost of the transformer investment, economic losses of the transformer life loss could be calculated. The final transformer’s overloading risk assessment combines life losses and the economic losses of the transformer faults. With the final risk assessment, the evaluation of this overloading could be got.