The converter transformer is affected by harmonic excitation in actual operation

and its loss and temperature rise are significantly higher than those of the fundamental rated condition

which poses a serious challenge to the long-term operation stability of the equipment. Firstly

based on the actual structural parameters of the converter transformer

a three-dimensional simulation model is constructed to obtain the magnetic flux density distribution and loss characteristics of the transformer under fundamental and harmonic conditions. The thermal-flow coupling simulation model is built to reveal the distribution law of temperature field and flow field. The results show that the transformer loss and winding hot-spot temperature increase significantly after considering harmonics. Secondly

the loss suppression effect is studied by adding electrical shielding and magnetic shielding devices on the wall of the fuel tank

and the magnetic shielding suppression effect is more prominent. On this basis

the influences of the length

thickness and height of the shielding device on the loss of the fuel tank are analyzed. The response surface method combined with the satisfaction function is used to construct a multi-objective optimization model

and the parameters are globally optimized. After optimization

the total loss of the fuel tank wall and the winding hot-spot temperature are significantly reduced. Finally

based on the optimized transformer tank wall

the correlation between the measuring points in each region and the hot-spot temperature of the winding is calculated

and the GA-BP neural network model is introduced to invert the points with high correlation. The results show that the temperature inversion of different combinations of high correlation measuring points are carried out

and the inversion accuracy of the best combination of measuring points is more than 0.99. This study provides a reliable theoretical support for the selection of monitoring points on the wall of transformer tank.