In order to improve the effectiveness of evaluating the state of transformer winding

it is necessary to solve the problem of large dispersion of output data flow in the process of online monitoring the transformer short-circuit reactance in engineering applications. Firstly

we analyzed the existing offline short-circuit reactance algorithm and online short-circuit reactance algorithm

and established an uncertainty mathematical model of the short-circuit reactance input characteristic. Thereby

we summarized the effects of offline measurement and online measurement on the output short-circuit reactance under different combinations of transformer accuracy in the measurement system

and built a test platform in the laboratory for actual measurement verification. The research results show that the accuracy of the voltage transformer is a key factor that affects the dispersion of the short-circuit reactance in online monitoring

and the accuracy of the current transformer has a relatively small effect. The boundary constraint formula of the input current signal can effectively reduce the dispersion of the output data of the short-circuit reactance. This method is applied to online monitoring of short-circuit reactance

which can effectively improve the reliability and practicability of transformer winding deformation evaluation.