Many major transformer accidents arise from minor internal faults

and traditional transformer protection lacks sensitivity to early faults and cannot effectively identify them. A novel protection scheme utilizing harmonic signals of transformer leakage magnetic fields to detect early faults is proposed for this issue. Firstly

the harmonic mechanism of transformers is analyzed

and an analytical model considering the harmonic distribution of leakage magnetic fields in three-phase windings is established. Then

a method for measuring and extracting the third harmonic of leakage magnetic flux

which comprehensively utilizes the triangle method

short-time Fourier transform

and median filtering

is proposed. Fault feature quantities based on normal

transient

and steady-state information of harmonic signals are constructed

forming leakage magnetic field harmonic differential protection that quickly and sensitively identifies and distinguishes between inter-turn short circuits and winding waveforms. Finally

the effectiveness and superiority of the scheme are verified through simulations and dynamic modeling experiments

providing a new method for real-time detection of early faults in dry-type transformers.