In the actual operation status of dry-type air-core reactor

the focus and difficulty of the fault diagnosis method is to reduce the false alarm rate and the missing alarm rate. About it

this paper proposes a weighted Na?ve Bayes state evaluation method for a dry-type air-core reactor. First

a simulation model of a reactor inter-turn short circuit fault is established using multi-physical field coupling

and the effectiveness of the simulation method is validated by constructing a reactor operation test platform. Second

simulation and analysis of reactor current amplitude

current harmonics

impedance angle

and hot spot temperature under multiple operating conditions are performed to obtain the reactor's normal and known fault sample sets. Finally

a weighted Na?ve Bayes state evaluation model is developed using multi-state feature quantities. The example demonstrates that this method is effective for reactor operation state recognition and classification because it has high classification accuracy and requires fewer training samples.