In order to realize the on-line fault diagnosis of large scale PFN circuits for super-high level electromagnetic emission

a diagnosis method based on equivalent circuit and parameter identification is proposed.After the circuit stray parameters and component distribution parameters are taken into consideration

the circuit is equivalent to a specific electrical parameter circuit. A detailed mathematical model is established to complete the preliminary identification of electrical parameters through clever mathematical transformation and geometric application. The mathematical model is transformed into a multi-parameter object optimization problem

and the genetic algorithm is applied to achieve high precision parameter identification. Based on the actual model

the relationship between the deviation of electrical parameters and fault diagnosis is analyzed

and a fast fault diagnosis method is obtained. The Simulink simulation model is built in case of considering various kinds of interference. The simulation results show that the maximum discrimination error of inductors is 14.5%

and the maximum discrimination error of capacitance and resistance is 1%

fully meeting the fault diagnosis requirements. Based on an actual electromagnetic launch system

the proposed algorithm is applied to the PFN charging stage in the actual launch and fault diagnosis experiment is carried out. Experimental results show that the proposed algorithm can diagnose all 3 capacitor faults and 2 charge isolated diode short breakdown faults on line

which is useful in engineering application.