The operating mechanism of high voltage circuit breaker is the key part to ensure reliable operation of circuit breaker

and accurate judgement of the operating mechanism status is of great importance. A non-intrusive multi -information detection method based on image

vibration and sound is proposed

the multiple correspondence analysis of the extracted image feature vectors (opening time

closed time

opening speed

closed speed

opening stroke

closed stroke)

vibration feature vectors (feature time points

equal energy 5-subsection energy entropy)

sound feature vectors (feature time points

equal energy 5-subsection energy entropy) and status of the operating mechanism (normal

pedestal looseness

shaft jam

performance degradation of the buffer and insufficient lubrication of the lever) are analyzed

and the relationship between the state of the mechanism and the corresponding eigenvectors has been found. A hierarchical diagnosis strategy based on the relationship between the status of the mechanism and the corresponding eigenvectors is proposed. Throughout fault simulation and diagnosis of the experimental laboratory

the correct rate of the joint feature diagnosis of the fault of pedestal looseness

shaft jam and buffer performance degradation

compared to single signal diagnosis

is improved to 90%

80% and 90% respectively. The the accurate rate of the joint feature diagnosis of insufficient lubrication of lever

compared to the single signal diagnosis

is not improved obviously. The general effectiveness of the method is proved through the actual measurement at site and disassembly analysis of the equipment.