In order to deeply study the failure mechanism of sealed relay contact

it is necessary to investigate the correlation analysis between corrosion topography and type. The contact DEAX SEM diagram of sealed relay was selected as the analysis sample. Based on the analysis of possible types of corrosion reaction

it extracted irregular contours on the extraction distribution of corresponding elements

thereby eliminating the non-corrosive part and restoring the 3D shape of the corrosion area. The eigenvalues of the corrosion morphology were analyzed quantitatively and matched the possible corrosion type. Finally

BP neural network algorithm was used to verify the relevance. The results showed that the percentage of Ag and Cl atoms in the contact of sealed relay was approximately 1:1

so it was highly possible to produce AgCl compounds in certain areas. At the same time

the mapping relationship between contact topography and corrosion type was determined. It is concluded that the corrosion topography characteristics of each reaction still exist under the combined action of various corrosion reactions

and this paper provides a method to explore the relationship between corrosion types and corrosion morphology characteristics.