Abstract: Fire in power places is an important threat to production safety and personal and property safety,especially in highvoltage power places,where the voltage level is high,the current load is large,and the equipment is highly concentrated,resulting in a higher fire risk.Before the occurrence of electric fire,there are often symptoms such as abnormal temperature rise,thermal decomposition of insulation layer,and the generation of fire particles.Due to the different types of fire,the generation speed and concentration change of fire particles also reflect different characteristics.By simulating the occurrence of several typical fire risks in the cable corridor,the fire particles generated were detected,data analyzed and curve characteristics were studied,and a fire identification method based on the change characteristics of fire particles was porposed to distinguish fire types and provide support for fire prevention and control management in power places.