基于YOLO-v3的水轮发电机转子外圆结构件智能监测OA

The rotating parts of hydro-generator feature high speed and frequent changes in working conditions,so production accidents often occur as a result of loose bolts,abnormal deformation of outgoing lines,etc.At present,the personnel of power plants inspect and maintain important parts inside the generators mainly through manual patrol inspections.However,manual patrol inspections shall be carried out after shutdown,and the accuracy of pre-failure judgment and data measurement during patrol inspections vary from person to person.This paper uses an image recognition system for automatic detection and judgment of rotor outer circle structural parts,thus replacing conventional manual operation and maintenance.The system uses non-contact detection methods to eliminate the shutdown for identification and detection and avoids data errors caused by human factors.It can achieve extremely high measurement data consistency,measurement accuracy,efficiency,and reliability.