Research On Infrared Image Intelligent Diagnosis Method And System For Power Equipment

Infrared detection is an effective means of state detection of power equipment.The traditional manual infrared diagnosis is heavy in workload,low in efficiency and insufficient in intelligence level.On the other hand,diagnosis requires rich professional knowledge,and the experience and knowledge of existing experts cannot be fully utilized.In order to improve the efficiency of power equipment fault diagnosis,reduce the workload of personnel and improve the level of fault diagnosis,it is necessary to study new intelligent diagnosis methods.Firstly,the deep learning method is used to identify the equipment type and divide the structure of the infrared image of the power equipment.On this basis,the temperature information of the equipment structure is extracted to realize the intelligent diagnosis of the thermal fault of the power equipment.The method of type recognition and structure division of infrared image of power equipment is studied.labelme software is used to make the data set.Firstly,Mask＿RCNN deep learning algorithm is used to identify the type and divide the structure of power equipment.Then,a two-layer network model based on Inceptio-V3 and Mask＿RCNN is proposed for complex background images.Inceptio-V3 model is used for power equipment type identification,and Mask＿RCNN model is used for device structure classification.The results show that the based on Inceptio-V3 and Mask＿RCNN overall m AP value of RCNN’s two-layer network model for the division of power equipment structure can reach 0.9072,which effectively improves the situation that the traditional Mask＿RCNN model has more false detection of similar structures and complex background images of power equipment.It provides a good basis for temperature extraction and thermal fault diagnosis in the structure area of power equipment.The method of extracting the temperature of each structure of equipment in infrared image is studied and the temperature criterion of equipment fault diagnosis is constructed.Based on the type identification and structure division of power equipment,the linear equation between gray value and temperature was established according to the numerical temperature information of infrared image of equipment and the gray value of temperature width strip,and the temperature information of equipment structure area was calculated.Then the operating state of each structure of equipment was judged according to the application specification of infrared diagnosis of live equipment,and the thermal fault diagnosis of power equipment was completed.The infrared fault diagnosis system of power equipment is designed by Qt Designer tool of Python.Then,a case based diagnosis method for power equipment based on deep learning was proposed.Firstly,establish a typical equipment fault case database,and then use the improved VGG-16 model to extract features from the case images to form a feature library.Finally,extract the corresponding case information from the case database based on the image feature matching results.And use Python’s Qt Designer tool to design a power equipment case diagnosis system.The results show that using deep learning methods to retrieve images greatly reduces computational complexity and the average retrieval time is only 0.255 seconds,making it suitable for image retrieval of large-scale data.In the thermal fault diagnosis method of electric power equipment,deep learning is used to replace manual work to complete part of the diagnosis,which greatly reduces the labor intensity of staff and improves the intelligent level of fault diagnosis of electric power equipment.Moreover,this method deals with infrared image recognition and structure division of electric power equipment separately from later state diagnosis,effectively avoiding the problem of insufficient fault samples in the traditional intelligent diagnosis method.It has good feasibility.The example-based diagnosis method of power equipment can make full use of the accumulated historical experience and knowledge of experts,solve the problem of the lack of follow-up treatment suggestions in the thermal fault diagnosis of power equipment,realize the auxiliary analysis and decision of fault causes and fault treatment methods,and has an important reference significance for on-site maintenance personnel.