| The leading equipment for pellet production by rotary kiln pelletizing method is composed of a sintering machine,rotary kiln,and ring cooler,in which the sintering machine plays an important role.In the industrial field,if the grate bar of the sintering machine is missing,there will be a gap at the bottom of the trolley,which will affect the sintering quality,and even lead to the falling of the sinter and secondary accidents.If the trolley axle of the sintering machine fails,the trolley wheel will fall off,and even the sintering machine system will be shut down.At present,the detection of defective parts of the sintering machine trolley still mainly adopts manual inspection and maintenance,which is inefficient and lags behind.Therefore,it is of great significance to realize the automatic detection of defective parts of the sintering machine trolley.This dissertation mainly studies the lack of grate bar and axle failure of the sintering machine trolley.The main work of this dissertation includes:Research on the lack of grate bar of the sintering machine trolley based on the YOLOv5 network model.Firstly,the grate missing data set is made.Based on YOLOv5,the grate missing characteristics of the sintering machine trolley are learned in the environment of the Pytorch deep learning framework.Through training,the m AP value of the model for detecting the missing grate bar can reach 99.5 %,and the weight after training can detect the missing grate bar data image.Finally,a scheme for detecting the grate bar from the left and right angles at the sintering machine head is proposed.Research on the axle fault of the sintering machine trolley based on the characteristics of the wheel swing.Firstly,the video image of the wheel running of the sintering machine trolley is obtained,and the frame is intercepted,then the difference operation is carried out on the two adjacent frames,then the noise points are removed by using the image processing algorithm,and finally,the interesting area of the wheel of the two adjacent frames is located and obtained in combination with the YOLOv5 target detection algorithm,and mapped into the binary image obtained after image processing.In the binary image,the new region of interest corresponding to the wheel is obtained by obtaining the rules,and the value of the prime phase point corresponding to the new region of interest of the wheel is compared with the predetermined threshold to determine whether the wheel of the sintering machine trolley swings,so as to determine whether the axle of the sintering machine trolley is faulty.Finally,based on Py Qt,the interfaces of the above-mentioned sintering machine trolley defect parts detection system are constructed respectively,and the corresponding detection algorithms are called to combine them.Select the corresponding function reading module through the visual interface to detect the loss of grate bar and axle fault of the sintering machine trolley,and display them on the window interface in real-time.Finally,the detection results are stored in the My SQL database for easy reference and management.This scheme can realize the automatic detection of the defects of sintering machine trolley parts,and provides a new scheme for the establishment of an intelligent chemical plant. |
