Detection Method For Surface Crack Defects Of Turbine Blades In Aircraft Engines

Turbine blades,as key components of aviation engines,work under high temperature,high pressure,and harsh environmental conditions.After long-term service,they are prone to defects such as cracks,pitting,burns,and scratches,which further lead to blade failure and engine failure.Therefore,both domestic and foreign engines require regular maintenance of the blades during service.Although traditional manual visual inspection and commonly used non-destructive testing methods have certain detection effects on specific blade defects,they have certain limitations in terms of detection range,accuracy,and efficiency,and cannot be widely promoted.In order to realize automatic and intelligent detection of turbine blade surface crack defects qualitatively and quantitatively,this paper studies the key technologies of turbine blade surface crack defect detection,and proposes a detection method based on the combination of depth learning surface defect recognition technology and line structured light 3D measurement technology.The main research content of this article is as follows:(1)Based on the analysis of the characteristics and causes of crack defects in turbine blades,a detection scheme for surface crack defects in turbine blades was constructed.Based on the developed detection plan,the system hardware structure was designed,and the selection of important components such as industrial cameras,lenses,and line lasers in the detection system was completed.(2)Produce a dataset and conduct experiments on crack defects in turbine blades to verify the feasibility of YOLOv5 algorithm for automatic identification of crack defects.A method for capturing clear images of blade crack defects based on adaptive adjustment of blade posture and autofocus technology has been proposed.Combining digital image processing technology and automatic threshold segmentation algorithm to extract pixel regions of blade crack defects.(3)Analyze the mathematical model of the detection system and complete its parameter calibration experiment.Conduct experimental comparisons on commonly used light stripe center extraction algorithms,and determine the light stripe center extraction method in this paper based on detection requirements.A mapping method from crack pixel regions to 3D point cloud data was proposed to extract geometric features of crack defects.Finally,a method for measuring the geometric features of crack defects was studied.(4)Developed a software system for detecting surface crack defects on turbine blades.Apply the software system to the established experimental platform for system accuracy evaluation experiments,blade contour point cloud scanning experiments,and blade surface crack defect detection experiments.The experimental results indicate that the method studied in this paper can accurately identify and measure crack defects on turbine blades,and significantly improve the detection efficiency and accuracy of surface crack defects on turbine blades.