Research On Ultrasonic Phased Array Automatic Scanner Based On Machine Vision

Liquefied petroleum gas(LPG)is widely used in the chemical industry.As a large spherical steel pressure welding vessel for storing liquefied petroleum gas,the welding quality of LPG spherical tank is very important to production safety.Because the spherical tank is subjected to the high pressure of internal petroleum gas,the welding stress of tank wall is large,and it is corroded by H2 S and other gases in liquefied petroleum gas,resulting in weld defects,Seriously affect production safety.At present,aiming at the defect detection of the weld on the wall of large LPG spherical tank,it is mainly checked by manually carrying ultrasonic testing equipment.The disadvantages of this operation method are very obvious,such as long operation cycle,high labor intensity and potential safety hazards.Therefore,it is urgent to improve the existing weld inspection method.With the continuous development of industrial automation and intelligence,taking the wall climbing robot as the research object instead of manually carrying ultrasonic testing equipment,this paper proposes a design scheme of ultrasonic phased array automatic scanner based on machine vision.Aiming at the problem of automatic tracking and testing of spherical tank weld,the following research is carried out:Firstly,according to the actual requirements of ultrasonic phased array detection,the basic tasks of the automatic scanner are analyzed,the technical requirements of the automatic scanner are clarified,and the overall system scheme design of the automatic scanner is completed;At the same time,according to the mechanical structure design of the automatic scanner,the selection of the scanner image acquisition module and motion control module is completed;The software design of upper computer and lower computer of weld tracking system is completed.Secondly,aiming at the noise interference problem of spherical tank weld image,the noise types are analyzed,and a complete set of image processing flow is proposed.The optimal processing algorithm is determined through experimental comparison,and the weld feature is extracted from the weld image based on the algorithm flow to obtain the weld centerline.Thirdly,the industrial camera is calibrated by using the open CV library function and checkerboard calibration board,and the internal parameters of the industrial camera are obtained.According to the coordinate transformation principle,the weld centerline is transformed into the automatic scanner coordinate system,and the distance from the ultrasonic phased array probe to the weld is obtained.Finally,the automatic seam tracking system based on sliding mode control is studied.Based on the Simulink module of MATLAB platform,the simulation model of weld tracking system is built,and comparative experiments are carried out to verify that the sliding mode controller has good response ability,high tracking performance and strong anti-interference;The actual image acquisition and processing experiments and weld tracking experiments are carried out to determine that the system meets the requirements of ultrasonic phased array defect detection.