The Research Of Pipe Weld Inspection Robot Based γ-ray And Its Key Technology

With the development of the long-distance pipeline transportation, the weldingquality for pipeline directly influences the reliability of the transportation. As the newequipment for welding test, the research and application of pipeline inspection robothas become one of the hot topics recently both at home and abroad. At present thetheoretical research and practical application of pipe robot are still in the developmentstage and will be available in the future and the research results mostly are onlylimited to the experimental prototype, which should be done before industrializationthrough systematic research and pilot plant scale. Supported by GeneralAdministration of Quality Supervision, Inspection and Quarantine fund （Grant No.201210044-03）, the paper analyzes the existing problems that the pipeline robot is tooheavy, positioning is not accurate, the ability of adapting pipe is poor, andsystematically put forward a new type of mechanism of pipe robot, analyzing andstudying profoundly the detecting principle, the bulk structure and the localizationmethod.Considering the particularity of field operations, how to make robot effectivelyreduce dependence on external energy has become one of the key factors. According tothe facts that pipe robot adopts mostly X-ray to detect the weld by dragging cable, thepaper intends to make a contrastive analysis of the characteristics of different γ-raysources and proposes the test scheme to use Se75. The effective weld detection range isascertained in level AB through experiment. The experiments about the detection ofpipe welds show that the image reached AB-level in the range between4mm and30mm by Se75and meet the national standards and design requirements.The mechanical self adapting characteristics of robot are one of key performancesand the novel ontological structure is designed scientifically. The stretching leg structure whose energy and the adjusting mechanism of working platform are suppliedwith accumulator is created, which greatly improves the ability to adapt to pipeline andensures the reliability of the weld inspection. Based on the establishment of the motiondynamics model, the paper analyzes the motion state of robot. The simulation resultsreveal that the suitable diameter of the pipeline robot in bend pipe ranges from440mmto858mm. The self stability of the robot mass centroid exerts significant influence onthe test results, the passing capacity and the obstacle negotiation performance. Thepaper analyzes the causes of the eccentricity and vibration of robot and establishes themovement state formulas. The analysis shows that the motion law of the centroid is anellipse trajectory in bending pipe, and verifies the correctness of the mathematicalmodel by ADAMS.The paper analyzes the current positioning scheme of robot in pipe and problems,and puts forward the location plan by which the odometer method is mainly locatedand the visual recognition method is accurately positioned accurately. In view of thetraditional visual position based on gray level difference which is difficult to extractthe weld characteristics, the recognition method of image for welding seam based onthe color characteristics is applied, and the improved vector median filtering and theimage enhancement algorithm based on based on normalized improvement arepresented. The measures improve the color difference of multi-channels effectively,enhancing the image contrast and realizing the accurate positioning of the weld centerline. The analysis shows that error bound is1.84~3.76mm in the measurable pipelineand improves location accuracy greatly.