MFL intelligent pigs are widely applied for pipeline inline inspection because of possibility of complete pipe and weld coverage. Such types of pipeline defects as pitting corrosion and transverse weld seams cracks can be detected and estimated only y MFL pigs. At the same time MFL remains an indirect method of defects estimation with a lot of influence factors, which obviously decrease estimation accuracy. Nevertheless using of a priori information of typical defects and sections particularities enables appreciable increasing of estimations reliability. Introduced subclasses of metal loss defects enable to increase accuracy of field inverse problem solution within limited training sample. At the same time it proves to be possible to apply achieved training results to the pipes with a wide variety of their diameters and wall thickness. Some other problem concerns evaluation of real corrosion agglomeration especially within automated inspection data processing.Keyworks:MFL,pipeline inspection
展开▼