Structural assessment of pipelines with corrosion defects has been studied since the late 1960s. From the principles of fracture mechanics, semi-emprical mathematical expressions have been developed for predicting burst pressure of corroded pipes. Subsequently, these expressions have been modified and calibrated based on the results from finite element analyses and laboratory burst tests. The main methods are ASME B31G, Modified B31G (version 0.85 dL), Effective Area and DNV RP-F101 for single and complex shaped defects. This thesis was done utilizing some laboratory tests from Produt 600536 Project. For this study, five specimens with 3.0 meters each were removed from Petrobras pipeline (Orbel I), during its rehabilitation. The pipe material was API 5L X46, 18 diameter and 0.25 wall thickness. These specimens had real internal corrosion with very long defect length, at the 6:00 o`clock position. The corrosion defects were mapped by manual ultrasound, at 20 mm intervals, and mecanized ultrasonic measurements CSCAN at 5 mm intervals. For each pipe, the yield strength and ultimate strength were determined by tension tests of 4 specimens, 2 removed from transverse and 2 longitudinal. Every pipe specimen was monitored by strain gages and pressurized up to the point of rupture. For each specimen different profiles were determined depending on the length defined for the defect and for the intervals of measurements. The burst pressure was predicted by the following methods: ASME B31G, Modified B31G (version 0.85 dL), Effective Area and DNV RP-F101 for single and complex defects. These predicted rupture pressures were compared with the real burst pressures. The results confirmed the conservatism of the ASME B31G method and demonstrated that the Effective Area and DNV RP-F101 for complex shaped defects methods can be considered good for evaluating longitudinal corrosion defects, considering only internal pressure load.