Since the late 1960s’ the Battelle Two-curve (BTC) model is the standard method applied in setting up design requirements with regard to the prevention of long-running ductile fracture in pipelines. It is a straightforward tool employing Charpy-V notch (CVN) toughness as key-measure for material resistance against crack propagation. On basis of pipe dimensions, material strength, and under consideration of decompression behavior of the transferred media, it enables to set up requirements for a minimum CVN toughness level to achieve crack arrest. Overall applicability of the BTC model is based on calibration of the underlying equations to a sound data-base, including both full-scale burst test results and small-scale laboratory testing data involving typical line-pipe grades at that period, i.e. up to grade X70 steels with below 100 J upper-shelf CVN toughness. Now over the last decades, mechanical behavior of line-pipe steels was improved significantly. Responding to market demands, higher grades were designed and also toughness levels were raised as outcome of R&D efforts within the steel industry. Unfortunately, stepping outside the original material data-base from BTC model calibration, this method did forfeit its reliability. At the beginning, mispredictions were mainly related to higher grade steels and elevated operating pressures. But more recent full-scale tests did reveal discrepancies in application of the BTC model also for so-called new vintage steels, i.e. grades actually being inside the original data base for model calibration but from current production routes.
With regard to applicability/reliability of BTC model based predictions for crack arrest, the origin of uncertainty has particularly been traced back to the involved material toughness measure. Nowadays, it is common sense that the CVN upper-shelf toughness value inadequately describes the resistance against running ductile fracture. More recent thoughts coherently argue towards closer involving stress-strain response and plastic deformation capacities of the material. On basis of results for grades X65, X80 and X100, the general relation between ductility and toughness is discussed. Finally, an elastic-plastic fracture mechanics related analytical approach is introduced which enables to quantify the resistance against ductile fracture propagation. The objective is to provide a reliable procedure for crack arrest prediction in line-pipe steels.
Two types of the crack arrest during full-scale pneumatic testing of main gas pipelines
