The ANSI/ASME B31G guideline has been useful to pipeline operators in assessing the integrity of corroded line pipe. Because large safety margins have had to be incorporated, the guidelines can be excessively conservative, which in turn can force costly repairs and replacements that may not actually be necessary. On the other hand, because the current guidelines consider only pressure loading and neglect bending and axial compression, they could give nonconservative failure predictions when combined loading exists. Therefore, a study was initiated to develop a theoretically sound methodology for assessing the integrity of corroded line pipe subjected to combined loading. A key step in the successful application of this methodology is the development of a sophisticated three-dimensional finite element procedure that can accurately simulate full-scale pipe tests under conditions of combined loading. This paper describes thirteen full-scale failure tests on artificially corroded pipes subjected to simultaneous internal pressure, bending, and longitudinal compression and presents a detailed account of the finite element analysis procedure that was developed to simulate these tests numerically. Additional finite element analyses that were conducted to investigate the effect of key parameters on failure, and to expand the corroded pipe failure database, are also discussed.

The inner wall of a nuclear pressure vessel can become embrittled during service due to irradiation effects. As a result, a small flaw could become critical during the thermal shock that would accompany a loss-of-coolant-accident. To assess the consequences of this possibility, elastodynamic computations were performed on the subsequent rapid unstable crack run-arrest events. One difficulty is the lack of reliable dynamic fracture toughness data as a function of crack speed and temperature. However, using plausible estimates, computational results were obtained for comparison with conventional quasi-static analyses. These indicate that the neglect of the dynamic effects is mildly anticonservative for this application.

A method of evaluating the J-integral for a circumferentially cracked pipe in bending is proposed. The method allows a J-resistance curve to be evaluated directly from the load-displacement record obtained in a pipe fracture experiment. This method also permits an analysis for fracture instability in a circumferential crack growth using a J-resistance curve and the tearing modulus parameter. The influence of the system compliance on fracture instability is discussed in conjunction with the latter application. The results suggest that a compliant piping system containing a crack can exhibit ductile fracture instability after some stable crack growth. The importance of using a J-resistance curve that is consistent with the type of constraint for a given application is emphasized.

The possibility of a pipe fracture emanating from a stress corrosion crack in the heat-affected zones of girth-welds in Type 304 stainless steel pipes was investigated. The J-resistance curve—tearing modulus parameter for the prediction of crack initiation, stable growth and fracture instability—was employed. To evaluate the analysis, a pipe fracture experiment was performed using a spring-loaded four-point bending system that simulated an 8.8-m (29-ft) long section of unsupported 102-mm- (4-in-) dia pipe. An initial through-wall crack of length equal to 104 mm (4.1 in.) was used. Fracture instability was predicted to occur between 15.2 and 22.1 mm (0.60 and 0.87 in.) of stable crack growth at each tip. In the actual experiment, the onset of fracture instability occurred beyond maximum load at an average stable crack growth of 11.7 to 19 mm (0.463 to 0.750 in.) at each tip.

A method of analyzing internal surface circumferential cracks in ductile reactor piping is presented. The method utilizes an alternate but equivalent definition of the J-integral based on nonlinear structural compliance. The analysis is valid for situations where the cross section containing the crack is fully yielded. Results are obtained for radial and circumferential crack growth for pipes subjected to bending. The stability of radial crack growth (wall breakthrough) is assessed using the J-integral-based Tearing Modulus approach. The analysis is shown to be in good agreement with experimental results on the stability of surface crack growth in Type 304 stainless steel pipes. Example quantitative results for fracture instabiltiy assessments for nuclear piping are presented to demonstrate the practical application of the approach.

Previous work has shown that a speed-independent dynamic fracture toughness property can be used in an elastodynamic analysis to describe crack initiation and unstable propagation under impact loading. In this paper, a further step is taken by extending the analysis from simple laboratory test specimens to treat more realistic crack-structure geometries. A circular cylinder with an initial part-through wall crack subjected to an impulsive loading on its inner surface is considered. The crack is in a radial-axial plane and has its length in the axial direction long enough that a state of plane strain exists at the center of the crack. Crack growth initiation and propagation through the wall is then calculated. It is found that, once initiated, crack propagation will continue until the crack penetrates the wall. Crack arrest within the wall does not appear to be possible under the conditions considered in this paper.

Previous work at Battelle-Columbus on the development of a theoretical model for unstable crack propagation and crack arrest in a pressurized pipeline is extended in this paper by including the effect of backfill. The approach being developed involves four essential aspects of crack propagation in pipelines. These four components of the problem are: 1 – a shell theory characterization of the dynamic deformation of a pipe with a plastic yield-hinge behind an axially propagating crack, 2 – a fluid-mechanics treatment of the axial variations in the gas pressure acting on the pipe walls, 3 – an energy-based dynamic fracture mechanics formulation for the crack-driving force, and 4 – measured values of the dynamic energy absorption rate for pipeline steels. Comparisons given in the paper show that the steady-state crack speeds predicted by the model are in reasonably good agreement with the crack speeds measured in full-scale tests, both with and without backfill. The analysis further reveals the existence of a maximum steady-state crack-driving force as a function of the basic mechanical properties of the pipe steel and the pipeline goemetry and operating conditions. Quantitative estimates of this quantity provided by the model offer a basis for comparison with the empirical crack-arrest design criteria for pipelines developed by AISI, the American Gas Association, the British Gas Council, and British Steel. These are also shown to be in substantial agreement with the predictions of the model developed in this paper.

In a previous paper, a simplified dynamic-shell theory representation was formulated for steady-state motion in a pipeline without backfill. The present work extends this model by (1) incorporating a gas dynamics treatment to determine the axial variation in the pressure exerted by the gas on the pipe walls, and (2) incorporating a plastic yield hinge behind the crack tip. Solutions to the governing dynamic equations are obtained for these conditions and used to calculate the steady-state dynamic energy release rate as a function of crack speed. In the single full-scale experiment in which an independent estimate of the dynamic fracture energy is available for a pipe without backfill, the model predicts a steady-state speed of 780 fps. This can be compared with measured speeds which ranged from 725 to 830 fps in the test. Because the calculated steady-state dynamic energy release rate exhibits a maximum, it is suggested that this approach may offer a basis for crack arrest design of pipelines.

A solution for an edge loaded, semi-infinite elastic solid on an elastic foundation is presented. This solution, arising in the development of a nonlinear atomic model for cleavage crack propagation, was obtained by the Fourier method. Example results are given for a uniform load on a finite interval and for a concentrated force, Fig. 1, in terms of the sine and cosine integral functions and their auxiliary forms.