The geometry dependence of the Jr curve for circumferential growth of through-wall cracks in cylindrical pipes subject to bending loads

Proceeding from the experimental observation that circumferential crack growth in circular cylindrical pipes subject to bending loads is associated with a constant crack tip opening angle, the J_R curve is predicted in terms of the pipe and crack parameters. The predictions compare favorably with experimental results for 4 in (˜10cm) diameter pipes containing cracks with different sizes.     

The instability of circumferential crack growth in pipes—II. A pipe with freely supported ends

This paper addresses the question of the instability of growth of a through-wall circumferential crack in a pipe that is freely supported at its ends, the pipe being subject to bending deformation caused by a displacement applied to some part of the system. The effects of both crack position and displacement application point position on the instability criterion are assessed. The first important conclusion is that the tendency toward instability is not a maximum when the displacement application point and crack location coincide, a result which complements the results obtained from the author's earlier analysis for a pipe that is built in at both ends. The second conclusion is that as far as instability is concerned, the uniform bending condition considered by Tada, Paris and Gamble is, in some cases, less severe than the loading conditions considered in this paper; this conclusion does not accord with their belief.     

Effect of residual stresses on the linear-elastic KI-T field for circumferential surface flaws in pipes

This paper describes the effect of four realistic residual stress patterns on the crack-front constraints for circumferential surface flaws in the wall of a pipe, characterized by the linear-elastic K_I-T stresses. The numerical representation of the residual stresses combines a thermal-induced eigenstrain field with a local mechanical traction applied over the interface between the heat affected zone and the remaining pipe. The verification study confirms the accuracy of this approach based on the analytical K_I solution for a 2-D crack under prescribed residual stresses. This study demonstrates low crack-front constraints near the deepest crack front for three typical residual stresses.     

The effect of nonuniform bending on the stability of circumferential growth of a through-wall crack in a pipe

The Tada, Paris and Gamble analysis of the stability of a circumferential through-wall crack in a circular cylindrical pipe is extended, by incorporating nonuniform bending effects. As a result of analyzing specific situations, where a pipe has built-in ends, it is shown that nonuniform bending can, in some cases, have an adverse effect on the crack stability criterion. Care must therefore be taken to correctly account for nonuniform bending effects when assessing the integrity of piping systems.     

The instability of circumferential crack growth in pipes—I. A pipe with built-in ends

This article focuses on the instability of growth of a through-wall circumferential crack in a pipe with built-in ends, the pipe being subject to bending deformation caused by a displacement applied to some part of the system. The effects of both crack position and displacement application point position on the instability criterion are assessed. An important conclusion is that the tendency towards instability is not necessarily a maximum when the displacement application point and crack location coincide.     

Assessment of the susceptibility to hydrogen embrittlement. II: Effect of specimen geometry on delayed failure time of 38 NiCrMo 4 steel

Delayed failure tests, in which suitable samples were cathodically charged with hydrogen (current density 8 mA·cm^?2) in 0.1N sulphuric acid, while under sustained tensile loads, were carried out on a low alloyed Ni-Cr steel (UNI 38 NiCrMo 4). The effect of the geometry on the delayed failure time was assessed using specimens with different diameters (2,3,4 mm) and different lengths (13, 23, 33 mm). The experimental results showed that the specimen geometry does not influence the material behaviour: only a minor effect on the threshold stress value was noted due to diameter variations.     

New ‘ηpl'' and ‘γ'' functions to evaluate J–R curve from cracked pipes and elbows. Part II: experimental and numerical validation

Experimental evaluation of J–R curve in a crack growth situation requires ‘η_pl' and ‘γ' functions that are specific to a cracked geometry and loading condition. In Part I [Engng. Fract. Mech., in press] of this paper, new η_pl and γ functions, which are not available in the literature, for pipe and elbow geometry with various crack configurations under different loading conditions have been derived. In this paper, some of these newly proposed η_pl and γ functions have been validated experimentally through comparison of crack initiation load and J–R curve. In few cases, numerical validation has also been provided by comparing the J-integral values calculated through η factor approach and finite element method.     

The adaptive da/dN method: a solution to the problem of processing crack growth data

The standard fatigue data‐processing procedure, published in ASTM E647, is not adapted to the use of modern crack length measurement techniques. Because the use of this standard is usually required by journals, the raw data are often reduced to only a few data points. This way valuable information is simply thrown away and mathematical errors are unintentionally made. More importantly, the fact that no satisfactory reduction method exists, has led to destandardization of the processing procedure. Therefore, a new standard processing method is desired. In this paper a new data‐processing method, referred to as the ‘adaptive da/dN method’, is proposed and discussed. This method is suitable for both optical and modern (electrical or automated) measurement techniques as well as modern (computer‐assisted or ‐controlled) processing techniques. The adaptive da/dN method is validated both by data generated with a certain amount of scatter as well as actual experimental data. It shows a more accurate behaviour than the ASTM standards for all data types.     

Dynamics of Polymeric Fluids: Part II The Molecular Theory of Die Swell: Correlation of Ultimate Die Swelling Effect to the Molecular Parameters and the Operational Variables

A general expression for the correlation of the simple shear (tan(?)) to the molecular parameters and the shear rate (γ) was deduced. It shows that the simple shear (tan(?)) may be resolved into free recoil (recoverable strain) and viscous heating (unrecoverable strain). The magnitudes of the simple shears for recoil (tan(?)E) and (tan(?)V) for viscous heating not only depended on the molecular parameters and the operational variables, but also on the exponential fractions of the recoverable (1-Wγ) and unrecoverable (W-γ) conformations for recoil and viscous heating. Therefore the magnitudes of the simple shears (tan(?)E) for recoil and (tan(?)V) for viscous heating are, respectively, expressed as the partition function to the (1-Wγ)th power and the partition function to the (Wγ)th power. Thus correlations of the total recoil and the ultimately recoverable strains to the molecular parameters [n', a,η0, GN0 and (1-Wγ)] and the operational variables (γ, (L/D) and tr) were deduced respectively, which show that at very different shear rates (0≤γ≤∞) the polymeric liquids may exhibit a very different viscoelastic behaviors. After introducing the uniform two-dimensional extension, the definition of swelling ratio and the ratio of L to D [De=(L/D)], two expressions for the ultimate die swelling effect and the ultimate extrudate swelling ratio BEVT5 to the molecular parameters [n', a,η0, GN0 and (1-Wγ)] and the operational variables (γ, (L/D) and tr) were obtained. The two correlation expressions were verified by the experimental data of high-density polyethylene (HDPE) which shows that the two correlation expressions can be used to predict the correlations of the ultimate extrudate swelling behaviors of polymeric liquids to the molecular parameters and the operational variables.     

Dynamics of Polymeric Fluids：Part Ⅱ The Molecular Theory of Die Swell： Correlation of Ultimate Die Swelling Effect to the Molecular Parameters and the Operational Variables

A general expression for the correlation of the simple shear (tan ) to the molecular parameters and the shear rate ( ) was deduced. It shows that the simple shear (tan ) may be resolved into free recoil (recoverable strain) and viscous heating (unrecoverable strain). The magnitudes of the simple shears for recoil (tan E) and (tan V) for viscous heating not only depended on the molecular parameters and the operational variables, but also on the exponential fractions of the recoverable (1- ) and unrecoverable ( ) conformations for recoil and viscous heating. Therefore the magnitudes of the simple shears (tan E) for recoil and (tan V) for viscous heating are, respectively, expressed as the partition function to the (1- )th power and the partition function to the ( )th power. Thus correlations of the total recoil and the ultimately recoverable strains to the molecular parameters [ , a, η0, GN0 and (1- )] and the operational variables ( ), (L=D) and tr) were deduced respectively, which show that at very different shear rates ( ) the polymeric liquids may exhibit a very different viscoelastic behaviors. After introducing the uniform two-dimensional extension, the definition of swelling ratio and the ratio of L to D [De=(L/D)], two expressions for the ultimate die swelling effect and the ultimate extrudate swelling ratio BEVT5 to the molecular parameters [ , a, η0, GN0 and (1- )] and the operational variables ( , (L/D) and tr) were obtained. The two correlation expressions were verified by the experimental data of high-density polyethylene (HDPE) which shows that the two correlation expressions can be used to predict the correlations of the ultimate extrudate swelling behaviors of polymeric liquids to the molecular parameters and the operational variables.