In the framework of the cooperative Dutch research project BROS II (Continued Research on the Behaviour of Cracks in Thickwalled Steel Constructions), Subprogram 6 [i], a study was made of slow stable crack growth during monotonic loading of precracked steel St.52-3 SENB-and CNT specimens.In order to obtain fracture resistance data (J-curves) from such K specimens, it is necessary that the amount of stable crack extension is known.The most commonly used procedure is the multi-specimen heat-tint method.After termination of the test, the unloaded specimen is heated for several hours at approximately 300 deg C, which causes marking of the cracked area due to oxydation.After breaking, the specimen at liquid nitrogen temperature (-196 deg C), the crack extension can be measured with a traveling microscope, using the nine-point averaging procedure as standardized by the ASTM [2].Through the above procedure, only one crack front is obtained per specimen; and in order to reduce the number of specimens needed to obtain sufficient fracture test toughness data, recorse must be made to single specimen test procedures, e.g., potential drop, unloading compliance, key curve method.When using a single specimen test procedure, the amount of crack extension is determined indirectly, and hence it would be desirable to mark the crack front several times during the loading of the specimen in order to obtain a number of actual crack extensions against which the accuracy of the single specimen test method can be checked.As this cannot be done by fatigue cycling (it can be expected to alter the near tip stress and strain fields as well as sharpen the actual crack tip), recourse can be made to crack front marking by dye penetrants (colourants).The investigation of numerous combinations of dye venetrants led to a combination of three dyes, the specifications of which are summarized in Table i. The dyes are listed in order of subsequent application.These dye penetrants and their solvents were seen to meet the following requirements in order to obtain interpretable "colour fronts": i. Capacity for penetration up to the very tip of the loaded crack.For each of the three penetrants listed in Table i, a separate SENB specimen was loaded up to a certain amount of crack extension. After injection and drying of the dye, the specimen w a s u n l o a d e d and the crack was extended further by fatigue cycling.From the halves of the broken specimens, it was found that for each of the applied dye penetrants, the crack had been entirely marked up to the subsequent fatigued area.2. The dyes should not be soluble in each other's solvent in order to prevent blurring of previous markings by partial redissolvement in a subsequent solvent.The first dye penetrant was sprayed on the fracture surfaces of two specimen halves (originating from earlier fracture mechanics tests, and used here in order to approach actual testing conditions).After measuring the average distance to a Int Journ of Fracture 27 (1985) R94 reference line (using the nine point averaging v...

No abstract

Recent work in elastic-plastic fracture mechanics has demonstrated that estimates for the J-integral and specimen load line displacement A can be obtained from a combination of a linear elastic and a fully plastic solution J = Jel(ae,P) + Jpl(a,P,n) A = Ael(ao,P) + Apl(a,P,n) (i) where P, n, a, and a stand for load, strain hardening exponent, crack length, and a plasti~ zone corrected crack length, respectively [i], [2]. In order to obtain meaningful estimates, the material's stressstrain curve must be approximated by the Ramberg-Osgood relation = + a(olo )n (2) ~I~o °IOo o where c , o , n, and a stand for a reference strain, a reference stress O the strain ~ardening exponent, and a proportionality coefficient, respectively. Fully plastic solutions for various geometries are tabulated for both plane stress and plane strain in an Elastic-Plastic Handbook [3]; while the elastic solutions can be obtained, e.g., from [4].This note will discuss two applications of this so-called estimation scheme viz. prediction of the load-load line displacement curve of a specimen from an available experimental J-R-curve, and the inverse procedure, i.e., determination of a J-R-curve from the load-load line displacement curve of a single specimen.Both applications will be discussed for single-edge-notched-bend (SENB) specimens of St.52-3 and compact tension specimens of A542 steel.Specimen dimensions and material data as well as the Ramberg-Osgood approximations of the respective stress-strain curves are given in Figures 1 and 2. The ~llowing experimental J-R lines and blunting lines were obtained for both specimens u~ing the ASTM standard [5]: SENB: J = 956 Aa ; J = 223 Aa + 175 (3) CT : J = 1240 Aa ; J = 111 Aa + 132 (4) Int Journ of Fracture 23 (1983) R92If values for J and ka according to (3) and (4) are used as input into the estimation scheme, corresponding values for P and k can be calculated from (i) for a given Ramberg-Osgood approximation and initial crack length a . o In Figure 3, the estimated behaviour of the SENB specimen for three different Ramberg-Osgood approximations is compared with the experimental curve.Also included are results of finite element calculations allowing for crack growth [6]. The reason for using different RambergOsgood approximations follows from Figure i: the material exhibits a large yield plateau in its stress-strain curve and hence cannot be easily modeled by a Ramberg-Osgood relation. Figure 3 shows that a higher degree of assumed strain hardening (lower values for n) results in a more conservative estimation.The differences between the three estimated curves become marginal, however, for higher values of &, beyond crack initiation when ligament stresses as computed by the finite element calculations are prevailingly in the range of 450 -550 N/mm 2. In this range, the Ramberg-Osgood approximations are seen to be very close to each other in Figure i.Similar results for the CT specimens are given in Figure 4. The finite element calculations, however, were performed with a stationary crack.Clo...

The unloading compliance technique for JR-curve testing is not well established for single-edge notched specimens loaded in three-point bending (SENB) to date. This paper investigates various effects hampering the successful application of the method of SENB specimens. It is shown that the specimen's compliance is significantly affected by rotation of the specimen halves during deformation and stable crack extension. By contrast, the effects of roller indentation, outer roller movements and friction, and of crack front curvature are relatively minor. Based on these findings, a general procedure is proposed for correcting the experimentally measured compliance of SENB specimens for the above effects. Crack extensions predicted for three steels from SENB unloading compliance experiments incorporating the proposed corrections are seen to correlate within ±10% with dye penetrant beachmarks made at several points during the tests.