Compact tension and J integral tests

The previous Connect article in this series, number 76, dealt entirely with the CTOD test and illustrated the use of a single edged notched bend (SENB) specimen. This test was developed at TWI as a cost effective method of determining fracture toughness in a metal that exhibits some degree of plasticity - plane stress conditions - before fracture, the analysis of the results being carried out using elastic plastic fracture mechanics (EPFM).

The failure mode is a function of the material properties, the rate of loading, the temperature and the material thickness. The lower the material toughness or temperature, the faster the loading rate or the thicker the specimen the more likely it is that brittle fracture will occur. With carbon manganese steels toughness is generally sufficiently high that it is difficult to achieve plane strain conditions except at low temperatures ie on the lower shelf, or with thick plate.

The SENB specimen that is used for the CTOD test can also be used in linear elastic fracture mechanics (LEFM), a situation where the failure mode is accompanied by little or no plastic deformation - plane strain conditions. Any displacement that occurs is essentially elastic, a situation that pertains when brittle fracture occurs. Both types of fracture, brittle or ductile, can therefore be characterised by the SENB test.

The compact tension (CT) specimen and the J integral test, the two test methods briefly described in this article, may also be used to characterise fracture behaviour by using the appropriate calculation techniques irrespective of the failure mode.

The compact tension specimen is, in some respects, similar to the SENB specimen in that it is a proportional specimen of full plate thickness containing a fatigue crack. The sides of the specimen are approximately twice the specimen thickness as illustrated in Fig.1. The specimen has a notch machined into one face in the area - weld, HAZ etc - to be tested and a fatigue crack is then grown from the tip of the machined notch to give a total 'crack' length approximately equivalent to the specimen thickness. The specimen is tested in tension with deformation measured by means of a clip gauge mounted across the mouth of the notch. Load and deformation are recorded and crack length is measured on the broken test piece. A decision may then be made as to the failure mode and the appropriate analysis tools then used to calculate toughness.

The compact tension test has the advantage compared with the SENB test that the specimen is more economical in material and this can be important when thick plates are to be tested. The specimen is, however, more expensive in machining costs and the method of loading tends to give lower toughness results than the SENB specimen. For this reason the CT test is favoured by the nuclear industry, where safety is crucial and lower bound results are preferred.

The J integral is a third method of determining toughness and is based on the amount of energy required to propagate a crack. Both CTOD and J can be measured on the same specimen by using two clip gauges, one to measure CTOD, the other to measure J. To determine J the specimen is loaded at successively higher loads and the displacement and crack length at each load is measured. The area beneath the load/displacement curve gives the amount of energy required for fracture propagation to occur. Analysis of the results enables a J factor to be calculated as a measure of fracture toughness.

Qualitative tests

All the above tests - the SENB test, the CT test or the J integral - enable critical defect sizes to be calculated and decisions made about fitness for service of a structure or the required level and sensitivity of NDE. They are therefore quantitative tests. There are however a number of qualitative tests that have been developed where the test regime attempts to simulate service conditions and the test gives a 'go - no go' result. Typical of this family of tests is the NDT or drop weight test.

This test was developed in the USA for the testing of naval steels where the temperature at which the failure mode of a plate subjected to impact loading would change from ductile to brittle behaviour. The sample size is standardised with three sizes depending upon plate thickness. A 20mm thick plate, for instance, would require a specimen measuring 125mm x 50mm and full plate thickness.

A brittle crack starter weld is made along one side of the sample, often using a hard facing electrode. This weld deposit is notched and the sample laid, notch down, across two supports. A standard weight hammer is then dropped onto the sample and this initiates a crack in the hard facing as illustrated in Fig.2 which shows weld metal being tested. The test is carried out on a number of samples at progressively lower temperatures until the test piece breaks. This temperature is known as the 'nil-ductility temperature' (NDT). A further two tests are carried out at a temperature 50°C above the NDT to demonstrate that complete failure does not occur and that a crack will arrest provided the test sample is above NDT.

The test may be used to characterise and compare weld metals and plates or as an acceptance criterion by specifying the NDT temperature.

There are a number of other tests available some of which may be called up in application standards or in contract specifications. One such test is the dynamic tear test - a test that is, in principle, similar to the Charpy test. The test piece, however, is 15.8mm thick, 38mm deep and 180mm in length with a notch pressed into the edge instead of being machined. The test results are absorbed energy, temperature and, if requested, the amount of crystallinity on the fracture face. One other test worth mentioning is the Wells wide plate test, developed here at TWI in the early 1960's and illustrated in Fig.3. This uses a full size plate with a machined notch and/or fatigue crack. Service conditions, residual stresses etc are simulated as closely as possible. Large scale and very expensive tests such as this have been almost entirely replaced by the more cost effective SENB and compact tension specimen methods.

Relevant specifications are given in the Table.

Related specifications

This article was written by Gene Mathers.