1 Dr Afshin Motarjemi ( email@example.com) is a senior integrity engineer, undertaking a range of FFS assessments and on-site and laboratory investigations.
2 Julian Speck ( firstname.lastname@example.org) is TWI's structural integrity department manager, responsible for FFS and RBI projects and development activities.
Paper published in Inspectioneering Journal, March 2005.
TWI's Members recently requested an evaluation of the instrumented indentation technique (IIT). IIT is claimed to be capable of determining tensile properties from a local indentation. TWI subsequently investigated the capability, usefulness and limitations of the IIT and some of the findings are reported here. The leading manufacturer of IIT equipment is FRONTICS in Korea. In this investigation, FRONTICS kindly offered to collaborate with TWI on the project.
IIT is different from a conventional hardness test. For the purposes of determining tensile properties, a spherical indenter is driven at approximately constant speed into the surface of the material to be measured, Fig.1. The load required to reach the penetration depth is recorded. At specific depths, the indenter is unloaded and withdrawn slightly and then reloaded to a greater depth in a series of increments until the test is complete.
Indentation load-depth curves are obtained from this measurement procedure for each repeated loading and unloading cycle, Fig.2. There are three stages of deformation in the indentation process: elastic, elastic-plastic and fully plastic. The load/unload data points are defined by measured indentation parameters, eg. contact depth, indentershape, and the morphology of the deformed surface (ie. elastic deflection and the plastic material pile-up around the contacting indenter). A true stress-true strain curve is then derived by an algorithm that fits a curve to the load/unload data points, given knowledge of the generic type of material.
IJ readers will recall that if the true stress, based on the actual cross-sectional area of a tensile test specimen, is used in a conventional tensile test, it is found that the stress-strain curve increases continuously up to failure. If the strain measurement is also based on instantaneous measurements, the resulting curve is known as a true-stress-true-strain curve. For most metals this curve, in the region of uniform plastic deformation can be expressed by the simple power curve relation σ = Kεn , where n is the strain-hardening exponent and K is the strength coefficient.
This fitted curve is extrapolated to the yield and ultimate tensile regions. The predicted value of yield strength is calculated from the fitted equation by extrapolating strain to zero. Then, the intercept of the elastic-line (with the slope equal to the slope of the measured unloading curves) with the fitted curve gives the predicted value of yield strength. The predicted value of the ultimate tensile strength is defined as the point at which elongation is equal to the work hardening exponent.
A comparison was made between the tensile properties obtained from the following: IIT; conventional tensile tests; conventional hardness-tensile correlations; and micro-flat tensile tests. The trials investigated several grades of steel (including rebar) and aluminium alloy parent material; and welds made by gas metal arc welding, friction-stir welding and laser beam welding. The hardness-strength correlations were applicable to steel only, and were based on the validated equations for estimating tensile properties proposed in ISO/TC 164/SC 4.
The 'essential variables' in IIT are considered to be:
- The test location, where the properties are required (eg. a weld HAZ, when it is helpful to etch the surface for accurate location);
- Surface preparation, as results are affected by variable surface roughness and heavy surface work or heating (it is necessary to smooth weld cap profiles prior to testing to ensure components have flat surfaces);
- Secure attachment of the IIT unit (eg. for plates or small diameter pipes, flat or curved magnets can be used; for large pipes a chain lock system may be used);
- The choice of indenter (the diameter of the indenter to be used depends on the thickness of the material and width of the location of interest, eg. a 0.5mm diameter indenter is suitable for a 1-2mm wide weld HAZ);
- The test procedure, which although it is semi-automatically controlled by the unit, requires correct set-up and execution by the operator; and
- Data processing, specifically the setting of the curve fitting parameters, since changes to these parameters will shift the fitted curve up- or downwards.
Some useful results
One of the groups of specimens investigated were API 5L X60 carbon-manganese steel pipes, with 20mm wall thickness and 508mm outside diameter, some of which were girth-welded by the GMAW process. Figure 3 shows the IIT results for the parent material. Both weld and parent materials exhibited Lüders plateaus in their stress-strain curves.
As was the case for all the specimens within the group, the IIT tests resulted in lower tensile values than those from conventional tensile tests, especially in the areas near the yield and the Lüders plateau. Yield strength was underestimated by 10-23% (mean underestimate of 16.5%) and ultimate tensile strength by 2-7% (mean underestimate of 4.5%). The level of underestimation was greater for the yield strength than that for ultimate tensile strength.
In summary, IIT was found to be capable of estimating the stress-strain curve to within a reasonable and safe range of that obtained from a conventional tensile test. The technique gives repeatable results (within the band of normal variability) so long as surface finish, test piece clamping and machining have been carried out correctly.
Potential uses for IIT
IIT can be used in the workshop and in the field, Fig.4, for pre-qualification or quality control and in-service assessment of ageing structures, respectively.
IIT is useful to estimate tensile properties without the need for sample removal (non-destructively), and in the event that a record of tensile properties is needed for a fitness-for-service (FFS) study. This would include determination of a full stress-strain curve required for a Level 3 (material specific assessment) FFS assessment to API RP579 or BS 7910. The failure assessment diagram (FAD) in this procedure may then be constructed based on the IITfull stress-strain curve. Similarly, IIT is useful for Level 3 FFS assessment of corroded or locally-thinned pressure components that use plastic-collapse finite element methods, where knowledge of yield or ultimate tensile strengths is required.
IIT has advantages over hardness testing because: (a) it can be used for materials where a hardness correlation is not available; (b) it can be used for materials where the hardness is outside the validity range of the correlation; and (c) it can measure the strain hardening coefficient directly. For material where a valid hardness correlation is available, the advantages of IIT are marginal and it is unlikely to supersede hardness testing as the established technique for estimating tensile properties.