By Isabel Hadley
Flaw assessment procedures such as BS 7910 and R6 are based on sound fracture mechanics principles but require validation against full-scale test data and numerical analyses in order to demonstrate that they are safe, user-friendly and practically useful.
This report presents the results of a set of wide plate tests which have been analysed in accordance with both BS 7910 and R6. Some of the newer tools, such as assessment of constraint and use of the Master Curve, are included in the assessment, along with consideration of biaxiality, which neither BS 7910 nor R6 currently address in detail.
- All wide plate tests analysed using the basic Option 1/Option 2 fracture assessment procedures of BS 7910:2013 produced failure points outside the Failure Assessment Line (FAL), confirming the safety of the procedure for the assessment of known flaws.
- A similar assessment using R6 also produced failure outside the FAL, albeit with a smaller safety margin, because of the use of different limit load solutions.
- Stress biaxiality affected the failure conditions of both centre-cracked tension (CCT) and surface-cracked tension (SCT) specimens throughout the temperature range considered (between –163 and +70°C).
- In the collapse dominated region of the Failure Assessment Diagram (FAD), assuming other factors (test temperature, flaw size, section thickness) to be similar, failure load (and hence Lr at failure) was higher in equibiaxially loaded (k=1) plates than under uniaxial loading (k=0), and higher still for the condition k=0.5. This is manifested in high modelling error (ie failure well outside the FAL, denoting a conservative assessment) when biaxially loaded plates are assessed without taking biaxiality specifically into account.
- In the fracture-dominated region of the FAD, crack tip constraint effects predominated, with uniaxially loaded specimens failing at higher loads (higher values of Lr/Kr) than their biaxially loaded counterparts. This held for both SCT and CCT specimens.
- For the SCT plates tested under biaxial loading in the collapse-dominated region, a simple correction factor can be applied to the basic (uniaxial) definition of Lr in order to derive a more realistic value. When the appropriate correction factor (1.1 for k=1, 1.15 for k=0.5) was used, the failure assessment point of the plates still fell outside of the FAL (except in one marginal case), when a modified BS 7910 collapse solution was used. A modified R6 collapse solution proved to be non-conservative.
- Use of handbook methods in conjunction with a simple biaxiality correction factor for the calculation of Lr in SCT plates led to results similar to those derived by direct finite element modelling, ie Option 3 assessment.
- The Master Curve method for assessment of fracture toughness (Pf=5%) was shown to produce results comparable with Minimum of Three Equivalent (MOTE) results at temperatures within ±50°C of the reference temperature, T0.
Summary of previous fracture validation. Each symbol/colour denotes a different set of test data.