Ludovic Vincent

Fatigue microcrack detection with digital image correlation

A digital image correlation method is proposed to detect and quantify automatically microcracks on the surface of a specimen during a fatigue test. The proposed procedure allows for a fast scanning of the entire surface with all possible (pixel-wise) locations of microcrack centres and the detection of cracks having a sub-pixel opening. An experimental test case is presented as an illustration of the method and a comparison with a replica technique is performed.

Regularised digital-level corrections for infrared image correlation

Global digital image correlation (DIC) applied to infrared images is considered for a test where very significant and localised laser heating is applied. Because of temperature changes and high gradients, digital levels are to be corrected in addition to the displacement registration. Digital-level corrections and kinematics are decomposed over an unstructured mesh to account for steep gradients. A regularisation strategy is used to mend ill-posedness penalising rapid variation of corrections and displacements. High gradients motivate further softening of the regularisation (i.e. reduction of the regularisation length scale) based on prior estimates. The methodology is tested on different experimental cases in which DIC residuals are drastically reduced and systematic displacement errors are cut down to pixel after grey-level corrections (performed on numerical cases). Additionally, the regularisation improves the convergence speed as well as the quality of the sought fields. The capability of the code to account for blur corrections is illustrated on a practical case.

Probabilistic Modeling of Crack Networks in Thermal Fatigue

A probabilistic model is proposed to predict the formation and propagation of crack networks in thermal fatigue. It is based on a random distribution of sites where cracks form and on the shielding phenomenon corresponding to the relaxed stress field created around propagating cracks. The stress gradient that arises under a surface submitted to thermal shocks is accounted for as well as multiaxial stress states. Experiments using digital image correlation have been performed to introduce the hypotheses made herein, and others to identify the parameters of the crack density as a function of the stress range and the number of cycles. Simulations of the formation of crack networks in a heterogeneous biaxial state of stress are carried out. A good qualitative agreement is obtained with existing thermal fatigue experiments especially considering the effect of the stress range on the crack densities as well as on the distribution of crack sizes.© 2008 ASME

A Conservative Damage Accumulation Method for the Prediction of Crack Nucleation Under Variable Amplitude Loading for Austenitic Stainless Steel

The application of Miner’s rule using a loading issued from a mock-up of a RHR (removal heat system) of PWR plant, made of 304 steel gives a very important non-conservative fatigue lifetime in strain control when strain fatigue curve is used. A large number of test in strain and stress control are performed in different laboratories. Two modeling of literature Smith-Watson-Topper (SWT) and Fatemi-Socie (FS) have been used to simulate these tests. Much better responses than Miner’s rule are obtained. However these models need an elastic-plastic constitutive law which is difficult to propose in the presence of high cycle secondary hardening observed in austenitic stainless steels. So a conservative model for fatigue damage accumulation under variable amplitude loading is proposed for austenitic stainless steels (AISI 304, 316) in strain control, which does not need a constitutive law. Linear damage accumulation is used, while, sequence effect is taken into account using the elastic-plastic memory effect through cyclic strain stress curves with pre-hardening. This modeling is based on the fact that for stainless steels, pre-hardening is detrimental for fatigue life in strain control while it is beneficial in stress control. In the case of materials that do not demonstrate load sequence memory the modeling is identical to Miner rule. In the presence of low mean stress, the modeling is approved based on a large number of tests. Moreover the modeling permits to explain the larger detrimental effect of a tension mean stress in strain control tests than in stress control tests. To extend the modeling to higher values of mean stress it is proposed to divide mean stress effect into maximal and “real” mean stress effects. Extending this work to the case of significant mean stress is ongoing.Copyright