LAUSR

Abstract The theory of critical distances is based on the definition of a material-dependent length L. Here, we investigate the statistical properties of L deduced from the crack threshold or a suitable notched specimen geometry. Monte Carlo simulations are done for best-fitting analytical functions to express mean, standard deviation and skewness of L. Standard-deviation-to-mean ratio is the lowest for the threshold-derived L estimation and decreases with notch sharpness. The minimum notch severity to achieve the desired accuracy in L estimation is identified. The impact of these statistical properties on the prediction of independent notched and cracked configurations is evaluated.