Estimation of the accuracy of determination of the Williams coefficients under the conditions of normal cleavage

Evaluation of random errors in Williams’ series coefficients obtained with digital image correlation

“…As can be seen from table 2, the evaluated values of RMS of random errors of the crack tip coordinates and the first Williams' coefficient (SIFrelated) are less than 0.1%. Small values of random error of the crack tip coordinates and the first Williams' coefficient showed their minimal noise sensitivity, which matches the numerical simulation results [22]. A large error value in coefficient A I2 evaluation can be explained by its relatively small contribution into the stress field (<1% of component σ xx range).…”

Improving the accuracy of derivation of the Williams’ series parameters under mixed (I+II) mode loading by compensation of measurement bias in the stress field components data

“…The influence of random component of errors in measured data on the accuracy of derived Williams' series parameters is significantly reduced by processing a large amount of data obtained by full-field measurement methods [5]. However, the influence of bias in measured field components data on the accuracy of the parameters derivation can be greater than the influence of random errors [15,16]. This influence can be reduced by means of compensation of bias in the measured data, but it is complicated to obtain the required values of…”