ν12 ¼0.261; thus κ ¼3.24). Then a series of iDIC analyses were performed using different E1 settings ranging from 80 to 120 % of the target value. As expected, errors in stress estimation increase almost linearly with ε(E1), but some points are to be noted: – Even in the worst case, corresponding to a stress field oriented parallel to the principal axis, the scale factor is less than 0.5, thus a 9 % error in the stress estimate corresponds to 20 % error in the Young modulus estimation. – Looking at the hydrostatic stress state, the σy estimates are almost constant (and correct), thus errors on E1 do not affect the estimation of stress in the orthogonal direction (i.e. there is little or no cross-talk). – Maximum error for the shear test case is less than 4 %. Figure 43.4 shows the iDIC behavior when assuming an erroneous location of the center of the hole. Three different tests were performed, using either an erroneous x-coordinate, or an erroneous y-coordinate or both. Apart from the absolute value of the error (less than 5 %), the behavior of the cy curve (error direction orthogonal to the principal direction of the material) is quite interesting, because after an initial increase, errors decrease to less than 1 %. Note that the influence of the y-displacement is so large that even when both x and y locations are incorrectly estimated a similar behavior is observed. Finally, Fig. 43.5 shows the influence of the incorrect estimation of the radius of the hole: it is apparent that an overestimation of the radius induces small or no errors in the final result, whilst underestimating the radius induces large errors. However, there is a clear saturation effect and this parameter does not influence the estimation of the τxy. 1 1.005 1.01 1.015 1.02 1.025 1.03 1.035 1.04 1.045 800 805 810 815 820 825 830 1 1.00625 1.0125 1.01875 1.025 1.03125 1.0375 exact center location: (800,800) estimated stress / expected value assumed center location [pixel] center location error (relative) cx cy cx and cy Fig. 43.4 Influence of the errors in the identification of center location. The three curves correspond respectively to errors in the x direction only (cx), in the y direction only (cy) or both (cx and cy). Behavior is symmetrical with respect to the center, thus only half of the range is shown 360 A. Baldi
RkJQdWJsaXNoZXIy MTMzNzEzMQ==