Strains and Stresses

The strain gauge measurements were carried out in several locations on the surface of the structure concentrated around weld A and B. In the following Figure 7.15, three sets of measurements on the as-welded structure are compared to the numerically obtained stresses. The location of the measurement series on the structure is indicated in Figure 7.14 together with the triads representing the direction of the stresses compared in each series.

FIGURE 7.14 Locations of the strain gauge measurement series and the direction of the stress components..

The numerical model is specified in a coordinate system corresponding to that of e.g. series #3. To reflect the stress state in series #2 and #3, the calculated stresses in the Cartesian coordinate system are subjected to a coordinate transformation involving directional cosines.

As seen on Figure 7.15, a qualitatively good agreement exists between measurements and calculated stresses in all three series (#1, #2 and #3). It is noticeable that the calculated residual stresses in direction 3 corresponding to the longitudinal direction in relation to the weld are generally higher than measured.

This is opposite to the stresses in the other direction examined where the strain gauge measurements generally show higher values. At least two possible reasons exist for this general tendency.

First, the numerical analysis is based on a two-dimensional plane strain assumption that tends to over-constrain the structure in the out-of-plane direction even though a generalised plane strain assumption is adopted, incorporating the extra degrees of freedom this method allows. Secondly, the strain gauge measurements have been carried out on a 180 mm wide piece of the 2.5 m frame box section (original

measurements carried out on the full section failed and the structure was cut out by mistake before this was realised). This cutting will redistribute and to some extend release the high longitudinal stresses. In order to evaluate how far from the cut edge this has an effect on the stresses, a series of measurements in a certain distance parallel to the weld has been accomplished on the oblique plate next to weld A.

Figure 7.16 shows the stresses measured from the centre of the 180 mm thick section at 90 mm outwards to the edge. As expected, the stresses decrease closer to the edge, but a constant level in the centre of the plate cannot be observed indicating that the stress release has had an effect all through the 180 mm section.

FIGURE 7.15 Strain gage measurements and numerical calculations of residual stresses on the as-welded frame box structure. The three series are located as indicated in Figure 7.14.

-150E+6

0,27 0,29 0,31 0,33 0,35 0,37 0,39

distance [m]

CH A P T E R 7 . AN A L Y S I S O F WE L D E D EN G I N E FR A M E BO X

Another consideration in the evaluation of the strain gauge measurements is the fact that this measuring technique is only suitable for stress ranges up till approximately 70% of the yield stress and if absolute stress values with good accuracy is preferred, 50% of the yield stress should not be exceeded in the material under inspection. In the present case, several measurements lie close to the yield stress, still showing the same tendency as the numerically calculated stresses, but it is also in these areas the largest discrepancy is seen.

FIGURE 7.16 Residual stresses measured from the centre of a 180 mm thick section outwards to the edge.

As indicated in Figure 7.14, a series #4 was made on the cut plane of the section across the guide bar. The residual stresses measured here were expected to be influenced by the slow running band saw but on the other hand in a relative low and homogeneous stress area where the measuring technique should yield good results.

Figure 7.17 shows the comparison between measured and calculated stresses. If the deformation pattern from the numerical analysis of the structure is considered as shown in Figure 7.18, a bending moment of the guide bar can be expected leading to tensile stresses at the running surface and compressive stresses at the surface against the web plate and the oblique plates. This is confirmed by the measurements in Figure 7.17, indicating that the slow running band saw has only little effect on the residual stresses that can be measured by the Hole-Drilling Strain Gauge method.

0 20 40 60 80 100 120 140 160

0 10 20 30 40 50 60 70 80 90

distance [mm]

stress [MPa]

measurement, s11 measurement, s33

FIGURE 7.17 Strain gage measurements and numerical calculations of residual stresses on the cut plane of the frame box section across the guide bar.

FIGURE 7.18 The deformed structure as calculated, shown with a deformation factor of 20.

One of the main purposes with the numerical analysis is to evaluate the stress state around the weld toes and roots. The above comparisons of experimentally and numerically obtained mechanical stress fields together with engineering experience give a strong indication of the numerical model yielding correct estimates of the residual stresses from welding in the structure. Therefore, calculated contour plots of the stress state perpendicular to the predefined root errors are expected to show a stress state in agreement with the actual stress state in the test specimens. Such

-200E+6 -150E+6 -100E+6 -50E+6 000E+0 50E+6 100E+6 150E+6 200E+6 250E+6 300E+6

0 0,01 0,02 0,03 0,04 0,05 0,06

distance [m]

stress [Pa]

numerical analysis, s11 numerical analysis, s22 measurement, s11 measurement, s22

CH A P T E R 7 . AN A L Y S I S O F WE L D E D EN G I N E FR A M E BO X

contour plots are presented for weld A and B in Figure 7.19 (see also Figure 7.5 for illustration of predefined root errors). In both welds, the crack slit is in compression due to tensile stresses in the weld. This is a favourable stress distribution from a fatigue resistance point of view.

FIGURE 7.19 Contour plots of the stress state calculated perpendicularly to the root error. In both welds the “crack” is in compression due to tensile stresses in the weld.