The measurement of residual stresses by laser annealing and electronic speckle pattern interferometry has been successful for moderate to high stress levels. However, for lower stress levels, say less than eighty percent of yield; sub-micrometer deformations are encountered that result in sub-fringe interferograms. Sub-fringe measurements are not a serious problem with modern optical procedures such as phase shifting interferometry. However, such sub-fringes are comparable to noise levels resulting from air turbulence, environmental thermal variations and so forth. Stress relief from local heating is also subject to theoretical lower stress limits that result from the competing effect of thermal expansion in the case of tensile stresses. Thus, for both practical and fundamental reasons the technique was modified such that several fringes will result at lower stress levels. These experimental techniques are shown to expand the stress measurement range for this technique.

Previous attempts [1-6] to measure residual stresses by laser annealing and electronic speckle pattern interferometry have been successful for moderate to high stress levels. The method uses an infrared laser for relieving stress in a small spot. A dab on temperature indicating paint is applied to the spot and a specklegram of the spot and the surrounding area is captured. The paint is then heated with a laser until it melts. The heat is transferred from the paint into the metal resulting in a small amount of localized stress relief as the yield stress of the material drops below the stress levels surrounding the spot. Once the spot and area around it have cooled a second specklegram is captured and the images are processed to determine the in-plane strain. The amount of stress relief depends on the melting temperature of the paint since yield stress is a function of temperature.

The measurement of local stress relief by heating is subject to limitations that result from thermal expansion competing with the reduction in yield stress of the spot at the elevated temperature. That is, as the spot is heated it tends to temporarily reduce the stress in the region surrounding the spot as it expands into this surrounding region. This limits the amount of stress relief that can occur. This can be overcome to some extent by using higher temperature paints, which in turn lowers the yield stress in the heated spot. At some point, however, the thermal expansion overtakes the surrounding stress field and can even drive it into compression. Furthermore, for tension levels on the order of eighty percent or less of the yield stress, the sub-micrometer deformations result in less than a single fringe. The strains indicated by such sub-fringes are comparable to noise levels that occur from air turbulence, environmental thermal variations and so forth.

Thus, for both fundamental and practical reasons the technique was modified to increase the fringe count at lower stress levels. We have successfully performed two separate experiments to raise the fringe count. One method was simply to start observing the fringes (or strains) immediately after annealing. A clear relationship was observed between number of fringes and the tensile stress levels. The other approach was to cool an area surrounding the region of interest and then observe the net strain after thermal equilibrium is reestablished. Both methods have shown the ability to detect lower tension levels than were measurable by the earlier procedure.

Experimental Setup: The residual stress measurements for the heating experiments were performed with the apparatus described in Reference 2. A 50 Watt CO₂ laser was used to heat a 2 mm diameter spot on the tension specimens for 5 seconds. The spot was painted with 482°C (900°F) melting point paint. A speckle pattern interferometer was set up to measure the in-plane strain along the load direction. Fringes were observed at 5 and 15 seconds after heating was initiated. However, precise timing at the 5 second mark was not possible, so that the strain results at this shorter time are probably prone to larger errors than at the 15 second mark. Experiments were performed with 304L and 21-6-9 stainless steel ASTM sheet tensile (type E8) specimens.

Experimental Results: The experimental results for 304L stainless steel are shown in Figures 1 through 6. Figures 1 and 2 show the fringe patterns at 5 and 15 seconds after heating respectively for the case of a tensile stress of 50% of the room temperature 0.2% off set yield stress. Figures 3 and 4 are similar figures for 70% of the yield stress and Figures 5 and 6 are for 90% of yield stress. There were no carrier fringes introduced in these interferograms because the sense of the stress and hence the strain are known a priori.

Notice that there is a higher fringe density at the shorter time period in all cases. This is because the specimens are warmer and more thermal expansion is present. At long times (@ 10 minutes) the specimen cools to room temperature and the only remaining fringes are those due to plastic deformation of the heated spot. There is also a clear correlation of the number of fringes with the initial stress level. The sensitivity of the interferometer is 0.44785 m m per fringe order. The center fringe with the top missing is the zero order fringe. The top of the fringe is partly missing because the liquid temperature indicating paint evaporates from the heated spot and then some of it is deposited in the region just above the heated spot as the vapor rises. This causes decorrelation of the speckle pattern in this region. The width of the specimen is 0.5 inches. A summary of the results from these interferograms and similar ones for 21-6-9 stainless steel are shown in Tables 1 and 2 below:

The time in these tables is the time after the laser heating started. It can be seen that there is a direct correlation between the fringe count and the applied stress. In the previous work, tensile stress levels much below about 80% of the yield stress were not detectable They clearly are with this method. The lower limit of stress detection is not known because the number of specimens were limited. Notice also that the fringe density is higher near the heated spot. This fact can also be used to extend the lower limit of detectability even further.

The idea behind cooling is to create a region of contraction around the spot in which the residual stress is being measured so the local stress is increased sufficiently to increase yielding of the spot. A copper block was immersed in liquid nitrogen for a long time and then brought into contact with the specimen for ten seconds while the specimen was under tension. The contact area (Figure 7) was on the same side as the interferometer illumination. A 1.27 by 1.27 cm² (½ inch ´ ½ inch) square on the central portion of the 1.27 cm (½ inch) wide specimen does not come into contact with the copper block. The contact region is 5.1 cm (2 inches) along the length of the specimen on each side of the region of interest. A speckle pattern is captured and then the cooling process is performed. The second speckle pattern is captured after thermal equilibrium with the surroundings has been reestablished. Difference fringes after filtering and enhancement are shown in Figures 8, 9 and 10. The approximate temperature of the cooled region is indicated in each figure caption. A thermocouple on the cooled side of the specimen was used to determine this temperature. Notice also from the captions that the load increased as the specimen was cooled as was expected. Only 304L stainless steel results are presented. A clear correlation between the fringe count and the stress level can be seen. In these experiments there is a large variation in the fringe pattern cross the face of the specimen. This probably results from the uneven initial cooling of the specimen.

The results presented in this paper clearly demonstrate that the dynamic range of this measurement technique can be improved substantially over the earlier experiments. It is just as clear that a more systematic study must be performed to quantify these improvements and to generate usable calibrations. These results are also encouraging in the sense that this technique may now be appropriate for other materials with higher thermal diffusivities.

This work was sponsored by the United States Department of Energy, Contract Number DE-AC09-96SR18500.

Figure 9.  70% Pretension
Room temperature 19.9°C
Minimum temperature - 45°C
Tension reading crossed 90% of yield