Abstract
Thermoelastic stress analysis (TSA) is a non-contact and full-field type method of experimental stress analysis. Since TSA provides only the sum of the principal stresses, many studies have been carried out for developing techniques of stress separation. Unfortunately, most of the techniques developed hitherto are rather complicated and require much labor in practical application. Recently, the authors have developed a simple technique by utilizing the nonlinearity of the thermoelastic effect. In this technique, principal stresses are determined from two sets of temperature amplitude data measured under sinusoidal loadings of same amplitude and different mean values. However, the difference between temperature amplitudes obtained at different mean loadings is very small and, therefore, the result of stress separation becomes very noisy. In this paper, several smoothing techniques are applied to the temperature amplitude data. It is found that some simple smoothing techniques are effective for improving the accuracy of the stress separation.
