تجزیه و تحلیل حساسیت از روش های زمینه کامل برای اندازه گیری تنش پسماند

The hole drilling technique is a well known experimental method for residual stress investigation. This technique is usually used in combination with electrical strain gauges but there is no reason to enforce this choice and other approaches, in particular some full-field optical techniques, can be advantageously used. Since all these techniques give full field data, it becomes important to properly use this redundant information content to increase the robustness and reliability of the analysis. In this work, various well known approaches to the hole drilling/full-field data analysis will be investigated using a two-step approach. In the first one, a sensitivity analysis will be performed on the simpler algorithms and then the reliability of the methods will be estimated by Montecarlo analysis using a known displacement field as a reference.

The hole drilling method is a well-known technique for residual stress analysis [1], [2], [3], [4], [5], [6] and [7]. The “standard” version consists in measuring the strain components produced by drilling a flat-bottomed hole in the centre of a specifically designed strain gauge rosette. Strain gauges are extremely practical measuring devices, but the finite dimensions of the sensitive elements impose several limitations on their use. In fact, the acquired signal is obtained by integrating the strains over the strain gauge surface, whereas it is practically impossible to perform measurements in the vicinity of the hole edge (which is actually the region where deformation gradients are higher). Moreover, strain gauges are particularly sensitive to off-centre drilling errors [8]. Full-field optical measurement techniques (moiré interferometry, speckle interferometry, holographic interferometry and shearography) can overcome these limits as they are able to acquire the whole displacement (or strain) field along the sensitivity direction. The quantity of available data, which largely exceeds the required minimum number of three values, therefore poses the problem of their effective utilization to exploit the intrinsic redundancy of the data so as to reduce measurement errors. Numerous approaches to solving this problem have been proposed in the literature, the most well-known being the measurement of three displacement components at a single point [9] (repeated, if necessary, in different points), measurement of a single component [10], Fourier analysis of the displacement field around a circumference [11] and the least squares fit of the displacement field [12]. In this work, we analyse the performance of the above techniques in terms of accuracy and robustness to noise, using known stress fields for the classical case of a thin plate with through-hole and constant stress [13], [14], [15] and [16]. For this purpose, we generated synthetic data fields which were then perturbed by adding various levels of noise. The resulting “experimental” data were then analysed using the different algorithms and their results compared with the expected values. The final aim of the work is to determine the optimum combination of experimental technology and analytical methodology for measuring residual stress using optical methods.