No abstract

Integration of the microcomputer into acoustic emission instrumentation has brought AE monitoring of fatigue tests into the realm of practicality. On-line processing makes available a selection of software tools, enhancing classical techniques for eliminating the background noise which usually blanked out the desired data. Fatigue tests monitored for acoustic emission were carried out at room temperature on Incoloy 901 material specimens, over a stress-ratio range of -1 < R < .2. Valid AE data were obtained even when the load cycle passed through zero. The AE data permitted specific identification of the various phenomena occurring on the way to final failure. These included initial plasticity, crack nucleation and propagation phases. The AE findings were supported by microscopic examination. Based on the experimental data, a preliminary damage-prediction model was formulated.

No abstract

A new class of large, high‐speed seagoing ferry‐boat is under development for service around the world. The ships, which are built entirely of aluminium‐alloy plate and stiffeners, show a propensity for fatigue cracking of the welded structure. Cracks may occur in both the hulls and the superstructure early in their 20‐year service life. Early appearance of fatigue cracks is shown to result from the combined stress and strain fields set up in weld zones by the static residual stresses and cyclic loads, beyond the effects of weld and detail geometry. A numerical example demonstrates that conventional methods of fatigue analysis overestimate the lifetime of the welded aluminium structure, while damage tolerance analysis based on fracture mechanics leads to improved prediction.

A further modi cation of the Bolotin method for the determination of the natural frequencies and mode shapes of isotropic and orthotropic rectangular plates with various types of boundary conditions is given. Unlike the Bolotin method (BM) or the modi ed Bolotin method (MBM), the present approach does not postulate the formula for the eigenfrequency, but rather is based on the condition that the frequency obtained from the governing differential equations has to be equal to that yielded by Rayleigh's method. This modi cation is shown to be more straightforward and faster in computation, and the mode shapes derived are valid on a larger portion of the plate. Furthermore, the proposed modi cation easily provides a solution for boundary conditions for which the BM and MBM cannot provide a solution. Problems with two different sets of boundary conditions were solved in this study: a rectangular orthotropic plate with all edges clamped and rectangular isotropic and orthotropic plates clamped along one pair of opposite edges and free along the other pair. The results obtained for the rst set compared favorably with those yielded by the MBM and Rayleigh methods, whereas in the second case the BM and MBM failed to predict the beam-like modes of vibration, while the present modi cation treats the problem satisfactorily. Nomenclature= lengths of the side of the plate along the x and y axes, respectively c 1 , c 2 , C 1 , C 2 = real numbers= exural rigidity from the coupling of x and y directions, D x y = D x m y D y = exural rigidity in y direction, D y = E y h 3 / 12 (1 ¡ m x m y ) D 66 = torsional exural rigidity, D 66 = G h 3 / 12 E x , E y = Young's moduli along the x and y axes, respectively G = rigidity modulus H = D x y + 2D 66 h = plate thickness K = kinetic vibration energy m, n = integer numbers p, q = real numbers s 1 , s 2 , S 1 , S 2 = real numbers t = time w = transverse displacement X = mode shape function in x direction x, y = rectangular coordinates Y = mode shape function in y direction h= real or imaginary numbers, roots of characteristic equation m x , m y = Poisson's ratios for the material where E x m y = E y m x P = potential vibration energy q = mass density s = direction normal to the contour of the plate x = eigenfrequency or cycle frequency