New principles for the calculation of welded joints

For the calculation of the strength of welded joints, use has been made of a critical surface of the ultimately allowable values for the average stresses p either in the smallest longitudinal section (throat section) o r cross section of the weld.

In statically loaded structures the shape of the critical limit curve for loads perpendicular to the longitudinal axis of the weld can be deduced from the experiments of the Netherlands Foundation for Applied Scientific Research (T.N.O.). For the shape of the surface of limits, a surface of revolution (pearoid) has been assumed, of which the meridian section forms the above-mentioned limit curve.

Starting from the average stress in the throat section of the weld no theory of rupture or plastic flow has, up to now, given a satisfactory explanation of all phenomena of rupture observed in end fillet welds at the moment of fracture. This is to a great extent caused by the uneven distribution of stresses.

In comparison with Kist’s method of calculation in use at present in the Netherlands, the principles here presented allow not only some simplification but above all economy in the case of weld sections predominantly loaded in shearing or compression.

In dynamically loaded structures a symmetrical surface of revolution has been adopted as the shape of the critical surface of allowable effective stresses. This effective stress is to be obtained by multiplying the numerically greatest value of the oscillating stress by a coefficient γ, which is consistent with the fatigue diagram of the material and the type of weld. The values of γ are given both for butt and fillet welds for two frequency ranges of load oscillations. A full explanation based on both theory and experimental work is given.