Lifetime calculation under variable amplitude loading with the application of artificial neural networks

Abstract Numerous calculation concepts are available for predicting the fatigue life of components. These analytical methods are generally based on a comparison of the stress caused by external loads with the strength of the component. With the application of these concepts, the damage is evaluated on the basis of a linear damage accumulation in accordance to Palmgren and Miner. An alternative procedure is the application of artificial neural networks (ANN). In previous works good results have been achieved with the estimate of S–N -curves within the scope of the nominal stress concept. A further step involves a direct comparison of stress and strength with the use of ANN without the need of a damage accumulation and without the associated problems.