Abstract: Fatigue is the main failure mode of metal components under cyclic loading, thus the fatigue life prediction of metal components is very necessary to ensure the safety and reliability of structures. As a method suitable for low- and high-cycle fatigue life prediction, energy-based method is to create an effective explicit energy damage parameter combining with appropriate damage accumulation methods. Aiming at the fatigue life prediction of materials, a fatigue life prediction method based on energy method and neural network algorithm is proposed. In order to reflect the influence of different loading paths, two loading-path-dependent parameters are introduced based on the moment of inertia concept. Strain-controlled fatigue experimental results of nine materials are used to train and test the proposed neural network model. It can be found that the model not only has good prediction accuracy for both the training and test data, but also can effectively predict uniaxial loading, multi-axial loading, high cycle fatigue and low cycle fatigue life, which indicates the model has a wide range of applicability in multiaxial fatigue life prediction.

Key words: fatigue, moment of inertia, neural network, life prediction