Multi-damage life assessment under continuous damage mechanics

Materials in the service process will produce multi-damage, such as brittle damage, creep damage and fatigue damage. Traditional linear superposition principle, considering that different kinds of damages are independent, is not always accurate. In order to evaluate reliability of the material life under the multi-damage, a new method is presented in the paper. The method is based on continuous damage mechanics(CDM) framework, considering effects on grain size and void damage, proposing creep and fatigue continuous damage model. Then the BP neural network is constructed to train creep-fatigue damage and finally validated the rationality of the model. The method can accurately evaluate the reliability life, which provides a new way to study multi-damage of the material failure behavior under CDM framework based on BP neural network.