Study On Load Sequence Effect In The Unified Fatigue Life Prediction Method

Fatigue is one of the most commonly occurred failure modes in marine structures, such as ships and offshore platforms, as a result of variable loading and random wave environment in service. Complex loads on structures usually cause significant load sequence effect and have notable influence on the law of fatigue crack propagation, which result in the difficulty of predicting fatigue life accurately. But as yet there is no universal method on the fatigue life prediction adequately considering the load sequence effect. Recently, the idea of a unified fatigue life prediction method has been proposed based on fatigue crack propagation theory, which aims at more precisely estimating fatigue strength of metal structures. Hence, as a key step of the application of unified fatigue life prediction method, the main purpose of this thesis is the establishment and the application of a fatigue crack growth rate model considering the load sequence effect with cycle-by-cycle integration procedure.Firstly, a crack growth rate model based on the concept of the partial crack closure, has been proposed, which has good accuracy and wide applicability to predict the fatigue crack growth rate for wide range of engineering metallic materials under constant amplitude loading. The model will be a basis of considering load sequence effect. Till now, the mechanism of load sequence effect has not been sufficingly understood by researchers. It is assumed here that the essential cause of the load sequence effect is the variation of crack opening level with variable loading history, which leads to the change of the effective value of the external loading. Thus a coefficient,φ, is introduced to modify the term of the crack opening level in the crack growth rate model under constant amplitude loading in order to extend the applicable range of the model to the condition of variable amplitude loading. Then, a cycle-by-cycle integration procedure based on the improved crack growth rate model is established to predict crack growth rate under variable amplitude loading, in which the model parameters is one of most major input information. The engineering estimation methods of model parameters and transformation of different types of material test data to crack growth rate curve have been developed to satisfy the requirement of engineering application. Finally, comparisons between the predicted crack growth rate curves with their corresponding experimental data for several metal materials are carried out, showing that the improved model is able to predict fatigue life under constant amplitude loading and variable amplitude loading including single/multiple overload, underload, the combination of overload and underload and block loading. The establishment of the fatigue crack growth rate model by a cycle-by-cycle integration accounting for the load sequence effect is a key step of engineering application of the unified fatigue life prediction method.The original contributions of this thesis can be summarized as follows:(1) Adopting the viewpoint that the essential cause of the load sequence effect is the variation of crack opening level, the load sequence coefficient is first introduced to modify the term of the crack opening level in the unified model in order to denote the variation of crack growth driving force with various loading sequence.(2) Based on the unified model, a feasible method has been proposed to explain these phenomena of the retardation due to overload and the tiny acceleration due to underload, even the minor retardation due to overload followed by underload. Then the flow diagram for the cycle-by-cycle integration procedure considering the load sequence effect has been developed to be convenient for numerical calculation.(3) Systemic description of engineering estimation approaches of the parameters in the improved crack growth rate model have been carried out and transformation of different types of material test data to crack growth curve according to available experimental data has also been proposed to save time and cost.