Parameter estimation and comparison for linear and modified goodman diagrams

An accurate model for constant amplitude description can be used as a tool to obtain a more reliable life prediction of wind turbine composites under variable amplitude loading. Such a fatigue model describes fatigue behavior under all possible loading conditions. Experimental data are typically only available for selected load types. The accuracy of the model in describing constant amplitude fatigue data can be investigated by testing the fit of the model to the available fatigue test data. Popular ways to model constant amplitude fatigue life of a material are using the S-N curve (load versus life for a particular load case) and combining S-N curves and static properties in the Constant Life Diagram (CLD, describing fatigue life as a function of alternating stress and mean stress). The objective of this paper is to describe parameter estimation from the large fatigue dataset that comes from the UPWIND/INNWIND projects and compare the parameter sets of the models. Parameter estimation is done using regression analysis and statistical evaluation of the results. Two forms of the CLD, namely Linear and Modified Goodman diagrams are coupled with linear-log and log-log S-N curves to obtain a parameterized description of constant amplitude fatigue behavior. The best fit parameter sets of the models to the fatigue test data of UPWIND/INNWIND UD glass/epoxy reference material are generated and compared. Discussion and conclusion on the implementation of the models to the fatigue data of UPWIND/INNWIND reference material is presented.