Uncertainty Modelling And Model Validation In Health Assessment For Typical Mechanical Equipments And Components

Typical mechanical equipments such as aircraft engine are complex systems under the environmental condition of high temperature, high pressure and high-speed. Any failure of critical components is likely to lead to the occurrence of the flight accident or even more serious consequences. Thus it is necessary for the key components of aero-engine to implement the task of health assessment. Due to the increasingly complex mechanical structure and varying working condition for the typical equipments, there are a large number of errors and uncertainties in the material properties and geometric characteristics, which have been primary challenges to the health assessment of typical equipments including condition assessment, damage monitoring and reliability assessment. Therefore it makes sense to accomplish the tasks of uncertainty modeling and model validation for equipments and components.Taking on the fan shaft of a certain aircraft engine and pressure vessel as application examples, this thesis investigates, on the one side, parametric modeling of such components, on the other side it provides a preliminary study of model validation under uncertain conditions based on the finite element analysis. What’s more, evidence variables are adopted in dealing with uncertain parameters during the fatigue life prediction, then the impacts on the fatigue life prediction are analyzed so as to making decisions for the health assessment.The main contributions of this thesis are shown as follows:(1) Realizing the 3D parametric model and finite element analysis(FEA) model:The 3D models are constructed with Pro/E Version 2010 based on actual dimension of aircraft engine fan shaft and that of pressure vessels, then the FEA model derived from such 3D models are established using ANSYS Workbench Version 14.5.(2) Determining dangerous positions by stress-strain FEA under multiple loads:The elastic FEA of fan shaft is performed by considering different loading conditions, i.e. axial force, centrifugal force and torque loads and so on. By the effects of loading conditions and FEA models, the information of total deformation, the stress state of the fan shaft and two dangerous places for fatigue failure are obtained.(3) Investigating the general methods of model validation under uncertainties.The example of pressure vessel proposed in the “2014 Model V&V Challenge Problem” is considered in this thesis for exploring the general model validation method. By means of FEA and Kriging method, the maximum stress of pressure vessels can be predicted in terms of confidence interval of the maximum equivalent stress subject to a specific confidence level.(4) Analyzing the impact of evidence variables on the prediction of fatigue life with insufficient experimental data:.The Kriging response surface model of the fan shaft is put forward and the maximum equivalent stress is obtained under a certain confidence, by converting uncertain parameters of the fatigue life into evidence variables in the case of insufficient data. Then the impact of the evidence variables is discussed and the prediction result are compared with that of probabilistic model.