透過您的圖書館登入
IP:3.145.74.54

摘要


Under the condition of elastic-plastic deformation, aero engine casings experience local stress and strain concentrations along with associated variations in load paths and stiffness. The accurate prediction of such behaviour is clearly necessary for design optimisation, potentially leading to beneficial weight savings. The present research seeks to tackle the objective of accurate characterisation of elastic-plastic casing behaviour. The objective is to develop approximate techniques for predicting the elastic-plastic behaviour, for both generalised load-displacement responses (i.e. for global response) and notch stress-strain responses. Accurate prediction of the stress-strain distribution at a notch is difficult and existing notch prediction techniques can only be used for small strains. This paper seeks to develop novel techniques for predicting large elastic-plastic notch strain and associated stresses, with application to aero engine casing notches. The repeated local joints at the spoke-shell casing are of particular interest as they are the most likely sites for plastic deformation and possibly crack initiation. These local joints incorporate realistic notch-type features and the load cases cover a range of loading combinations, to develop insight and understanding of the elastic-plastic behaviour. This work analyse a double edgenotched flat bar with semicircular notches and a representative case of actual aero engine casing-type structures in a more simplified form. The investigation was carried out for structures for which stress and total strain are related by a power law. The equivalent stress at a notch can be estimated, given the value of n, by a linear interpolation between the stresses for a cases n=1 and n=0. The application of the notch stress-strain prediction method is illustrated through use of examples of notch components. The predictions are compared with results obtained using finite element analyses and approximate methods proposed by Nueber and Glinka.

延伸閱讀