An investigation into prying models in tension clips

Abstract

Models of the prying phenomena can be split into two types; simple analytical and semi-empirical. Simple analytical models are based on assumptions that allow for rough sizing, but do not address important prying phenomena. Semi-empirical models attempt to holistically address the various phenomena arising from prying, but are pragmatically focused on matching tested data and arriving at simple design formulae rather than accurately modelling the physical phenomena. A comparison of these two types of modelling philosophies is presented, addressing both their advantages and disadvantages. Since most tested data is from the civil engineering-side of the structural fraternity, they apply to end-plate connections with relatively ’thick’ bases (> 10 mm). The aerospace industry uses ’thin’ ( <3 mm) bolted tension ’T-clips’ extensively to transfer ’light’ loads. Accurately predicting the load in the fastener, as well as the bending and shear distribution throughout the clip, is therefore important; especially for fatigue considerations. To achieve this objective, a method is proposed that assumes the clip to rest on an infinitely stiff, tensionless Winkler foundation. The corner radius is modelled using Castigliano's 2nd theorem to accurately predict the shear and moment distribution throughout the clip. The fastener is modelled as both an axial and torsional spring, allowing for extension and rotation of the fastener, and to establish the point of first contact of the clip with the foundation. Fastener preload is also included in the model. Finally, the maximum applied force at which the clip initiates plastic hinges is also predicted. The proposed method predicts the centroid of the prying load distribution to be a function of base thickness, with significantly larger prying force reacted by the fastener in ’thin’ bases. Predicted results for ’thin’ (<3 mm) aluminium clips match those of 1D Nastran beam models (CBARs) on CGAP elements, as well as a 3D Marc non-linear contact model, closer than any other model found in the literature.