ESDU Engineer

Investigation into the Transverse Rupture of Lugs by Finite Element (FE) Analysis

Mr Neil Dev-Anand, Senior Aerospace Engineer – IHS ESDU, London, England

With the ever-increasing need to reduce aircraft weight, engineers are looking more closely at even the simplest components to find weight savings without compromising safety and performance. One such component is the lug. As there are numerous lugs built into an airframe, many quite substantial, the potential for weight saving through reducing the size of lugs is significant.

Lugs are most frequently used in aerospace structures to transfer loads and to connect major structural components (e.g. elevator to tail plane, fuselage to landing gear). They tend to be highly loaded structural details used in the primary and secondary structures. As they are usually part of a single load path, they can be critical to structural integrity and performance. For this reason, methods to design, analyse accurately, and predict the failure of lugs are necessary.

Since the work published by Melcon and Hoblit of the Lockheed Company in 1953, no comprehensive or conclusive studies, experimental or numerical, have been undertaken to investigate the parameters influencing the rupture of lugs under transverse loading. Owing to this lack of reliable design data, many aircraft lugs are very conservatively designed, with high reserve factors, compromising aircraft weight.

The strong requirement of the aerospace industry for reliable design data for lugs prompted ESDU to embark in 2003 on an investigation into the transverse rupture of lugs. This investigation has led to the development of new design data that are now facilitating more efficient lug design.

In the investigation, both material and geometric non-linear finite element analyses of lugs were performed using a general-purpose finite element program – MSC Visual NASTRAN for Windows. As a first step, the FE results were validated by comparison with test data from Hatfield Polytechnic, now known as the University of Hertfordshire, in order to give confidence in the lug model and also the analyses. A mesh convergence study was also undertaken and the results of the finest mesh deemed necessary were used.

Two material models were used in the investigation into the material parameters influencing the transverse rupture factor. They were (a) an elastic-perfectly plastic model and (b) a moderately ductile material model. The elastic-perfectly plastic material stress-strain curves were modelled as bi-linear curves in order to investigate whether or not the ratio of the ultimate total tensile strain to the total proof strain influenced the transverse rupture factor.

A range of stress-strain curves for the moderately ductile material were produced by varying the parameter that determines the shape of the plastic region of the curves. The various stress-strain curves were derived using a variant of the Ramberg-Osgood model using the 0.2 per cent proof stress and the ultimate tensile strength and their corresponding total strains as the reference stresses and strains. These were used rather than the 0.1 and 0.2 per cent proof stresses in order to give more accurate representations of the stress-strain curves at high strains. Also, the use of a variety of stress-strain curves for the moderately ductile material allowed the influence of the Young’s modulus on the transverse rupture factor to be examined.

Examination of the results produced by the FE analyses led to the conclusion that any difference between Young’s moduli of the pin and lug materials had only a small effect on the plastic stress concentration factor; as a consequence, the design data obtained by FE analyses may be applied to assemblies with any realistic combination of lug and pin materials.

As mentioned above, the FE results were validated by comparison with experimental data from Hatfield Polytechnic. The picture shows the smooth shear strain contour plot of a Hatfield Polytechnic lug model under transverse loading. Comparison between the FE and experimental data revealed that the FE predictions were conservative and within the limits normally expected when comparing analytical and experimental results.

The investigation revealed that the ratio of the ultimate total tensile strain to the total proof strain, the shape of the material stress-strain curve, and the lug geometry parameter all had a significant influence on the transverse rupture factor. It was also found that the transverse rupture factor was an almost linear function of the ratio of the 0.2 per cent proof stress to the ultimate tensile strength.

In addition, the FE analyses demonstrated that the transverse rupture factors for straight and tapered lugs were the same provided the lug geometry parameter remained unchanged. In conclusion, FE analysis was successfully employed to produce improved design data and information for the prediction of the rupture load of lugs under transverse load; those data, along with comprehensive supplementary information, are presented in ESDU 06021, “Strength of lugs under transverse load”, which is part of the ESDU Aerospace Structures Series.

Further information on this investigation is given in a paper presented at NAFEMS World Congress 2007, Vancouver, Canada by Neil Dev-Anand, Senior Aerospace Engineer, IHS ESDU, London, England. A copy of the paper can be obtained by contacting Neil via neil.dev-anand@ihs.com