DePuy is the oldest manufacturer of orthopaedic implants in the United States. The company was founded in 1895 when Revra DePuy, a salesman, revolutionized the fracture management industry by introducing wire splints to replace the makeshift wooden splints then in use for stabilizing fractures. DePuy is one of the world's leading orthopaedic companies, with a reputation for innovation in new product development.

DePuy researchers were on the forefront of the total hip replacement concept introduced in the early 1960s, and it was this development that was primarily responsible for creating the orthopaedic market as we know it today. The graphic to the right illustrates a typical hip replacement.

The subject of this work was to analyse the insertion of a Duraloc Un-cemented Acetabular Hip socket into the pelvic bone. The results that were required from the analyses were to establish the force required to insert the socket into the pelvic bone given a certain interference fit. Once the assembly load had been evaluated it was also necessary to establish the force required to remove the socket axially or the torque required to remove it via twisting. The photograph to the left shows the Duraloc socket which is made from Titanium with a porous coated acetabular shell and is available in diameters ranging from 48mm through to 66mm.

The cup, which is a shell/liner dome, is implanted into bone, once the bone has been hollowed out using a grater. The cup was to be modelled in the bone with a 1.5mm interference fit. This was simulated using non-linear contact elements and anisotropic material behaviour with respect to the bone. The smooth cup assembly analysis was conducted in both two dimensions, an axisymmetric analysis, as well as in three dimensions i.e. an 180o section. In all analyses the implant cup has been modelled with the material properties of Titanium.A literature search was conducted in order that the mechanical material properties for both cancellous and cortical bone could be established. The contact definition between the cup and bone was modelled to include a coefficient of friction.

The graphic (below) illustrates the axisymmetric finite element model used with the different colours referring to the different bone materials and the titanium cup. For both the two and three dimensional analyses parametric models were created in order that different bone and implant cup geometries, material properties and boundary conditions could be evaluated. The assembly conditions involved inserting the cup into the bone to overcome interference, allowing the frictional effects to hold the cup in place and then to subsequently pull-out or twist-out the cup from the bone to establish dis-assembly loads.

The graphic (right) shows the stresses induced in the assembly due to the interference fit for the two dimensional axisymmetric analysis. The areas coloured in grey illustrate the region of the bone which could be expected to yield during the assembly process.

The graphic (left) shows the equivalent three dimensional 180° sector finite element model. Two dimensional axisymmetric analyses are unable to evaluate the twist out loads for the assembly and hence the equivalent model needs to be analysed in three dimensions. The graphic (right) illustrates the

stress distribution in the assembly after the interference has been taken up. These stress distribution plots can be created in the ANSYS program for any point in time during the non-linear solution. It is also possible to create graph plots of load versus time history in order that the assembly and dis-assembly loads can be visualised.An example of this type of load versus time history can be seen (right). All of the analysis work described here was performed using Intel based personal computers running the ANSYS Revision 5.6 program. The analysis run times for the problems vary from a few minutes for the two dimensional analyses up to a few hours for the three dimensional problems.

DePuy are users of ANSYS and the parametric models created here have been supplied in order that subsequent analyses of different configurations may be performed in-house.

The main benefits that accrue to DePuy from this form of modelling are the ability to evaluate different configurations of implant design on computer rather than by physical testing. Physical testing is time consuming and expensive in comparison to numerical modelling and also there are limitations on the amount of physical testing that can be done with real bone materials due to availability. There are some synthetic and naturally occurring materials that can be used in testing, however, their material properties do not precisely match that of human bone materials. Numerical modelling also allows DePuy to view detailed stress and deflection distribution plots and load versus time history plots which cannot easily be created from physical tests. Comparisons between the results obtained here and those obtained from previous testing reveal a close correlation and hence this modelling approach has been proven to be viable. As a result of this DePuy is considering extending this type of design evaluation to other orthopaedic implant products due to the success of this project.