Probabilistic analysis of an uncemented total hip replacement |
| |
| Authors: | Carolina Dopico-González Andrew M. New Martin Browne |
| |
| Affiliation: | 1. National Centre for Advanced Tribology at Southampton (nCATS), University of Southampton, Highfield, Southampton SO17 1BJ, UK;2. Bioengineering Science Research Group, Faculty of Engineering and the Environment, University of Southampton, Southampton SO17 1BJ, UK;3. Southampton Orthopaedic Centre for Arthroplasty and Revision Surgery (SOCARS), Southampton General Hospital, Tremona Road, Southampton SO16 6YD, UK;1. Eastern Health Clinical School, Monash University, Victoria, Australia;2. Department of Physiotherapy, Eastern Health, Victoria, Australia;3. Department of Rehabilitation, Nutrition and Sports, La Trobe University, Victoria, Australia;4. Imaging Associates Box Hill, Victoria, Australia;5. Department of Epidemiology, School of Public Health and Preventive Medicine Monash University, The Alfred Centre, Victoria, Australia;6. Department of Orthopaedics, Eastern Health, Monash University, Victoria, Australia;1. Mechanical Engineering Department, University of Larestan, Lar, Iran;2. Mechanical Engineering Department, Shiraz University, Shiraz, Iran |
| |
| Abstract: | This paper describes the application of probabilistic design methods to the analysis of the behaviour of an uncemented total hip replacement femoral component implanted in a proximal femur. Probabilistic methods allow variations in factors which control the behaviour of the implanted femur (the input parameters) to be taken into account in determining the performance of the construct. Monte Carlo sampling techniques were applied and the performance indicator was the maximum strain in the bone. The random input parameters were the joint load, the angle of the applied load and the material properties of the bone and the implant. Two Monte Carlo based simulations were applied, direct sampling and latin hypercube sampling. The results showed that the convergence of the mean value of the maximum strain improved gradually as a function of the number of simulations and it stabilised around a value of 0.008 after 6200 simulations. A similar trend was observed for the cumulative distribution function of the output. The strain output was most sensitive to the bone stiffness, followed very closely by the magnitude of the applied load. The application of latin hypercube sampling with 1000 simulations gave similar results to direct sampling with 10,000 simulations in a much reduced time. The results suggested that the number of simulations and the selection of parameters and models are important for the reliability of both the probability values and the sensitivity analyses. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
