Clarke S G, Phillips A T M, Bull A M J
Structural Biomechanics, Department of Civil and Environmental Engineering, Imperial College London, London, UK.
Comput Methods Biomech Biomed Engin. 2013;16(7):717-24. doi: 10.1080/10255842.2011.633906. Epub 2011 Dec 8.
To enable large-scale multi-factorial finite element (FE) studies, the FE models used must be as computationally efficient as is feasible, while maintaining a suitable level of definition. The present study seeks to find an optimum level of model complexity for use in such large-scale studies by investigating which model attributes are most influential over the chosen model outputs of principal stress and strain in the intact acetabulum. A multi-factorial sensitivity study was carried out using 128 FE models, representing combinations of the following variables: bone stiffness distribution, imposed muscle loading, boundary condition location, hip joint contact conditions and patient's bone anatomy. The relative sensitivity of each input factor was analysed, and it was concluded that the optimum level of model definition must include CT-dependent trabecular bone properties and a sliding interface at the hip joint. It was found that it was not essential to describe the ligamentous sacroiliac and pubic symphysis joints; these could be rigidly fixed in space; and for the normal walking load case, muscle forces may be neglected. It was also concluded that a variety of bone anatomies should be included in a multi-factorial analysis if results are to be inferred for a wider population.
为了开展大规模多因素有限元（FE）研究，所使用的有限元模型必须在保持适当定义水平的同时，尽可能提高计算效率。本研究旨在通过调查哪些模型属性对完整髋臼中主应力和应变这两个选定的模型输出影响最大，来找到适用于此类大规模研究的最佳模型复杂程度。使用128个有限元模型进行了多因素敏感性研究，这些模型代表了以下变量的组合：骨刚度分布、施加的肌肉负荷、边界条件位置、髋关节接触条件和患者的骨骼解剖结构。分析了每个输入因素的相对敏感性，得出的结论是，模型定义的最佳水平必须包括依赖于CT扫描的小梁骨属性以及髋关节处的滑动界面。研究发现，描述骶髂关节韧带和耻骨联合关节并非必要；这些关节可以在空间上进行刚性固定；对于正常行走负荷情况，可以忽略肌肉力。研究还得出结论，如果要将结果推断到更广泛的人群中，多因素分析应纳入多种骨骼解剖结构。
