股骨骨折一般醫療方法為植入骨板與骨釘，為了減少患者之負擔與醫院所花之時間成本，本研究依據實體骨板與骨釘樣品建立模擬模型，並進行最佳化分析來減少可能損耗之成本。模擬模型使用COMSOL Multiphysics多物理場耦合軟體進行有限元素分析；實體模型使用實際Sawbones樣品與MTS Criterion 43靜態萬能材料試驗機進行實體試驗，考慮緩慢下壓(3、4、5 mm/min)之受力形式，從中探討含植體之股骨位移與應變之力學研究，並將模擬結果與實驗結果進行驗證，結果證明模擬模型具有一定可靠性。 利用此模擬模型進行股骨骨折之骨釘數量與位置的最佳化分析，其方法分為基因演算法與部分因子設計法之迴歸分析，因子為股骨髁處六根骨釘，其編碼符號為A至F，並依據醫師的經驗法則與降低應力屏蔽效應設置限制條件與目標函數。根據基因演算法之分析結果可知，拿掉骨釘A與D時可得骨板最大應力的最小值；迴歸分析法之結果可得一個多項式來取代繁雜的有限元素分析，並可知各骨釘對骨板應力之顯著性，其預測需拿掉之骨釘也與基因演算法相同。從上述結果可知此骨折模型之最佳骨釘位置與骨板的應力分布，由此提供醫師術前規劃的初步參考資訊。 ;A common medical treatment for femoral fracture is the implantation of bone plates and screws. To reduce patients’ burden and hospitals’ time cost, this study constructed a simulation model according to bone plate and screws samples and performed optimization analysis to reduce potential costs. The simulation model is the CAD model of Sawbones, and the finite element analysis was performed using COMSOL Multiphysics software; the experimental model is the actual Sawbones sample, and the solid mechanics was performed by using MTS Criterion 43. Considering the force of downward pressure (3,4,5 mm/min) on femur, this study investigated the mechanics regarding the displacement and strain of femur with implants. Verification through the simulation and experimental results confirmed that the simulation model possessed satisfactory reliability. The proposed simulation model was used to optimize the number of bone screws required in the femoral fracture model. The methods used were the genetic algorithm and regression analysis based on the fractional factorial design. The design variables are the six screws at femoral condyle. Their symbolic coding are A to F. The constraints and objective functions were set up by the surgeon’s rule of thumb and the stress shielding reduction. According to the results of the genetic algorithm, the minimum stress of the bone plate can be found when the screws A and D are removed. The results of the regression analysis can obtain a simple polynomial to replace the complicated finite element analysis, and the significance of each screws to the stress of the bone plate can be known. This method predicts the screws which need to be removed is the same as the genetic algorithm. According to the analysis results of the femoral fracture model, the optimal number and location of screws at femoral condyle and the stress distribution of the bone plate can be known. It can provide preliminary reference information for clinicians before surgery.
