Xiang Shen1, Hongfei Zhu2, Jiabao Jiang2, Yongquan Deng2, Song Ji2. 1. Department of Mechanical Engineering, Jiangsu University, No. 301, Xuefu Road, Zhenjiang, China. sx@ujs.edu.cn. 2. Department of Mechanical Engineering, Jiangsu University, No. 301, Xuefu Road, Zhenjiang, China.
Abstract
PURPOSE: Recent studies suggested that suboptimal delivery and longitudinal stent deformation can result in in-stent restenosis. Therefore, the purpose of this paper was to study the effect of stent geometry on stent flexibility and longitudinal stiffness (LS) and optimize the two metrics simultaneously. Then, the reliable and accurate relationships between metrics and design variables were established. METHODS: A multi-objective optimization method based on finite element analysis was proposed for the investigation and improvement of stent flexibility and LS. The relative influences of design variables on the two metrics were evaluated on the basis of the main effects. Three surrogate models, namely, the response surface model (RSM), radial basis function neural network (RBF), and Kriging were employed and compared. RESULTS: The accuracies of the three models in fitting flexibility were nearly similar, although Kriging made more accurate prediction in LS. The link width played important roles in flexibility and LS. Although the flexibility of the optimal stent decreased by 13%, the LS increased by 48.3%. CONCLUSIONS: The obtained results showed that the multi-objective optimization method is efficient in predicting an optimal stent design. The method presented in this paper can be useful in optimizing stent design and improving the comprehensive mechanical properties of stents.
PURPOSE: Recent studies suggested that suboptimal delivery and longitudinal stent deformation can result in in-stent restenosis. Therefore, the purpose of this paper was to study the effect of stent geometry on stent flexibility and longitudinal stiffness (LS) and optimize the two metrics simultaneously. Then, the reliable and accurate relationships between metrics and design variables were established. METHODS: A multi-objective optimization method based on finite element analysis was proposed for the investigation and improvement of stent flexibility and LS. The relative influences of design variables on the two metrics were evaluated on the basis of the main effects. Three surrogate models, namely, the response surface model (RSM), radial basis function neural network (RBF), and Kriging were employed and compared. RESULTS: The accuracies of the three models in fitting flexibility were nearly similar, although Kriging made more accurate prediction in LS. The link width played important roles in flexibility and LS. Although the flexibility of the optimal stent decreased by 13%, the LS increased by 48.3%. CONCLUSIONS: The obtained results showed that the multi-objective optimization method is efficient in predicting an optimal stent design. The method presented in this paper can be useful in optimizing stent design and improving the comprehensive mechanical properties of stents.