参数化人体有限元模型的体表样本选取方法
A selection method of body surface sample for parametric finite element model of human body
吕华溢 1许述财 1宋家锋 2王波 1张金换 1赵翔3
作者信息
- 1. 清华大学 智能绿色车辆与交通全国重点实验室,北京 100084
- 2. 清华大学 智能绿色车辆与交通全国重点实验室,北京 100084;吉林大学 工程仿生教育部重点实验室,长春 130022
- 3. 中国地质大学(北京)工程技术学院,北京 100083
- 折叠
摘要
为了准确选取到参数化方法建立的中国女性人体有限元模型的体表样本,设计了一种用于人体体表建模的样本选取方法.通过采用样本估计总体的方法确定体表样本的选取数量;采用计算正态分布区间估计参数和按照数据分布比例方法生成体表样本的身高、BMI和年龄3 个统计特征参数;利用拉丁超立方抽样方法按照体表样本统计特征的参数分布生成各样本的选取参数;通过Pearson相关系数检验各特征参数向量间的相关性.基于所设计的体表样本参数生成方法,在置信度为95%、误差不大于10%的条件下,所生成的97 例女性体表样本各特征参数所构成向量之间无显著相关性,各特征参数分布与中国女性人体的总体分布相一致.所设计的体表样本选取方法可有效选取到代表中国女性人体特征的体表样本.
Abstract
A sample selection method for human body surface modeling is designed for accurately selecting body surface samples used for parameterization methods to build finite element models with female Chinese body characteristics.The method of population estimation based on samples is applied to determine the reasonable number of body surface samples;three statistical characteristic parameters of height,body mass index and age of body surface samples are generated by the interval estimation parameter calculation of normal distribution and the data distribution proportion;the selection parameters of each sample are generated by Latin hypercube sampling method according to the distribution of statistical characteristic parameters of body surface samples;Pearson correlation coefficient is employed to test the correlation between the vectors of characteristic parameters.Based on the designed generation method for body surface samples parameters,under the condition of 95%confidence and an error less than 10%,there is no significant correlation among the vectors formed by each characteristic parameter of 97 female body surface samples and the distribution of each characteristic parameter is consistent with the overall distribution.The designed selection method for body surface samples effectively selects body surface samples representing the characteristics of female Chinese.
关键词
样本选取方法/拉丁超立方抽样/中国人体特征/参数化建模Key words
sample selection method/Latin hypercube sampling/body characteristics of Chinese/parametric modeling引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据