【摘要】 为提升麻纤维力学性能评价效率，以红麻、大麻为例，采用湿化学分析法和单纤维强度测试法对27个麻纤维样本进行化学成分含量和纤维强度测定，采用主成分分析、聚类分析对以3个主要化学成分替代整体化学成分实现对麻纤维强度响应的可行性分析，分别以整体化学成分和3个主要化学成分为自变量，采用支持向量回归模型对纤维强度进行预测，并对预测效果进行评价。结果表明：主成分数量为3时的累计贡献率达94.48%,分别以主成分纤维素、半纤维素、木质素3个主要化学成分为分类依据，所得聚类结果与以全部化学成分为分类依据所得聚类结果的一致性分别为96.3%和92.3%,支持向量回归模型校正样本集的内部交叉验证效果好，对于未知麻纤维样本的预测相对误差均值分别为1.78%和2.19%;利用麻纤维3个主要化学成分可替代全部化学成分实现基于支持向量回归模型的麻纤维强度的预测。更多还原

【Abstract】 Objective In order to improve the efficiency of evaluating the mechanical properties of bast fibers, wet chemical analysis and single fiber strength testing methods were used to determine the chemical composition content and fiber strength of 27 kenaf and hemp fiber samples. The differences in chemical composition content and mechanical properties of these bast fibers were analyzed.Method The feasibility of replacing the overall chemical composition with three main chemical components to achieve strength response of bast fiber was analyzed using principal component analysis and cluster analysis. The overall chemical composition and three main chemical components(cellulose, hemicellulose, and lignin) were used as independent variables, and support vector regression model was used to predict fiber strength. The prediction effect of bast fiber strength was evaluated.Results The results of principal component analysis showed that when the number of principal components was 3, the cumulative contribution rate of principal components reached 94.48%, which basically reflects the response of fiber chemical composition to fiber mechanical properties in the original population sample data. Cluster analysis was conducted on bast fiber samples using all chemical components, principal components, and the main chemical components of cellulose, hemicellulose, and lignin as indicators. After classification, significant differences were observed in the mean mechanical properties of each type of fiber. The consistency between the clustering results based on principal components and those based on all chemical components was 96.3%. The consistency between the clustering results obtained based on the main chemical components cellulose, hemicellulose, and lignin as indicators and the clustering results obtained based on the classification of all chemical components was 92.3%. A support vector regression model was constructed with the overall chemical composition and three main chemical components as input variables, and bast fiber strength as output variable. The model performed well in internal cross validation of the corrected sample set, with mean relative prediction errors of 1.78% and 2.19% for unknown bast fiber samples, respectively.Conclusion The research results proved that cellulose, hemicellulose, and lignin as the three main chemical components, are able to replace all chemical components to reflect the mechanical properties of bast fibers. The use of the three main chemical components of bast fibers can replace the overall chemical composition to achieve the prediction of bast fiber tensile strength based on support vector regression model.更多还原