Weibull Statistic and Artificial Neural Network Analysis of the Mechanical Performances of Fibers from the Flower Agave Plant for Eco-Friendly Green Composites

ABSTRACT

The research conducted focused on examining the unique properties of Agave Americana Flower Stem fiber (AAFS), particularly its behavior under quasi-static tensile conditions. A total of 200 AAFS fibers were subjected to tensile tests using a standard gauge length of 40 mm. Tests spanned seven groups with quantities (N) ranging from 30 to 200. The study aimed to understand the fibers’ mechanical traits, as tensile resistance and modulus of elasticity, and to see how different test quantities influence these properties. A significant observation was the dispersion of the tensile characteristics of AAFS fibers, a common trait of natural fibers. To understand this, we applied rigorous statistical tools, including the Weibull distribution at a 95% confidence interval and one-way ANOVA. A mathematical model was produced utilizing data from experiments regarding the tensile behavior of AAFS fibers. The ANN provided correlation coefficients (R²) of 0.9897, 0.9971, 0.9993, and 0.9939 for training, validation, testing, and all datasets respectively, which were able to accurately predict the experimental data. The proposed model would be of tremendous assistance to engineers and designers in obtaining green composite materials that are based on natural fibers and thereby more durable. These methods illuminated the patterns in our results, enriching our understanding of AAFS fiber mechanics.

摘要

本研究的重点是研究美国龙舌兰花茎纤维（AAFS）的独特性能，特别是其在准静态拉伸条件下的行为. 使用40 mm的标准标距长度对总共200根AAFS纤维进行拉伸试验. 测试分为七组，数量（N）从30到200. 这项研究旨在了解纤维的力学特性，如抗拉强度和弹性模量，并了解不同的测试量如何影响这些性能. 一个重要的观察结果是AAFS纤维的拉伸特性的分散，这是天然纤维的一个常见特性. 为了理解这一点，我们应用了严格的统计工具，包括95%置信区间的威布尔分布和单因素方差分析. 利用有关AAFS纤维拉伸行为的实验数据建立了数学模型. 人工神经网络对训练、验证、测试和所有数据集的相关系数（R2）分别为0.9897、0.9971、0.9993和0.9939，能够准确预测实验数据. 所提出的模型将极大地帮助工程师和设计师获得基于天然纤维的绿色复合材料，从而更耐用. 这些方法阐明了我们结果中的模式，丰富了我们对AAFS纤维力学的理解.

KEYWORDS:

• Agave americana
• tensile behavior
• statistical methods
• Weibull statistic
• hazard function

关键词:

• 美洲龙舌兰
• 拉伸行为
• 统计方法
• 威布尔统计
• 危险功能

Acknowledgments

The authors are thankful to the Deanship of Scientific Research at Najran University for funding this work under the Research Groups Funding Program (grant code: NU/RG/SERC/12/24).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The work was supported by the Deanship of Scientific Research at Najran University [NU/RG/SERC/12/24].