Characterization and Modeling of Corn Stalk Fibers tied with Clay using Support Vector Regression Algorithms

Abstract

Several research groups are recently focusing on natural fibers as components of construction materials, contributing to the search for sustainable solutions that reduce the ecological footprint of buildings. Many of these fibers are proposed as acoustic absorbers to replace man-made fibers widely used to reduce reverberation in rooms, such as fiberglass and stone wool, which consume a lot of energy in their production and are not biodegradable. In this article, the acoustic absorption of fiber panels composed of corn stalk fibers and clay – both environmentally friendly materials – is studied, considering samples of 6 mm, 12 mm, and 24 mm thickness. Three percentages of water were used for the kneading of the clay. A support vector machine model has been calculated to predict the behavior of this composite material. 24 mm sample with 6% of water returns values of the acoustic absorption coefficient between 0.6 and 0.8 in the frequency range from 750 to 1600 Hz. 6 mm samples with 16% and 26% of water result in values of the acoustic absorption coefficient near one at 4500 Hz and 4750 Hz, respectively. The simulation performed with the support vector machine model returned Pearson’s correlation coefficient values of 0.997, demonstrating excellent generalization and prediction ability of the model.

Keywords

• 天然材料
• 吸声系数
• 向量机回归模型
• 声学测量