Recognition of patterns: classification and identification of damage in composite materials of sugar cane and bamboo fiber

Abstract

The objective of this work is to demonstrate a method of pattern recognition, formed by the angle, Of the force (N) and deformation (mm), of the mechanical test of the sugarcane and bamboo fibers mercerized with liquid epoxy bicomponent resin. The methodology is based on an algorithm that uses a reference angle and comparesother angles of other mercerized sugarcane and bamboo fibers, with and without resins, tested from the mechanical test experiment. The pattern recognition algorithm is based on parametric statistical theory and shows graphical results of linear regression and correlation of the signals. The result presents two important points of view: the first and most important, the classification and quantification of mercerized sugarcane and bamboo fiber groups, with and without resin, by pattern recognition; the second point of view makes the pattern recognition procedure a method of experimental characterization. The intrinsic result of the classification of the standard groups also shows a possibility of deepening scientific research, that is, this method proposes a concept of opening qualitative secondary scenarios, after classification. Therefore, the objective of this work has been met, and can be applied in the areas of materials, acoustics, vibration, civil and mechanical.