Classification and prediction of flexural properties of bamboo slices made from flattened bamboo with a gradient structure based on GA-BP neural network model

Abstract

Bamboo slices (BS) have been successfully used in bamboo winding composites due to their excellent flexural properties. However, BS with a gradient structure is prone to breakage easily during the winding process, particularly when varying thicknesses are involved. This study investigated flexural properties and size effects of BS, as well as prediction of their flexural properties. A mathematical relationship was established between fiber content of BS and its radial position, revealing an exponential decay function with an average fit quality of 0.9. The flexural strength, flexural modulus and radius of curvature of BS increased with higher fiber content. However, for BS with a thickness of 0.7 mm, the radius of curvature exhibited an inverse relationship with fiber content when the load was applied to the side of BS with fewer vascular bundles. Analysis indicated that BS with thicknesses of 0.3 mm and 0.5 mm can be considered homogeneous materials, while BS with a thickness of 0.7 mm retained the gradient structure and properties of bamboo culm wall. Additionally, BS showed a significant size effect, where thicker BS displayed lower strength due to defect effect, variation of length-thickness ratio and hoop effect. At last, a GA-BP neural network model was developed and validated as an effective tool for predicting BS flexural properties based on their radial position, achieving an accuracy of over 95 %. This study provides valuable insights into the flexural properties and size effects of BS, providing a scientific foundation and technical support for the development of bamboo winding products.