Reliability analysis for mechanical properties of structural bamboo

Abstract

Bamboo is a good natural nonhomogenous organic substance exhibiting large variations of physical properties along the length of its members, such as external and internal diameters, dry density, and moisture content. This chapter presents a pilot study carried out to examine the variation of compressive strength against several physical properties of bamboo members—external diameter, wall thickness, dry density, and moisture content.Some of these physical parameters vary significantly along the length of bamboo members, and it is highly desirable to establish any correlation between the physical and the mechanical properties of bamboo members. The backpropagation neural network (BPNN) has been successfully applied to analyze the relationship between the physical and the mechanical properties for Mao Jue. A total of forty pilot test data were complied and used to train a neural network (NN) model under the control of ten randomly selected test data. The trained NN model with an independent production set, the sensitivity and reliability analyses were performed on the NN model to gain insight on how physical properties of bamboo influence its compressive strength and Young's modulus, and to facilitate the design of bamboo structures at a known level of confidence against failure. This new NN-based approach has been proven to be a reliable and cost-effective means for studying mechanical properties of structural bamboo and facilitating the development of associated design methods.