Fault diagnosis is of great significance for ensuring the safety and reliable operation of rolling bearing in industries. Stack autoencoder (SAE) networks have been widely applied in this field. However, the model parameters such as learning rate are always fixed, which have an adverse effect on the convergence speed and accuracy of fault classification. Thus, this paper proposes a dynamic learning rate adjustment approach for the stacked autoencoder network. First, the input data is normalized and enhanced. Second, the structure of the SAE network is selected. According to the positive and negative value of the training error gradient, a learning rate reducing strategy is designed in order to be consistent with the current operation of the network. Finally, the fault diagnosis models with different learning rate adjustment are conducted in order to validate the better performance of the proposed approach. In addition, the influence of quantities of labeled sample data on the process of backpropagation is analyzed. The results show that the proposed method can effectively increase the convergence speed and improve classification accuracy. Moreover, it can reduce the labeled sample size and make the network more stable under the same classification accuracy.
In this study, nine square concrete‐filled steel tubular (SCFST) columns considering longitudinal stiffeners as parameters were produced to investigate the axial compression mechanical properties of SCFST columns in the condition of constant steel consumption. This included one specimen without stiffeners, three specimens stiffened with one stiffener on each tube wall, two specimens stiffened with two stiffeners on each tube wall, two specimens stiffened with three stiffeners on each tube wall and one specimen stiffened with four stiffeners on each tube wall. Simultaneously, numerical simulations were performed under identical conditions using ABAQUS software. The results showed that the longitudinal stiffeners can act as an out‐of‐plane restriction support to delay the development of local buckling of the steel tube, which indirectly improved the constraint effect of steel tube on core concrete. First, both the number of local buckling and the ultimate displacement of the specimens grew correspondingly when the number of stiffeners increased. As the sub‐wall width‐to‐thickness ratio was increased, the steel tube became prone to local buckling. Second, the installation of stiffeners decreased the confinement effect of steel tube on the core concrete, in which the sub‐wall width‐to‐thickness ratio and the stiffener width‐to‐thickness ratio had a great influence on the ultimate bearing capacity of the specimens. Further, the experimental outcomes indicated that the bearing capacity of stiffened specimens was less than that of the unstiffened specimen. The bearing capacity of the specimens stiffened with three and four stiffeners was higher than that of the specimens stiffened with one and two stiffeners. Third, the stiffener width‐to‐thickness ratio had a considerable impact on the ductility of the specimens. The ductility of specimens stiffened with one and two stiffeners was greater than that of the specimens stiffened with three and four stiffeners. Finally, the finite element simulations were consistent with the experimental process. The confinement factor of stiffened specimens became enlarged when either the number of stiffeners in the same thickness condition increased or the thickness of steel tube wall under condition of the same number of stiffeners grew.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.