Four-limb CFST latticed columns seismic performance: experimental and ANN predictions

四肢钢管混凝土格构柱的抗震性能:实验和人工神经网络预测

阅读:1

Abstract

This paper investigates the seismic performance of four-limb concrete-filled steel tubular (CFST) lattice columns with different slenderness ratios (10.8, 10.8, 18.4, and 27.9) and axial load ratios (0.2, 0.3, 0.2, and 0.2), conducting horizontal low-cycle reciprocating load test. Based on the test results, an artificial neural network (ANN) is used to improve the prediction of the seismic performance of four-limb CFST latticed columns, considering ANN's defects of low accuracy prediction and the poor fitting for the load, unload and extreme point of hysteresis curves, the sparrow search optimization algorithm (SSA) is adopted to optimize the ANN's weights and thresholds. Quantum computations are proposed to improve the SSA's iteration convergence and avoidance of local optima, and its convergence curve is compared to circle chaotic sparrow search algorithm and tent chaos sparrow search algorithm (Tent-CSSA). The damage variables are calculated and compared with the predicted results from the SSA, the sparrow search algorithm based on quantum computations and multi-strategy enhancement (QMESSA) and the test results based on the energy damage model and Park's model. The results demonstrate that the load-displacement hysteresis curves of Specimen 1 and Specimen 2 are bow-shaped, which show a strong plastic capacity. The hysteresis curve of Specimen 3 appears in an inverse S-shape, which was due to the slip caused by the low bond strength between concrete and steel pipe. The hysteresis curve of Specimen 4 is pike-shaped, which has a high shear span ratio and obvious bending performance. QMESSA effectively optimizes the weights and thresholds of the ANN, and its predicted damage variables are consistent with those of the conventional damage model. This indicates that QMESSA can effectively predict the load-displacement hysteresis curves of four-limb CFST lattice columns under low-cycle reciprocating loading.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。