关键词: FRP混凝土组合构件, FRP管壁厚, 管径, 混凝土强度, 空心率, 含钢当量, 人工神经网络

Abstract:

This paper conducted an axial compression test on 32 group of FRP concrete composite structure with 3 kinds of section forms (FRP pipe-concrete-steel section, FRP pipe-concrete-steel pipe and hollow FRP pipe concrete). The artificial neural network was used to analysis the experimental dates. The increase of ultimate bearing capacity values was used as mechanical performance evaluation indicators. The effects of five parameters on the mechanical performance, such as the thickness of FRP pipes, the pipe diameter, the hollow ratio, the concrete strength and the steel ratio, were the emphasis of the investigation. The research results show that the most influential parameters of mechanical performance are the thickness of FRP pipes and the pipe diameter, followed by concrete strength, the difference between the three is insignificant, and the influence of hollow ratio and steel ratio are the lowest. Both the increase of the thickness and diameter of FRP pipe, The mechanical performance is significantly improved. The overall trend of ultimate bearing capacity values is decreasing with increasing the concrete strength. But under different conditions, the trend of ultimate bearing capacity values is also increasing or increased at first and then decreased, which effected by the concrete strength. With the increase of the hollow ratio and steel ratio, the overall mechanical performance decrease and the downward trend is relatively slow. The mechanical performance of small wall thickness member is decreased significant with the increase of hollow ratio. But the impact of large wall member components is not obvious. With the increase of steel ratio is significantly decreased the mechanical performance of small thickness and high concrete strength member, but the mechanical performance of big thickness and high concrete strength member is significantly increased.

Key words: FRP concrete composite structure, The thickness of FRP pipes, Pipe diameter, Concrete strength, Hollow ratio, Steel ratio, Artificial neural network