基于粒子群算法优化Kriging模型的桥梁钢筋锈蚀可靠度分析

Reliability Analysis of Bridge Steel Corrosion Based on Kriging Method Improved by Particle Swarm Optimization Algorithm

桥梁结构在不利影响因素作用下性能会受到影响,对桥梁的可靠性进行分析具有重要意义。对桥梁钢筋锈蚀开裂可靠度计算方法进行研究,首先,结合实际测试数据提出了基于粒子群算法优化Kriging模型的方法,以解决钢筋锈蚀和保护层开裂与各主要影响因素(钢筋直径、混凝土抗压强度、保护层厚度和沿筋方向裂缝宽度)间的不确定关系。其次,建立了钢筋锈蚀正常使用极限状态的功能函数,并利用Monte-Carlo方法计算桥梁钢筋锈蚀适用性的失效概率和可靠度。结果表明:通过粒子群算法优化后构建的Kriging模型能够更准确地预测钢筋锈蚀程度;桥梁钢筋锈蚀可靠度指标随着钢筋直径、混凝土抗压强度和保护层厚度的增加而增加,随裂缝宽度的增加而减少。

Under the influence of adverse factors,the performance of the bridge structure is seriously affected,it is of great significance to analyze bridge reliability.A new reliability calculation method for bridge steel corrosion was studied in this paper.Firstly,a Kriging model improved by particle swarm algorithm was proposed to simulate and predict bridge steel corrosion combining with the actual test data,and the uncertainty relationship between bridge steel corrosion and the main influence factors(steel bar diameter,concrete compressive strength,the thickness of the protective layer and crack width along the direction of steel)was solved by the Kriging model.Secondly,the performance function of steel reinforcement corrosion was established,and the failure probability and reliability of the bridge steel reinforcement corrosion applicability was calculated by the Monte-Carlo method.The results illustrated that the Kriging model can predict the steel reinforcement corrosion degree more accurately,and the reliability index of steel corrosion increased with the increase of steel diameter,concrete compressive strength and protective layer thickness,but decreased with the increase of crack width.