基于GSA-SVR的超宽钢箱梁顶推施工全过程局部最大应力预测

Prediction of Local Maximum Stress of Ultra Wide Steel Box Girder during Incremental Launching Construction Process Based on GSA-SVR

钢箱梁顶推施工中将局部应力控制在一定范围之内,有利于保证施工安全和质量。目前常根据是否产生结构失稳等施工隐患来考虑是否对局部进行加强。采用有限元计算超宽钢箱梁局部应力建模难度较大。为实现局部最大应力的科学预测,采用万有引力搜索算法(GSA)对支持向量回归(SVR)算法进行改进,实现了对惩罚参数和核参数的优化计算。基于GSA-SVR算法构建钢箱梁局部应力预测模型,可实现局部应力的有效预测。将算法和预测模型应用于济南凤凰黄河大桥钢箱梁顶推施工中的局部最大应力预测。结果表明:该模型预测有较高的精度,采用预测模型获得局部最大应力相比有限元计算获得方式效率更高。

The local stress in the construction of steel box girder is controlled within a certain range,which is beneficial to ensure the safety and quality of construction.At present,local reinforcement is often considered according to whether there is structural instability and other potential construction hazards.It is difficult to model the local stress of ultra wide steel box girder by finite element calculation.To achieve scientific prediction of local maximum stress,the gravitational search algorithm(GSA)is adopted to improve the support vector regression(SVR)algorithm,and the optimization calculation of the penalty parameter and the kernel parameter is realized.The local stress prediction model of steel box girder is constructed based on GSA-SVR algorithm,which can realize effective prediction of local stress.The algorithm and the prediction model are applied to the prediction of local maximum stress in the steel box girder incremental launching construction of Jinan Phoenix Yellow River Bridge.The result shows that the model has high prediction accuracy,and the use of the prediction model to obtain the local maximum stress is more efficient than the finite element calculation method.