%0 Journal Article
%T 大跨径混凝土斜拉桥线形优化预测研究<br>Study on predication of line optimization of large-span concrete cable-stayed bridges
%A 于景飞
%A 吴炎奎
%A 苏醒&lt
%A br&gt
%A YU Jingfei
%A WU Yankui
%A SU Xing
%J 铁道科学与工程学报
%D 2018
%X 大跨径混凝土斜拉桥施工工序复杂，施工过程中受到诸如拉索垂度、温度变化、混凝土收缩徐变效应等非线性因素影响，使立模标高的设定存在较大误差，理想的成桥状态难以实现。为使成桥线形准确，结构受力均衡，综合分析影响施工立模标高的因素，建立基于粒子群优化算法的BP神经网络立模标高修正参数的预测模型，对修正值进行预测。研究结果表明：所建模型性能稳定，具有较好的预测泛化能力。由预测结果得到的线形更接近理想成桥线形，主梁结构受力合理，能够实现较好的成桥状态，为斜拉桥主梁线形优化方法提供参考。<br>The large-span concrete cable-stayed bridge has rather complex construction procedure. It is affected by non-linear factors such as cable sag, temperature variation as well as shrinkage and creep of concrete etc in the construction process, which then leads to relatively large error of setting of formwork erection elevation. In this case, the ideal completed bridge state can hardly be achieved. In order to guarantee accurate bridge line and balanced structure stress, factors influencing the formwork placing elevation of construction are analyzed comprehensively. A predictive model of the modified value of formwork placing elevation based on the PSO-BP neural network is established, so as to predict the modified value of the formwork placing elevation. The simulation results indicated that the established model with stable performance has very good predication and generalization abilities. Lines obtained from forecasted results were closer to ideal completed bridge lines. The main girder structure with reasonable stress achieved favorable completed bridge state, which also provides certain reference for line optimization methods of cable-stayed bridges
%K 立模标高
%K 斜拉桥施工
%K 粒子群算法
%K 神经网络
%K 线形优化&lt
%K br&gt
%K formwork placing elevation
%K cable-stayed bridge construction
%K particle swarm optimization
%K neural network
%K line optimization
%U http://www.jrse.cn/paper/paperView.aspx?id=paper_317530