大跨径连续梁桥静载试验变形的快速预测技术研究  

Research on Fast Prediction Technology of Static Load Test Deformation of Long-Span Continuous Beam Bridge

作　　者：王培金 WANG Pei-jin(Shandong Provincial Communications Planning and Design Institute,Shandong 250031,China)

机构地区：[1]山东省交通规划设计院,济南市250031

出　　处：《北方交通》2020年第3期33-36,共4页Northern Communications

摘　　要：大跨径PC连续梁桥静载试验过程中的结构分析是比较耗时耗力的工作,容易出现错误。根据已有数据,利用模型树机器学习算法,开发相应程序并训练得到相应的预测模型。利用该预测模型,可以快速预测静载试验状态下的桥梁跨中下挠的理论值,预测结果可以作为“校核参考值”,用于结构分析结果的校核工作。测试数据表明,该算法预测误差较小,算法本身具有很好的工程应用价值。Structural analysis of long-span PC continuous girder bridge for static load test is a time-consuming and labor-consuming work,which is prone to errors.This paper,by using existing data,uses model tree machine learning algorithm to develop the corresponding programs and train the prediction model.Using the prediction model,the theoretical value of mid-span deflection of a test bridge can be predicted quickly,and the predicted value can be used as a"check reference value"to check the results of structural analysis.The test data show that the prediction error is small,which indicates that the algorithm itself has good engineering application value.

关 键 词：模型树 连续梁桥 静载试验 桥梁工程 机器学习 

分 类 号：U448.215[建筑科学—桥梁与隧道工程]