论文部分内容阅读
为明确大跨度悬索桥扁平钢箱梁温度和温度梯度特征,以南溪长江大桥正交异性钢箱梁为研究对象,基于健康监测系统中温度传感器的长期实测数据,采用分段函数描述环境温度和日照辐射共同作用下钢箱梁日温度变化曲线。在此基础上,采用高斯混合模型描述钢箱梁年温度多峰概率分布,并引入赤池信息判别准则(AIC)和贝叶斯信息判别准则(BIC)确定最优高斯分量数。统计钢箱梁一年日温差极值并进行参数评估,得到钢箱梁年温差极值分布模型。对年温差极值分布函数进行外推,得到设计基准期温差的极值分布函数并计算温差标准值。引入相关系数分析法对各温差组进行相关性分析,剔除实际不存在的温差模型。研究结果表明:相比正弦函数,分段函数能更准确地描述太阳辐射作用下箱梁截面日温度变化特征;当高斯分量数为3时,混合高斯模型拟合钢箱梁年温度概率分布最优;外推设计基准期模型能够较好地计算设计基准期温差标准值;通过相关性分析剔除了4组不存在的温差模式;得到顶板和腹板各8组温差模式;最后与《公路桥涵设计通用规范》(JTG D60—2015)中钢混结构竖向正温差设计值进行对比,一、二级温度梯度与设计规范分别相差4.2℃和2.3℃。
In order to clarify the temperature and temperature gradient characteristics of the flat steel box girder of the long-span suspension bridge, taking Nanxi Changjiang River Bridge as an example, based on the long-term measured data of the temperature sensor in the health monitoring system, Daily Temperature Variation Curve of Steel Box Girders under Radiation. On this basis, the Gaussian mixture model is used to describe the multi-peak probability distribution of annual temperature of steel box girders. The AIC and BIC are introduced to determine the optimal Gaussian components. Statistical analysis of steel box girder extreme temperature difference of one year and parameter evaluation, obtained annual distribution of steel box girder extreme temperature difference model. Extrapolation of extreme temperature distribution function of annual temperature, the extreme distribution function of temperature difference in design base period is obtained and the standard value of temperature difference is calculated. Correlation coefficient analysis method was introduced to analyze the correlation of each temperature difference group and remove the temperature difference model which did not exist in practice. The results show that, compared with the sine function, the piecewise function can describe the daily temperature variation characteristics of box girder section more accurately. When the Gaussian component is three, the probability distribution of the annual temperature distribution of the composite box girder with Gaussian mixture model is the most The extrapolated design base period model can well calculate the standard value of temperature difference during design base period. The four non-existing temperature difference patterns are eliminated through correlation analysis. The eight sets of temperature difference patterns of roof and web are obtained. Finally, (JTG D60-2015), the first and second temperature gradients are different from the design specifications by 4.2 ℃ and 2.3 ℃, respectively.