基于多元线性回归桥梁温度响应预测研究

Prediction Study of Bridge Temperature Response Based on Multiple Linear Regression

摘要: 为准确预测桥梁温度响应，提高桥梁的整体稳定性和耐久性，利用多元线性回归，开展了桥梁温度响应预测研究。首先，布置多组传感器，采集桥梁温度与应变数据；其次，计算桥梁温度作用，得出温度变化对桥梁结构的影响程度和范围；在此基础上，基于多元线性回归建立桥梁温度响应预测模型，结合桥梁当前温度状态，预测桥梁未来的温度响应。实验结果表明，提出方法应用后，拟合效果和显著性优势显著，在所有时间点的温度响应预测均方误差均较小，具有较高的预测精度和稳定性。

Abstract: In order to accurately predict the temperature response of the bridge and improve the overall stability and durability of the bridge, a study on the prediction of the temperature response of the bridge was carried out by utilizing multiple linear regression. Firstly, multiple sets of sensors are arranged to collect the bridge temperature and strain data; secondly, the bridge temperature action is calculated to derive the degree and range of the influence of the temperature change on the bridge structure; based on this, a bridge temperature response prediction model is established based on multivariate linear regression, which combines with the current temperature state of the bridge and predicts the future temperature response of the bridge. The experimental results show that after the application of the proposed method, the fitting effect and significance advantage are significant, and the mean square error of temperature response prediction at all time points is small, with high prediction accuracy and stability.