高层建筑物沉降预测方法及对比分析

Settlement Prediction Methods and Comparative Analysis of High-Rise Buildings

对高程建筑物进行沉降预测对于保证人民群众生命财产安全具有重要意义。针对广州市某高层建筑物的历史沉降数据开展试验,分别采用多项式拟合法、灰色GM(1,1)模型和BP神经网络模型等方法进行数据处理和沉降预测分析,并从算法实时性、小样本情况下的预测性能和低信噪比条件下的稳健性等3个方面对方法的适用性进行了对比分析。结果表明,多项式拟合法算法简单,实时性最好;灰色GM(1,1)模型对训练样本数的要求最低,在小样本情况下能够获得最高的预测精度;BP神经网络的非线性预测能力最强,具有最强的噪声稳健性。本文的对比分析结果可供工程实践中参考使用。

It is of great significance to predict the settlement of high-rise buildings to ensure the safety of people’s lives and property.Based on the historical settlement data of a high-rise building in Guangzhou,this paper uses polynomial fitting method,grey model and BP neural network model for data analysis and settlement prediction.On this basis,three evaluation indexes are proposed to compare and analyze the application effect of the above methods in practical engineering practice from different dimensions.The results show that BP neural network has the strongest nonlinear prediction ability and the strongest noise robustness;the grey model has the lowest requirement for the number of training samples,and which can obtain the highest prediction accuracy in the case of small samples;the polynomial fitting algorithm is simple and has the best real-time performance.The comparative analysis results of this paper can be used for reference in practical engineering application.