等维BP神经网络在建筑物沉降预测中的应用探讨

The Research on the Equal Dimensional BP Neural Network Applied in Sedimentation Forecast of Structure

基于神经网络在建筑物沉降预测中的局限性,许多学者对此进行了探讨和研究,文中采用灰色理论中的等维新息思想构建训练样本,建立了等维BP神经网络,通过采用MATLAB数学工具编程实现,对实际的沉降量进行变形预测。结合具体工程实例进行分析,实验结果表明该模型比灰色GM(1,1)模型具有更好的预测效果,能够满足实际应用的需要。

Based on the limitation of neural network in settlement prediction ,many scholars have discussed and stud-ied ,this paper establishes a new neural network model ,which based on the concept of equal dimension and new informa-tion in grey theory , and applies this model in the practical deformationing forecast whith the mathematical instrument MATLAB.Combining with specific engineering example analysis ,the experimental results show this model has a more for-casting precision than grey GM （ 1,1） model,which can meet the demand of application .