高速公路路基沉降预测的神经网络模型

软土地基的沉降控制是保证高速公路建设质量的一个关键技术.论文主要介绍了一个对高速公路路基沉降进行预测的神经网络模型.对神经网络的BP算法进行了改进,提高了BP算法的学习收敛速度和网络性能的稳定性.神经网络法预测路基沉降的难点之一是合适的训练样本构造问题,论文提出了新颖独特的“训练样本”构造方法,且应用效果良好.利用路基沉降量实测资料直接建模,采用BP网络计算的改进算法,可较为准确地预测大约4个月之后的沉降量,预测值与实测值吻合较好.

Study on the Neural Network Based Model for Predicting the Settlements of Soft Ground of Highway

One of the key techniques of building highway is to control the settlements of soft ground.The model based on the neural network for predicting the settlements of soft ground of highway is introduced in this paper.Some improved steps for the BP neural network are introduced.The improved BP neural network can increase the convergence speed,and can improve the performance of BP neural network.Neural network can be used for predicting the settlements of soft ground of highway,but one of the difficulties is to construct the structure of the samples.One method to construct the structure of the samples is put forward by the author,and the method is novel and unusual.The neural network based method for predicting the settlements of highway is come true.Making use of surveying data of settlements of highway,the neural network based method can predict the settlements of highway about 4 month later well and truly.