不同预测方法的路基弹性模量的对比研究

Comparative Study on Subgrade Elastic Modulus Based on Different Predicting Methods

路基弹性模量是道路设计的主要技术指标之一,快速准确评价其数值对道路性能具有重要的意义。利用BP神经网络方法预测路基弹性模量,在此基础上,对路基干湿状态及AASHTO等不同方法也进行了研究,得到了相应的路基弹性模量预测结果。研究内容包括:路基弹性模量数据调查及影响因素分析,不同预测方法的路基弹性模量预测,预测结果与实测结果的对比分析。通过对不同预测结果的对比分析,论证了BP神经网络方法的科学合理性,探讨一种更合理的快速准确预测不同季节路基弹性模量的方法。研究认为,BP神经网络方法可以比较准确地预测不同季节的路基弹性模量,预测结果与实测结果的误差在合理范围内,与其他方法相比,具有一定的优势。研究成果可以取得一定的经济效益与社会效益。

Subgrade elastic modulus is one of the main indexes of road design,the quick and accurate evaluation on its value is of great significance to the road performance. The subgrade elastic modulus has been predicted by BP neural network,on the basis of it,the methods such as wet and dry states of subgrade and AASHTO have also been discussed,and the corresponding predicting results of subgrade elastic modulus have been obtained. The research contents include: analysis on data investigation and influence factors of subgrade elastic modulus,prediction of subgrade elastic modulus based on different predicting methods,comparative analysis on predicting results and testing results. Through the comparative analysis on the different results,BP neural network method has been demonstrated rational,the objective is trying to explore a more reasonable quick and accurate method of predicting subgrade elastic modulus in different seasons. The project has suggested that BP neural network method can accurately predict subgrade elastic modulus in different seasons,and errors of the results are within a reasonable range,it has the certain advantages compared with other methods. The research results can get a certain economic benefit and social benefit.