论文部分内容阅读
正确选用计算参数对提高土工数值计算的准确度至关重要。先基于ADINA有限元分析法得到不同路堤力学参数与路堤沉降关系的人工神经网络训练样本,然后用BP神经网络对样本进行学习训练,采用训练后的BP神经网络对实测沉降进行网络仿真,得到了路堤的参数。通过对比分析,认为采用BP神经网络进行力学参数反演具有较高准确性,为正确预测路堤沉降提供了可靠依据。
Correct choice of calculation parameters to improve the accuracy of geotechnical numerical calculations is essential. Firstly, the artificial neural network training samples with different embankment mechanics parameters and embankment settlements were obtained based on ADINA finite element analysis method, and then the BP neural network was used to train the samples. The trained BP neural network was used to simulate the measured settlement. Embankment parameters. Through comparative analysis, it is considered that BP neural network is more accurate for the inversion of mechanical parameters, which provides a reliable basis for the correct prediction of embankment settlement.