Application of artificial neural networks (ANN) to evaluate the influence of internal friction angle (φ) and over consolidation ratio (OCR) of soil on K0 value

摘要

The prediction of coefficient of earth pressure at rest (K_0) of soil is of major importancernin a wide variety of geotechnical problems. Evaluation of the effect of soil mechanical properties,rnsuch as internal friction angle (φ), and stress history, such as over consolidation ratio (OCR), on K_0rnvariation looks to be necessary. Numerous investigations have been carried out and many researchersrnhave presented their relationships based on substantial database of soil tests. This study designs anrnANN model to predict the coefficient of earth pressure at rest of soil in comparison with a database.rnThen, the ANN model has been trained and tested with database. By validating the model, accuracyrnof ANN has been accepted. The sensitivity analysis of model to input variations is the other purposernof this study.