سامانه همایش های دانشگاه آزاد اسلامی واحد نجف آباد

صفحه اصلی / International Conference on Artificial Intelligence; City, Industry and Health

Application of ANN artificial network in slope behavior evaluation using machine learning technique

نویسندگان :

Marziyeh Tourani¹ Hadi Bahadori²

1- PhD candidate, Department of Civil Engineering, Urmia University, Urmia, Iran 2- Professor, Geotechnical Engineering Group, Department of Civil Engineering, Urmia University, Urmia, Iran

کلمات کلیدی :

Machine learning،artificial network،slope stability،safety factor،ANN

چکیده :

Ensuring safety has always been a challenge in geotechnical engineering, as soils are inherently subject to variability and uncertainty. Evaluating slope behavior under load is an important aspect of geotechnical engineering, as it directly affects the safety and stability of slopes in various construction and mining projects. The integration of artificial intelligence (AI) and machine learning (ML) in geotechnical engineering has evolved significantly. Machine learning techniques, including artificial neural networks (ANN), are used in various geotechnical applications, from estimating the factor of safety (FoS) to predicting slope stability. In this study, the capabilities of artificial neural networks (ANN) in predicting the factor of safety (FOS) of slopes were investigated. The data used included 349 samples from previous studies, of which 259 valid data remained after removing invalid data. Initially, the input parameters including soil specific gravity, cohesion, internal friction angle, slope angle, slope height and pore pressure ratio were selected. However, initial investigations showed that the correlation of these parameters with the factor of safety is weak. Therefore, based on the principles of fluid mechanics, the dimensionless parameters c / γh, (tan(∅)) ⁄ (tan (β)) and p / γh were introduced. Finally, the results showed that the artificial neural network (ANN) performed well because it provided the lowest MAE (62%), SSE (0.21%), MSE (0.27), and RMSE (0.46), and the coefficient of determination (R2) reached 0.946