Showing 2 results for Gep
Maryam Nikooee, Ali Noorzad, Kaveh Ahangari,
Volume 7, Issue 2 (3-2014)
Abstract
Determination of stress in earthfill dams is one of the most important parameters in dam safety studies. Stress monitoring can be done using total pressure cells which are typically installed during construction. The cell is installed with its sensitive surface in direct contact with the soil to measure total stress of soil and in combination with piezometers to measure pore-water pressure acting in the soil mass. Total pressure cells needs to be installed with care to get reasonable measurements. However, measurements are often incompatible with the theoretical predictions such that pressure cell results usually have some inaccuracies. There are several parameters effecting pressure cell errors. However, in the present paper it is only focused on the height of embankment and the duration of dam construction. For this purpose, a case study, namely Alborz embankment dam located in northern part of Iran was studied. It is an earth dam with clay core with a height of 78 m. Using the monitoring data and considering the effect of embankment height and construction period parameters, a model is presented to predict the pressure cells error with Gene Expression Programming (GEP) procedure by GeneXProTools 4.0 software. The computed coefficient of correlation (R2) for the proposed model is 0.98 showing a good agreement with the monitoring data. The obtained results indicate that the ratio of height difference to time difference for Alborz dam has a significant role in dam pressure cells errors
Sayed Rahim Moeinossadat, Kaveh Ahangari, Danial Behnia,
Volume 9, Issue 1 (6-2015)
Abstract
The present study aims to employ intelligent methods to predict shear wave velocity (Vs) in limestone. Shear wave velocity is one of the most important rock dynamic parameters. Direct determination of this parameter takes time, cost and requires accuracy as well. On the other hand, there is no precise equation for indirect determination. This research attempts to provide some simulations to predict Vs using the information obtained several dams located in Iran, using different approaches, including adaptive neuro-fuzzy inference system (ANFIS) and gene expression programming (GEP). 136 datasets were utilized for modeling and 34 datasets were used for evaluating its performance. Parameters such as Compressional wave velocity (Vp), density (g) and porosity (n) were considered as input parameters. The values of R2 and RMSE were 0.958 and 113.620 for ANFIS, where they were 0.928 and 110.006 for GEP respectively. With respect to the accuracy of the intelligent methods, they can be recommended for future studies
