Evaluating the predicting performance of indirect methods for estimation of rock mass deformation modulus using inductive modelling techniques

In current era of research in the field of rock engineering the researchers gaining more interest in estimation of deformation modulus through different empirical models rather than in situ tests. Because the empirical models are simple, cost effective and required limited input data for estimation...

Full description

Saved in:
Bibliographic Details
Published in:Journal of Himalayan earth sciences 2018, Vol.51 (1), p.61-74
Main Authors: Hussain, Sajjad, Khan, Mujahid, Ur Rahman, Zahid, Mohammad, Noor, Raza, Salim, Tahir, Muhammad, Ahmad, Ishaq, Sherin, Saira, Khan, Naseer Muhammad
Format: Article
Language:English
Subjects:
Aquifers
Civil engineering
Classification
Deformation
Economic models
Elasticity
Geology
Laboratories
Mechanics
Methods
Mining engineering
Neural networks
Researchers
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In current era of research in the field of rock engineering the researchers gaining more interest in estimation of deformation modulus through different empirical models rather than in situ tests. Because the empirical models are simple, cost effective and required limited input data for estimation of rock mass deformation modulus. The 109 (75% of total data sets) were used for training of both the models while 37(25% of total data sets) were used for validation and testing of models. Since there were three input variables for Nicholson and Bieniawski and Carvalho while two input variables for all the remaining models as described in Table 2, therefore the same were used as input variables for development of inductive models. The individual as well as the combined plots showed that ANN based models performance is better as compared to MLR based model because the points for earlier are located close to the 45 degree line as compared to the later one. [...]it supplements the result that we obtained from the tables. 6.Conclusions The performance of empirical models used for predicting the deformation modulus should be verified by inductive modelling techniques because, accurate prediction of deformation modulus for any rock mass environment is very essential for accurate numerical analysis of stability and pre-failure mechanical behaviour of rock masses. [...]on comparing the inductive models within itself, it was concluded that for all the data sets the performance of ANN based models was better to the MLR based models resulting in high values of R2 and smaller values if RMSE and AAE.
ISSN:1994-3237
1994-3237