Prediction of transmissivity of aquifer from geoelectric data using artificial neural network

2nd International Conference on Science and Sustainable Development "The Role of Science in Novel Research and Advances in Technology"

自然科学;经济学

Sanuade, Oluseun Adetola^1 ; Oyeyemi, Kehinde David^2 ; Amosun, Joel Olayide^3 ; Fatoba, Julius O.^3 ; Hammed, Olaide S.^4

Geosciences Department, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia^1

Department of Industrial Physics, Covenant University, Ota, Ogun State, Nigeria^2

Department of Geophysics, Federal University Oye-Ekiti, Ekiti State, Nigeria^3

Department of Physics, Federal University Oye-Ekiti, Ekiti State, Nigeria^4

关键词: Basement complex;    Coefficient of correlation;    Geoelectric;    Linear relationships;    Root mean square errors;    Transmissivity;    Transverse resistance;    Vertical electrical sounding;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/173/1/012025/pdf
DOI  :  10.1088/1755-1315/173/1/012025

学科分类：自然科学（综合）

来源: IOP

【 摘 要 】

We have used artificial neural network (ANN) for the prediction of transmissivity (T) from geoelectric parameter in an aquifer of a typical basement complex, Southwestern, Nigeria. The study area is composed of migmatite-gneiss, charnockite and granite gneiss. A geophysical investigation involving Vertical Electrical Sounding (VES) was carried out in the study area. Twenty VES were acquired using Campus Ohmegaresistivity meter. VES curves were interpreted quantitatively by using the technique of partial curve matching and with 1D forward modeling by means of computer using WinResistsoftware. The interpreted data were used to determine transverse resistance, TR. Transmissivity were measured in the boreholes drilled at the 20 VES locations. TR and measured T were subjected to ANN analysis using MATLAB 2017a software in order to predict T of the aquifer. Root mean square error (RMSE) was used to test the performances of our model. The results show that TR range from 133.00 -381.20 Ωm2, T also vary from 0.4 to 4.1 m2/day. However, ANN model was able to predict T values with coefficient of correlation (R) values of 0.97, 1.00, 0.99 and 0.94 for training, test, validation and all network models respectively. RMSE value for the ANN model was found to be 0.085 which implies high performance of our model. A linear relationship was suggested for the ANN analysis to predict T. It can therefore be concluded that with ANN model, it is possible to predict T of aquifer in the study area where geoelectric data such as TR is known and T values unknown.