Comparison of Fuzzy Inference Models and Artificial Neural Networks for Estimating the Depth of Rock bed of Aquifer Case study: South Khorasan, Birjand Aquifer

Document Type : Research Paper

Authors

Abstract

Groundwater is one of the most important water sources in the arid and semi-arid areas. With regard to reducing the level of water tables due to overdraft in most of Iran's plains, the wells' flow rate has been greatly  decreased and this issue necessitates the attention to planning the water resources.   Determining the  aquifers' thickness and type of alluvial and materials in aquifer is necessary for the development of the city planning and its infrastructures . with respect to the importance of  estimating the  depth of bed rock of aquifers for estimating the volume and planning the water resources in this study, the efficiency  of artificial neural network models and neural fuzzy inference system were studied on the  bed rock depth and its  zoning in different parts of the aquifer.In this study, the parameters of geographical latitude and longitude, salinity, water table level, and the ground level were used as input, and tried to determine a suitable  model for predicting the bed rock. Results showed that the neural network with  R2 =0.835 RMSE [1]=49. 488 meter with the inputs of  geographical latitude and longitude and the groundwater level has a higher accuracy of ANFIS [2] models.



[1] -Root mean square error


[2] -Adaptive neuro fuzzy inference system

Keywords

Files

History

  • Receive Date: 28 June 2016
  • Accept Date: 28 June 2016

Share

How to cite

Statistics