DSpace Repository

ARTIFICIAL NEURAL NETWORK MODEL FOR GROUNDWATER LEVEL PREDICTION

Show simple item record

dc.contributor.author Ananya.M.M
dc.contributor.author Dilsha.C
dc.contributor.author Gargi.A
dc.contributor.author Asha Joseph, (Guide)
dc.date.accessioned 2021-09-23T06:51:52Z
dc.date.available 2021-09-23T06:51:52Z
dc.date.issued 2020
dc.identifier.uri http://14.139.181.140:8080//jspui/handle/123456789/1134
dc.description.abstract Groundwater level is an indicator of groundwater availability, groundwater flow and the physical characteristics of the groundwater system. It is an important natural resource for human survival system and major sources of irrigation. Groundwater quality is deteriorating day by day. So the measurement and analysis of groundwater level is needed for maintaining groundwater availability. The accurate prediction of groundwater levels is essential for sustainable utilization and management of vital groundwater resources. For management of groundwater level a model is required which can predict the groundwater level in future with the current available information. The Artificial Neural Network (ANN) technique has been found to be very much suited to the modelling of non-linear and dynamic systems such as water resources systems. The main advantage of the ANN technique over traditional methods is that it does not require the complex nature of underlying processes to be explicitly described in mathematical form. After proper training, ANN models can yield satisfactory results for many prediction problems in the field of hydrology. In this study different ANN models are developed to evaluate the groundwater level fluctuations. One month ahead prediction models were also developed to extend the possibility of forecasting groundwater levels in coming future. Models were developed with different combinations of transfer function and number of hidden layers. All these were developed using MATLAB 7.0 software which is a multi-paradigm programming language and numeric computing environment developed by MathWorks. The best model for predicting groundwater levels were selected on the basis of coefficient of correlation(R) and RMSE value. Models with higher R value and lesser RMSE value is found to be the best performing one. Sreekrishnapuram region near Pattambi was selected as the study region. Input parameters for groundwater level fluctuations were identified and the monthly data of the same were collected for a period of 20 years from Jan 1999- Dec 2019. The model was trained, validated and tested for randomly chosen parameters. The developed ANN models for predicting groundwater level fluctuations shows good correlation coefficient ranging from 0.6-0.78. And the developed ANN models for one month ahead prediction also showed better values of R with an R value of 0.68 for the best model. All the models developed showed comparatively lesser RMSE value. Thus it can be determined that ANN provides a feasible method in predicting groundwater levels. en_US
dc.language.iso en en_US
dc.publisher DEPARTMENT OF IRRIGATION AND DRAINAGE ENGINEERING en_US
dc.relation.ispartofseries ;P 509
dc.title ARTIFICIAL NEURAL NETWORK MODEL FOR GROUNDWATER LEVEL PREDICTION en_US
dc.type Thesis en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account

Context