A prediction model of shallow groundwater pollution based on deep convolution neural network. (23rd July 2021)

• Record Type: Journal Article
• Authors: Jiang, Zhongfeng; Gao, Hongbin; Wu, Li; Li, Yanan; Cui, Bifeng
• Is Part Of: International journal of environmental technology and management. Volume 24:Number 3/4(2021)
• Page Start: 278
• Page End: 293
• Publication Date: 2021-07-23
• Subjects: deep convolution neural network -- water pollution -- water quality -- prediction model; Environmental management -- Periodicals; Green technology -- Periodicals; 333.705
• Languages: English
• ISSNs: 1466-2132

Abstract

In order to solve the problems that the shallow groundwater pollution is affected by water quality in the prediction process, resulting in the low prediction index and water quality index of shallow groundwater pollution, a prediction model of shallow groundwater pollution based on deep convolution neural network is proposed. The index system of shallow groundwater pollution is constructed, and contents of dissolved oxygen, oxygen demand, ammonia nitrogen and pH in shallow groundwater are determined. With the help of gradient descent method and Guss-Newton method, the weight of index content is modified; the modified content value of pollution index is entered into the depth convolution neural network for optimisation, and the optimised value is obtained to complete the shallow groundwater pollution prediction model. The experimental results show that the maximum prediction index of shallow groundwater pollution is about 0.99, and the maximum value of water quality index is close to 1.