期刊
APPLIED SOFT COMPUTING
卷 92, 期 -, 页码 -出版社
ELSEVIER
DOI: 10.1016/j.asoc.2020.106298
关键词
Deep learning; Numerical simulation; Groundwater; Internet of Things
资金
- Ministry of Science and Technology, Taiwan (R.O.C.) [MOST 108-2625-M-008-007, MOST 108-2511-H-019-002, MOST 108-2511-H-019-003, MOST 107-2116-M-008-003-MY2]
- Soil and Groundwater Pollution Remediation Fund in 2018
- Water Resources Planning Institute, Taiwan (R.O.C.) [108706]
- Soil and Groundwater Pollution Remediation Fund in 2019
Deep learning for enhancing simulation IoTs groundwater flow is a good solution for gaining insights into the behavior of aquifer systems. In previous studies, corresponding results give a basis for the rational management of groundwater resources. The users generally require special skills or knowledge and massive observations in representing the field reality to perform the deep learning algorithms and simulations. To simplify the procedures for performing the numerical and large-scale groundwater flow simulations, we apply the deep learning algorithms which combine both the numerical groundwater model and large-scale IoTs, groundwater flow measuring equipment and various complex groundwater numerical models. The mechanism has the capability to show spatial distributions of in-situ data, analyze the spatial relationships of observed data, generate meshes, update users' databases with in-situ observed data, and create professional reports. According to the numerical simulation results, we revealed that the deep learning algorithms are high computational efficiency, and we can enhance precise variance estimations for large-scale groundwater flow problems. The findings help users to best apply the deep learning algorithms in an easier way, get more accurate simulation results, and manage the groundwater resources rationally. (C) 2020 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据