中南大学学报（自然科学版）Issue(5)：1707-1714,8.DOI:10.11817/j.issn.1672-7207.2015.05.019
大地电磁人工神经网络反演
Magnetotelluric inversion using artificial neural network
摘要
Abstract
The artificial neural network(ANN) was studied to improve the magnetotelluric inversion operation efficiency and calculation accuracy. The input of neural network was apparent resistivity of the known models, and the output of neural network was the geoelectric parameters. The neural network was trained by back propagation algorithm, which contains the methods of information forward propagation and error back propagation. The best weight of network calculated by the training work was used to test the unknown model patterns, which is the inversion process. The networks for two-layer model and three-layer model were set up respectively. The results show that the neural network inversion is feasible and effective.关键词
大地电磁/神经网络/反演Key words
magnetotelluric/neural network/inversion分类
天文与地球科学引用本文复制引用
王鹤,蒋欢,王亮,席振铢,张道军..大地电磁人工神经网络反演[J].中南大学学报（自然科学版）,2015,(5):1707-1714,8.基金项目
国家自然科学基金资助项目(41304090)；大洋“十二五”重大项目(DY125-11-R-03)；中央高校科研业务费资助项目(2012QNZT074)@@@@(Project(41304090) supported by the National Natural Science Foundation of China （41304090）
Project(DY125-11-R-03) supported by Ocean Major during “Twelfth Five Year Plan” （DY125-11-R-03）
Project(2012QNZT074) supported by Scientific Research Expenses Fund of Central University) （2012QNZT074）
