Neural Network for Mapping Full Time Apparent Resistivity Solution from TEM Electromotive Force Data
The apparent resistivity of the induced electromotive force observed by the transient electromagnetic (TEM) in center loop is proposed by using the neural network.According to the characteristics of increasing of the induced electromotive force with the resistivity of the transient electromagnetic sounding, the neural network input and output relationship and the network structure of the transient electromagnetic induction electromotive force response are designed according to the uniform half space central line observation mode.The nonlinear model of transient electromagnetic induction electromotive force and apparent resistivity is fitted by the neural network to obtain the full apparent resistivity value of the measured induced electromotive force data at a certain sampling time, and the aim of resistivity and imaging is achieved quickly.Through the calculation and verification of the simulation model of the transient electromagnetic in grounding grids, the apparent resistivity section can be obtained, and the position of the breakpoint can be clearly determined, and the effect of solving the inverse problem is achieved.The calculation of the verification model shows that the method makes the calculation time of the transient electromagnetic apparent resistivity greatly shorten, which is a practical algorithm.
Transient electromagnetic Full time apparent Resistivity Central loop Apparent resistivity Neural networks Backpropagation
Ming-hui BAO Shan-qiang QIN Guo-jun HE Jun-qiang Li Zhi-hong FU
Electric Power Research Institute, Chongqing Electric Power Company Chongqing, China Department of Electrical Engineering Theory and New Technology, School of Electrical Engineering, Ch Chongqing Triloop Prospecting Technology Co., Ltd, Chongqing, China
国际会议
上海
英文
326-332
2018-04-15（万方平台首次上网日期，不代表论文的发表时间）