Qingdao , China. Qingdao , China. Beijing , China *Corresponding author

Transcription

1 017 3rd International Conference on Applied Mechanics and Mechanical Automation (AMMA 017) ISBN: Comparison of Simulated Results of Deployed and Towed Undersea Dipole Sources in Marine CSEM Method Lei DONG 1,, Guang-yuan CHEN,3,*, Ming DENG 4, Hai-jing HE,3 and Zhuo LEI,3 1 Shandong Technological Center of Oceanographic Instrumentation, Qingdao 66001, China Shandong Provincial Key Laboratory of Ocean Environmental Monitoring Technology, Qingdao 66001, China 3 Institute of Oceanographic Instrumentation, Shandong Academy of Sciences, Qingdao 66001, China 4 School of Geophysics and Information Technology, China University of Geosciences, Beijing , China *Corresponding author Keywords: Marine CSEM, Deployed undersea dipole sources, Towed undersea dipole sources, Navigation error. Abstract. We developed a -dimension nonlinear conjugate gradient (NLCG) inversion program to compared the responses generated by deploy and towed horizontal source dipoles in marine CSEM method. We constructed different geoelectic models for different detecting objective and drew the conclusions. For the resistivity buried shallower than 500 m and transverse resistance lower than Ω m, the advantage of deployed mode is more obvious. The lateral resolution is much higher and the false anomalies are much less than the towed one. When the rotate or tilt angle of source dipole over 10 degrees or the deviation more than 100 meters, the error should be marked which can be corrected once by the deployed mode. Hence, considering the more time consumption and more complicated procedure, the deployed mode should be used as a delicate survey method in the range already scanned by the towed mode. Introduction Marine controlled-source electromagnetic method (CSEM) research arose in the last century [1-] and has been applied to the study of marine lithosphere structure as a replacement method for magnetotelluric methods for high-frequency signals at the outset. Nowadays, this method has become an important supplementary method in marine hydration exploration, and presents a substantial potential. A horizontal electrical dipole is a common dipole device used in Marine CSEM that was first proposed in publications by Bannister [3]. The electrical dipole is towed near the seafloor (about m above) and emits a multi-standard impulse current in order to excite electromagnetic anomalous signals for detection [4-5]. In this method (towed modes), the seawater layer between the transmitter and seafloor makes a negative impression on the resolution, because it leads to the amplitude decrease and phase distortion of the electromagnetic signals. According to the experiment of Scripps Institution of Oceanography in 009 [6], this paper designed the horizontal electrical dipole to be deployed directly on the seafloor referred as the deployed mode here. Considering the differences between the deployed and towed mode in which the source dipole is 50 meters above the seafloor, the revision of launch source would be made, such as reduction of electrical dipole distance. But we only focused on theoretical simulation study here and the dipole moment is set as 1 Am to simplify the calculation [7]. We developed a two-dimensional electromagnetic inversion program based on nonlinear conjugated method in order to simulate the inversion result made by the deployed and towed mode. Comparing items contain response strength, lateral resolution and navigation error effects. 195

2 Nonlinear Conjugate Gradient Method NLCG was first proposed by Flectcher and Reeves in 1964 [8] and was brought into nonlinear problem solution in geophysical research in 1987 [9]. Zhdanov applied it into electromagnetic shift image in 00 [10]. NLCG algorithm merits in no necessary to calculate and store Jacobian matrix explicitly and only needs to calculate gradient of objective function. Though the searching in conjugate gradient direction, the objective function reach to the best model iteratively. We adopted the objective function that Willian Rodi and Randall Mackie used on magnetotelluric inversion research to develop the new inversion code. The objective function is shown in equator (1). T 1 T T ψ ( m) = ( d F( m) ) V ( d F( m) ) + λm L Lm (1) The regularization parameter, λ, is positive. The positive-definite matrix V indicates the variance of the error vector e. The matrix L is a second-difference operator related with the grid of model. 1-D CSEM Experimental Design Study According to the theory of electromagnetic induction, it is easily to understand the response strength of the deployed mode is greater than that of the towed mode. While how great it is and what kind of effects that impresses on anomalies detection? Those should be researched deliberately for the real exploration. Comparison of Inversion Effects of Deployed Mode and Towed Mode There are not sufficient field data can be obtained for our research, so we constructed various models to compare the theoretic results of the deployed and towed mode. The forward calculation adopted the two-dimensional program opened publicly by Kerry [11]. 50 receivers were spanning the 5-km line and the responds were mixed with error of 5% of datum amplitude. Comparison for Varying Anomaly Parameters Figure 1. Two-dimensional geoelectric model. Figure. NLCG inversion results of resistivity with 4000 Ω transverse resistance. In electromagnetic detection, the product of anomalous layer resistivity and thickness is defined as transverse resistance, and an anomalous with the same transverse resistance have the same 196 m

3 electromagnetic response. Therefore, transverse resistance is here taken as an influencing factor to study the difference in electromagnetic response features between the deployed and towed mode. Fig shows the inversion results of model in fig1 correcting the transverse resistance to 4000 Ω m. That means if the thickness of anomaly is 500 meters, the resistance is 80 Ωm m. From the imaging generated by the deployed mode (in fig (a)), we can figure out the resistivity easily and the borders are reflected correctly, while for the towed mode, the anomaly cannot be detected. Comparison of Resolution for Resistivity In this part, research is focused on comparison of resolution further. For the model in fig1, subdivision nets are up to 5018 in order to elevate calculation precision. In the simulation involved only one anomaly, the deployed mode performed better in horizontal boundaries reflection (shown in fig3 (a)) and there are many fake anomalies reflected by the towed mode (shown in fig3 (b)). Figure 3. NLCG inversion results of one resistivity buried in shallow layer. The difference on lateral resolution manifests more obviously for the model with two anomalies as fig4. From the reflection of the towed mode in fig5 (b), two resistive anomalies immerged into one image positioned in the middle of them in the original model. For resistivity of this scale, the towed mode can not reflect the real electrical distribution, while the deployed mode could distinguish the two anomalies clearly as shown in fig5 (a). Figure 4. Two-dimensional geoelectric model for two horizontal resistivies. Figure 5. NLCG inversion results of two horizontal resistivies. 197

4 Comparison of Effect of Transmitter Navigation Error Flows under the sea level make the rotation, tilt and deviation of the horizontal transmission dipole as fig6 which lead to the response offset from the ideal horizontal condition. Compared with the deployed mode, correction of this effect of towed mode turns to be more complicated. (a)dipole deviation (b)dipole tilt (c)dipole rotation Figure 6. The dipole posture changing of navigation error [1]. For the model in fig1, assuming source dipole parallels to the direction vertical to strike. In order to discuss the impact of posture of dipole on the inversion result, we use the developed NLCG inversion program to simulate the conditions when the angle between the dipole and survey line vary from 10, 0, 30, 40, 50, 60 in horizontal and vertical direction respectively, then consider the circumstance under which the dipole deviated from the survey line by 100 meters, 00 meters and 300 meters. The inversion calculations here covered the rotation, tilt and deviation of source dipole. We draw conclusions of the effect of transmitter navigation error as follows. When the rotating angle changing from 10 to 60 as shown, the resistance of resistivity reflected decreases gradually and the fake conductive anomalous increases. The error cannot be ignored for the angle over 0, according to the NLCG inversion. The same kind of phenomenon also appears for the dipole tilting and the impacts are more obvious. The effects of navigation error should be focused when the angle is over 10. While the deviation affect mostly present on evaluation of resistance and scale of anomalous, the error is not that great as deflections. If the deviation over 100 meters that is common in field work, the error should be corrected carefully. Conclusions Marine CSEM allows for noninvasive survey of an area and provides extensive lateral coverage. This paper develops a two-dimensional inversion program on the basis of NLCG method and uses it to compare the simulation results of the deployed mode and towed mode in resistivity detection. The conclusions are followed. (1) The deployed mode has greater and electromagnetic signals broader working frequency scope. It makes the deployed mode could use more data for more accurate results. () For shallower and small scale reservoir, the deployed mode can indicate the anomalous under some conditions while the towed mode cannot. (3) The boundary resolution of shallow resistive anomalous of the deployed mode is better than that of the towed mode. The reflections are clearer and the anomaly shapes reflected are more regular. (4) To some extent, navigation error cannot be ignored. The immediacy of rotation of source dipole made data process of the towed mode more complicated than the deployed one. Acknowledgment Thank Professor Deng Ming, and the Key Laboratory of Geo-detection, Chinese Ministry of Education for providing me an excellent study and research environment. I also thank Dr. Jing for his guidance and help with this research. 198

5 References [1] Weidelt, P. Canada gas hydrate observatory, Geophysical Prospecting, 61 (4): 84. Guided waves in marine CSEM, Geophys. J. Int, 007, 171, [] Cox, C. S. On the electrical conductivity of the oceanic lithosphere, Physics of the Earth and Planetary Interiors, 1981, 5(3), [3] Bannister, P. R. Determination of the electrical conductivity of the sea bed in shallow water, Geophysics, 1968, 33(6), [4] Deng, M., Wei, W. B., Zhang, W. B. et al.. Electric field responses of different gas hydrate models excited by a horizontal electric dipole source with changing arrangements, Petroleum Exploration and Development, 010, 37(4), [5] Edwards, R. N. and Chave, A. D. A transient electric dipole-dipole method for mapping the conductivity of the sea floor, Geophysics, 1985, 51(4), [6] Kannberg, P.K., Constable, S., Weitemeyer, K.A. et al.(011). Resistivity structure at the summit of South Hydrate Ridge. American Geophysical Union, Fall Meeting 011, abstract #OS13C [7] Swidinsky, A., Edward, RN., Jegen, M. et al. the marine controlled source electromagnetic response of a steel borehole casing: applications for the Neptune, 013. [8] Fletcher R, Reeves C, Function minimization by conjugate gradients. Comput. J. 1964, 7, [9] Flectcher R, Practical method of optimization I: unconstrained optimization.1987, Wiley, New York. [10] Kerry K and Jeffrey O. A parallel goal-oriented adaptive finite element method for.5-d electromagnetic modelling. Geophys. J. Int., 011, [11] Liu Y H, Yin C C, Wen A H et al, Attitude effect for marine CSEM system. Chinese J. Gephys.(in Chinese), 01, 55(8),