Three Component Time-domain Electromagnetic Surveying: Modeling and Data Analysis

Transcription

1 PIERS ONLINE, VOL., NO., 75 Three Coponent Tie-doain Electroagnetic Surveying: Modeling and Data Analysis Chow-Son Chen 1, Wei-Hsuan Chiu 1, and Ching-Ren Lin 1 Institute of Geophysics, National Central University, Taiwan Institute of Earth Sciences, Acadeia Sinica, Taiwan Abstract A coparative nuerical odeling using single coponent (easuring the vertical coponent of the field, Hz) tie-doain electroagnetic easureents (TEM) receiver and three coponent (easuring Hx, Hy and Hz) TEM receiver was undertaken. A forward odeling approach was used to copute the voltage response of half-space containing one or ore conductive bodies excited by a bi-polar square wavefor. Although this was a conductor scattering, it was particularly useful as a practical use for the unexploded ordnance (UXO) detection. Unlike the single coponent data, results fro the three coponent data are unabiguous as to the location and orientation of conductors. Measuring adding of the horizontal coponents of the secondary agnetic field leads to not only provide the best indication of target location, but also can be used to deterine size, orientation, and characteristics of the targets, especially for the horizontal extending target. A three-coponent TEM field experient at a well-docuented wells site, NCU capus, were consistent with the effects predicted by our theoretical odeling. As a result, the 3-coponent TEM survey is a ust for the high resolution EM in the engineering purpose. 1. INTRODUCTION In the traditional frequency-doain electroagnetic (FEM) ethods of exploration the ground is energized by passing an alternating current (ac) through an ungrounded loop situated usually on the surface of the earth. The priary field of the loop will induce eddy currents in all conductors present in the earth. The secondary electroagnetic (EM) fields due to these induced currents, together with the priary EM field, are recorded with a suitable receiver at various points in space. In general, the secondary EM field at the receiver, which contains all the inforation regarding the underground conductors, ay be several orders of agnitude saller than the priary field. Under these conditions the separation of the easured total EM field into its priary and secondary parts is difficult. This fact led to the idea of using tie-doain electroagnetic easureents (TEM), often referred to as transient EM techniques []. In TEM easureents a strong direct current (dc) is usually passed through an ungrounded loop. At tie t = this current is abruptly interrupted. The secondary fields due to the induced eddy currents in the ground can now be easured with a suitable receiver in the absence of the priary field. For poor conductors, the initial voltages are large but the field decays rapidly. For good conductors the initial voltages are saller but the field decays slower. We thus have a siple criterion of recognizing and differentiating the effects of various conductors. Near-surface TEM surveys have traditionally been siply one or ore Hz loops (easuring the vertical coponent of the field) for the purpose of unexploded ordnance (UXO) detection. However, in this paper, we offer a nuerical odeling of the TEM three-coponent (Hz, Hx and Hy) responses of vertical or horizontal conductors in the earth. Based on nuerical odeling, we devise a three-coponent TEM field experient at a well-docuented wells site, NCU capus. We find three-coponent TEM anoalies which correlate with known wells, and which are consistent with the effects predicted by our theoretical odeling. We conclude that the 3-coponent TEM survey is a ust for the high resolution EM in the engineering purpose.. THEORETICAL BACKGROUND AND MODELING In order to understand the TEM responses with a rectangular transitter loop excited by a bi-polar square wavefor with exponential rise-tie and rap turnoff, a forward odeling approach was used. To copute the voltage response of half-space containing one or ore conductive bodies is the bases of the case of etal detection. Of course, the theoretical responses derived fro these targets are inductively coupled. Based upon frequency doain, integral equation thin sheet theory,

2 PIERS ONLINE, VOL., NO., 7 a progra LEROI copute the cobined EM effects of ultiple conductors in a layered host was eployed [5]. Figure 1 shows the calculated profile plots fro a plate (conductance 3 S) in conductive host with the resistivity of 1 oh-. Figure 1(b) is the survey ap showing the location of the plate and survey lines. Figure 1(c) shows the calculated TEM transient voltage (in µv) along line (y = ) of Hz, Hx, and Hy coponent, top down respectively. The conductive features are clearly anoalous with strong negative values relative to background for Hz profile; the calculated Hx anoalies are dipolar with a negative/cross-over/positive anoaly as the target is approached and crossed; the calculated Hy anoalies are siilar to Hz but the polarity of the anoaly is reversed. In each case, the cross-over point is centered over the target. Figure 1(d) shows the calculated TEM transient voltage (in µv) along line 1 (y = 1). The doinant difference is of no Hy anoaly since y coponent is parallel to the plate, which is assued to be of no thickness. Figure 1(e) (a) (b) (c) (d) (e) Figure 1: Coputed three-coponent transient voltages for the TEM profiles of a vertical plate (conductance 3 S) in a conductive host (1 oh-). (a) The perspective displays of the three-diensional odel. Transitter loop position and survey lines are displayed on top surface at ground level. Target plate (in red color) suggests depth and three-diensionality. (b) The survey ap showing the location of the odel plate (in black bar) and survey lines. (c) The calculated TEM transient voltage (in µv) along line of Hz, Hx, Hy coponent, and the earth odel, top down respectively. (d) Line 1. (e) Line.

3 PIERS ONLINE, VOL., NO., 77 shows the calculated TEM transient voltage (in µv) along line (y = ). The profile is quite siilar to the profile, but Hy coponent in opposite polarity. As the target conductor becoes horizontal, the calculated profile plots and plane plots are created as in Figure. After exaine and coparing aong the plots, the horizontal coponent is better for horizontal target identification than the vertical coponent does, since only horizontal coponents can display the anoaly of conductive target (Figures (c), (d), and (e)). 3. THREE-COMPONENT TEM FIELD EXPERIMENTS A coparative survey using three-coponent (easuring Hx, Hy and Hz) TEM receiver was undertaken at a test site; we concentrated on the six wells in the Groundwater Research Well Field on NCU capus, four of the are stainless steels casing and the other two are plastic wells (Figure 3(a)). The capus is located at the Taoyuan Tableland; their deposits consist of an upper sequence of a red-brown lateritic soil over rounded sandstone clasts (up to boulder size) in an (a) (b) (c) (d) (e) Figure : Coputed three-coponent transient voltages (in µv) for the TEM profiles of a horizontal plate (conductance 3 S) in a conductive host (1 oh-). (a) The perspective displays of the odel. Transitter loop position and survey lines are displayed on top surface at ground level. Target plate (in red color) suggests depth and three-diensionality. (b) The survey ap showing the location of the odel plate (in red color) and survey lines. (c) The calculated TEM transient voltage (in µv) along line of Hz, Hx, Hy coponent, and the earth odel, top down respectively. (d) Line 1. (e) Line.

4 PIERS ONLINE, VOL., NO., 7 unceented atrix of sand, silt, and clay, and a lower sequence of interbedded, well-consolidated sand, silt, and clay. The lower clay layer coposed of laterally continuous clay fro 3 to the wells depth 35 [3] Field Experients For ours three-coponent TEM survey, a 5% duty cycle, tie doain, square wave electric current (5 aps) was transitted into the ungrounded transitter square loop of wire easuring on each side. In this survey, these agnetic fields are detected with three-orthogonal ungrounded Roving Vector Receiver. The easureents are ade at to ties (or windows ), fro about.5 to 3 s, after the transitter is turned off in order to easure the decaying response of the background earth as well as secondary fields fro strong conductors (such as etallic objects) after the background earth response has decayed to near-zero values. SIROTEM TEM [1] was used for this purpose. An area easuring eters by eters around the central six wells in the Groundwater Research Well Field was surveyed in detail to establish the extent of the subsurface etal, as well as to detect any other nearby anoalies. 3.. Data Analysis and Interpretation Other than the profile display as in the previous nuerical odelling, we organized all the experiental profiles and plot the results as a plane view, so as to exa the anoaly in two diensional distributions. Figure 3(b) shows the Hz survey results in plan view at 1.7 s after the transitter is turn-off. At tie 1.7 s, the subsurface etal casings in the wells site are clearly anoalous, it close correlated to the wells site. As tie increased to 5.7 s, all the six anoalous of the associated with well casings is clearly seen that extend below to the deeper depth (not shown here), although well 3 response weaker, and we strongly suspected the lacking of etal casing for this well during the drilling works. Figure 3(c) shows the wells TEM responses of the horizontal coponent, calculated fro the Hx and Hy data. Only etal casings, well 1,, and, are close correlated the TEM anoalies to the well sites fro shallow to depth; the plastic casings, well 5 and, are response TEM signal weak. This deonstrates that the iportant and superior of the horizontal TEM coponent in vertical target detection; it adds the capability of the illuinations of the orientations and characters of the etal target. However, the horizontal signal is weaker than that of the vertical, this is the reason why we transitted the higher current to 5 A in this experient. This behavior is predicted by our theoretical odelings (Figure 1 and Figure ). In the experient of extending the survey to surrounding areas, about 5 to the west, a siilar anoalies were encountered that were evident in the horizontal coponent data but not in the Hz Well locations z:1.7s xy:1.7s 1:Metal 5:Plastic 1 :Metal :Plastic 3:Metal 1 1 :Metal 1 (a) 1 (b) 1 (c) Figure 3: Measured three-coponent transient voltages (in µv) for the TEM ap around the wells site on NCU capus. (a) The survey ap showing the location of the TEM soundings (black dots) and the location of wells: the bigger cross for etal casing and the saller one for plastic casing. (b) Vertical coponent TEM survey results in plan view around the well field at delay tie 1.7 s. Evidence is strong anoaly near well locations. (c) Horizontal coponent TEM survey results in plan view for the sae area at delay tie 1.7 s. Copared with the vertical coponent, the horizontal coponent TEM anoalies are ore response to the casing characteristics, etal and plastic.

5 PIERS ONLINE, VOL., NO., 79 data. Figure (a) shows the location, and Figure (b) is an east-west trending anoalous zone at early tie crossing through the grid at y =, oving fro west to east along the survey line. This linear anoaly is the result of a buried power line that coes into the survey grid fro the west, connecting to the electrical access cover (x =, y = ), and continues out of the grid to the east toward another light pole. Because the horizontal power line is very shallow, about 1 below the surface, the anoaly disissed as tie increased to.7 s (not shown here). We note that the strong anoaly at the coordinate (, ), which is last to the later tie is the electrical access cover on the surface. We also note that there is a weak TEM response for the vertical coponent easureent (Figure (c)); if only vertical coponent easureent for this survey as traditional exploration approach, then a great risk would be encountered for this case. Power line location 1 Electrical access cover 1 1 xy:.7s 1 1 z:.7s 1 (a) (b) (c) Figure : Measured three-coponent transient voltages (in µv) for the TEM ap around a horizontal pipe. (a) A buried power line at depth about 1 that coes into the survey grid fro the west, connecting to the electrical access cover (x = 13, y = ), and continues out of the grid to the east toward another light pole. (b) Horizontal coponent TEM survey results in plan view at tie.7 s. An east-west trending anoalous zone crosses through the respected grid. (c) Vertical coponent TEM survey results in plan view for the sae area at tie.7 s: weak response for pipes, while strong response for surface etal cover.. CONCLUSIONS AND FUTURE WORK In this paper, we have deonstrated nuerically that three-coponent TEM easureents can potentially give access to conductor inforation in conductive earth. We analyze 3-axis TEM data fro known well site and detect transient voltage anoalies which are consistent with our theoretical odeling and which can be correlated with well locations in the conductor host. Fro this and other surveys, it is apparent that there is uch useful inforation in the horizontal coponents of near-surface TEM surveys. The Hz data, which displays only uni-polar, negative anoalies, are usually stronger data, and provide the best indication of target location. Hx and Hy data are by nature weaker for the horizontal transitter, but can be used to deterine size, orientation, and characteristics of the conductive targets, especially for the horizontal extending target (i.e., power line, sewer line, etc.). Carlson and Zonge [] reached the siilar conclusions. We are encouraged by the correspondence between the TEM anoaly and the conductor inforation, and believe that the use of ulti-coponent data to locate conductors will be a ust for a high resolution EM ethod for engineering purposes. Undoubtedly soe possible factors ay cause this technique to fail: (1) High cultural background noise levels overwheled the anoaly. To increase the signal to noise ration, higher agnetic dipole oent of the transitter is a good choice; these include high transitter currents, large transitter loop, or using ultiple-turn transitter loops. () Using wrong tie windows would lose the target signals; for exaple, ours horizontal target (Figure ) detection experient would be failed if skip the tie window around.7 s. Recently, we are doing deeper exploration by using agnetotellurics (MT) study for Taiwan s tectonic structure. Professors Xie and Li s AGILD MT and CSMT odeling and inversion, as in reference [ 9], are useful tools, and the new study results will be reported in the next paper.

6 PIERS ONLINE, VOL., NO., REFERENCES 1. Buselli, G. and B. O Neill, SIROTEM A new portable instruent for ultichannel transient electroagnetic easureents, Austral. Soc. Expl. Geophys., Vol., 7, Carlson, N. R. and K. L. Zonge, Minerals exploration ethods odified for environental targets, Exploration Geophysics, Australian Society of Exploration Geophysicists, Vol. 3, 1 119, Chen, C.-S. and C.-C. Chang, Use of cuulative volue of constant-head injection test to estiate aquifer paraeters with skin effects: field experient and data analysis, Water Resources Research, Vol. 3, No. 5, 15, 1.9/1WR3,.. Nabighian, M. N., Electroagnetic Methods in Applied Geophysics: application, Society of Exploration Geophysicsts, Vol., 5, Raich, A. P. and F. Sugeng, Predicting the transient EM response of coplex structures using the copact finite eleent ethod, Exploration Geophysics, Vol., 51 55, Xie, G., J. Li, E. Majer, D. Zuo, and M. Oristaglio, 3-D electroagnetic odeling and nonlinear inversion, Geophysics, Vol. 5, No. 3,,. 7. Xie, G., J. Li, and F. Xie, Advanced GILD EM odeling and inversion, PIERS Online, Vol. 1, No. 1, 15 19, 5.. Xie, G., J. Li, and F. Xie,.5D AGILD electroagnetic odeling and inversion, PIERS Online, Vol., No., 39 39,. 9. Li, J., G. Xie, L. Xie, and F. Xie, New stochastic AGLID EM odeling and inversion, PIERS Online, Vol., No. 5, 9 9,.