Numerical investigation of coal seam gas detection using airborne electromagnetics

Abstract

The use of airborne electromagnetic (AEM) techniques has been mostly utilized in the mining industry. The various AEM systems enable fast data acquisition to detect zones of interest in exploration and in some cases are used to delineate targets on a production scale. For coal seam gas (CSG) reservoirs, reservoir thickness and the resistivity contrast present a new challenge to the present AEM systems in terms of detectability. Our research question began with the idea of using AEM methods in the detection of thin reservoirs. CSG reservoirs resemble thin reservoirs that have been and are currently being produced. In this thesis we present the results of a feasibility analysis of AEM study on coal seam reservoirs using synthetic models. The aim of the study is to contribute and bridge the gap of the scientific literature on AEM systems in settings such as CSG exploration. In the models we have chosen to simulate both in 1-D and 3-D, the CSG target resistivity was varied from a resistive to a conductive target (4 ohm.m, 150 ohm.m, and 667 ohm.m) to compare the different responses while the target thickness was fixed to resemble a stack of coal seams at that interval. Due to the differences in 1-D and 3-D modelling, we also examine the differences resulting from each modelling set up. The results of the 1-D forward modeling served as a first order understanding of the detection depths by AEM for CSG reservoirs. Three CSG reservoir horizontally layered earth model scenarios were examined, half-space, conductive/resistive and resistive/conductive. The response behavior for each of the three scenarios differs with the differing target resistivities. The 1-D modeling in both the halfspace and conductive/resistive models shows detection at depths beyond 300 m for three cases of target resistivity outlined above. After the 300-m depth, the response falls below the assumed noise floor level of 5% response difference. However, when a resistive layer overlies a conductive host, the resistive/conductive model, the signal is reduced for the resistive target cases, but the response is unchanged for the conductive target layer. For a better understanding of the responses from more complex reservoirs, a 3-D model was developed to incorporate additional geology. The 3-D models were based on the 1-D models and the modeling parameters were not altered except for the finite extent of the layers. The system properties such as the transmitter waveform, moment and time gates did not change. For the 3-D coal seam reservoir models, the same level of response is not observed for the 240 x 240 m areal extent target. For the halfspace and conductive/resistive model, the AEM response is small. Also noticeable is the decreased response below 50-m target depth. For the assumed noise floor level, the different targets would not be detectable in tthese instances beyond 50-m when compared to detection depths of up to 300-m in the 1-D scenario. If, however, a resistive overburden exists, i.e. the resistive/conductive model scenario, the 3-D response for the conductive case target is strong compared to the other target cases due to the preferential current flow. In this scenario, a conductive target seam can be detected at a depth of 150-m and possibly deeper depending on the thickness of the overburden layer. In contrast, for the case of the resistive targets, the anomalous body would be undetectable beyond 50-m depth. I apply the same modeling techniques to a more complex model adopted from the Queensland Surat Basin CSG reservoir. I simulate responses in both 1-D and 3-D. The 1-D responses show promise for detecting targets at up to 500 m deep. The 3-D models with an embedded a target with an areal extent of 240 x 240m display small responses and indicate shallow detection depths. However when I increased the target's areal extent to 480 x 480 m, a stronger response is observed that is larger then the 5% noise floor level for all three target cases. This is a good indication that the size of the CSG target is important for AEM application. From this study, a few conclusions can be draw; if the target is not large enough in lateral extent, it is unlikely that we would be able to detect it. Additionally, if production of a coal seam group does not produce a resistivity contrast that is measurable, the AEM method such as modeled here would also not be beneficial. As for the exploration case, with laterally continuous targets, AEM might may prove useful if the resistivity contrast is large enough between target and host. In 1-D the responses were significant and indicated a good possibility of using AEM methods for this task at depths reaching 500 m. However, the 3-D responses show limited detection depths of 150 - 200 m for laterally confined targets. To the best of my knowledge, this study is a first pass on the topic, it is recommended that an even more thorough investigation of the scenarios be examined in greater detail with the hope of providing quantifiable results that would aid in the better understanding of these reservoirs.