The advancement of instruments over the last decade has made airborne EM a major geophysical method in mineral exploration and near surface applications. The new systems and the data sets generated by them offer many interesting and challenging research topics. For instance, the amount of data from an airborne EM survey can be hundreds or thousands times greater than that from a ground survey. Focusing on processing, modeling and interpretation, I have worked on data sets from most of the commercial airborne EM systems, including frequency-domain (DIGHEM, RESOLVE), time-domain (VTEM, HeliGEOTEM/HeliTEM, MegaTEM, SPECTREM, SkyTEM, AeroTEM, HeliSAM), and natural source (ZTEM) systems. The applications include mineral exploration (porphyry, massive sulphide, uranium, kimberlite, etc.) and environmental problems (groundwater). 3D modeling and inversion are featured in my research.
Detailed geological interpretation relies on high-resolution 3D geophysical models, which are usually time-consuming to compute. For some really large surveys, the increase of computational cost can be exponential. I have been developing a new adaptive framework, referred to as survey decomposition that better scales the computation in a parallel computing environment. Survey decomposition uses two essential techniques. First it allows each datum to be simulated on its own mesh (with its own time stepping or frequency), eliminating the need of a massive global 3D mesh in the forward modeling. Second the number of randomly selected data (subproblems or local meshes) used in an inversion is adaptive to the degree of regularization. Together, they save significant amount of over-computing. Modeling of data on separate and customized meshes also enables the utilization of large-scale parallelization.
The fluid transport in reservoirs can be traced by the electrical or electromagnetic signals associated with the fluid. A promising technique is monitoring the fluid, for examples, in hydraulic fracturing or carbon sequestration, using surface-based measurement. It is made possible by the steel cased wells that channel significantly enhanced current down to the depth of reservoirs. My focus on this project is to build the numerical capability of modeling the multiple casings existing in oilfields together with arbitrary 3D earth and frac models. The modeling program is then used to better understand the electrical interaction between casings and the earth, which leads to more effective monitoring strategies.
Multiple data sets can be generated from different types of surveys at the same site. The sensitivity unique to each survey allows the earth's physical properties to be better inferred. The goal is to obtain a unified model that is supported by all the data sets. To achieve this, I have been working on developing the algorithms and workflows within the context of 3D inversion. A high-profile massive sulphide mining site Lalor served as a good testbed for my multi-data research. The site has been surveyed by more than 15 types of electrical and EM methods, producing a large amount of data at different scales from airborne, surface and borehole platforms.
Copyright © 2018 All Rights Reserved.