Multi-scale Data Integration In Crosswell EM Imaging And Interpretation

From a physics point of view Crosswell electromagnetic (EM) is a natural extension of the single well induction logging method, whereby the transmitter and receiver are located in separate wells. However, the data processing required to produce an image of inter-well resistivity is significantly different than standard induction log processing, involving integration of multiple data sets to reduce the non-uniqueness of the EM inversion problem. Additional multi-scale data assimilation is required to extend the interpretation process away from that of a 2D or 3D resistivity model to parameters of more interest to a reservoir engineer, such as map-estimates of water saturation. After a brief description of the measurement as well as an assessment of resolution provided by the cross-well data alone, a description of the processing flow is provided highlighting how different data sets are employed to reduce non-uniqueness provided by the combined data acquisition/inversion process, and ultimately yield a higher resolution image of resistivity. Next, ideas for employing additional data to yield maps/cross-sections of parameters that are more important to reservoir managers such as water saturation are suggested. The benefit of employing a priori data to construct geologically reasonable starting models will be demonstrated on a data set collected at the University of Texas'' Bureau of Economic Geology Devine Test Site.