A textbook description of history matching could indicate that it is a process wherein changes are made to an initial geologic model of a reservoir, so that the predicted reservoir performance matches with the known production history. A practitioner recognizes that this is an overly simplistic description:

  1. Does not recognize subsurface uncertainty and how we may explore this uncertainty with multiple geologic models.

  2. Offers no guidance on the requirements of spatial resolution, and how this choice may vary depending the stage of field life or reservoir processes, e.g., a pressure history match versus the modeling of a secondary or tertiary process.

  3. Offers no guidance on the nature of the geologic model: its characteristics, its level of detail, or its structure.

  4. Provides no description of the history matching process itself, nor the choices available to a practitioner.

We have been supplied with the historical field data for a deepwater turbidite reservoir, which is still under active development. We have used this reservoir to explore a number of history matching strategies. This project involved an integrated seismic to simulation study, wherein we interpreted the seismic data, assembled the geological information, performed petrophysical log evaluation and well test pressure transient analysis before creating the 3D geologic model.

In the matching process we examined the trade-off between exploring a wide number of models versus calibrating a single model. We examined the scale at which the geologic model was constructed, and how the scale of simulation could be determined. Finally, we studied both the large discrete steps in the process, and the smaller local "assisted" parameter calibration. The results provide general guidance on workflow sequence, guidance on model selection, and guidance on the scales of static and dynamic modeling. We expect many of these conclusions to have general utility, not restricted to this specific study.

You can access this article if you purchase or spend a download.