Reconciling high-resolution geologic models with pressure and multiphase production history is one of the most time-consuming aspects of the workflow for geoscientists and engineers. Although significant advancements have been made in this area over the last decade, current industry practice largely involves iterative trial and error methods and often utilizes arbitrary permeability multipliers that lead to loss of geologic realism and consequently poor performance predictions.

In this paper, we demonstrate the application of an efficient inverse modeling scheme for history matching a waterflooded, structurally complex and faulted offshore turbiditic oil reservoir. The field has an estimated 500 MMSTB and is located in a prolific offshore hydrocarbon basin. Permeability and fault transmissibilities are the main uncertainties. More than 10 years of production data from up to 8 producing wells are available for history matching. Our approach combines the efficiency of the streamline-based sensitivity computations with the versatility and accuracy of finite difference models for matching the water cut at the producers.

Specifically, a finite-difference model is used for flow simulation and to generate velocity fields based on which we compute streamlines and sensitivities for updating reservoir permeability via fast inverse modeling. These sensitivities relate the reservoir properties to production data and can be obtained using a single flow simulation resulting in an efficient inverse algorithm for history matching. A unique aspect of the field study here is the novel approach used for streamline tracing and sensitivity calculations across faults and non-neighbor connections in the geologic models via local grid refinement.


Streamlines techniques offer an attractive combination of properties for history matching1–7:

  • They can provide a fast discretized flow domain (the mapping from the model parameters to the simulated data) and

  • fast and simple calculation of approximate sensitivities (i.e. partial derivatives of simulated data with respect to the model parameters).

In this contribution we exploit the fact that ii) can be executed independently of i), i.e. no (often approximate) streamline simulation must be applied to calculate streamline-based sensitivities. These sensitivities provide the fundamental relationships that allow us to efficiently invert the production data, measured at the wells, into modified reservoir properties between the wells. The major steps are:

  1. flow simulation using a finite-difference simulator

  2. streamline-based analytic sensitivity computations

  3. quantification of the production data mismatch and,

  4. updating of reservoir properties via inverse modeling.

The intercell fluxes (or velocities) are extracted from the finite difference calculation, and are then used to trace the streamlines and calculate time of flight. An outline of the streamline-based automatic history matching workflow is given in the flow chart Figure 1. The details will be presented in the next section.

The specific implementation of the inversion method used in the present contribution:

  • Uses liquid rates and water breakthrough times as input data and is well suited for waterfloods such as the field considered here.

  • Is restricted to water cut inversion but is also able to handle three-phase flow models (no GOR data was available to be used in the inversion).

  • Inverts absolute permeabilities and no other model parameters.

  • Performs streamline tracing and sensitivity calculations following a rigorous formulation, which can handle highly non-orthogonal cells and non-neighbour connections.

This content is only available via PDF.
You can access this article if you purchase or spend a download.