StreamSim/Stanford University Joint Technology Development Project

This project proposes to fast track very promising initial streamline-based history matching results obtained at Stanford University into a deployable, commercial grade, software product over a three year period. The project will be coordinated by StreamSim Technologies, which will be responsible for the final commercial implementation. Stanford University will be responsible for the development and extension of efficient streamline-based history matching algorithms.

If you are interested in participating in this industry consortium, please contact Streamsim Technologies at info@streamsim.com.  
 

The objective of this project is to combine StreamSim Technologies' leadership in streamline-based flow simulation with Stanford Universities' extensive knowledge on geologically consistent history matching to produce a commercially deployable software tool to help engineers in the area of history matching using streamlines.

This project offers a unique opportunity to fast track recent developments in the area of streamline-based history matching into commercial-grade software. This project is outside of the traditional academic affiliates programs offered by Stanford (SUPRI/SCRF consortia) as it delivers a commercial end-product by the end of year 3 based on research done at Stanford.
 

A simple measure of geological continuity is provided in geostatistics through the variogram model. The variogram model is often estimated from well-log data. It provides a constraint on the spatial extent and direction of permeability anisotropy. The variogram, however, cannot capture strong reservoir connectivities or curvi-linear features.

In Caers et al. an iterative procedure is proposed for injecting the variogram continuity information into the streamline-based history matching method outlined above. An iterative Monte Carlo technique is used to constrain the current reservoir model kl(u) to the streamline effective permeabilities ?k(u) at the same time honoring the correct geological constraints. It is a Gauss-Markov method where each cell is visited randomly, and an attempts is made to change each reservoir grid-block such that the reservoir model honors better the streamline effective permeability constraint without destroying the geological continuity provided by the variogram.

Figures 1 and 2 display typical results from this approach. Figure 1 shows the reference field and 3 different matches constrained to the same data. It is a confined five spot with an injector in the center and producers at each corner. The mobility ratio is 5, and only 150 days of data are used to obtain a match, Figure 2. The predictive abilities of each of the 3 matched permeability fields are rather good even at producer 1 that has not broken through by 150 days. Further details are provided in Caers et al..

Reference: Caers, J., Krishnan, S., Wang, Y. & Kovscek, A. R. 2002. "A Geostatistical Approach to Streamline-based History Matching" SPE Journal. 7(3) September, 250-266.