Express assessment of the choice of optimal parameters of the development system under the conditions of geological uncertainty

Background. In most cases, the 3D Hydrodynamic Model (3HM) of a field is used to choose an optimal development system. Despite of high level of detail and accuracy of the model, significant computational resources and time are required to create and apply the 3HM. Therefore, the task to reduce the resource expended on project decisions making while minimizing the loss in the quality is relevant.

Aim. The approach proposed to optimize the existing methods for solving the problem of choosing optimal parameters of the development system is to make the rapid assessment of NVP and КИН based on multivariate modeling using a Hydrodynamic Two-Dimensional simulator and sensitivity analysis with machine learning methods under conditions of geological uncertainty.

Materials and methods. A Hydrodynamic Two-Dimensional simulator is used to solve the problem of choosing optimal parameters of the development system. Machine learning methods based on the Python programming language are used to process, visualize and conduct sensitivity analysis of the results.

Results. The algorithm of choosing optimal parameters of the development system based on multivariate modeling using machine learning methods is proposed. The analysis of the sensitivity of oil recovery factor and NPV to changes in the input parameters of the model was carried out. Based on the results, a three-dimensional matrix of results was created. The matrix of results allows selecting the optimal density of a well pattern using the triple values of the effective reservoir thickness, permeability and length of the horizontal section. The practical significance of the matrix of results lies in the possibility of forming a hybrid project grid of wells in clustering zones according to the values of permeability and effective reservoir thickness. Also, the matrix helps to adjust project decisions for choosing optimal parameters of the development system under conditions of geological uncertainty in a short time.

Conclusion. Based on multivariate modeling, the alternative approach for choosing combined grids using machine learnin’g methods is proposed to field development planning.

1. Muskat M. The flow of homogeneous fluids through porous media. McGrawHill, New York, 1937.

2. Kanevskaya R.D., Matemeticheskoe modelirovanie razrabotki mestorozhdeniy nefti i gaza s primeneniem gidravlicheskogo razryva plasta (Mathematical modeling of the development of oil and gas using hydraulic fracturing). Moscow: NedraBiznestsentr Publ., 1999, 212 p. (In Russ.).

3. Willhite G.P. Waterflooding, SPE Textbook Series, 1986.

4. Baykov V.A., Zhdanov R.M., Mullagaliev T.I., Usmanov T.S. Selecting the optimal system design for the fields with low-permeability reservoirs, Neftegazovoe delo, 2011, no. 1, pp. 84–97 (In Russ.).