Application of two classes of inverse problems for optimization and management of the natural hydrocarbon field development process

Abstract

The article considers two classes of inverse problems in the modeling of hydrocarbon field development: optimization and adaptation of the development history. They jointly form the basis for field development management in a closed loop, based on a constantly updated 3D geological and technological model. The problems of model adaptation (identification) arise due to the point-by-point measurements and the inaccuracy of determining the initial data on porosity and permeability of the host rocks and other parameters of the deposit when building the 3D model. Such problems are regarded as the inverse ones, because the consequence – field measurements – is used to determine their cause. The problems under consideration are related to the identification of both porosity and permeability of the reservoir and other key parameters that determine the dynamics of development. Geologically consistent formulation ensures that the adapted model retains the original geological principles of the distribution of reservoir properties. The second class of problems under consideration is the problems of optimization (regulation) of development. One of the sought-after formulations is related to the optimal redistribution of a given target level of production from the field to all production wells in time. In this case, the technological limitations of surface equipment are taken into account. The maximized criterion reflects the economic effect of the asset development. In this article, without going into details of mathematical realization, the authors consider practical examples of solving these problems using full-scale 3D multiphase filtration models of the real fields. The algorithms based on the current methods of optimal control theory are used, which were developed and implemented by the researchers of Oil and Gas Research Institute of the Russian Academy of Sciences in the SimMatch software package and confirmed their effectiveness on numerous synthetic and real examples. Application and further development of these methods within the framework of a single loop, taking into account continuous information from the entire set of sensors, will significantly increase the efficiency of intelligent wells and field control systems.