# Industrial context

The Simkarts project has been initiated in 2012 by Pr. Han Pingchou, PKU, and Bernard Montaron, Schlumberger China, in order to understand multi-phase flow dynamics in deep oil reservoirs located in the Tarim basin, West China.

Karst reservoirs of Tarim Basin in Xinjiang Province of People’s Republic of China have been determined as an area with large quantities of oil and gas and are exploited by Sinopec and Petrochina Tarim Oil Company since the 1990’s.

This carbonate formation of Ordovician age is buried at more than 6000 meters and represents a complex network of connected caves full of oil and water large of 20 x 20 km.

* The Tarim **Basin A fracture/cave reservoir *

In order to optimize oil production, it is necessary to understand and predict oil flow inside the reservoir. Indeed, the fluid dynamic inside the reservoir is complex. In the Tarim basin, an aquifer is situated below the reservoir. Therefore, as oil is pumped up, water goes up by capillarity effect. Moreover, in many reservoir, water or gas are injected in order to maintain pressure inside the reservoir.

If geologists can find oil deep in the ground by using variety of methods such as sismology, surface features, soil types, sniffers or satellites images, numerical simulators are needed in order to understand fluid dynamic inside reservoirs.

Very few oil reservoirs in the world produce oil from caves. In traditional reservoirs oil is contained in microporous rocks with typical pore size smaller than 30 microns.

Existing reservoir simulators currently used by the oil and gas industry are designed to simulate fluid flow in porous media governed by Darcy's law.

These simulators are not appropriate for simulating fluid flow in cave networks where Darcy's law does not apply. Moreover, Navier-Stokes model is impracticable because of the complexity of the cave's geometry, the scale of the reservoir and massive calculation and data burden that it would involve.

Thus, the ultimate objective is to design a 3D reservoir simulator that can make realistic multi-phase flow simulations (gas, water, and oil) in large and complex cave networks made of thousands of caves distributed in 3D space and connected according to a given pattern.

Bernard Montaron, Schlumberger China, proposed to investigate the feasibility of using cellular automata technology for such a simulator, and therefore initiated the SimKARST Project in 2012 with Pr. Han Pingchou, Peking University PKU.

The first part of the project started in October 2012 with two PKU students who developed a prototype cellular automata software to simulate 2D flow of water and oil in a cell with a simple geometry.

A cellular automaton is a discrete model where particules are constrained to move on points of the lattice. Each particle is in a certain state, which is be in this case oil, water, gas or rock. Movement of particules between neighboring lattices points are governed by prescribed rules based on gravity for the vertical state and on probablility for the horizontal state. For instance, a rule stipulate that oil blocks will always move upwards when water blocks are on top of them.

If this method is a fast simlulation technique and give reasonable kinetic results, it is not based on any fluid dynamics and does not respect the physical properties of fluids. Moreover, velocities are limited by lattice spacing and time. Therefore, this method needs to be compared with experiments and simulations based on physics laws, in order to validate its accurency.

* Representation of the reservoir by the automata simulator*

It order to validate and to improve the automata simulator, two ENSEEIHT students, Nicolas Sobecki and Thibault Moreau, realized a summer intership in China in order to conduct two-phase flow experiments on a specific geometry. The acquisition of this experimental data allowed to compare the experimental flow behaviour observed with flow simulated with the cellular automata software.

At the same time, Tinting Han, another ENSEEIHT student, realized CFD studies with fluent in order to simulate two-phase flow on the same geometry, in order to compare the results with experiments and data from the automata cellular. As CFD simulations are based on the Navier and Stokes equations, they respect physics law and give therefore accurate results.

*Two-phase flow experiment CFD Simulation with Fluent*