Numerical simulation and X-ray imaging validation of wormhole propagation during acid core-flood experiments in a carbonate gas reservoir

https://doi.org/10.1016/j.jngse.2016.02.036Get rights and content

Highlights

  • We tried to simulate acid flow through a core sample from a carbonate gas reservoir.

  • Data from core-flow experiment was used to justify simulation results.

  • A developed 2-D continuum model was solved in 3-D domain.

  • The flooded core, was x-rayed to understand prediction capability of the model.

  • A very good agreement between experiments and simulation was observed.

Abstract

Acidizing is a sophisticated stimulation treatment that significantly improves production rate particularly in the carbonate reservoirs. However, performance evaluation of the treatment requires extensive laboratory experiments. This includes investigation of various parameters such as temperature, pressure, mineralogy of the rock and the effect of chemical diverters in the zones with high contrast permeability (Safari et al. (2014). A tuned mathematical model, capable of describing the acidizing in core-scale, can play a vital alternative role due to high expenses of the required experiments. In this paper, we used a simulation model, tuned by experimental data, to be used as an alternative tool for modeling acidization. Since wormhole propagation is strongly geometry dependent, the model was solved in a 3D domain rather than its original dimension, i.e. 2D in Panga (2005) model, to simulate the propagation more accurate.

Mass transfer, continuity, and energy equations were combined to develop a model, which simulates the acid flow in porous media accurately. The model, a cross-link between two different scales: pore-scale and Darcy-scale considers the effect of convection, diffusion, and particularly chemical reactions.

In this study, a complete package including acid core-flood experiment, X-ray imaging and numerical simulation of the wormhole propagation was performed to study acid flow in the porous media. The acid core-flood experiment was performed on a core sample from a gas reservoir in south of Iran. In addition to X-ray imaging technique employed to visualize the result of acid core-flooding, a simulation model was developed in finite element analysis software. The model has the dimensions of the core sample. The required chemical reaction equations were implemented in the simulation model to simulate the wormhole propagation process. A very good agreement was achieved when the simulation results were compared with experimental results and X-ray imaging outputs.

The model, as an invaluable tool could be used to study the different aspects of acidizing treatment such as acid efficiency curve, simulation of parallel flooding, etc. This then helps to determine key parameters and optimized conditions for achieving a successful acidizing treatment.

Introduction

Generally, all applied sciences are based on conducting experiments and interpreting their results. These experiments are to investigate the impact of various parameters affecting performance of the system. One of the most common methods is to employ an appropriate mathematical model to evaluate effect of individual system variables and their mutual effects. In this study, we used a model that was developed by Panga (2005) to simulate wormhole propagation in the acid-flood experiments in a carbonate gas reservoir.

Many experimental studies such as Fredd and Fogler, 1998, Bazin, 2001, Safari et al., 2014 investigated the wormholing process and acid efficiency curve. The majority of the studies focused on limestone formation, however, for instance, Fredd and Fogler (1998) showed a significant difference in pore volume to breakthrough curve for the dolomite core used. This kind of difference occurs especially at low temperatures, in which reaction rate in dolomite is much slower than in limestone. The pore scale heterogeneity, which exists because of the different structure of the microbial carbonate, is also reported important by Ziauddin and Bize (2007). Many other experimental studies investigated various parameters that can affect reaction between acid and rock matrix during any acid-based treatment (Mumallah, 1991, Mumallah, 1996, Mumallah, 1998, Sjoeberg and Rickard, 1984, Taylor et al., 2006, Taylor and Nasr-El-Din, 2009).

Since majority of studied core's mineral composition was consisted of calcite (73%) and dolomite (25%) plus very little amount of siderite (>1%), following reactions were the main mechanism of change while acid injection:I)CaCO3+2HClCaCl2+CO2+H2OII)CaMg(CO3)2+4HClCaCl2+MgCl2+2CO2+2H2OIII)FeCO3+2HClFeCl2+CO2+H2O

Besides the laboratory activities, many other studies were performed to develop models simulating the flow of the acid in porous media.

Steefel and Lasaga (1994) described a numerical model for computing coupled multi-component chemical reactions, multi-species chemical transport, hydrodynamic flow, and heat transfer. Their model employs a new algorithm which solves simultaneously for multi-component reactions and solute transport in one and two dimensions and which uses kinetic formulations for mineral dissolution and precipitation reactions, making the a priori assumption of equilibrium between water and minerals unnecessary.

Liu et al. (2013) used a model which couples a two-scale continuum model simulating wormholing in the invaded zone and a reservoirflow model for the compressed zone was used to study the wormhole propagation behavior under reservoir conditions. Their results showed that the normally distributed porosities simulate wormholing better.

Generally, these models can be categorized into four main groups: namely fractal, capillary tube, pore network, and continuum (average) models.

Relating important dimensionless group to experimental observation is a solution that is followed in fractal models (Fredd and Fogler, 1998, Fredd, 2000, Daccord and Lenormand, 1987, Frick et al., 1994, Pichler et al., 1992).

Capillary tube models are capable of evaluating effect of mass transfer, mutual relation of the wormholes, fluid loss, and reaction on wormhole growth. Interaction and competition between wormholes are key outputs of capillary tube models. Although these models are quite simple, they do not consider the effect of chemical reaction in pore scale as well as transport mechanism. This simplifications may affect the optimum conditions for acidizing (Hung et al., 1989, Huang et al., 1999, Huang et al., 2000, Buijse, 2000).

Pore network model represents porous media in form of many paths that are connected in nodes. Studies performed by Fredd and Fogler (1998) and Hoefner and Fogler (1989) are well-known examples of such modeling. Hagen–Poiseuille relationship describes acid flow in these paths for laminar regime. The acid reaction is considered in form of increase in diameters of these paths. However, scaling up of the result to the field conditions is an issue for these models. Applying such models require large computational power while including heterogeneities effect in pore scale is not usually straightforward.

Due to the deficiencies of first three categories of models, it is generally believed that continuum models are more precise in systematic study of acidizing process. There are three continuum models available: Liu et al., 1997, Golfier et al., 2002, and Panga (2005). While acid is entering the pores, if the reaction rate is very slow compared to the mass-transfer rate, the concentration gradients are negligible. In this case, the reaction is considered to be in the kinetically controlled regime, and a single concentration variable is sufficient to describe this situation. However, if the reaction is very fast compared to the mass transfer, steep gradients develop inside the pores. This regime of reaction is known as the mass-transfer controlled regime. The Liu et al. (1997)’s model does not take account for mass transfer in reaction and considers chemical reaction at local equilibrium. This model has some limitation to the kinetic regime. The developed model by Golfier et al. (2002) is also available in mass transfer controlled regime. Panga et al. (2005) developed a model that overcome drawbacks of two aforementioned models and was infrastructure for many other studies in this area pointed out here. For instance, Izgec et al. (2010) explored the effects of heterogeneity on vuggy carbonate acidizing with high resolution computerized tomography imaging, image processing, geostatistical characterization, acid core-flood experiments, and numerical simulations with a continuum approach modeling. De Oliveira et al. (2012) developed a methodology to numerically represent the acid treatment in a test plug, as well as to reproduce the different existing dissolution patterns and to obtain the corresponding values of pore volumes to breakthrough. They used experimental data published by Fredd and Fogler (1998) to justify their simulation results.

Ratnakar et al. (2013) used an empirical rheological model to describe the relationship between viscosity of in situ cross-linked acids and temperature, shear rate and pH. They presented a two-scale continuum model to describe reactive dissolution of carbonates with in situ cross-linked acids and to analyze wormhole formation in single and dual core set-ups.

Due to the limitations of the first two models, i.e. Liu et al. (1997) and Golfier et al. (2002) that already discussed, we applied Panga's model for our case. This continuum model, extended to a 3D domain, was solved using a finite element based software and tuned by matching the simulation results with output of X-ray imaging of an acid flooded core from a carbonate gas reservoir in Iran.

Section snippets

The experimental study summary and analysis

Fig. 1 illustrates a schematic of the setup used for performing the experiments. Formation water and acid were already put in the accumulators. After putting core sample in Viton sleeves, it was inserted in core-holder. Backpressure, which was supplied by Nitrogen, kept the produced CO2 during reaction in solution. While injecting, injection rate, temperature, overburden pressure, and back-pressure were kept constant and pressure drop along the core was recorded with digital and analog

Numerical approach

The Panga's model simulates the linear acid flow in two dimensions. This two-scale continuum model (TSCM) is extended to radial flow in three dimensions in this study while including both the Darcy and pore scale physics in the model. Fig. 2 illustrates schematics of the different length scales and geometry used in this study. The mathematical model describes the phenomenon of reactive dissolution as a coupling between processes occurring at the Darcy scale and the pore scale. This model is

Simulation outputs

In following sections, some parameters are discussed that could provide great understanding of acid wormholing in the core.

Validations

In this section, using available data and tools we tried to evaluate that whether the simulation is close to the experiment or not.

Summary and conclusion

Pang's model, a continuum model, was used to model acid flow through core samples. Panga's 2D model then was extended to simulate 3D radial behavior of acid wormhole propagation of samples from a carbonated gas reservoir. Different boundary conditions, initial conditions, and core and fluid properties, adopted from the core-flow experiment, were applied to simulate the acid flow behavior using a finite element based software. The flooded core samples were irradiated to x-ray to validate the

Acknowledgment

This work was part of a M.Sc. thesis entitled “Study, Evaluation, and modeling of wormhole propagation during matrix acidizing carbonate reservoirs”. We would like to thank Amirkabir University of Technology and Research Institute of Petroleum Industry (RIPI) for their support and permission to publish the study.

Nomenclature

a0
Initial interfacial area per unit volume [1/m]
av
Interfacial area per unit volume [1/m]
Cini
Initial acid concentration [mole/mˆ3]
Cf
Acid concentration [mole/mˆ3]
Cs
Acid concentration in liquid–solid interface [mole/mˆ3]
dh
Pore hydraulic diameter[m]
Dm
Molecular diffusivity of acid [mˆ2/s]
De
Dispersivity tensor [mˆ2/s]
DX
Longitudinal dispersivity [mˆ2/s]
DT
Transversal dispersivity [mˆ2/s]
kc
Mass transfer coefficient [m/s]
ks
Surface reaction constant [m/s]
K
Permeability tensor [mˆ2]
K0
Initial permeability of

References (29)

  • G. Falkovich

    Fluid Mechanics: a Short Course for Physicists

    (2011)
  • C. Fredd

    Dynamic model of wormhole formation demonstrates conditions for effective skin reduction during carbonate matrix acidizing

  • C.N. Fredd et al.

    Influence of transport and reaction on wormhole formation in porous media

    AIChE J.

    (1998)
  • Thomas P. Frick et al.

    An improved modeling of fractal patterns in matrix acidizing and their impact on well performance

    SPE Prod. Facil.

    (1994)
  • Cited by (34)

    • Development and application of a double action acidic emulsion for improved oil well performance: laboratory tests and field trials

      2021, Colloids and Surfaces A: Physicochemical and Engineering Aspects
      Citation Excerpt :

      From the X-ray CT images of selected rock samples we establish that after acidizing with Emulsion A, porosity of the rock samples increases on average by 20 % relative to the initial value, and the pore volume increases by 18 %. This result suggests that Emulsion A performs significantly better compared with conventional emulsions used worldwide, with which the average increase in porosity is in the range between 10–15% [19,48–51]. It is well established that carbonate rocks are chemically reactive with acids in general and with HCl solutions in particular.

    • Validation of a reactive transport model for predicting changes in porosity and permeability in carbonate core samples

      2019, International Journal of Greenhouse Gas Control
      Citation Excerpt :

      Kalia and Balakotaiah (2009) improved the 2-D model by exploring the effect of variations in initial porosity, although they noted that this technique did not extend to the description of fractured media. Safari et al. (2016) then extended the same model to 3-D volumes for simulation of wormhole propagation through a single core. Biswal et al. (2009) described a numerical workflow procedure for discretizing 3-D simulations of an oolitic carbonate rock volume at arbitrary resolutions and noted “reasonably accurate predictions of transport parameters at intermediate and higher resolution,” although no explicit comparison to experimental data was available.

    View all citing articles on Scopus
    View full text