Modeling of Both Near-Wellbore Damage and Natural Cleanup of Horizontal Wells Drilled With Water-Based Drilling Fluids

Abstract

Summary Prediction of formation damage that occurs in horizontal wells, often openhole completed, is a critical point for optimizing an oilfield development. The economic impact of near-wellbore induced drilling damage and cleanup efficiency has led to significant progress in both experimental and numerical studies designed to assess the wellbore flow properties during oil production. In a previous paper, a methodology combining both experimental and numerical approaches was presented to evaluate the natural cleanup of horizontal wells drilled with an oil-based mud (OBM). This paper presents an extension of the methodology for simulating both (a) near-wellbore invasion and permeability damage generated with a water-based mud (WBM), and (b) natural cleanup during oil backflow when the well is put into production. There is a fundamental difference between WBM and OBM invasions. In an oil-bearing formation, the displacement of the oil in place with an OBM filtrate is a miscible displacement process, while the displacement with a WBM filtrate is a two-phase flow process (imbibition), generating high wetting-phase saturation in the invaded zone. Then, during oil backflow, a portion of the wetting phase is trapped, leading to residual wetting-phase saturation greater than the initial one. Even in the absence of chemical interaction between filtrate and fluids in place, this induces an adverse water/oil relative permeability effect, which is an additional permeability impairment. This paper describes a numerical approach to model the formation damage with WBM and to predict well performance for natural cleanup when the well is subject to a pressure drawdown. The kinetics of fluid filtrate invasion, the filter-cake properties, and the filtrate/oil relative permeability curves in imbibition and drainage, together with damaged and return permeabilities, are obtained from specific drilling fluid damage laboratory tests. Using these data, the fluid filtrate invasion during the drilling phase is simulated, leading to a cone-type invasion depth along the horizontal well. This approach has allowed us to study the impact of various parameters related to fluids or cake properties, drilling conditions, and natural cleanup processes on the well performance. Introduction It is well recognized that near-wellbore flow properties are altered by drilling-fluid and fluid-filtrate invasion during overbalanced drilling operations. The degree of alteration, generally called "formation damage," depends upon a large number of parameters, such as nature and characteristics of the drilling fluid, formation properties, and operating conditions (shear rate in the drilling fluid, overbalance pressure, temperature, etc.). Formation damage caused by drilling-fluid invasion may create substantial reductions in oil and gas productivity in many reservoirs.1,2 Productivity losses are especially critical for long horizontal wells which are often "openhole" completed.3 In such a case, the near-wellbore damage is not bypassed by perforations and may lead to very large skin values. Therefore, prevention of formation damage generated by a drilling fluid may not always be possible because, first, the drilling time of the horizontal segment in the producing zone is usually many times greater than in a typical vertical well, leading to a much deeper filtrate invasion; and second, the very low drawdown pressure that is needed to produce from a typical horizontal well reduces the viscous forces available to cleanup near-wellbore damage. Generally, the data obtained from production logging of many horizontal wells show severe damage across a large portion of the horizontal wellbore.4 In such a case, the numerical modeling of the full process of fluid invasion and oil backflow may give a better prediction of formation damage and an evaluation of the well performance during production. In a previous paper,5 a simplified numerical approach for modeling the natural cleanup process has been presented in the case of a horizontal well drilled with an OBM. In this paper, this modeling is extended to wells drilled with a WBM and also includes the simulation of filtrate invasion. In the case of a WBM, the two main damaging mechanisms are caused by both particulate invasion during the initial spurt loss period, and by filtrate invasion through filter cakes. Even in the absence of physicochemical interaction between filtrate and formation fluids (compatible rock/fluids systems), there is a fundamental difference between OBM and WBM displacement processes. In an oil-bearing formation, the displacement of oil-in-place with a WBM filtrate is an imbibition process that generates high wetting-phase saturation in the invaded zone, while the OBM filtrate is almost a miscible displacement process. In addition, WBM filtrate is mainly composed of polymer molecules that can deeply invade the reservoir6,7 even if the larger molecular weight species are retained in the filter cake.7 Depending on their molecular weight and filtration conditions, the polymer chains can be stretched by the flow, go through the filter cake, adsorb within the porous media, and even plug rock pores. Polymer chains associated with water increase the capillary retention of water, leading to residual wetting-phase saturation after oil backflow, higher than the initial ones. This induces an additional damaging effect (water blocking) caused by drastic reduction of oil relative permeability.8 Generally, for rating the performance of various drill-in fluid formulations, the permeability damage is quantified through oil return permeability measurements and flow-initiation pressures (FIP), performed at relevant flow rates on core samples damaged during dynamic fluid filtration tests.9–14 But only a few attempts were made to transfer these laboratory data into a near-wellbore model to evaluate the impact of the permeability damage on the well performance. Lane15 and Semmelbeck et al.16 simulated filtrate invasion for improving log interpretation, but their impact on well performance was not investigated. Other researchers17–19 have studied well performance using representative formation damage with nonuniform skin along the well, but laboratory tests were not integrated in their study. In our work, the full process of near-wellbore damage followed by natural cleanup is modeled. The WBM filtrate invasion is simulated using standard waterflooding concepts. This led to a cone-type invasion depth along the horizontal well. Filtrate/oil relative permeability curves (imbibition curves for invasion and drainage curves for backflow) are used as input parameters. In addition, filter-cake properties (thickness and permeability) and final oil permeabilities, obtained from specific laboratory measurements, are used to model the cleanup process.

On the filtrate drilling fluid formation and near well-bore damage along the petroleum well

“…During this period, an internal filter-cake is created because of the deposition of solid particles. These solid particles are retained on the wall wells and it prepares the thin filter-cake layer which controls the rate of filtrate invasion (Ding et al, 2004).…”

“…It is straightforward to emphasize that the well-bore damage deals with the drastic drop of permeability, whose effect can be seen as the oil productivity loss. The degree of damage depends on several factors such as the characteristics of the drilling fluids, the proprieties of heterogeneous formation and the pressure conditions (Ding et al, 2004;Gentzis et al, 2009;Khodja et al, 2010b;Bennion et al, 1996).…”

“…6 A smart simulation model of the damage associated with the laboratory experiments can be utilized as a benchmark guide to prevent and weaken the damage formation. In the open literature, different models for damage prediction have been proposed for the longitudinal and radial petroleum wells (Jiao and Sharma, 1994;Thomas and Sharma, 1998;Suri and Sharma, 2004;Ding et al, 2004;Parn-Anurak, 2003;Parn-Anurak and Engler, 2005;Civan, 2007;Iscan et al, 2007;Jin, 2009;Windarto et al, 2011Windarto et al, , 2012. 7 However, the above-addressed models do not take into account the inertia effect for the well-bore damage prediction.…”

Coupled simulation of near-wellbore and reservoir models

“…Modelling of near-well flow requires very fine gridblocks around the well, which is not at the same scale as the reservoir flow simulation. Some models have been developed to study the near-wellbore formation damage mechanism and the stimulation procedure (see, for example, Minssieux et al, 1998;Veerapen et al, 2001;Ding et al, 2004;Parn-anurak and Engler, 2004;Ding and Renard, 2005;Bedrikovetsky et al, 2006;Moghadasi et al, 2006;Civan, 2007;Cohen et al, 2008;Byrne et al, 2009;Lohne et al, 2009;Wu et al, 2009), but most of them are standalone for the modelling around a single well and are decoupled from reservoir and production scenarios.…”

Alteration of permeability by drilling fluid invasion and flow reversal

“…The polymer was found to be partially responsible for the damage in the cores. Ding et al (2004) studied the near wellbore damage and natural clean-up of horizontal wells. The nearwellbore properties were altered by drilling fluid, fluidfluid interaction, and fluid-filtrate invasion during overbalanced drilling operations.…”