Search

A water treatment pilot skid called WaOω has been developed by TotalEnergies to test the efficiency of the centrifugation technology in treating the produced water containing back produced polymer. In case of success, this technology would be implemented on field and the water quality targeted by the technology must allow re-injecting the treated produced water in matrix flow regime for pressure maintain and sweep efficiency. The same interest was expressed by OMV and a partnership project has been built. It was also agreed that OMV builds a second pilot skid called PRT that allows carrying out core flood tests onsite to assess the formation damage and related permeability decline that could be induced by the treated produced water. Both pilot skids have been implemented, connected to each other, and tested during more than one year on the OMV's Matzen oil field nearby Vienna where degraded polymer is already back produced by wells and present in the produced water. More than seventy core flooding tests have been performed in different centrifugation conditions in terms of speed and water qualities, some of them on high permeable sand packs representing the field targeted by TotalEnergies and some others on consolidated sandstone samples of lower permeability representing OMV reservoirs. The effect of adding fresh polymer to the treated produced water for EOR purposes has also been investigated. Some complementary core flood tests have also been performed in TotalEnergies labs using reconstituted sand packs and produced waters with and without polymer to understand the contribution of the degraded polymer alone, the produced water quality alone and both to understand the formation damage and some uncommon results observed with the PRT pilot skid. Core flood tests data often obtained on long injection periods revealed of a high quality, reliable and reproducible. They also showed that even if centrifugation seems to be a good technology, the very clean and transparent water that it delivered induced surprisingly some core permeability declines the origin of which would be discussed in this paper. However, it was clearly established that the presence of degraded polymer has a cleaning effect and limits the formation damage induced by the produced water injected on cores if the Total Suspended Solids in the treated water remains at an acceptable level. Adding fresh polymers limited even more the formation damage because their cleaning effect is more pronounced than with degraded polymer.

Formation damage is still a controversial subject within the industry especially when it takes place during Produced Water Re-Injection (PWRI). However, there is a consensus that formation damage results from a complex interaction between water quality, petro-physical and rock mechanical properties of the reservoir as well as the well architecture and completion. A lot of experimental work has been made in these last decades to understand and quantify the main mechanisms, parameters governing formation damage due to PWRI and the literature in this domain is plentiful. Nevertheless, a lot of questions are left unanswered particularly when it is a question of unconsolidated sands. This paper describes a core-flooding program performed with sand packs at different permeabilities, variable water qualities and injection conditions. Experiments were performed by injecting well-characterized system that contains either solid particles, mono-sized oil droplets or both of them at various sizes and proportions. It was shown that for mixture of solid and oil particles, the damage approach based on flow of solid suspensions alone doesn't apply and solid particles-oil droplets interaction is to be considered. This paper also details the role of each of the produced water components, their interaction and deposition within the core as well as the subsequent damage they induce. More generally, answer to some practical questions would be provided like: Do solid particles act by their size or concentration? Is oil in water content a key factor? How and where solids and oil deposit? Is there a critical concentration of solid particles and/or oil allowing injecting always in matrix regime and avoid fracturing unconsolidated reservoirs which often lie at shallow depths and in which fracturing can induce cap rock integrity and injection confinement issues? Would it be possible to control formation damage by controlling the injection conditions (flow rate, water salinity, etc…)? The answers provided here would give practical rules to help in adequately designing and optimizing PWRI projects.

The evolution of the residual oil saturation as a function of the trapping number $$N_\mathrm{t}$$ (capillary number plus Bond number), is generally known as the capillary desaturation curve (CDC) and constitutes an important input parameter in chemical enhanced oil recovery flooding. However, less importance has been paid to the investigation of the influence of oil ganglia evacuation on relative permeabilities. We report on an experimental investigation dealing with the effect of flooding parameters, fluid interfacial properties and rock structure on the CDC and on the water relative permeability. Experiments were performed on a set of water-wet sandstone plugs with different petrophysical properties, and X-ray computed tomography (CT-scan) imaging was used to accurately measure the local oil saturation. Oil ganglia size distribution as well as pore scale geometrical properties was also quantified at the scale of the micrometer using high-resolution micro-computed tomography (MCT). Results showed that the CDC depends on the pore structure and specifically on the average throat radius and the inverse of the relative permeability. Oil ganglia size distribution obtained by MCT follows a typical power law as suggested by percolation theory. We also showed that CDC obtained from macroscopic measurements can be predicted from the measured oil ganglia size distribution and rock structure geometrical parameters. In parallel, we observed that for low trapping numbers, water relative permeability is independent of the trapping number. However, for intermediate trapping numbers, a strong dependence of the water relative permeability on the latter can be noticed. In this range, we showed that water relative permeability has a specific scaling with the trapping number.

At the end of the secondary recovery in water-wet reservoirs, the remaining oil saturation might be high due to capillary trapping of the oil ganglia within the micro-structure of the rock. However, this trapped oil can be produced if the capillary barrier is suppressed by surfactant injection. To characterize this phenomenon, the Capillary Desaturation Curve (CDC : the residual oil saturation in a rock sample as a function of the trapping number Nt (sum of the Capillary and Bond numbers)) needs to be evaluated experimentally. This measure constitutes an important input parameter in chemical EOR flooding for screening and simulation. In this work, we investigated experimentally the relation between CDC and microscopic properties at the pore scale : the oil ganglia size distribution and the porous structure. Experiments were performed on a set of water-wet sandstones with different petrophysical properties. Computed Tomography imaging (CT-Scan) was used to accurately measure the mean residual oil saturation and the local saturation distribution at the plug scale (macro scale). Oil ganglia size-distribution as well as pore geometrical properties were also quantified using high resolution micro-CT images at the pore scale. The effect of trapped oil on water relative permeability is also discussed. Results showed how the CDC depends on the pore structure and is strongly related to the oil ganglia size distribution at residual oil saturation. We show that based on those microscopic properties the Capillary Desaturation Curve measured on the macro plugs can be directly predicted. We show also that in this regime of oil ganglia mobilization the water relative permeability has a specific scaling with the trapping number.

Surfactant flooding is one of the effective technologies to enhance oil recovery of water flooded petroleum reservoirs. Several previous studies have demonstrated that the release of oil from a rock submitted to waterflood depends on the competition between displacing forces (viscous and buoyancy) and capillary forces. This competition is expressed by two dimensionless numbers, the Capillary number, defined as the ratio of viscous to capillary forces and the Bond number, defined as the ratio of gravity to capillary forces. At the core scale the evolution of the oil residual saturation as function of the capillary number is better known as the Capillary Desaturation Curve (CDC), which constitutes an important input parameter in EOR processes reservoir simulation. In this work we present a new experimental workflow to investigate the effect of rock structure on the CDC. This workflow combines core flood experiments and CT-scan imaging to accurately measure the mean residual oil saturation at different capillary number values as well as the local saturation distribution along the core plug. The CDC was measured on a set of water-wet sandstone with different permeability and degree of heterogeneity. The capillary number was varied by injecting a surfactant solution at different flow rates. Using this methodology, we first study the effect of core flood orientation (Horizontal injection and vertical downflow injection). The resulting local saturation curves show that in a horizontal configuration the oil migrates upwards in the sample inducing an oriented front strongly deviating from the piston like displacement, and very different from the one observed in the vertically placed sample. This behaviour is manly attributed to buoyancy forces that are no more negligible compared to capillary forces when surfactant is injected. This experiment shows that when measuring a CDC the capillary number needs to be corrected to take into account the buoyancy forces. In a second step we have investigated the effect of local heterogeneity of core plugs on the CDC with Clashach and Fontainebleau sandstones. Two samples of Clashash sandstone with equivalent mean properties were studied (mean permeability of 380 and 426 md and mean porosity of 13.4 % and 14.1% respectively). The resulting CDC curve exhibits almost two decades difference in the critical capillary number. This discrepancy is explained in terms of the local variation of the porosity that induces important differences in the local saturation profile.