Determination of Water Drive Mechanism Activation Using Geological Model for Heavy Oil Field in North of Iraq

Abstract


Introduction
Water drive is an important source that heve a direct effect on production of oil reservoirs when it is considered active aquifer.However, aquifer properties such as size, porosity, and permeability are typically uncertain and making predictions of field performance to be a big challenge (Glegola et al., 2012).Aquifer charecteristics are determined by analyzing the petrophysical properties in the areas close to the contact region between oil and water (Darwesh and Al-Tamimi, 2019).The observed pressure decline is in direct propotion to the activation of water influx, and large uncertainties in the actual reservoir state are common (Jean et al., 2021).
Aquifer characterization and water influx into reservoirs (oil and gas) can be estimated via several mathematical approaches, known as aquifer models, as they proceed to simulate changes in pressure within the aquifer and between the aquifer and the reservoir.The main approaches in use are related with the type of flow within the reservoir/ aquifer; either steady-state or unsteady-state (Jean et al., 2021).
Water influx is commonly applied when the aquifer attached to an oil reservoir is small and the reservoir pressure decline under primary depletion acts as a significant limitation on the recovery factor.It is important, therefore, to quantify the strength of the aquifer to assess if future waterflood is needed.If the initial pressure is significantly higher than the bubble point pressure, it might be possible to learn something about the aquifer influx through a period of primary production without inducing any of the problems associated with secondary gas (Chappell, 2020).Aquifer flow modeling has been used at many reservoir models with varying degrees of success.The ability to reliable predict the rate and direction of aquifer flow is critical in planning and implementing (Bear, 1992).
At present time, and due the development technologies for handling and processing large amounts of reservoir data which lead to construct more reliable geological and reservoir models, as evidenced by the work on the automation of the construction of aquifer models which are usually coming with big uncertainties in the data (Abraham et al., 2019), although it is coming out with acceptable results that can model the reservoir for next time step with confidence and accuracy.
This study will handle a case of modeling the aquifer of a heavy oil field in North West of Iraq named Qaiyarah Oil Field to investigate the aquifer Petrophysical properties and ensure its activation.This field have not developed properly so far although it is discovered in 1918, the production from this field till today is less than 10% of the total reserve (MoO, 2020).
After completing the full geological model for the field and due to the lack in the availability of reliable subsurface information and production data, it found out that there is unceranity in determining of the aquifer support to the oil reservoir.
Qaiyarah Field, Tertiary Reservoir, is an anticline trap forming part of a bigger structure composed of 4 anticlines: Qaiyarah, Najmah, Jawan, and Qassab.The first three structures are successive individual culminations on a single fold axis, which trends approximately northwest from the western bank of the Tigris at Qaiyarah.Qassab is a parallel anticline, with two domes, which is offset to the northeast, and separated from Jawan by a broad, shallow synclinal saddle (Al-Mubarak, 1978) (Fig. 1).

Materials and Methods
The purpose of this study is to investigate the activation and strength of the aquifer in maintaining the production from Qaiyarah field depending on the static model (geological model) that build using PETREL software and will be used to simulate the flow model using PETRELRE (eclips) software.The dimensions of grid cells (size of the simulation model) was chosen based on the model accuracy, raw data quality and time.Then simulation setup combined with the history matching was performed for reservoir model calibration through incorporating observed well production data to determine if the aquifer is active for pressure support or not since its scenario provided the best historical match for pressure and production data.
For a long time, heavy oil flow in carbonate reservoirs has been a challenging issue in scientific research (Zhang et al., 2017).Accurate flow simulation is essential for the efficient exploitation due to the heterogeneity and complexity of the porosity-permeability and fracturing system (Al-Dabbas et al., 2020;Mahmood and Al-Fatlawi, 2021;Farouk and Al-haleem, 2022).
This work comprises full field simulation.The objectives at this stage are to build a full field model of the tertiary reservoir and to match the field performance of pressure and production, where the main affecting parameters required will be explained in the next sections.

Modelling Petrophysical Properties
The Petrophysical characteristics must be determined properly to model the reservoir performance (Jameel et al., 2020;Wahid et al., 2022).The followings are the Petrophysical properties that were evaluated for the reservoir model simulation in Qaiyarah oil field:

Reservoir rock porosity modelling
The total porosity is calculated by using of the Neutron-Density logs.The Neutron porosity was corrected of shale effect using equation 1 (Tiab and Donaldson, 2004).
Where; ØD = shale effect corrected density derived porosity (fraction). ma = density of matrix; (2.71 gm/cm 3 ) for limestone, (2.87 gm/cm 3 ) for dolomite.b = formation bulk density recorded by density log (g/cm 3 ).f = fluid Density (mud filtrate); (1 g/cm 3 ), for fresh water or (1.1 g/cm 3 ) for salt mud.Density derived porosity although was corrected of shale effect using equation 3 to remove the effect of shale from the calculation of porosity in intervals with a shale volume more than 10%. Where; ∅N corr.= Shale effect corrected density derive porosity (fraction).∅D = Raw density derive porosity (fraction).
Effective porosity ∅  was computed by removing the shale related porous from the total porous network using equation 5.
∅ effc.= ∅ t × (1 − V sh ) (5) Where; ∅  = effective porosity.∅  = total porosity. ℎ = shale volume.In Qaiyarah oil field (Tertiary Reservoir), estimates have been made of the total/effective reservoir porosity and reservoir matrix porosity using the log interpretation tool in the PETRELRE.Since no correction was applied for Vclay the total porosity and effective porosity estimates are the same (Fig. 2).

Reservoir water saturation modelling
Water saturation have been computed using Archie formula as in equation 6 (Archie, 1942).
Where for Qaiyarah Field, Formation Resistivity "Rw" is 0.28, the cementation factor "m" is vary in between (1.5 to 2.5).Several SCAL (Special Core Analysis) of "m" as shown in Table 1 (NOC,2006), where clearly indicated that porous and fractured, dolomitized grain stone facies defined as good and very good facies in the Jeribe and Euphrates formations yields as low values for "m" as 1.57.
On the other hand, in very upper part of Jeribe formation, the dense muddy wackestones and strongly cemented grainy facies which are regarded as poor reservoir facies, with non-connected vuggy porosity and complex pore geometry, yields a value higher than 2. Some intermediate values are recorded for variably and fractured and/or vuggy dolomites and dolomitic Limestone.Significant errors can occur in the calculation of fluid saturation if the wrong value of "m" are used, then the computed water saturation (Sw) will show erroneously high saturation.A high cementation factor of 2.4 indicates poor pore space connectivity and an expected low permeability.This is in direct contrast to the SCAL data and available core descriptions.
In this study, the Archie equation constants used are cementation factor "m" = 1.8, saturation exponent "n" = 2 and constant "a" = 1.0 (NOC,2006) and average water saturation is calculated at 31.5%, Fig. 3 shows the water saturation distribution in the Qaiyarah Field.

Reservoir rock-fluid modelling (relative permeability and capillary pressure)
The relative permeability is a dimensionless term that is the ratio between effective to absolute permeability and it's a dimensionless functions of saturation with values generally ranging between 0 and 1 (Gilman, 2003).Relative permeability is important for estimating the flow of reservoir fluids.It is often recommended that laboratory relative permeability experiments for reservoir performance predictions should be conducted under simulated reservoir conditions (Eleri et al., 1995).
Since the fluid flow in reservoir occurs for more than one fluid, which means the overall ability of any fluid to flow will be affected by the presence of any other fluid in the same porous media.In order to develop estimates of fluid behavior in reservoirs, the phenomena need to be quantify in some way (Kantzas et al., 2012).Many heavy oil carbonate reservoirs in the world showed uncertainity in determination of the capillary pressure and relative permeability charecterstics in the reservoir areas near the contact between the oil and water (Al-Sudani, 2014).Due to the lack of SCAL information in Qaiyarah field, only one well's core analysis report was available and used in this study to model the relative permeability and capillary pressure.
The distribution of relative permeabilities and capillary pressure curves within each formation has been modified to match the performance of the fluid flow in the reservoir.
The effect of relative permeabilities on the water cut depends mainly on the slope of the relative permeability curves at low water saturations.And due to the lack of the information from the core analysis which is not sufficient enough to investigate them confidently by the numerical simulation.Although the available data and specially after refining it gave some valuable indication of the wettability system for the reservoir rocks included all the formation those investigated in this study (Jeribe, Dhiban and Euphrates).
The analysis indicated that the rocks are almost water wet which is considered a special case, since globally most of the heavy oil reservoirs are moderate to fully oil wet (Neshat and Pope, 2018;Al-Ameri and Mazeel, 2020).Corey formulation was used to compute the sets of relative permeabilities of (Jeribe, Dhiban and Euphrates) curves and results are showed in Table 2 (Corey, 1954).These modified saturation curves were used for the reservoir simulation.Fig. 4 and 5 shows the results of relative permeability and capillary pressure for the formations those considers the hydrocarbon bearing formations.

Aquifer Modelling
The challenge in aquifer modelling is to simplify reality in a way that does not adversely influence the accuracy and ability of the model output to meet the intended objectives.Despite their efficiency, models can be complicated and produce wrong results if they are not properly designed and interpreted (Baalousha, 2011;Ekwe et al., 2022).
The following observed data set is history matched to the Qaiyarah field: flowing bottom hole pressure (9 wells), static pressure (9 wells) (20 current producers).The field was producing for very short periods in last four to five decades, and lack of well test and pressure records, as well as the absence of water production, created confusion regarding the extent to which the aquifer is supporting the reservoir.A simulation was run to approve and discover if the reservoir is not supported by an aquifer, the key of that was depending on the pressure data in combination with the production history from the field, Fig. 6 from the field history match showed a sharp decline in the pressure with periods of production, which is not the case in the reality, and the reservoir pressure have been recorded in 2018 is 389 psi and simulation showed around 243 psi. the unmatched production data in the late of 2019 is due to well-46 which found out that the well is abandoned in early 2018 and cemented to the surface while errors in recoding production data occur in recording the well information as it is shown in Figs. 6 and 7. Based on that and to comply with observed well pressure drop, an aquifer is modeled and adjusted by increasing the porosities in outside cell range located in Southern part and Western flank of the structure.Therefore, aquifer strength was improved in the southern area where porosities, and consequently permeabilities, were not decreased with structural depth but left with high oil zone values.In Northern and Eastern flanks, limited low values for porosities and permeabilities were considered to reduce aquifer size and strength, this was accomplished by implementing many trails to match the history with an adage aquifer activation.The aquifer types and parameter have been tested in order to restitute the measured reservoir pressure.Slandered deviation of the difference between measured and computed pressures was used as a criterion to appreciate the best restitution.Among linear, hemispherical, bottom and radial aquifer types the last one was selected as the most convenient.Water drive parameters are related to dimensionless time and water influx by equations 7 and 8: Where; Ri = equivalent oil pool radius (Kilometers).Ce = total aquifer compressibility (psi -1 ).
The final aquifer characteristics that satisfied the pressure fit of the reservoir which is used to model the aquifer are: aquifer permeability (Kw) = 1500 md, constant aquifer (C) = 0.001 to 1,000,000 bbl, reduced time factor (A) = 0.1 Day -1 , oil pool radius (Ri) = 3.5 Kilometers.

Results
After determining that pressure profile for the reservoir have not matched the calculated from the simulation with conditions that all other parameters were put with high level of accuracy.Even though, the solutions of numerical reservoir simulation are pressures, production rates and fluid saturations; the fluid-saturations/fluid-contacts was not included in the history-matching process.History matching to all the available pressures, production rates and fluid saturations/fluid-contacts should increase reliability of a simulation model for forecasting.This history-matching of only pressure and production concept was applied to a matured waterflood reservoir.
The full field model is made of 179,860 cells, included three main formations from bottom to top, these formations are corresponding respectively to Euphrates, Dhiban and Jeribe.The pay zone has been represented by a regular grid size of (250 x 250) meters.Gilman referred to the fluid flow in reservoirs naturally fractured could be model using Effective Fracture Media in which contribution degree of both matrix blocks and fractured network to provide fluid storage and permitting fluid flow to be happened (Gilman, 2003).The results led to model the aquifer that correspond to the measured field parameter to match the production and pressure data and relying on the collected data from the field history to construct the base case of the reservoir behavior to forecasting and prediction scenarios of the field future production in its primary depletion stage, where below conditions were applied.
• Bottom open to water influx corresponding to bottom water drive conditions.
• No upper boundary flow.
• Oil and water flank influx computed accordingly to the history pressure imposed at the outer limit.
Depending on the history of the pressure profile from the producer wells and crude oil analysis from north east and south west located wells, it is being notice that the producers in at the north east flank never had water cut in the production history and with more pressure drop comparing to the producers from south east flank that been noticed of water traces in the analysis and lower pressure drop from the producer wells.This information was the key to model the aquifer as edge water supporting the reservoir pressure.
A history match simulation was run to approve the idea using same production and pressure history that was used in the case of the bounded reservoir, the results of the history match is shown in Fig. 8 and 9 below that approved activation of the edge water support from the south east flank.As well as it can be noticed that well 46 is corrected and it can be seen that the unmatched production data have disappeared.Table 3 shows results of the simulation case and compare it with the available data from the field.

Conclusions
• Base case simulation showed that reservoir pressure has dropped very fast and pressure profile has not matched the measured from the field.• Relative permeability models showed the three produced formations in Qaiyarah field are water wet, which considered an advantage point to model the aquifer correctly.• Because of missing 3D seismic interpretation and no data of the fracture system in the field below conditions were applied to adjust and satisfy the history match of production and pressure at the same time: -Vertical permeabilities equated to horizontal permeabilities -To compensate the fracture system participation in the flow, the permeability of horizontal and vertical directions has multiplied by 100 in all three formations.• Many trails of aquifer model simulation led to correct aquifer model for Qaiyarah field and simulation results showed perfect match with the field measured production and pressure data.

Recommendations
• Further investigation is required at the edges of the reservoir to confirm the water derive activation through continuous production from the field for long term.• Accurate fracture model is required to collect valuable information regarding the fractures orientations, that will have a major impact on selecting the type of the wells and their completion for future water injection plan when reservoir pressure goes down eventually.

Table 2 .
Corey Parameters for modified relative permeability curves

Table 3 .
Comparision results of the simulation outcomes summary and available data from the field