Alusta, Gamal Abdalla (Heriot-Watt University) | Mackay, Eric James (Heriot-Watt University) | Collins, Ian Ralph (BP Exploration) | Fennema, Julian (Heriot-Watt University) | Armih, Khari (Heriot-Watt University)
This study has focused on the development of a method to test the economic viability of Enhanced Oil Recovery (EOR) versus infill well drilling where the challenge is to compare polymer flooding scenarios with infill well drilling scenarios, not just based on incremental recovery, but on Net Present Value as well.
In a previous publication (Alusta et al., 2011, SPE143300) the method was developed to address polymer flooding, but it can be modified to suit any other EOR methods. The method has been applied to a synthetic scenario with constant economic parameters, which has demonstrated the impact that oil price can have on the decision making process.
The method was then applied and tested (Alusta et al., 2012, SPE150454) with varied operational and economic parameters to investigate the impact in delaying the start of polymer flooding to identify whether it is better to start polymer flooding earlier or later in the life of the project. Consideration was also given to the optimum polymer concentration, and the impact that factors such as oil price and polymer cost have on this decision. Due to the large number of combined reservoir engineering and economic scenarios, Monte Carlo Simulation and advanced analysis of large data sets and the resulting probability distributions had to be developed.
In this paper the methodology is applied to an offshore field where the choice has already been made to drill infill wells, but where we test the robustness of the method against a conventional decision making process for which there is historical data. We do this by performing calculations that compare the infill well scenario chosen with a range of polymer flooding scenarios that could have been selected instead, to identify whether or not the choice to drill infill wells was indeed the optimum choice from an economic perspective.
We conclude from all the reservoir simulations and subsequent economic calculations that the decision to drill infill wells was indeed the optimum choice from an economic perspective.
Inorganic scales precipitate in oilfield systems - downhole in the reservoir, in the production flow tubing, and in surface facilities - as a consequence of thermodynamic changes that affect the flowing brines. These changes may be induced by temperature or pressure changes, or by mixing of incompatible brines. While much work has been performed to study the effect of thermodynamic changes such as pressure decrease or temperature increase on scale precipitation, it is only recently that a body of work has been developed on the impact that the dynamics of brine mixing in the reservoir has on scale precipitation in situ. Much of this work has been conducted using finite difference simulators, which are handicapped with regard to these calculations in that numerical dispersion effects can be orders of magnitude greater than physical dispersion.
The introduction of chemical reaction calculations into streamline simulation models presents a very significant opportunity for improving the accuracy of such calculations. While numerical dispersion effects for immiscible calculations (eg water displacing oil) can be countered by pseudoisation of the relative permeability functions, in finite difference models it is difficult to control numerical dispersion for miscible displacements - eg seawater (with a SO4 concentration) displacing formation water (with a Ba concentration), which may lead to scaling in the reservoir (BaSO4 precipitation). Streamline simulation reduces the numerical errors for both miscible and immiscible displacement - so making the scaling calculations much more accurate.
The objective of this paper is to study the application of a streamline simulator that has the appropriate chemistry modelling capabilities to simulate realistic reservoir scenarios. The workflow will consist of two stages:
1) Study of synthetic systems to identify the impact of brine mixing in simple scenarios (eg single layer and multi layer quarter five spot patterns)
2) Application of the technique to full field reservoir systems where produced brine chemistry data are available.
The calculations performed will demonstrate where, and under what conditions, scale precipitation takes place in situ in the reservoir, and what the resulting impact on the chemical composition of the produced brine will be. This information is key in the planning of management of oilfield scale, especially in deepwater developments where options for scale mitigation may be limited. As a result of this work, it is expected that guidelines will be drawn up for the use of streamline simulation, particularly during the FEED stage of new projects, and when considering infill well drilling in mature projects. These guidelines will be illustrated by reference to field examples.
Introduction to Streamline Simulation
Streamline simulations approximate 3D fluid flow calculations by the sum of 1D solutions along streamlines. The choice of the streamline directions for 1D calculation makes the approach extremely effective for modelling convection-dominated flows in the reservoir. This is typically the case when heterogeneity is the predominant factor governing the flow behaviour. The geometry and the density of the streamlines reflect the impact of geology on fluid paths providing better resolution in regions of faster flow.
The volume of CO2 that can be stored in the Captain Sandstone saline aquifer in the North Sea was investigated by building a geological model and performing numerical simulations. These simulations were also used to calculate the best position for the injection wells, and the migration and ultimate fate of the CO2.
The overall migration of CO2 and the pressure response over the entire saline aquifer was studied by the calculated injection of 15 million tonnes CO2 per year. The injection rate was restricted to a maximum of 2.5 million tonnes CO2 per year for each of a possible 12 wells considered. An important objective was to predict how to avoid flow of the injected CO2 toward potential leakage points, such as the sandstone boundaries and faults. The migration of injected CO2 towards existing oil and gas fields was also a determining factor.
The summary conclusions are:
- The Captain Sandstone saline aquifer has significant potential CO2 storage capacity. Even with all boundaries closed to flow, the probable storage capacity is calculated to be about 358 million tonnes, giving a storage efficiency of 0.6% of pore volume, with an expected operating life-span of 15-25 years.
- The possible storage capacity of the formation may be at least four times greater if the aquifer boundaries are open. This increase would be a result of displacement of salt water, and not CO2.
- The storage capacity if the sandstone is closed to flow may be increase from 358 to 1668 million tonnes of CO2 by significant additional investment in 15 to 20 water production wells.
- Injection of up to 2.5 million tonnes CO2 per year in one well has an impact on the pressure throughout the entire formation, and thus interference between different injection locations must be considered.
Keywords: CO2, CCS, Storage Capacity, Saline Aquifer
There may be various drivers to implement Produced Water Re-Injection (PWRI). However, re-injecting produced water from the same field cannot replace the voidage created by production, especially early in the life of the field, since most of that voidage is created by hydrocarbon extraction. Thus seawater may have to be considered to "top up?? PWRI. This raises the question of what are the implications for scale control of mixing potentially incompatible brines before injection, compared to the conventional injection scenario where the mixing takes place in the reservoir.
A study was set up to consider scale management during the life cycle of four offshore fields. The available data included analysis of formation and produced water and seawater compositions, and the time evolution of the produced water - seawater split in the injection system. The tools used included thermodynamic scale prediction and reservoir simulation calculations. Thus the evolution of the scale risk over the entire water cycle - from injection, through the reservoir, to production could be evaluated.
The produced water compositions and the results of the calculations show that the scale risk at the producers is much lower than if only seawater had been injected. Calculations were also performed to identify whether bullhead application of scale inhibitor would provide adequate protection for the wells. This was important, as some of the wells are subsea completions. The clear conclusion was that any residual scale risk at the producer wells could be managed by bullhead squeezing.
However, the corollary is that the scale risk at the injectors is much higher. The trigger for scale precipitation in this scenario is brine mixing, but instead of that happening in the reservoir, here it occurs before injection. Thus the location of greatest scale risk is moved much further upstream in the flow process.
Keywords: Oilfield scale, barium sulphate, PWRI, reactive transport simulation
Gomes, Roberto (Petrobras S.A.) | Mackay, Eric James (Heriot-Watt U.) | Deucher, Ricardo Huntemann (Petrobras) | Bezerra, Maria Carmen Moreira (Petrobras S.A.) | Rosario, Francisca Ferreira (Nalco Company) | Jordan, Myles Martin
Evaluation of the scaling risk at production wells is generally carried out using thermodynamic prediction models. These models are generally very accurate in terms of predicting the type of scale that may form, the degree of supersaturation, and
the mass of scale that will deposit by the time the system reaches equilibrium - provided the brine composition or compositions involved are well known, and the pressure and temperatures conditions are accurately specified. However, in
performing these calculations, engineers and chemists often fail to take account of reactions occurring in the reservoir, and assume that brines reaching the production wells have not reacted in any way prior to entering the wellbore. This often leads
to a significant overestimate of the scaling risk.
The work presented in this paper addresses this issue by studying data from various fields to identify what can be learnt from the produced brine compositions. A new technique to estimate the range of scaling tendencies that takes account of
reservoir precipitation is developed, and the results are displayed in a 3D response surface. This is illustrated for barium sulphate scaling tendency, accounting for different levels of ion stripping.
In order to calibrate some simulation parameters, and to identify the more important equations that should be inserted in the reservoir simulation, studies were performed based on the observed data. Different reservoir simulations were used and
compared, with a focus on scale management to identify positive and negative aspects of each one.
This work has identified that in fields with reservoir temperatures above 120°C and calcium concentrations above 7000 mg/l, significant sulphate stripping occurs due to anhydrite precipitation. This effect is increased where ion exchange
leads to a reduction in magnesium and an increase in calcium concentration as the injected brine is displaced through the reservoir.
Oilfield scale formation represents a very significant flow assurance challenge to the oil and gas industry, with increasing water production worldwide and higher oil prices. Scale Inhibitor (SI) squeeze treatment is the most widespread method to
combat downhole scaling. In order to predict SI squeeze treatments accurately for further optimisation, it is necessary to simulate the SI retention in the formation, which may be described by pseudo-adsorption isotherms. While these are often
derived from core flood experiments, sometimes they are not appropriate for modelling well treatments because the core tests on which they are based cannot fully represent field scale processes. In practice, the parameters of an analytic form of the isotherm equation are modified by trial and error by an experienced practitioner until a match is obtained between the prediction and the return profile of the first treatment in the field.
The main purpose of this paper is to present a Stochastic Hill Climbing Algorithm for automatic isotherm derivation. The performance of the algorithm was evaluated using data from three field cases. Two success criteria were defined: firstly,
ability to match a single historical treatment and secondly, ability to predict subsequent successive treatments. To test for the second criterion, a candidate isotherm was derived from the first treatment in a well that was treated with the same chemical package on consecutive occasions, and then the predictions using the suggested solution were compared with the observed scale inhibitor concentration return profiles from the subsequent treatments. In all the calculations, performance of the isotherms suggested by the Hill Climbing algorithm and isotherms derived by trial and error were compared. The results demonstrate that the Hill Climber Algorithm is a very effective technique for deriving an isotherm to enable accurate modelling of scale inhibitor squeeze treatments.
The injection of seawater into oil bearing reservoirs to maintain reservoir pressure and improve secondary recovery is a well-established, mature operation. Moreover, the degree of risk posed by deposition of mineral scales (carbonate/sulphate) to the injection and production wells during such operations has been much studied. The current deepwater subsea developments offshore West Africa, Gulf of Mexico and Brazil have brought into sharp focus the need to manage scale in an effective way.
In recent years there has been some consideration given to deployment of scale inhibitor within the fluids associated with the completion of production wells, prior to the start up of production. Until now, effective scale control in frac packed wells at low water cuts has been achieved with phosphonate-based inhibitors applied as part of the acid perforation wash and overflush stages, prior to the actual frac packing operation itself. The deployment of these inhibitors has proved effective in controlling barium sulphate scale formation during initial seawater production, and eliminating the need to scale squeeze the wells at low water cuts (<10% BS&W). Recent developments allowing inclusion of scale inhibitor in the linear and cross linked gel stages has highlighted the need to be able to model this process effectively, thereby enabling optimal use of the chemical and improved squeeze designs.
This paper outlines simulation work carried out using the Petroleum Experts REVEAL software to assess introduction of scale inhibitor into frac pack operations, and identify the most suitable stage of the well completion process during which to apply the inhibitor, to maximise treatment life. Simulation results and field data from these treatments are compared to demonstrate the opportunity this technique presents, and to highlight the importance of chemical placement and the post stimulation flow regime to squeeze life.
In a previous publication we introduced a methodology to assist in choosing between Enhanced Oil recovery (EOR) and infill well drilling (SPE 143300). Operating companies are often reluctant to use EOR techniques when they have the option of infill well drilling instead. Reasons for this include how operating companies assess and manage risk and uncertainties. The methodology developed includes performing reservoir calculations to evaluate additional recovery using both techniques, and then using data generated as input to economic analysis. In the previous work, polymer flooding for 10 years after two years of waterflooding was studied using a synthetic reservoir model. The technique involved running a range of reservoir simulation scenarios to test possible recovery outcomes; these outcomes then provide input data that will be used in the probabilistic economic evaluation tool to be introduced as a follow up in this paper.
This current paper presents the results of the impact of operational factors, such as delaying the start of polymer flooding. This involves assessing the best possible timing for polymer injection to achieve optimal economics. This type of assessment is possible because the economic model developed and presented here allows input from multiple reservoir simulation sensitivity calculations. Monte Carlo Simulation (MCS) is then performed to establish confidence in the method, and test economic uncertainties and the risks associated with implementation of polymer flooding. Defining variables with a probability distribution can establish more precisely the economic value of the polymer flooding project. The analysis of uncertainty involves measuring the degree to which input contributes to uncertainty in the output.
MCS is a statistics based analysis tool that yields probability impact on Net Present Value (NPV) of the key operational parameters included in the project (oil, water and polymer production and injection costs, polymer concentration, timing, etc.) and various economic factors (oil price, polymer cost, etc).
Sequestration of carbon dioxide in geological formations has drawn increasing consideration as a potential method to reduce the level of CO2 in the atmosphere, and therefore mitigate climate change. In particular, saline aquifers can potentially provide a large storage volume world-wide. It is essential to assess the risk involved in storing CO2 in the subsurface, and simulations of CO2 injection play an important role. Detailed simulations using a compositional simulator, which solves the equation of state for the fluids and calculates the partitioning of fluids between phases, is time consuming. It is therefore advantageous to use a simpler method for simulation, such as a modification of a black-oil simulator (designed for use in the oil industry), where fluid properties are input using look-up tables.
In this study, we have tested the accuracy of flow simulations of CO2 storage in saline aquifers using a black-oil simulator (BOS) compared with a compositional simulator (CS). A range of models was investigated: 2D, 3D and radial models, horizontal and tilted, and homogeneous and heterogeneous. On the whole the results compared well, although accuracy of the BOS depended on the type of grid used, being less accurate for radial models, where discretisation effects were evident.
In agreement with other studies, we found that the black-oil simulations were, on average, a factor of four faster than compositional simulations.