Unconventional reservoirs such as the Eagle Ford have had tremendous success over the last decade, and while wells come on at high rates, they drop quickly and the recovery factors are low, which suggests the need for enhanced oil recovery. One method that has become popular is cyclic (huff-n-puff) gas injection. In this method, gas is injected down a well for some time, and then the injection well is turned back into a producer until the production drops, and the process is repeated. A few companies have successfully tested this in the field, and the production data clearly indicates that incremental oil is being produced, but the physical mechanisms for the additional recovery are not well understood.
In this paper, the relative significance of four proposed recovery mechanisms is examined: (1) oil swelling, (2) viscosity reduction, (3) vaporization, and (4) pressure support. A numerical flow simulation model is used to study these effects. A model of an unconventional reservoir is constructed where all these mechanisms are present and contributing to the recovery. To validate the model, it is history matched to a pilot gas injection project in the Eagle Ford. A primary case and a full gas injection case are completed. A run is also completed where all of the recovery mechanisms for gas injection are turned off, which provided a result that is similar to primary production. Then each mechanism is turned on and off (one at a time), and the model is re-ran to determine the relative contribution of each mechanism. This process is carried out for different reservoir fluids from low gas-oil-ratio (GOR) black oils to liquid rich gas condensate.
By evaluating the recovery mechanisms for fluids at various GORs, charts are created that show how important each mechanism is as a function of different reservoir fluids. All mechanisms provide some contribution, but their significance varies as a function of GOR. Pressure support provides similar small response for all fluid types. Vaporization is most important for gas condensate reservoirs (high GORs), but it plays a role for all fluid types. Oil swelling has a large impact for low GOR oils, but diminishes for higher GOR fluids, and viscosity reduction plays a minor role only for low GOR cases.
Most importantly, the impacts of the mechanisms on recovery are better understood for these processes. Currently huff-n-puff gas injection has been applied successfully in parts of the Eagle Ford, but this play has a wide range of in-situ fluid types; and likewise, other unconventional basins have a large variety of reservoir fluids. By better defining the recovery mechanisms for cyclic natural gas injection, EOR can be improved in existing unconventional plays and better designed for new areas.
Unconventional oil reservoirs such as the Eagle Ford have had tremendous success over the last decade, but challenges remain as flow rates drop quickly and recovery factors are low; thus, enhanced oil recovery methods are needed to increase recovery. Interest in cyclic gas injection has risen as a number of successful pilots have been reported; however, little information is available on recovery mechanisms for the process. This paper evaluates oil swelling caused by diffusion and advection processes for gas injection in unconventional reservoirs.
To accurately evaluate gas penetration into the matrix, the surface area of the hydraulic fractures needs to be known, and in this work, three different methods are used to estimate the area: volumetrics, well flow rates and linear fluid flow equations. Fick's law is used to determine the gas penetration depth caused by diffusion, and the linear form of Darcy's law is used to find the amount from advection. Then, with the use of swelling test information from lab tests, we are able to approximate the amount of oil recovery expected from cyclic gas injection operations.
During the gas injection phase, gas from the fractures can enter the matrix by both advection (Darcy driven flow) and diffusion. We estimate that over 200 million scf of gas can enter the matrix during a 100 day injection/soak period. Using typical reservoir and fluid parameters, it appears that 40% is due to diffusion and 60% is due to advection. Sensitivity analysis shows that these numbers vary considerable based on the parameters used. Analytical models also show that during a 100 day production timeframe, over 14,000 stock tank barrels (STB) of oil can be produced due to huff-n-puff gas injection.
Both gas injection and oil recovery amounts are compared to recent Eagle Ford gas injection pilot data, and the model results are consistent with the field pilot data.
By determining the relative importance of the different recovery mechanisms, this paper provides a better understanding of what is happening in unconventional reservoirs during cyclic gas injection. This will allow more efficient injection schemes to be designed in the future.
Over the last decade, unconventional resources like the Bakken formation have revolutionized the petroleum industry, but they have only produced by primary mechanisms, and recovery factors have remained low. The need for IOR processes is clear, but there has only been minor work in this area and no commercial field applications. Flow simulation models can be used to test different methods without interrupting field operations, but models have had a poor track record for unconventional IOR, partly because there is little field injection information to validate the models. In this work, we history matched the model to an IOR injection pilot location in Mountrail County, North Dakota that included both water and gas injection tests.
A county sized geologic model was previously constructed based upon available core, log and geologic information. The model allows for easy extraction of smaller segments for flow simulation. For the current study, a segment around the pilot injection area was isolated. The injection well and two offset producing wells were included in the model. Fluids were added into the model based on a nearby PVT report, and the hydraulic fracturing was captured with a dual permeability grid. The model was matched to the historical production and injection data. At the offset wells, breakthrough times, water cuts and gas oil ratios were also reproduced by changing the fracture and matrix properties.
By matching the injection data, the interwell connectivity is reproduced, which should improve predictions from the model. Various situations were then tested with the model including both gas and water injection scenarios. In the actual field pilot, gas was only injected for two months in the injection well, and there was only a minor response. In one scenario, therefore, we injected into all three wells in a huff-n-puff manner for ten years, and the results showed significant additional oil recovered – 30% more than the primary recovery. In other scenarios, water was injected in both a continuous and huff-n-puff manner. The continuous case had early breakthrough and poor sweep, but the huff-n-puff injection case indicated that oil rates would increase almost as much as the best gas injection cases.
This work shows that by reproducing the field injection data in unconventional reservoirs, more realistic models are created. We evaluated a large number of scenarios, and some of them did not show any increase in oil production, but the models that did show an increase helped us identify IOR techniques that have a better chance of success in the Bakken, which will improve designing the much needed next generation of field pilot tests.
Hoffman, B. Todd (Montana Tech)
The Eagle Ford formation has been an overwhelming success producing around 2 billion barrels of oil over the last seven years, yet its potential may be even greater. The projected recovery factor is only 5-10%, and using improved oil recovery (IOR) methods to increase recovery could result in billions of additional barrels of production. Significant research is required to access this oil, and while a number of companies have field tested an IOR method called huff-n-puff gas injection, most of the published results are from lab and modeling studies. This paper evaluates the results from these field tests and discusses the successes and opportunities.
The huff-n-puff process involves injecting a miscible gas into a well, and then after some amount of time, producing back from that same well. The first part of this paper evaluates the publically available data from the Texas Railroad Commission and other sources for these pilots. Analytical techniques are used to predict the amount of additional recovery and the pattern efficiency from this data. This is compared to pre-injection forecasts. All cases show increased production rates with injection, and in one pattern where the data was easiest to interpret, the incremental production has doubled since the huff-n-puff project started.
This paper also proposes methodologies for implementing second generation pilots for unconventional reservoirs. It is important to define clear objectives that characterize the value of the pilots. The significance of developing optimum drilling and completion strategies for primary and IOR success is also highlighted. Long term information collecting strategies are proposed along with methods to optimize the projects during the pilot, and contingency plans to deal with difficulties that may arise. Finally, we discuss how the location and infrastructure needs of a pilot are paramount to its success.
Using IOR to increase recovery from unconventional oil fields is important for the continued success of plays like the Eagle Ford. Pilot tests are an integral part of developing the best IOR techniques, and this paper provides a thorough analysis of implementing IOR pilots in the Eagle Ford. It also shows how and where it has been applied successfully and discusses ideas to further improve the likelihood of success in the future.
Improving oil recovery by low salinity waterflooding (LSWF) has gained a lot of attention in the last two decades. The effect of LSWF was demonstrated by coreflooding experiments in several core samples from sandstone and carbonate reservoirs around the world. Also, this effect has been shown at the field scale by some field trails.
While the exact mechanisms that cause increased recovery due to LSWF are not fully understood, most agree that changes in wettability and interfacial tension are the reasons that LSWF perform better than high salinity waterflooding (HSWF). Therefore, LSWF can be modeled by changing the property that determines the effect of wettability in fluid flow equations, which is the relative permeability. In this paper, coreflooding results from a carbonate reservoir are used to find the relation between the relative permeability curves for HSWF and LSWF. A numerical simulation model of the coreflooding experiment showed that the relative permeability for the LSWF in carbonate reservoirs can be estimated by changing only one parameter in the Corey-type relative permeability equation of the HSWF: residual oil saturation.
An application of this result was performed on a full-field simulation model to evaluate the effect of LSWF using simulation and economics. The field model was built for a carbonate reservoir in the Madison formation of Wyoming. The simulation results showed an increase in the recovery factor of more than 5% by using LSWF instead of HSWF. Furthermore, an economic analysis was performed to determine if the additional oil would justify the expense of making low salinity water. With proper assumptions of the construction and operating costs of a water desalination plant, a development plan with LSWF showed a higher net present value than a development with HSWF.
This research provides a practical approach to evaluate the effect of LSWF on certain fields using simulation. It provides a screening tool to evaluate quickly the oil gain from the LSWF before spending money on core samples testing for further research.
Over the last decade, there has been tremendous success is developing and producing unconventional oil resources such as the Bakken and Eagle Ford; however, flaring produced gases from these fields has become a problem that needs to be addressed. In many instances, pipeline infrastructure is not in place to transport gas away from the wells.
One possible solution is to reinject the produced gas back into the formation, which has numerous positive benefits. It has the potential to increase oil recovery from unconventional reservoirs because of the strong likelihood to achieve miscibility with reservoir oil. Furthermore, unlike flaring, the injected gas will be available for sales once gas gathering pipelines become available. Gas reinjection also should mitigate environmental concerns associated with greenhouse gas emissions.
We evaluated the recovery potential along with the costs of reinjecting gas to determine the economic value of this process. A dual-porosity, compositional flow simulator was used to model the gas injection process into a well surrounded by producers and to determine the amount of incremental oil and gas produced. We calculated the net economic value of the process by including the cost of gas compression and fuel gas for injection.
We have concluded that oil production rates can be significantly increased, and the economics of the process is very positive. Additionally, significant volumes of gas can be recycled, which alleviates environmental concerns of gas flaring and improves resource efficiency. As a result, we recommend pilot testing the gas injection process to assess the commercial application of the proposal.
The Bakken Formation in the Williston Basin is the most productive liquid-rich unconventional play. The middle Bakken member is the primary target for horizontal wellbore landing and hydraulic fracturing because of the better rock properties. Even with the new technology available, the primary recovery factor is believed to be around 10%. Previous studies have shown that gas injection could increase oil recovery factor for these types of reservoirs. In this study, the Elm Coulee Oil Field is selected as the area of interest, and CO2 is selected as the injection gas. The primary goal for this study is to build an EOS model for compositional simulation based on PVT tests from Bakken reservoir fluids and investigate different factors which might influence recovery for CO2 injection.
Liquid samples were collected as recombined fluid from the Elm Coulee Field; standard PVT tests were performed. By matching the reservoir fluid EOS model with PVT tests, an EOS-based compositional fluid model is created. The resulting EOS model is used to investigate production outcomes of CO2 injection for the Bakken Formation in the Elm Coulee Field.
It is believed that a logarithmically gridded reservoir model could represent hydraulically fractured reservoir better, and it is used in this study. The middle Bakken member could be divided into different lithofacies based on depositional environment; thus, different rock properties are assigned to different lithofacies. A homogenous reservoir model is used to compare the production outcomes with the heterogeneous cases.
Two producers and one injector are positioned in one section area. For base production run, producers produce without gas injection for 30 years; for gas injection cases, producers produce for 5 years without gas injection, and then production rate is simulated with CO2 injected for next 25 years. Without gas injection, oil recovery factor is around 12%, while gas injection could boost overall oil recovery factor by additional 8% to 14% of the OOIP. Lean gas and separator gas were also evaluated as injecting fluids, and it was discovered that these gases performed similar to CO2. Heterogeneous reservoir models result in slight higher oil recovery factors, but with earlier gas break-through and more gas production.
The Sanish Field, which is located in the Mountrail County, North Dakota, is the focus of the current study. The primary recovery factor of the Sanish Field remains low and has been estimated to be less than 15%. Other than horizontal drilling and multi-stage fracturing application, enhanced oil recovery is the essential process to increase the recovery factor and maximize the potential production from this field. Among several EOR options, CO2 flooding may be effective to increase the recovery factor. Earlier studies of Bakken in the Elm Coulee Field and in the Saskatchewan part of the Bakken indicated that the recovery factor could be increased by 10- 15% when using gas injection.
In this paper, a numerical reservoir simulator is used to evaluate the performance of CO2 injection for the Bakken interval in a sector of the Sanish Field. There are presently three 10,000 foot laterals in the 4 square miles sector. For modeling purposes, reasonable data values were chosen from known ranges, and well and completion information from the research area was included. A low primary recovery factor of 5.42% was obtained through flow modeling, and declining trends of the future production performance of wells in the research area were observed. Several different scenarios of gas injection are tested to analyze gas injection performance and evaluate its technical feasibility and effect. It appears that gas injection is suitable in such tight environments, as the recovery factors increased significantly for miscible CO2 injection.
Sensitivity analysis was ran by using different injection rates, by adding additional wells to the pattern, by comparing different fracture conductivities and by evaluating different injectants. Depending on the scenario, the recovery factor increases the most by 24.59% through adding four new horizontal injectors into the field sector. Moreover, gas injection was confirmed to be effective than water flooding. Maximum of 8000 psia injection pressure and maximum injection rate of 5000 Mscf/day along with more horizontal injection wells were estimated to be better options for gas injection in the study area.
This study can help to evaluate expected ultimate recovery (EUR) for future projects in the Sanish Field. It can also help to estimate the future economic viability of using gas injection and evaluate risks for the Sanish Field potential development. All these factors will directly impact the oil companies' interests and future unconventional resources development.
New technology in horizontal drilling and stimulation has caused production from ultra-tight oil formations to increase rapidly over the last decade. While initial rates are high, recovery factors for these types of reservoirs are predicted to be low, around 5-10%. Unlike conventional reservoirs, water flooding does not appear to be a viable secondary option due to its low injectivity. Recent analysis has shown that gas injection may be an effective alternative.
A 4-section area of the Elm Coulee field in eastern Montana is used to study the impact of different gas injection schemes: carbon dioxide, immiscible hydrocarbon and miscible hydrocarbon. This paper examines the difference in total recovery, production rate and efficiency using a flow simulation model.
Recovery efficiency is similar for both miscible hydrocarbon gases and carbon dioxide with recoveries increasing from 6% on primary production to around 20% with gas injection.
While the increased recovery is encouraging, both methods have some practical limitations. Carbon dioxide is currently unavailable in many basins, and while hydrocarbon gases are available in most oil fields, they are rarely used as injectants because they are marketable. We performed a cost-benefit analysis of selling the hydrocarbon gas versus using it to increase oil production. We assumed a $10 million investment in compression and facilities for the 4-section area, and used a $5/mscf cost for the gas and $80/stb revenue for the oil. The net present value for these criteria in this area is $68 million and the rate of return is 83%.
The value of this work is that it demonstrates that injecting gas (both immiscible and especially miscible) will appreciably increase oil recovery in very low permeability reservoirs. These types of reservoirs are becoming more prevalent, and a large prize is available to those who find ways to increase recovery from them. In the case of hydrocarbon gas, the economics appear to be favorable for current commodity prices.
There are a number of factors that can affect a well's production response that are not caused by the reservoir such as a fluid level in the wellbore, downhole pump inefficiencies, completion irregularities, a poor cement job, undetected formation damage and many others. The first part of this paper demonstrates the variability in the "measured?? production data when these issues arise, and shows that the impact can lead to a well that produces significantly less than well that is influenced only by reservoir behavior.
Some of the non-reservoir effects (NRE) are known and included in reservoir models, but many are not known and not included in the models. When reservoir properties are modified to match data that is a function of non-reservoir behavior, we show that model predictions do not perform as well as when the NRE are incorporated. The next section of the paper presents ways to account for the NRE with a conventional reservoir simulator. Techniques are developed to automatically change the simulation models based on historical impact and frequency of the NRE.
The techniques were tested on a real field example with 25 years of historical production. We assumed that only the first 5 years were known and matched just that data, and then predicted the next 20 years using both a traditional model and one that included non-reservoir effects. The model with NRE has the same look as the real data and the prediction error was reduced from 47% to 8%.
In petroleum operations, there are other aspects besides the reservoir that impact the measured production data, and this work shows how to account for those effects. When included, models more accurately predict actual reservoir behavior, which should lead to better decision making about the reservoir.