Hydrocarbon gas and nitrogen injection

CO2 is not the only gas that can be injected for EOR. Hydrocarbon gas, which every oil field produces, has been injected for pressure maintenance and miscible displacement since the 1940s, and nitrogen, separated from air, has supported some of the largest oil fields in the world. More recently, operators in tight oil plays have tested gas injection into horizontal shale wells, a very different application from a conventional flood. Ethan looks at all three, because Prairie Crest Energy Ltd. has a gas supply problem of its own: the Birch Coulee gas pool and the solution gas from the Cardium A pool have to go somewhere, and injecting them may be worth more than selling them.

Hydrocarbon miscible flooding

Hydrocarbon gas flooding comes in three forms, distinguished by the gas composition and the mechanism of Lesson 6. High-pressure lean gas, mostly methane, can develop miscibility by vaporizing intermediates from a light, intermediate-rich oil, but only at high pressure, commonly above 30 MPa (4,500 psia), so it suits deep reservoirs with volatile oils. Enriched or rich gas, methane with added ethane, propane and butane, develops miscibility at much lower pressure through the condensing-vaporizing mechanism; the operator can tune the enrichment to reach miscibility at the reservoir's pressure. Liquefied petroleum gas slugs, first-contact miscible, are driven by lean gas or water.

The trade-off is always economic. Every unit of ethane, propane or butane injected is a unit not sold, so rich-gas floods make sense when natural gas liquids have low local value, when the injected solvent is largely recovered, or when the oil is valuable enough. Alberta was a pioneer. In the 1960s and 1970s, operators ran hydrocarbon miscible floods in Devonian reefs such as Golden Spike, Rainbow Lake and the Swan Hills and Judy Creek fields, and in Cardium sands in the Pembina area. On the North Slope of Alaska, the Prudhoe Bay field has injected a miscible injectant, made from produced gas enriched with natural gas liquids, for decades. Hydrocarbon gas is also the natural choice offshore and in remote fields where there is no gas market, because injection both conserves the gas and improves recovery.

Gravity-stable gas injection

In thick, steeply dipping or reef reservoirs with good vertical permeability, the gravity override that spoils most gas floods can be turned into an advantage. Gas is injected at the crest and the gas-oil contact moves downward, pushing a bank of oil ahead of it toward producers completed low in the structure. If the downward velocity of the front stays below a critical value, gravity keeps the front flat and stable even though the mobility ratio is unfavourable, and both the sweep and the displacement efficiency can be very high. Dietz and later Hill derived the critical velocity by balancing viscous and gravity forces:

where Δρ is the density difference between oil and gas, α the dip angle and the relative permeabilities are endpoint values. Above the critical velocity, gas fingers through the oil. Gravity-stable miscible floods in Alberta's pinnacle reefs, where α is effectively 90 degrees, have achieved some of the highest recoveries reported anywhere, in some pools well above half of the original oil in place, at the price of slow, carefully controlled production.

# Gravity-stable gas injection rate and a simple huff-n-puff cycle balance for a tight horizontal well.
import math
SI = True      # True: metric units, False: oilfield units
BBL = 6.2898                       # bbl per m3
G = 9.81

# 1. Critical rate for gravity-stable downdip displacement of oil by gas in a reef
k = 10e-15                         # vertical permeability, m2 (10 mD)
d_rho = 500.0                      # oil minus gas density, kg/m3
mu_o, kro, mu_g, krg = 0.6e-3, 0.8, 0.03e-3, 0.4
u = k * d_rho * G / (mu_o / kro - mu_g / krg)   # m/s
area = 1.0e6                       # reef cross-section, m2
q = u * area * 86400
if SI:
    print(f"Critical front velocity {u * 86400 * 365:.1f} m/yr; maximum gas injection {q:,.0f} reservoir m³/d")
else:
    print(f"Critical front velocity {u * 86400 * 365 / 0.3048:.1f} ft/yr; maximum gas injection {q * BBL:,.0f} reservoir bbl/d")

# 2. Huff-n-puff cycles on a horizontal well producing 30 bbl/d (4.8 m3/d) on hyperbolic decline
qi, b, D = 30.0, 1.2, 0.0015      # bbl/d, -, 1/d
def base(t): return qi / (1 + b * D * t) ** (1 / b)
inj_days, soak_days, prod_days, gas_rate = 20, 10, 150, 1.0   # gas in MMscf/d
back = 0.75                        # fraction of injected gas produced back in the cycle
t = 0.0; tot_inc = tot_gas = 0.0
for c in range(1, 5):
    lost = sum(base(t + d) for d in range(inj_days + soak_days))
    t += inj_days + soak_days
    gain = sum(base(t + d) * 1.5 * math.exp(-d / 60.0) for d in range(prod_days))
    t += prod_days
    inc = gain - lost; gas = inj_days * gas_rate * 1e3          # Mscf
    tot_inc += inc; tot_gas += gas
    if SI:
        print(f"cycle {c}: incremental {inc / BBL:5.0f} m³, gas {gas * 28.317 / 1e3:4.0f} thousand m³, "
              f"gross {gas * 28.317 / (inc / BBL):5.0f} and net {gas * (1 - back) * 28.317 / (inc / BBL):4.0f} m³ gas per m³ oil")
    else:
        print(f"cycle {c}: incremental {inc:5.0f} bbl, gas {gas / 1e3:4.1f} MMscf, gross {gas / inc:4.1f} and net {gas * (1 - back) / inc:3.1f} Mscf per bbl")
print(f"Four cycles: {tot_inc / (BBL if SI else 1):,.0f} {'m³' if SI else 'bbl'} incremental oil")

Critical front velocity 2.3 m/yr; maximum gas injection 6,278 reservoir m³/d
cycle 1: incremental   218 m³, gas  566 thousand m³, gross  2599 and net  650 m³ gas per m³ oil
cycle 2: incremental   179 m³, gas  566 thousand m³, gross  3169 and net  792 m³ gas per m³ oil
cycle 3: incremental   152 m³, gas  566 thousand m³, gross  3719 and net  930 m³ gas per m³ oil
cycle 4: incremental   133 m³, gas  566 thousand m³, gross  4254 and net 1063 m³ gas per m³ oil
Four cycles: 682 m³ incremental oil

The first result shows why gravity-stable floods are slow. With a vertical permeability of 10 mD the front can move only about 2 metres a year without fingering, so the flood of a reef a hundred metres thick takes decades. It also shows why such floods suit reefs and other reservoirs with high vertical permeability: the critical rate is proportional to it.

Nitrogen and flue gas

Nitrogen is cheap, inert and non-corrosive, and it can be separated from air anywhere with cryogenic or membrane plants, so it needs no pipeline from a distant source. Its drawbacks are a very high MMP, so it is miscible only with light, volatile oils in deep, hot reservoirs, and a high formation volume factor, so large volumes are needed to fill the pore space. Nitrogen has been used most for pressure maintenance and gas cap expansion, notably at the giant Cantarell field offshore Mexico, where a large nitrogen injection project began in 2000, and for miscible flooding in a few deep fields such as Jay in Florida and the East Binger unit in Oklahoma. Flue gas, a mixture of nitrogen and CO2 from combustion, was tried in several fields but is now rare because of its corrosivity and the need to clean it.

Gas injection in tight oil

Horizontal wells with multistage hydraulic fractures recover only a small fraction of the oil in place in tight formations, often estimated at 5 to 10 percent, because the matrix permeability is measured in microdarcies or nanodarcies. The remaining oil is an enormous target, and since about 2012 operators have tested whether gas injection can recover some of it. Conventional pattern flooding does not work well: injected fluid races through the hydraulic fractures from injector to producer and contacts little matrix. The approach that has worked best is cyclic gas injection, often called huff-n-puff. A producing well is shut in and gas is injected down it for a few weeks, the well soaks for days to weeks while gas penetrates the matrix around the fractures, and then the well is put back on production. The cycle is repeated.

The mechanisms differ from those of a conventional flood. Gas raises the pressure around the fractures, re-energizing the depleted matrix. It dissolves into the oil by diffusion, swelling the oil and lowering its viscosity, and at high enough pressure it extracts intermediates and approaches miscibility. When production resumes, swollen oil flows back into the fractures. Because most of the injected gas comes back during the puff period, it can be separated and reinjected, so the net gas consumption is much smaller than the gross volume injected.

Huff-n-puff was developed at commercial scale in the Eagle Ford play of south Texas, where several operators have run hundreds of wells through cycles of produced gas injection since the mid-2010s. In the Bakken of North Dakota, the Energy and Environmental Research Center at the University of North Dakota has worked with operators on a series of CO2 and rich-gas injection tests, and in 2021 Liberty Resources reported a single-well test in Mountrail County that alternated water and gas and raised production by about 25 percent, using about 4 to 5 thousand cubic feet of gas per incremental barrel. Canadian operators in tight oil plays in Alberta and Saskatchewan have run similar trials. For Prairie Crest Energy Ltd., the question is whether a huff-n-puff pilot on the Cardium horizontal 03-22 would add oil at an acceptable gas cost.

The second half of the code models four cycles on a horizontal well in late decline. Each cycle loses about a month of base production during injection and soak, then produces at an elevated rate that decays over several months. The incremental oil per cycle falls as the near-fracture matrix is depleted, and the gas used per incremental barrel rises. The net gas use, after accounting for the gas produced back, is in the range reported from field tests, but the gross volume of gas that must be compressed is three to four times larger, which is the main facility cost.

Practical issues in tight oil gas injection

Field experience has exposed several problems. Injected gas can travel through the fracture network to neighbouring wells on the same pad, so operators often inject into groups of adjacent wells at the same time to keep pressure balanced. Containment within the target formation must be demonstrated to the regulator, especially where fractures may have grown into adjacent zones. Injection pressures are high, so compressors are large and costly, and wells must be checked for casing integrity before injection. Response varies widely between wells and cycles, and much of the industry's experience is unpublished or reported only in summary form. Ethan's notes describe gas injection in tight oil as promising, proven at scale in the Eagle Ford, and still being tested in the Bakken and Canadian plays.

Gas supply is often the deciding factor for hydrocarbon gas projects. Solution gas production falls as the oil rate declines, so a project that relies only on its own gas eventually needs make-up gas bought from a pipeline, which competes with sales value.