Cyclic steam stimulation

Prairie Crest's thermal history began not with SAGD but with a single vertical well that was steamed, shut in and produced, over and over again. Lena, the company's senior reservoir engineer, keeps the old well file on her shelf. 'Before we had SAGD at Wapiti Lake, we ran a cyclic steam pilot on the deeper, thinner sand east of the lease,' she tells Ethan. 'It taught us a lot about how this bitumen responds to heat. And next door at Cold Lake, cyclic steam has produced bitumen commercially since the 1980s.'

This lesson explains cyclic steam stimulation (CSS), also called huff and puff or steam soak: how a cycle works, why it gives high early rates, how its performance declines from cycle to cycle, and how its high injection pressures affect cap rock integrity. It compares CSS with SAGD and steamflooding and introduces the major projects that use it.

A cycle in three stages

In cyclic steam stimulation a single well is both injector and producer. Each cycle has three stages. In the injection stage, a large volume of steam, often several thousand tonnes, is injected over a few days to a few weeks. In the soak stage, the well is shut in for a few days to let the steam condense and give up its heat to the oil and rock around the well. In the production stage, the well is put on pump and produces heated oil and condensed water, at first at a high rate that declines over weeks to months as the heated zone cools and the pressure falls. When the rate has fallen too far, a new cycle begins.

CSS was discovered by accident. In 1959, during a steam injection test in the Mene Grande field in Venezuela, an injection well blew out and, after it was brought under control, it produced heavy oil at high rates. Operators soon realised that a well steamed and then produced could yield much more than a cold well, and the process spread quickly through California's heavy oil fields in the 1960s.

How CSS produces oil

Several mechanisms work together. The heat lowers the viscosity of the oil around the well, which raises its mobility and the well's productivity many times over. Steam injection raises the pressure near the well, and when the well is opened that pressure drives fluids back toward it; at high injection pressures, the formation may also be dilated or fractured, and when the pressure falls the sand compacts and squeezes fluid out. Gravity drainage of heated oil toward the well contributes, particularly in thick sands. Expansion of the rock and fluids on heating, steam flashing as the pressure drops and solution gas also help.

The size of the heated zone can be estimated with an energy balance. The heat delivered to the sand, after allowing for losses, is spread over a cylinder of rock and fluids that is warmed from reservoir temperature to near steam temperature. The radius of that cylinder sets how much oil is mobilized in the cycle.

Here Q_inj is the heat injected, E_h the fraction of it still in the target zone after the soak, h the heated thickness, M_R the volumetric heat capacity of the saturated rock and T_s and T_r the steam and reservoir temperatures. The radius is typically tens of metres, and grows only slowly with more steam, because it varies with the square root of the heat injected.

# Cyclic steam stimulation: heated radius and oil-steam ratio over several cycles
import math
SI = True   
h, phi = 15.0, 0.33                  # m, porosity of the target interval
Mr = 2600.0                          # kJ/(m3 K) volumetric heat capacity of the saturated rock
T_s, T_r = 300.0, 15.0               # C, steam and original reservoir temperature
steam_t = 9000.0                     # tonnes of steam per cycle (cold water equivalent)
h_inj = 2200.0                       # kJ/kg delivered at the sandface (hot water plus latent heat)
eff = 0.6                            # fraction of injected heat still in the target zone after the soak
Q = steam_t * 1000 * h_inj * eff     # kJ
r = math.sqrt(Q / (math.pi * h * Mr * (T_s - T_r)))
print(f"heated radius after one cycle ~ {r:.0f} m" if SI else f"heated radius after one cycle ~ {r * 3.2808:.0f} ft")
osr0 = 0.45                          # oil-steam ratio of the first cycle
for c in range(1, 7):
    osr = osr0 * 0.82 ** (c - 1)     # each cycle recovers less as the heated zone depletes
    oil = steam_t * osr
    if SI: print(f"cycle {c}: OSR {osr:.2f}  oil {oil:,.0f} m3  (SOR {1 / osr:.1f})")
    else:  print(f"cycle {c}: OSR {osr:.2f}  oil {oil * 6.2898:,.0f} bbl  (SOR {1 / osr:.1f})")

heated radius after one cycle ~ 18 m
cycle 1: OSR 0.45  oil 4,050 m3  (SOR 2.2)
cycle 2: OSR 0.37  oil 3,321 m3  (SOR 2.7)
cycle 3: OSR 0.30  oil 2,723 m3  (SOR 3.3)
cycle 4: OSR 0.25  oil 2,233 m3  (SOR 4.0)
cycle 5: OSR 0.20  oil 1,831 m3  (SOR 4.9)
cycle 6: OSR 0.17  oil 1,501 m3  (SOR 6.0)

The calculation gives a heated radius of about 18 m (61 ft) for the first cycle and an oil-steam ratio that falls from 0.45 in the first cycle to below 0.2 by the sixth. Lena explains why. In each cycle the well draws oil from a region that has already been partly drained, so there is less oil left in the heated zone to produce, and more of the heat goes into rock and water. Operators usually inject more steam in later cycles to reach unheated oil farther out, but the economics deteriorate until the well is converted to a different process or abandoned.

Pressure, fracturing and cap rock

CSS is usually operated at high injection pressure, often above the fracture pressure of the formation, so that large volumes of steam can be injected quickly and the sand is dilated. That makes the process productive, but it also makes cap rock integrity critical. If the injection pressure fractures or reactivates a pathway through the cap, steam and bitumen can escape toward shallower zones or the surface.

Alberta's experience shows why this matters. At Canadian Natural's Primrose and Wolf Lake project, bitumen emulsion reached the surface at several locations in 2013. The regulator's investigation found that high-pressure cyclic steaming had created fractures through the cap rock, likely along existing weaknesses, and it imposed restrictions on steaming in the affected areas. Thermal operators now carry out detailed cap rock studies and monitor injection pressures, and AER Directive 086 sets reservoir containment requirements for SAGD projects in the shallow part of the Athabasca area.

CSS in practice

The best known CSS project is Imperial Oil's Cold Lake operation in northeastern Alberta, which began commercial production in 1985 and is one of the largest in situ projects in the world. Its wells are drilled directionally from multi-well pads, and many have gone through more than a dozen cycles. To recover more of the bitumen between wells, Imperial has added follow-up processes, including steam flooding of the inter-well areas, and has started SAGD production from the Grand Rapids Formation and is applying solvent-assisted SAGD in its Leming redevelopment. Canadian Natural's Primrose and Wolf Lake project is the other major CSS operation in Alberta.

CSS has clear strengths. It gives a quick production response, needs only one well per drainage point and works in reservoirs that are too thin or too heterogeneous for SAGD. Its drawbacks are its declining performance cycle by cycle, an ultimate recovery that is typically only about 20 to 25 percent of the oil in place, far less than SAGD or a mature steamflood, and the high pressures it needs, with their containment risks. For that reason many CSS projects are designed with a follow-up process in mind, often a steamflood that uses the heated, pressure-depleted region left by the cycles.

Designing a cycle

The main design choices are the steam volume, the injection rate and pressure, the soak time and the length of the production stage. More steam heats a larger radius but with diminishing returns. A short soak avoids wasting heat to the cap and base rock but must allow enough time for the steam to condense; a few days is typical. The production stage is ended when the oil rate falls below an economic limit or the produced fluid temperature drops. Simulation and the history of previous cycles guide each decision, and operators often schedule the cycles of neighbouring wells together so that the pressure in one does not drive fluids into another.

Pads, deviated wells and cycle scheduling

Modern CSS projects are developed from multi-well pads, with directional or horizontal wells fanning out from each pad to cover a large area of the reservoir from a small surface footprint. Horizontal CSS wells, which contact far more of the sand than vertical ones, have become common where the pay is thin or where operators want to reach oil between existing vertical wells. Because neighbouring wells share the reservoir, the timing of their cycles matters. If one well is injecting at high pressure while its neighbour is producing, steam and hot fluids can flow between them, a problem known as interwell communication, which wastes heat and can damage equipment. Operators therefore schedule cycles in groups, injecting several wells at once and producing them together, and adjust the schedule as the reservoir pressure changes over the project's life.

Monitoring follows each cycle. Engineers record the steam volume and pressure, the soak time, the produced oil, water and gas, and the temperature of the produced fluids, and compare each cycle with the previous ones. A cycle that produces much less oil than expected, or one that returns far less water than was injected, may point to a fracture that has carried steam out of the target zone, and it is investigated before the next cycle begins.

Converting from cyclic to continuous steam

As cycles mature, the regions heated around neighbouring wells start to connect, and the reservoir between them is partly heated and depleted. At that point it often makes sense to switch to a continuous process. Some wells are converted to injectors and the others to producers, turning the CSS project into a steamflood; or, where horizontal wells are available, a SAGD-like gravity drainage process is used to recover oil from the upper part of the sand. The heated, depleted reservoir left by CSS is a good starting point for these follow-up processes, because less heat is needed to bring it up to temperature and its pressure is low enough for steam to enter easily.

The decision to convert is economic. Engineers compare the declining performance of continued cycles with the forecast of the follow-up process, including the cost of new wells or conversions, and choose the timing that maximises value. Many California fields made this transition decades ago; in Alberta, Cold Lake's operators have done the same, adding steamflood and SAGD areas to their CSS operations.