18.3 Constant BHP Circulating Kill Procedures for Horizontal Wells

Constant-BHP Circulating Kill Procedures for Horizontal Wells

A horizontal well uses the same pressure-balance principles as a vertical well. The long measured path changes volumes and friction; the surveyed vertical profile determines hydrostatic head. Plan the kill with both, rather than treating heel arrival as a switch between two guaranteed pressure patterns.

Follow the Influx Through the Well

Surface / controlled return system
  |
  | Vertical and build sections:
  | gas rises; changing absolute pressure can cause substantial expansion
   \\
    +---------------------------------------------- Toe / bit
    Heel              Horizontal lateral
                      small elevation change, possible segregation

All positions: follow the selected pressure-control reference,
track volumes and monitor formation/equipment/return-system limits.

Lateral Transit

In the ideal flat-lateral hydrostatic model, gas moving at constant TVD does not replace additional vertical mud head. That explains why the initial pressure differential may be small. It does not prove that gas pressure stays constant while circulating: friction, fluid distribution, temperature, dissolution and expansion can change the readings.

Select sufficient flow to sweep gas from the lateral within the pressure window. Compare actual pressures and returns with the prepared model. Do not hold the choke fixed merely because gas is thought to remain below the heel, or open it simply because casing pressure is not rising.

Build and Vertical Transit

As gas rises and its absolute pressure falls, expansion can increase the return volume and the surface backpressure required to maintain BHP. For a fixed amount of free gas: $\frac{P_1 V_1}{Z_1 T_1}=\frac{P_2 V_2}{Z_2 T_2}$ Use absolute pressure and temperature. This relation does not imply exponential expansion with time or a mandatory pressure spike at the heel.

The choke operator follows the selected drillpipe-pressure target or other approved phase reference, allowing for system response lag. Make controlled adjustments based on actual pressure response while monitoring the annular pressure window. A casing-pressure rise with pit gain is consistent with gas expansion, but also requires checking the current fluid positions and other possible causes.

Select the Kill Method for the Actual Well

Driller’s Method can begin circulation with the available original mud. Wait-and-Weight requires suitable kill mud and a prepared pressure schedule. Choose the approved method after assessing the pressure window, mud availability, gas movement, geometry and equipment; horizontal geometry alone does not establish a universal preference.

Compare Fluid-Front Timing

Wait-and-Weight can lower shoe pressure if kill mud enters the annulus before the gas reaches the controlling zone. Compare drillstring displacement with the annular route and the actual influx position/expansion. A lateral annular volume larger than drillstring volume does not remove that potential advantage; it may allow kill mud to reach the bit while gas remains below the shoe. Neither method guarantees a particular peak pressure without a fluid-profile calculation.

Construct the Geometry-Dependent Pressure Schedule

For Wait-and-Weight, use surveyed TVD at each fluid-front position, measured-length capacities, pump output and the relevant circulating losses. A straight ICP-to-FCP line is the standard uniform-vertical approximation, not an all-geometry schedule.

At a specified constant flow rate, the pressure balance at the standpipe is: $P_{DP,target}(s)=P_{BHP,target}-H_{DP}(s)+\Delta P_{downstring+bit}(s)$ Here $H_{DP}$ is the current drillstring hydrostatic head referred to the standpipe datum. The loss term includes the applicable path from that gauge to the bit. It changes as the two muds displace one another.

A standard segmented calculation using linear hydraulic-loss interpolation is: $P_{DP}(x)=ICP+(FCP-SCR)\frac{V_x}{V_{ds}}-0.052(KMW-OMW)TVD_x$ $V_x$ is drillstring capacity to the kill-mud front, $V_{ds}$ total drillstring capacity, and $TVD_x$ the front’s vertical depth. This is a calculation approximation; actual fluid rheology and hydraulic conditions must support the selected schedule.

SURFACE TO HEEL: hydrostatic head increases with surveyed TVD;
                circulating losses also change.
HEEL TO BIT:    a flat lateral adds no vertical head;
                circulating losses may still change.
BIT ARRIVAL:   the selected schedule reaches FCP.

Thus the hydrostatic part of the step-down ceases in a flat lateral, but the total pressure target need not be flat. In the interpolation above, heavier mud continuing down the lateral increases the loss term. Do not discard that term and impose the hydrostatic-only heel pressure as the final target.

For the standard gas-cleared second Driller’s circulation, use the appropriate casing-pressure reference until kill mud reaches the bit, then hold drillpipe FCP. Monitor the expected drillpipe response and pressure limits throughout; do not substitute the Wait-and-Weight control phase indiscriminately.

Extended-Reach Friction and Choke Management

Long, narrow annuli can have significant pressure losses even at a selected kill rate. Spm alone does not establish those losses: pump output, fluid properties, geometry and flow rate are required.

For upward returns to a surface choke gauge: $BHP=H_{ann}+P_{casing,flowing}+\Delta P_{return}$ $\Delta P_{return}$ includes the relevant annular and choke-line friction. Add the actual subsea line head and datum conversion where applicable. The flowing casing-gauge reading is not SICP.

On startup, losses develop; on slowdown, they disappear. Coordinate rate and choke backpressure to preserve the required BHP. Holding initial casing pressure is the standard surface-stack startup approximation only when annular losses are negligible. With significant annular friction or subsea CLF, use the appropriate compensated reference, reversing the adjustment on shutdown.

Anticipate Gas Return with Several Measurements

1. Calculate displacement estimates: $V_{ann,lateral}/Q_{pump}$ gives strokes to displace one lateral annular volume. Define the counter origin and include actual capacity segments.
2. Track the fluid fronts: The gas leading edge may arrive earlier because the influx already occupies part of the annulus, expands or slips. Dispersion can also leave gas behind the nominal front. The volume estimate is not an exact arrival or clearance count.
3. Brief before transitions: Use the planned positions and actual trends to keep the choke operator ready. Follow the method’s pressure target with controlled adjustments and response lag; do not apply a universal 90% trigger or prescribed choke opening percentage.
4. Monitor the return system: Check the MGS gas/liquid/venting envelope and liquid seal where used. If limits are approached, coordinate rate reduction and backpressure within the well pressure window. If safe handling cannot be maintained, secure the well under the approved contingency; verify the system and pressure reference before resuming.

Worked ERD Planning Example

ILLUSTRATIVE WELL DATA
 • Bit: 24,600 ft MD, 8,800 ft TVD
 • Heel / shoe: 10,600 ft MD, 8,750 ft TVD
 • Lateral measured length: 14,000 ft
 • Original mud: 13.0 ppg OBM; shoe integrity equivalent: 15.2 ppg
 • SIDPP 420 psi; SICP 425 psi; pit gain 32 bbl
 • Original SCR: 850 psi at 25 spm; pump output 0.10 bbl/stroke
 • Given average lateral annular capacity: 0.0459 bbl/ft

Under the standard communicating formation-pressure model: $P_{form}=0.052\times 13.0\times 8800+420=6368.8\text{ psi}$ The standard ICP estimate is $850+420=1270$ psi. Assess return-path losses before choosing the actual startup reference; the SCR measurement alone does not separate annular losses from drillstring losses.

The lateral displacement estimate is: $V_{lateral}=14000\times 0.0459=642.6\text{ bbl}$ $N=642.6/0.10=6426\text{ strokes}$ $t=6426/25=257.04\text{ min}=4.28\text{ h}$ This is one lateral annular volume, not a guaranteed first-gas arrival. The 50 ft heel-to-toe TVD difference also means the given lateral is slightly toe-down, rather than exactly flat.

The shoe integrity pressure is $0.052\times 15.2\times 8750=6916$ psi. In a static illustration with constant BHP 6368.8 psi and original mud between shoe and bit, shoe pressure would be: $P_{shoe}=6368.8-(0.052\times 13.0\times 50)=6335\text{ psi}$ The illustrative margin is $6916-6335=581$ psi. As fluids move and circulation losses change, calculate the actual shoe pressure from that current profile. Neither a future peak shoe pressure nor separator capacity can be proved from these data alone.

During circulation, follow the chosen control phase, compare predicted and measured returns and monitor limits. After expected displacement, make a controlled shutdown and verify pressures, fluid density, volumes and no flow. A nominal stroke count alone does not establish that dispersed gas has all been removed.