Model Simulates Gas Kicks in Nonaqueous Drilling Fluids
Nonaqueous drilling fluids, such as synthetic-based and oil-based mud (SBM and OBM, respectively), are used frequently to drill one or more sections of a well to reduce drilling problems such as shale sloughing, wellbore stability, and stuck pipe.
Nonaqueous drilling fluids, such as synthetic-based and oil-based mud (SBM and OBM, respectively), are used frequently to drill one or more sections of a well to reduce drilling problems such as shale sloughing, wellbore stability, and stuck pipe. However, solubility of formation gas in such fluids makes early gas detection, and the well-control process, very challenging. This is of particular concern in deep offshore wells. This paper presents a novel and comprehensive hydraulic model to simulate a gas kick in nonaqueous drilling fluids.
Early kick detection becomes cumbersome when a nonaqueous drilling fluid is used. This is because of dissolution of gas in the nonpolar base fluid. Although gas entrance into the well still will result in increased flow-out on surface, the flow signature will be less pronounced in comparison with a similar gas kick in water-based mud (WBM). Moreover, the emergence of gas from the solution may often only begin when the gas/drilling-fluid mixture is relatively close to the surface, which leaves the crew only a short period to react before gas reaches the surface. Constant-bottomhole-pressure/managed-pressure drilling (CBHP-MPD) offers the benefit of delaying the bubblepoint until the kick has passed the choke, thereby not allowing the kick to come out of solution in the actual well.
Gas-Kick-Simulation Multiphase-Flow Modeling. The drift-flux model (DFM) used in this work is a mechanistic model consisting of separate mass-balance equations and a combined momentum-balance equation, together with several closure algebraic equations and a slip law. To obtain the velocity of each phase, the slip law is used in conjunction with the combined momentum-balance equation.
Solubility in oil-based drilling fluids is directly proportional to pressure, gas specific gravity, and base-oil volume in the drilling fluid and is inversely proportional to temperature, drilling-fluid solids content, and brine and emulsifier volumes in the drilling fluid. For a given drilling-fluid volume, the relative volumes of base oil, brine, and emulsifier in the drilling fluid can be determined from a standard retort analysis. However, the most-reliable method to determine this crucial simulation parameter is to conduct pressure/volume/temperature (PVT) testing for each mixture under downhole conditions.
DFM Formulation. Mathematical modeling of a proposed 1D DFM to describe the multiphase-flow dynamics for nonaqueous drilling fluids is discussed in detail in the complete paper. The current formulation takes into account the mass transfer between the liquid (drilling fluid) and formation/injected gas. The model consists of two sets of partial differential equations for two different cases: when gas is completely dissolved in the liquid (no free gas) and when dissolved and free gas coexist in the system.
Results and Discussion
Input Data. Two simulation cases are presented here: gas influx during CBHP-MPD and conventional drilling. A deviated offshore well was selected to mimic a real-life scenario. The wellbore path [true vertical depth (TVD) vs. horizontal departure] is shown in Fig. 1.
To capture the effect of gas solubility on crucial well-control parameters better, simulation results of a 12.6-lbm/gal nonaqueous drilling fluid (SBM) with an oil/water ratio of 70/30 were compared with a 12.6-lbm/gal WBM. For simplicity, it was assumed that both drilling fluids had identical rheological properties. In addition, it was assumed that the influx consisted only of a natural-gas mixture with a specific gravity of 0.65. It should be noted that the solubility of natural gas in SBM is significantly higher than its solubility in WBM.
Simulation Results. Case 1: Gas Influx During CBHP-MPD. An influx during simulation was introduced by selecting a reservoir pressure higher than the circulating bottomhole pressure. In this simulation, it was assumed that the gas influx can be detected either after a 5% increase in flow-out or a 4-bbl increase in pit gain, whichever comes first. After the influx was detected, casing backpressure was increased by using an automated choke to arrest the influx. A timeline was followed during the simulation.
- 0–3 minutes, normal operation—Mud flow-in rate was set to 600 gal/min and initial surface backpressure was set to 140 psi by use of an automatic choke.
- 3–10 minutes (WBM)/3–26 minutes (SBM)—Reservoir pressure was set to 100 psi higher than the bottomhole pressure (BHP), which triggered a gas influx at the bottom of the well.
- 10–66 minutes (WBM)/26–93 minutes (SBM)—The influx was detected. Then, the proper action (increasing casing backpressure in this case) was taken to arrest the influx without shutting in the well.
Fig. 2 shows the mud flow-in/-out rate vs. time during the gas-kick event. The flow-in rate was kept at 600 gal/min. The flow-out rate exceeded the flow-in rate after the gas kick was encountered (Minute 3). However, because of the high gas solubility in SBM, the increase in flow-out was significantly less than that of WBM. In the case of WBM, the flow-out rapidly increased from 600 to 640 gal/min, followed by a sudden increase in pit gain, allowing the gas influx to be detected quickly. In the case of SBM, the gas influx was detected only after the pit gain reached the 4-bbl threshold. As a result, the amount of gas that entered the wellbore before the influx had been detected was higher for SBM significantly.
In the case of WBM, a sudden increase in pit gain was observed when the influx entered the wellbore, peaking at 15 bbl when the influx reached the surface. By contrast, the pit gain for SBM increased at a much slower rate before the influx was detected. Even though the amount of gas that entered the wellbore was significantly higher in the case of SBM, the maximum pit gain (10 bbl) was still less than that of WBM (15 bbl). This clearly highlights the importance of gas solubility on the observed pit gain during a well-control incident.
BHP decreased soon after the influx both for SBM and WBM because of a lower gravitational gradient of the mixture. Accordingly, this resulted in a higher pressure drawdown and a higher gas rate to the wellbore. The standpipe pressure followed the same trend as the BHP.
For both cases, the gas kick entered the wellbore after 3 minutes. Because the solubility of natural gas is negligible in WBM, free gas was observed in the wellbore under downhole conditions. By contrast, the gas influx was entirely dissolved in SBM for an extended period of time. The gas influx came out of the solution after 66 minutes, when it was very close to the surface at approximately 1,500 ft TVD. This resulted in a sudden increase in pit gain and an increase in flow-out.
Case 2: Gas Influx During Conventional Drilling. In this case, it was assumed that the pit alarm was set at a 15-bbl-influx threshold. When the pit gain exceeded this threshold, drilling was ceased and a flow check was conducted. After the influx was confirmed, the blowout preventer (BOP) was closed and shut-in drillpipe and casing pressure were monitored to determine the influx intensity. The simulation timeline was as follows:
- 0–3 minutes, drilling ahead—Mud flow-in rate was set to 600 gal/min.
- 3–13 minutes (WBM)/3–39 minutes (SBM)—Reservoir pressure was set to 200 psi higher than the BHP, which triggered a gas influx at the bottom of the well.
- 13–15 minutes (WBM)/39–41 minutes (SBM)—Pit gain exceeded 15 bbl, which triggered the influx alarm. Mud pumps were stopped at this point to conduct a flow check.
- 15–21 minutes (WBM)/41–47 minutes (SBM)—The gas kick was confirmed after the flow check. The BOP was used to shut in the well.
- 21–190 minutes (WBM)/47–235 minutes (SBM)—Kick removal was initiated at 200 gal/min (kill rate). The choke position was controlled automatically to keep the BHP 100 psi above the reservoir pressure during influx removal.
The initial mud-flow-in rate was set to 600 gal/min. Once the reservoir pressure exceeded the wellbore pressure, a gas influx entered the wellbore, leading to increasing flow-out. As with Case 1, the increase in flow-out rate for SBM was significantly less than that for WBM.
With negligible gas solubility in WBM, the pit gain reached 15 bbl in 10 minutes, which triggered the alarm. By comparison, it took 36 minutes before the alarm was triggered with SBM. During kick removal, the pit gain continuously increased for WBM while it remained relatively constant for SBM for an extended period. However, when gas came out of the SBM solution, a sudden jump was observed in the pit gain.
Initially, BHP and standpipe pressure decreased because of the lower gravitational gradient of the mixture. When the well was shut in, BHP and standpipe pressure increased and eventually stabilized. The kick-removal process was initiated thereafter, which resulted in further-increased BHP and standpipe pressure.
The casing peak pressures were approximately identical (1,100 psi) for WBM and SBM. This indicates that, because early kick detection is more difficult for SBM, larger gas volumes may enter the wellbore, which, in this particular case, resulted in a similar casing peak pressure.
Because gas solubility was negligible in WBM, free gas was observed during the entire circulation. The gas-void fraction increased significantly while waiting for shut-in drillpipe and casing pressure to stabilize. For SBM, the gas kick was dissolved entirely in mud under bottomhole conditions. However, it briefly came out of solution while waiting for shut-in drillpipe and casing pressures to stabilize.
In this paper, a novel DFM for nonaqueous drilling fluid is introduced. The model consists of partial differential equations of mass/momentum conservations with closure algebraic equations for density, slip law, and non-Newtonian frictional forces. The developed model is validated by using previously published data.
The new modeling approach can handle relevant complexities such as dynamic and conventional well control, multiple influxes from several formations, non-Newtonian drilling fluids, underbalanced drilling, multiple fluids in the system (such as cementing operations or the wait-and-weight method), an arbitrary wellbore path (including directional and horizontal wells), gas solubility in the drilling fluid, user-defined PVT input, and high-gas-void fractions, making it ideal for single- and multiphase hydraulic planning and monitoring.
The new modeling approach and a developed software tool were used to simulate two well-control scenarios: a dynamic well control for CBHP-MPD application and a conventional well control in which the driller’s method was applied for a realistic offshore-well scenario. The effect of gas solubility on the simulation results was investigated by comparing SBM results with WBM results, assuming gas solubility to be negligible for the WBM case.
For the presented cases, results show that gas influx was entirely dissolved in SBM under downhole conditions, which significantly delayed the kick detection. Two crucial indicators for early kick detection, namely flow-out and pit gain, were smaller for the SBM than for the WBM. For both well-control scenarios, gas broke out of the SBM solution near the surface, which resulted in a sudden jump in pit gain, flow-out, and casing pressure. Therefore, if a gas kick remains undetected in nonaqueous drilling fluids, the crew will only have a very short period to react before gas reaches the surface. These results, well-known from field application, stress the importance of additional caution and application of best well-control practices while using a nonaqueous drilling fluid, particularly in an overpressured gas-bearing zone.
In addition, the difficulty in early kick detection in SBM (in comparison with WBM) resulted in a significantly larger amount of gas in the wellbore before proper action could be taken to arrest the kick. Therefore, depending on the gas solubility in the mud, crucial well-control parameters such as peak casing pressure, flow-out, and pit gain for SBM could be lower, similar, or even higher than those for WBM. However, with the same amount of gas entering the wellbore (e.g., in a controlled experiment), these parameters are anticipated to be smaller for nonaqueous drilling fluids because of the higher solubility of gas.
The simulation results show that gas solubility in drilling fluid is a crucial factor and must be considered in both planning and in operation. The presented modeling approach offers valuable operational insights by enabling realistic simulation scenarios that can lead to superior preparation for well-control emergencies, offering the potential to improve rig safety and reduce nonproductive time and associated cost during drilling and completion.
This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 189606, “Gas Kicks in Nonaqueous Drilling Fluids: A Well-Control Challenge,” by Z. Ma, A. Karimi Vajargah, D. Chen, and E. Van Oort, SPE, The University of Texas at Austin, and R. May, J.D. MacPherson, SPE, G. Becker, SPE, and D. Curry, SPE, Baker Hughes, a GE Company, prepared for the 2018 IADC/SPE Drilling Conference and Exhibition, Fort Worth, Texas, USA, 6–8 March. The paper has not been peer reviewed.