Advanced Well Control Reduces Risk of a Blowout

Current well-control practices have significantly reduced the probability of loss of well control (Etkin 2009). Even so, some stakeholders are still concerned about the risk of a blowout. ExxonMobil is evaluating two concepts – polymer plugs and seawater injection – that can rapidly respond to a well-control incident to kill an out-of-control well. The concepts are applicable to any type well, onshore and offshore in both shallow and deepwater.

The polymer plug concept uses rapid cross-linking polymers formed by combining a suitable monomer (such as dicyclopentadiene (DCPD)) and a ruthenium-based catalyst (Nguyen et al., 1992; Grubbs, 2006)). The reaction of DCPD and related norbornenes with the ruthenium catalyst is shown in Fig. 1. These reactions are exothermic and highly adjustable depending on functionalization of both the monomers and the catalysts (Bielawski & Grubbs, 2007) allowing cure times that range from seconds to hours (Griffiths and Diver, 2015). The monomer formulations and the catalyst suspensions have rheological properties that allow them to easily flow under the conditions needed to rapidly fill a damaged blowout preventor (BOP) and form a polymer seal that stops additional hydrocarbon flow (Nedwed et al., 2019). Further, the resulting polymer is high strength with considerable toughness and good elongation properties (Woodson and Grubbs, 2000).

The seawater injection concept (known as Seawater Injection Method or SWIM) is another method to rapidly stop flow of reservoir fluids after a BOP seal failure is detected. The concept is to pump seawater into a leaking BOP below the failed seal. If seawater is pumped at sufficient rate, it will create enough backpressure as it passes through the leak to stop further influx of reservoir fluids into the wellbore. Further, replacing reservoir fluids with solids-free seawater will reduce the potential for additional erosion of the BOP seal compared to solids-laden mud or solids-laden reservoir fluids. It is applicable to wells drilled in all environments—a surface blowout through a failed BOP; an underground blowout where hydrocarbons are passing outside a wellbore to another formation; and an underwater blowout where fluids are passing outside a wellbore to the seabed. Further, the technique could be used for deepwater and shallow water drilling and both surface and subsurface BOPs. For shallow water drilling where there is the potential for gas release, seawater injection could eliminate gas loss from the well to keep it from reaching the surface.

This paper describes results from pilot-scale testing of the polymer-plug concept and numerical modeling of the seawater injection concept.

Fig. 2 shows one potential configuration of the polymer-plug concept. The figure depicts a scenario where loss of well control has occurred and hydrocarbons are escaping to the environment. The pipe that connects the BOP to the drilling rig (the riser) has disconnected. The left image shows that the BOP was activated to close the shear ram, however, the ram failed to completely seal the well and flow continues. The middle image depicts activation of pressure vessels located near the BOP that rapidly inject the liquid monomers and catalyst into the BOP below the leak point to initiate formation of a polymer seal. The right image shows the solidified polymer sealing the BOP to stop flow and also shows kill weight drilling fluid being pumped into the BOP below the polymer. The drilling fluid is pumped until it equalizes the bottomhole pressure allowing the well to return to static control (Nedwed and Meeks, 2017).

Fig. 3 shows a simplified drawing of the seawater-injection concept for the scenario of a failed shear ram in a BOP. The riser has disconnected and hydrocarbons are escaping to the marine environment. The left image shows that the BOP was activated to close the shear ram, however, the ram failed to completely seal the well and flow continues. After the leak is detected, seawater is injected into the BOP through the choke line (middle image) at sufficient rate to develop back pressure to stop the flow of hydrocarbons. At this point, a further increase in the flow of seawater will displace the drilling fluid and any reservoir fluids into the formationFig. . Once the well is dynamically controlled with seawater, kill-weight drilling fluid is pumped below the seawater injection point until reservoir fluids are displaced and the hydrostatic pressure exceeds the reservoir pressure to regain well control (right image).

Testing to determine polymer integrity, polymerization kinetics, and the rheology of both the polymer and catalyst are reported elsewhere (Nedwed et al., 2019). Additional testing of the kinetics was performed and is reported here. For these tests, approximately 150 ml mixtures of DCPD / norbornene monomers were placed in 250 ml beakers and heated to temperatures from room temperature up to 40°C. We did not heat the catalyst. After heating, a metal spatula was used to rapidly stir the monomers while simultaneously mixing in the catalyst at a ratio of 1:50. The time to complete polymerization based on observation was recorded.

In addition, testing of the polymer plug formation process was completed using a scaled simulated BOP system. The test system is illustrated in Fig. 4. The BOP was simulated using 1.5” ID drill pipe that was 3' long. The simulated BOP included either a 1/8” or a 1/4” diameter circular orifice located 2/3 of the way from the base. The orifice simulated a leaking BOP shear ram. Testing was performed using synthetic based drilling fluid as the fluid flowing through the BOP during the simulated blowout. The drilling fluid was stored in the “flowing fluid tank” and pumped through the BOP using the hydrocarbon pump. The “flowing-fluid tank” was equipped to heat the drilling fluid and temperatures between 40° and 60°C were tested. To perform a test, the hydrocarbon pump sent drilling fluid at a rate of 3 – 10 gpm and flowing pressures between 22 and 1980 psi through the BOP. Higher pressures were used for the 1/8” orifice tests and lower pressures for the ¼” orifice tests. The ¼” orifice tests at high pressure resulted in BOP residence times <2 seconds. After establishing drilling fluid flow, piston pumps holding the resin and catalyst were activated to simultaneously inject both in approximately 30 seconds. Before injection, the resin was heated to between 27 and 56°C. The resin and catalyst passed through a static mixer just prior to passing into the BOP. For most tests, the piston pumps were not able to deliver their entire load because polymer formed within the BOP to stop further flow. After polymer formation in the BOP (as determined by the inability of further resin / catalyst injection and the inability of the hydrocarbon pump to flow drilling mud), the hydraulic pumping unit (HPU) was started to pressure the BOP to 5000 psi – more pressure than the hydrocarbon pump could deliver. The HPU was maintained at up to 5000 psi for up to several hours to test the integrity of the polymer seal. The primary metric for these tests was to monitor the ability of the system to maintain the hydraulic pressure over time.

An analytical model was developed to evaluate the seawater injection method (SWIM) assuming the system was under static conditions, i.e., when the system reaches equilibrium after seawater injection is initiated and hydrocarbons are no longer flowing from the reservoir. The goal was to determine pumping requirements for the system. Static conditions require the highest pumping flow rates because the pressure inside the BOP and well are at a maximum. Wellbore pressures are lower when fluids are flowing up the well because of frictional losses.

Assumptions. The following assumptions were used for the analytical model:

1. SWIM flow path: The seawater was assumed to flow from a 4 inch diameter line connected from pumps on the surface into the leaking BOP, pass through a leak point in the BOP, and then into the seawater above the BOP. The flow in the 4 inch line was assumed to be turbulent with constant friction factor of 0.004.

2. Wellbore fluid column: The fluid below the BOP was assumed static which eliminates any frictional pressure losses in the wellbore when seawater injection just balances reservoir pressure. The fluid in the wellbore below the BOP was comprised of either drilling mud (12 ppg or 9 ppg depending on the well), oil (790 kg/m3), or gas (360 kg/m3) at reservoir conditions. The scenario with a gas filled wellbore is considered the most challenging and conservative because this will cause maximum pressure in the BOP.

3. Balanced Scenario: In the balanced scenario, the BOP was open to the seawater at the seabed and fluids below the BOP were static so the pressure below the BOP leak point (point A in Fig. 5) was impacted by both 1) seawater injection rate and 2) reservoir pressure after accounting for the hydrostatic head of the wellbore. The limiting pressure at the BOP leak point will be independent of the well geometry below the BOP and will only depend on the reservoir pressure, total vertical depth of the well, and the wellbore and seawater fluid density.

4. Leak Path: The leak itself was represented as a circular orifice. We varied the orifice diameter to cover a range of leak sizes and geometry for a BOP seal failure and determined the maximum leak size various pumps could handle. For these simulations, the orifice flow coefficient was assumed to be 0.84.

Equations 1 – 3 are solved simultaneously to determine pressure just below the leak point in the BOP (P_A), seawater injection pressure required at the discharge of a pump (P_inj), and the sea water mass flow rate (m_i) when all other parameters are specified.

We performed simulations where the wellbore was filled with drilling fluid (causing the greatest hydraulic pressure drop), crude oil, and free gas (causing the least hydraulic pressure drop) with assumption of single phase in the wellbore below the BOP. We modeled a leaking blind shear ram as the failure scenario for the BOP. The leak path was modeled as an equivalent orifice. We chose various orifice sizes to simulate various size leak openings.

We modeled 1) a normally pressured reservoir (0.44psi/ft; assumed bottomhole pressure of 6500 psi); and 2) an abnomally high-pressure reservoir (gradient = 0.85psi/ft; assumed bottomhole pressure of 11,750 psi) and evaluated the effect of wellbore fluid type (oil, gas or mud) on the seawater pumping requirements and compared them against the availabile pump capacities on a typical rig. Details of the well scenarios that were modeled are shown in Table 1.

We completed two types of tests evaluating the polymer plug concept. The first was to better understand the kinetics of the polymerization reactions. Depending on the well scenario, residence time of fluids in a leaking BOP could be from a few seconds to several tens of seconds. For example, considering that standard deepwater BOPs have an 18.75” internal diameter and are roughly 40' long, a 50K bbl/day flow of oil requires ~23 seconds to pass through a BOP. This means that polymerization reactions have to be at least this fast. As the reactions are exothermic, increasing the temperature of the reactants increases the reaction rates. We tested temperatures from 20°C to 40°C. Results are shown in Fig. 6. At 40°C, the reactions of the pure resin required 2.6 seconds to complete.

After completing the tests to understand reaction times, we completed experiments to form polymer seals in the simulated BOP illustrated in Fig. 4 under temperatures and pressures representative of real-world conditions. Table 2 shows the results of these tests. Testing of the 1/8” orifice opening was performed with the drilling mud flowing at a pressure of approximately 1500 psi. At this pressure, the residence time of drilling fluid in the simulated BOP was approximately 2 seconds. Increasing the orifice to ¼” significantly reduced the residence time to <0.5 seconds if the 1500 psi pressure was maintained.

As seen in Table 2, thirteen tests were performed with a 1/8” orifice and varying the circulating mud temperature from 40°C to 57°C and varying resin temperature from 27°C to 56°C. Two of these tests had equipment malfunctions. Three of the tests had either failure of a plug to form or the plug that formed had a slow leak. The remaining eight tests resulted in formation of a seal that held hydraulic pressures up to 4500 psi for up to ~18 hours. Longer pressure hold periods were not completed in order to continue testing.

Of the three tests performed using a ¼” orifice only one resulted in an integral seal. For the other two, the first one was performed at 1500 psi circulating pressure and this resulted in a flowing residence time of <0.5 seconds. This time was not long enough for the polymer to set before the resin and catalyst were pushed through the orifice. The second test had an equipment malfunction. The third ¼” orifice test resulted in formation of a seal. For this test, we reduced the mud circulating pressure to ~200 psi and increased the resin temperature to 54°C. The seal for this test held up to 5000 psi for ~1.5 hours before the test was terminated to allow additional testing.

Note that these were the initial tests of this BOP pilot system. Tests that were failures or had equipment malfunction were the result of unfamiliarity with the system. Additional testing is planned including tests with a BOP system that allows 10,000 psi of pressure. Results from these tests will be reported after they are completed.

The polymer system we used for these tests was initially developed for injection molding applications. Our testing in a simulated BOP was similar to the process of injection modling. In fact, after completion of tests, the BOP system was broken apart and the polymer plug seal was removed. Fig. 7 shows a picture of one of the seals. The plug appeared like an injection mold of the scale BOP with clear definition of the threads used for the end caps. The right side of the image is the base of the BOP and the left is the top. The dark color of the seal is because of the contamination from the drilling fluid in the system. The contamination was lower at the base of the BOP where a clearer polymer was formed. Prior tests of the integrity of the polymer found that it could withstand 15,000 psi differential pressure with over 20% (by volume) contamination of the resin with drilling fluid. Inspection of these plugs found that the more highly contaminated portions were somewhat softer and the clearer portion at the base was very solid to the touch.

The goal of the SWIM modeling was to determine the flow and pressure requirements necessary to dynamically stop reservoir flow in order to determine if commonly available pumps were capable. We didn't limit SWIM application based on concerns about well-control protocols. Here we assumed that well control was lost and hydrocarbons were escaping to the environment.

Fig. 8 and 9 show modeling results (pump discharge pressure versus seawater injection rate) for the low and high pressured reservoirs, respectively. The red (gas), green (oil), and brown (drilling fluid) solid lines show scenarios where we assumed the well bore was only filled with each of these fluids when seawater injection was initiated. The gray dotted line shows pumping requirements assuming a constant leak diameter. The pumping requirements (discharge pressure and rate) change for constant leak diameters as the wellbore fluids change so readings between the well bore fluid lines are scenarios where the well could contain mixtures of different well fluids. The blue lines show the operating envelopes for the various pumps that could be used to implement seawater injection.

Point A in Fig. 8 shows pumping requirements for a scenario where the normally pressured well is filled with natural gas only at the limit of capability for three mud pumps operating in parallel. For this scenario, the mud pumps could stop a leak equivalent to a 0.9 in diameter orifice. For a 1.2 inch equivalent leak diameter (which has almost double the flow area of a 0.9 in orifice), frac pumps are needed to meet the pumping needs. Changing the well bore fluid to oil, the pressure and rate requirements are shown by the green line. For this scenario, the mud pumps could handle the leak equivalent to a 1.25 in orifice diameter. The mud pumps can handle a 2 in equivalent leak diameter if we assume that the wellbore is filled with a low density mud (mud densities significantly above 9 ppg won't allow this well to be underbalanced).

Fig. 9 shows modeling results for the high-pressure reservoir. In this case, the rig pumps cannot stop leaks from openings as large as the low-pressure reservoir for all scenarios as expected. For a gas-filled wellbore, the mud pumps cannot handle any leak. Point A shows that cement pumps can handle an equivalent leak of ~0.2 inch in diameter because they produce higher pressure although at lower flowrates than the mud pumps. The mud pumps can handle a 0.8 in diameter orifice when the wellbore is filled with oil instead of gas (Point B in fig. 9). A leak slightly larger than a 1.1 in orifice can be handled by mud pumps if the well is filled with drilling mud (Point C). High capacity frac pumps are required to shut in flows with larger leak orifice sizes. Note that more than one frac vessel could be used to handle larger leak openings.

In addition, any seawater injection into the BOP will increase system pressure and thereby reduce the flow of hydrocarbons by roughly a rate equivalent to the seawater injection rate. So, even if complete shut-off isn't possible, it might still be worth injecting seawater to reduce hydrocarbon flow rates.

Two novel concepts to stop flow from an out-of-control well have been developed and tested. One concept relies on a rapid cross-linking polymer reaction to form a high-integrity plug seal when resin and catalyst are rapidly injected into a damaged BOP. The other concept utilizes the back pressure formed when seawater is injected into a damaged well. That is, seawater is pumped through a conduit from the surface and into a well bore. If the rate and pressure of the seawater pumping is sufficient to exceed the pressure that the reservoir is placing in the well bore, the well is dynamically killed.

The polymer plug concept is suitable for regaining control of a well that has failed because of a BOP seal failure. The seawater injection concept has broader applicability. That is, pumping seawater into a damaged and leaking well bore will cause an increase in the pressure anywhere in the well bore as long as there is hydraulic connection to the seawater injection location. This means that seawater injection allows regaining control of a well that has a leak anywhere in the well bore including a damaged BOP. Overpressuring the entire well with seawater will cause the seawater to displace any hydrocarbons escaping from the leak point.

Both concepts were evaluated for feasibility. The polymer plug concept was tested with a scale BOP system to determine if a polymer-plug sea would stop additional flow. The seawater injection concept (SWIM) was evaluated using an analytical model to determine the pressures and flowrates of seawater needed to pressure a well enough to stop reservoir flow.

For the polymer plug concept, polymer seals were formed within the scale BOP in ~2 seconds. Of thirteen separate tests performed in the 1.5” diameter scaled BOP with a 1/8” leak orifice at temperatures and pressure representative of real-world conditions, nine resulted in seals capable of holding up to 4500 psi of hydraulic pressure for over 17 hours. Three tests were performed with the same 1.5” diameter scaled BOP and a ¼” leak orifice. For these, one successful test was completed where the seal was capabile of holding 5000 psi of hydraulic pressure for 1.5 hours.

An analytical model of the SWIM concept was developed and used to evaluate a damaged well for a normally pressured reservoir and for a high-pressure reservoir. The results showed that the required seawater injection rate and pressure requirements go up as the reservoir fluid density decreases from drilling fluid to oil to gas. Another factor that can drive up the seawater injection rates are reservoir pressure. The modeling found that pumps available on drilling rigs, mud pumps and cement pumps, could be used to pump seawater into a damaged well as long as the leak opening was not too large. For larger leak openings, higher pumping rates are needed. To control these scenarios, larger pumps such as those available on fracking vessels may be required.

Both concepts could be used for deepwater and shallow water drilling and both surface and subsurface BOPs. For shallow water drilling where there is the potential for gas release, seawater injection could eliminate gas loss from the well to keep it from reaching the surface where it would interfere with well-control operations. Overall, both concepts provide another barrier to significantly reduce the risk from loss of well control.

These initial studies evaluated the feasibility of using seawater injection as a means of stopping the reservoir discharge for oil or gas wells. The next phase of study for the polymer plug concept will be to evaluate higher-pressure conditions using the simulated BOP system. The next step for the SWIM concept will include transient modeling with a range of reservoir pressures, reservoir fluids, and nozzle openings to provide a wider range of scenarios and an understanding of the timing required to shut in flow.