Prototype of well restriction tool has been designed and tested under static conditions; additional analyses in dynamic conditions still needed
Zach Owens, Brian Smyth, Nicoli Ames, John Pye, and Brun Hilbert, Exponent; Bob Brooks, RM Taber; Hector Mendez, Shell International Exploration and Production
In 2014, Royal Dutch Shell engaged Exponent to develop a prototype of a surface-activated, casing-integrated, flow-restriction tool capable of arresting flow during an uncontrolled blowout. This well restriction tool (WRT) is similar in function to a subsurface safety valve (SSSV) but is designed to be an integral component of a 9 5/8-in. casing string. Unlike an SSSV, the WRT is intended for use during drilling and can be installed deep in the well near the end of the casing string. It is intended to supplement the blowout preventer (BOP) by arresting the influx of hydrocarbons into the well, thereby increasing the chances that the BOP can successfully close. It also serves as a backup in cases where BOP closure is unsuccessful.
Actuation of the device is accomplished by shifting the opening sleeve, thereby releasing six spring-loaded, triangular fingers that come together to fully seal the wellbore. The actuation sequence is initiated when an encrypted wireless signal from the surface is relayed to a receiver on the WRT. When the signal is received, a charge of solid propellant, residing in the propellant cavity (Figure 1), is ignited. The high-pressure gas generated by the combustion of the propellant drives the opening sleeve forward, releasing the spring-loaded flapper (valve) fingers.
As the shift sleeve moves forward, grease from the grease cavity, is expelled into the wellbore through a series of vent holes. When the shift sleeve reaches the end of the flapper fingers and they start to protrude into the wellbore, any upward flow acts to enhance the drive of the fingers into the closed position. Notably, the flapper finger design ensures that the six triangular fingers will close together so that the sealing surfaces are appropriately mated.
The WRT is intended to be a one-time-use, emergency well control tool. To resume operations after actuation, a shift tool was designed to be run from the surface to engage the closing shift sleeve and rotate the valve fingers back into the stowed position, where they are locked into place.
If there is pipe across the WRT, it will not be able to completely shut-in the well. Thus, the most effective placement of the WRT is above the producing zones – but as deep in the wellbore as possible to minimize the chance that there will be interfering pipe present at the time of a well control event.
Design & Analysis
Finite element analysis (FEA) was utilized throughout the development of the WRT. Initially, static pressure analyses were conducted, following the Design by Analysis guidelines of the ASME Boiler and Pressure Vessel Code. The complete stress-strain curve for the tool material, including hardening up to the ultimate tensile strength, was obtained through tensile testing, and the complete elastic-plastic response of the material was used in all of the FEAs.
Figure 2 shows the stress field at the rated valve pressure of 5,000 psi. Other than hot spots at the flapper-housing bearing interface, the stresses in the tool are below the 80-ksi yield strength of the material. When the static pressure in the model was ramped up to 12,500 psi, large areas of plastic strain were observed, but the valve maintained its structural integrity and sealing capability.
The contact pressures from the static FEA indicated that there could be leak paths near the base of the flappers at the flapper-to-flapper sealing interface. The first round of static testing validated this leak path, and elastomers were introduced to improve the sealing of the flappers.
The valve closure event is expected to cause a dynamic fluid hammer pressure that will be more severe than the static pressure loading. To evaluate the valve integrity during rapid valve closure, dynamic FEA was conducted using the Abaqus Explicit dynamic FEA software.
The transient fluid hammer pressure was conservatively estimated using analytical methods and was applied to the valve FEA model. The first dynamic simulation indicated that the flapper pins might fail during a hammer event. The pin material was then upgraded from 80-ksi steel to a 150-ksi beryllium copper alloy.
The dynamic FEA showed that the stronger pins would likely survive the hammer event with minimal plastic straining. Additionally, the dynamic FEA calculated compressive plastic strains at the flapper-housing bearing interface, and tensile plastic strains of approximately 5% were calculated at the flapper pin holes. Despite these regions of plastic strain, the simulations showed that the structural integrity and sealing capability of the valve were maintained.
Once a satisfactory flapper design was achieved, FEA was used to evaluate the static and dynamic loads on the entire tool, including the housing, sliding sleeve and propellant chamber. This guided changes to sealing surfaces, O-ring locations, propellant pressure paths and the sliding sleeve.
The WRT is composed of two primary subsystems: (1) the remote actuation system, which receives the surface signal and translates the sliding sleeve to deploy the valve fingers, and (2) the valve mechanism itself, which consists of the six flapper fingers and associated housing assembly. Extensive analysis was performed for various conceptual designs of the remote actuation system, including an evaluation of the sleeve trajectory under various wellbore conditions and an evaluation of how the sleeve would interact with the flapper fingers during deployment. For the initial prototype testing, it was decided that only the flapper valve mechanism would be manufactured so that its performance could be well characterized prior to moving forward with the mechanical activation mechanisms. Designs for the remote actuation mechanism were completed but not manufactured during this phase of the development program.
The objective of the proof-of-concept testing of the valve sub-section was to determine its performance and to collect data for comparison against the FEA described previously. A test fixture was designed and manufactured specifically for testing the valve mechanism at various hydrostatic pressures, high temperatures, and different installation orientations. Figure 3 shows a schematic of the test fixture and a photo of the valve assembly.
A full-scale prototype valve assembly was manufactured and subjected to a range of pressures up to 10,000 psi, using water as the medium. This final test pressure is well in excess of the desired pressure rating of 5,000 psi. One of the flapper fingers was instrumented with strain gauges at four locations using triaxial rosettes, and a second flapper finger was instrumented with a single rosette to provide measurement redundancy.
Additional instrumentation to measure fluid pressure and temperature was installed on the high-pressure side of the valve. Testing at ambient temperatures was conducted with the low-pressure side of the WRT exposed, while testing at high temperature, up to 250˚F, was conducted with water on both sides of the valve.
Two variations of the WRT sealing surfaces were tested. The WRT was initially tested with metal-to-metal (MTM) contacts between the sealing surfaces of the flapper fingers, and between the perimeter of the flapper fingers and the housing. This design incorporated only a very small elastomer for the center plug.
One of the ways in which the MTM valve design was evaluated through FEA was by analyzing the deformations that occur under load and reviewing the resulting contact pressures at the mating surfaces. It was determined from the analysis that the MTM design would likely leak to some extent, but the design was explored because it offered the potential of a simpler sealing strategy and broadened the use cases.
There also was no difference between the basic geometry of the MTM design and the final elastomer-based prototype; therefore, testing the MTM design did not require manufacturing two separate valve assemblies. Preliminary testing was conducted on the MTM design, but ultimately the prototype sealing surfaces were unable to seal sufficiently. This was in large part due to real-world manufacturing tolerances and the complex mechanism geometry.
The MTM testing also provided preliminary data regarding the performance of the center plug. This data was used to revise the center plug for the elastomeric prototype. Additional research into alternative mating surface geometries may provide a viable path for an all-metal valve section, but it was decided to modify the MTM design to include elastomeric seals. The seals were incorporated into the final prototype.
Preliminary tests were conducted on the elastomer-based valve assembly, where slight variations on the center plug were tested until a design proved successful in sealing completely and reliably. Formal testing of the valve assembly took place after this design was established, and no additional changes to any part of the assembly were made after the center plug was finalized. This version of the prototype went through a more formalized test series, where it performed well over a wide range of pressures and elevated temperatures.
The prototype valve assembly required initial flow to energize the valve and allow the flapper fingers to seal completely. This was accomplished using pre-charged accumulators. The valve energized quickly, with very little flow, as observed by the limited pressure drop in the accumulators. The general test procedure was as follows:
• Install the valve assembly in the test fixture;
• Cycle the valve using a mock shift tool (used to open the valve, and then removed from the assembly, allowing the valve to close under only the spring load provided by the flapper finger torsion springs);
• Charge the accumulators to nominally 1,200 psi. Although lower pressures were also tested successfully, this was the typical initial pressure used throughout the majority of the testing;
• Start the data acquisition and cameras;
• Open the accumulator valve and monitor pressures for 5 minutes;
• Start pumps and pressurize system to pre-determined test pressure(s);
• Hold pressure for pre-determined duration while remotely monitoring the valve for any leakage; and
• Release pressure and stop the data acquisition and cameras.
Valve leakage during the initial prototype testing for the MTM design was monitored via remote cameras, monitoring system pressure and through fluid collection. Immediately after the initial MTM tests, a liquid gasket material was used so that leakage locations could be identified and evaluated through post-test inspection. These post-test observations were used to optimize the elastomeric design that followed.
The valve housing and flapper fingers were modified with grooves to accept the molded-in elastomers and a larger center plug. Two center plug materials were tested during this process, and a softer material was chosen. The final elastomeric design was tested, without modification, at each condition in Table 1. Only slight valve leakage was observed when priming the system, where the pressure was essentially zero, and prior to energizing the valve with the accumulators. Once the accumulators were primed, there was no leakage. Post-test, pressures were bled to zero while continuing to monitor performance of the valve. Leakage did not occur until the system pressure was below 100 psi.
Very small leaks were observed at the center plug location during two of the tests when the same elastomer plug was used for multiple tests. This leakage was minimal and occurred during two of the repeated cycles at 5,000 psi and below. It was observed after the test that the bottom of the elastomer was being plastically imprinted (deformed) at higher pressures. Leakage did not occur for any of the tests where the plug was replaced with a new part, as the WRT sealed completely whenever new elastomers were in place. Since the WRT is intended as a single-use tool, this did not cause concern.
Tests were conducted over a wide range of pressures to evaluate seal performance. The elastomeric prototype design performed well by sealing completely over a wide range of pressures and is deemed a successful proof of concept.
The tests proved that the concept is viable under static conditions and that additional analysis and testing under dynamic closing conditions are warranted. Dynamic testing will require manufacturing of the prototype housing and sleeve components that were designed, but not built, during the development efforts to date. The next step in the development of the WRT is to install a prototype in a flow loop so that it can be tested under flowing conditions, similar to what would be expected in the field. DC
This article is based on a presentation at the 2017 IADC Well Control Conference of the Americas, 29-30 August 2017, Galveston, Texas.