Making the Connection for Well Control on Floaters: Evolving Design Rationales for BOP Control Systems

2.2.1. Hydraulic communication, “surface—subsea” issues

Let us assume that the surface BOP stack has now been submerged for service subsea. Let us use the stack shown in Figure 4 on page 77. This basic stack shown, using today’s nomenclature (API Standard 53) [4] is a Class 4 A1-R3, interpreted this means a total of four preventers, one of which is an annular and the remainder are ram type preventers. The hardware at the base of the stack is a NT2 adapter which nipples up to the riser down on a jack-up. The NT2 adapter is not a hydraulic function on a surface BOP stack and is manually operated by a circular array of mechanical locking dogs [5].

And let us add that the stack has two hydraulically actuated BOP mounted valves (one on the choke line and the other on the kill line) [6, 7].

Therefore to now sum the quantity of supply and return hydraulic hoses required to control the functions of this stack if it were underwater would be:

Annular preventer2 each 1 in. nominal diameter

Upper pipe rams2 each 1 in. nominal diameter

Shear blind rams2 each 1 in. nominal diameter

Lower pipe rams2 each 1 in. nominal diameter

Choke high closing ratio valve (HCR)2 each ½ in. nominal diameter

Kill HCR2 each ½ in. nominal diameter

Figure 8 overleaf shows a scaled cross section of a hose bundle that satisfies the requirements to provide the above functions with hydraulic power. We can see, with some spare hoses surplus to requirements, the OD of the entire bundle is only ~6 in.

However, if we consider a BOP stack designed and built for subsea service (not a surface stack submerged!), the story is very different (Figure 8).

The subsea blowout preventer stack shown overleaf is an 18¾ in. nominal wellbore diameter, rated at 15000 psi maximum working pressure. This is denoted as “18¾—15 M.”

This particular blowout preventer stack is somewhat dated; “third generation” puts its age genre at around 15–20 years (Figure 9).

Given that the minimum outside diameter (OD) for the hose bundle is going to be around 7½ in. (with no spare lines in the bundled matrix) and the minimum critical bend radius (MBR) for this bundle, let us give the reel some arbitrary dimensions, as indicated in Figure 11.

With these dimensions, the first wrap on this drum would store around 365 feet. Two wraps then would cover the water depth requirement of 750 feet. However, for the “storm loop” hose allowance in the moon pool to accommodate rig heave, another 250 feet would be required. This necessitates three wraps on this reel assembly.

The end plates’ diameter would be in the order of 22 feet diameter. The reel assembly, its prime mover, and brake assembly are large scale!

Typically, hose reel assemblies are installed at an intermediate elevation above the moon pool weather deck elevation on a mezzanine deck. Two such typical arrangements are shown in Figures 12 and 13.

2.3.1. Introduction

The ingenious design of the first BOP control system, which is now presented as an overview, was developed in the first half of the 1950s when the maximum rated water depth for drilling offshore off floaters was still under 1000 feet (305 m) [8].

The reasons for the water depth limitation are varied and not directly attributed to the restraints of the BOP control system. Some of these were marine, drilling plant topsides’ limitations and to a lesser extent, and capabilities of marine drilling riser (the mechanical connection between the subsea BOP stack and the drilling installation).

2.3.2. Concept of the hydraulic pilot-operated control system

The immediate problems facing the pioneering design group:

How do we overcome the excessive dimensions of a simple closed hydraulic system deployed subsea?

How do we diminish the friction losses in the closed hydraulic system where the displaced fluid from the “other” side of the function slows the overall response times of the ram type and annular type preventers?

How do we build in redundancy to a point where the critical control system can satisfy the most stringent of regulators for reliability?

What can be done to minimize the risks of pollution to the marine environment?

Is there some way in which the volume/pressure drawdown effect can be reduced in the direct closed hydraulic system?

How should the topside equipment be arranged for optimum operation and account for rig motions?

What is required to configure the BOP stack to enable a disconnect while leaving the well secured in the coincidental event of a well influx?

Pictorially shown here are the basic principles that were proposed (Figure 14).

Pivotal to the success of the prototype design was the use of hydraulic relay valves which are installed in the newly conceived control pod(s) which are activated by a hydraulic pilot signal commanded from the surface. By the use of hydraulic relay valves and agreement that the displaced fluid volume from the “other” side of the function should exhaust directly through the “other side of the function” relay valve directly to the marine environment, it was immediately understood that the prior formidable size of the hose bundles could be greatly reduced, not least caused by using 3/16 in. pilot hoses in the bundle. The main hydraulic supply consequently consisted of one only nominal 1 in. diameter core hose within the bundle (see top right of the previous figure for details).

In order to have an open hydraulic system that exhausted hydraulic fluid directly to the marine environment, the hydraulic medium was changed from lightweight mineral oil to potable water dosed with additives in small percentages of dilution. This new hydraulic medium necessarily influenced careful material selection of both metal and rubber sealing components of the hydraulic valves, regulators, and other subsea control system components. Not only was the marine environment a factor in dispelling the consideration of the use of an oil-based hydraulic medium, but also differential pressure experienced across the thermoplastic wall of flexible hose at increasing hydrostatic pressure from the water column depth. This is discussed in the next sub-section.

100% redundancy was provided by furnishing two identical systems which became color-coded blue and yellow. The system is arranged whereby one to the two identical sides of the system may be used at any given moment but never both. The redundancy satisfied both operators, oil companies, and more importantly, class societies and legislative bodies. Since the early systems, levels of redundancy have been revised, and standard operating procedures adopted formerly have been revised reflecting greater caution and conservatism. This will be discussed in due course.

By the introduction of gas pre-charged hydraulic accumulators, nominally 11 US gallons capacity each, the early problems of system drawdown effects were satisfactorily banished as the 1 in. hydraulic supply in either hose bundle maintained full system working pressure in the stack-mounted accumulator bottles.

This system quickly became field proven and a number of proprietary vendors produced their own systems, however it has to be said that all were based on the principles put forward originally by Paul Koomey and his design team.

The remainder of this sub-section concentrates on the general arrangement detail of the system and its operational characteristics and finally limitations identified for this system.

2.3.3. The basic control pod

The hydraulic control system is always equipped with two control pods, designated as the blue or yellow pod. To maintain a fully redundant control system, both pods must be operational at all times.

Formerly, if a control pod becomes inoperable, drilling operations would be normally suspended and the BOP stack controlled with the working pod until repairs are completed and tested. This involved the retrieval to the surface of the defective pod, repair and test on surface before re-deploying subsea to latch back into its dedicated receptacle on the LMRP.

More recently with the advent of deep water drilling, the majority of oil companies will not allow continued drilling operations for the retrieval of one pod to surface for repair. If repairs are to be performed in the midst of a drilling program, drilling operations are suspended, and the well made safe and the entire LMRP retrieved to surface to repair the faulty control pod.

The active and selected control pod is normally alternated between the pods weekly or after a BOP stack test.

Koomey Shaffer introduced a 42 line retrievable pod which featured a double female receptacle design. The separate receptacles enable both the pod to be retrieved or else the entire LMRP (Figure 15).

Later, as the drilling contractors began to use BOP stacks with greater number of functions, Shaffer and others introduced a 64 line control pod, which, while featuring a different geometry (cubical rather than cylindrical) operated in the same manner and was intended for retrieval during drilling operations (Figure 16).

2.3.4. Control system hoses

2.3.4.3. Factors influencing time response

One of the basic concerns in regard to using thermoplastic hose to transport hydraulic fluids to great depths is differential pressure across the tube wall. At great depths, the external pressure may be sufficient to collapse the hose. The pressure at which collapse takes place is dependent upon the hose construction and the nominal diameter of the hose [4, 7].

The dominant property of differential pressure in this application arises from the differences between seawater and control fluid densities. Wherever in this type of system, there exists a degree of density difference across the hose tube wall, a chance of invoking hose collapse is possible. For instance, at a depth of 5000 feet, a thermoplastic hose containing a typical mineral oil as the hydraulic medium found in surface stack control systems will experience an overburden of around 220 psi, which is quite sufficient to collapse a hose. Pressures as little 30 psi can cause collapse of hoses with nominal diameters in the range of 3/8–½ in.

API specification 17E: specification for subsea production control umbilicals, states for collapse pressure [9]:

“The minimum value of external collapse pressure shall be 150% of the difference in the static head due to hydrostatic pressure at the maximum design depth less the static head at that depth due to the service fluid (hydraulic medium).”

Further unwanted differential pressure will be generated if the hydraulic lines are not 100% fluid filled. If any entrapped air is present in the tube length, the hydrostatic pressure will dominate and tube collapse will occur. This is easily eradicated by thorough purging and venting of all lines in the subsea umbilical hose bundle. The presence of air, however small, also dramatically increases response times due to the compressibility of gases [3].

The differential problem is overcome by choosing a hydraulic medium which has a specific gravity that is close to seawater.

Seawater has a gravity of ~1.03 and water is 1.00. Providing that the hydraulic medium is water-based with additives that only change the specific gravity to a new value remains close to that of the specific gravity of seawater then the possibility of hose collapse is virtually obviated.

2.3.4.3.1. Viscoelasticity

Hoses, being composites with polymeric constituents are found to behave in a time dependent viscoelastic manner when applied load is a hydraulic charge as found within a pilot line hose.

The result of the viscoelasticity manifests itself in a pressure decay after initial pressurization. This is not detrimental for the pilot signals in this application since the hydraulic pilot-operated relay valves subsea “fire” and “vent” at pressures well below the nominal pilot pressure of 3000 psi. Figure 17 shows the pressure decay versus time. The typical time constraints are well beyond time of the hydraulic relay valves “firing” in this control system.

2.3.4.3.2. Hose geometry changes during pressurization

Extensive laboratory testing has been performed to assess the changes in hose geometry subjected to a step positive change in internal pressure. The changes in geometry were measured using strain gauges, both axially and circumferentially affixed to the outer surface length of the hose under test.

The axial gauges measured any bending strains incurred and the circumferential gauges monitored hoop stresses. It was found that the hose length shortened with pressurization and this is explained by the layers of hose braiding attempting to establish a neutral lay angle during the buildup of pressure. The effect is almost instantaneous and remains constant, and hence the axial strain is not responsible for the viscoelastic effect.

Measured hoop strains correlate to observed pressure responses and shown typical viscoelastic behavior. In tandem with strain measurements, volume measurements have been recorded to estimate the variation in wall thickness. Such measurements have been quantified using two equations which account for the bulk modulus and pressure decay following initial viscoelastic expansion of the hose under test.

Results from these tests showed that both the internal and external diameters of the hose increased with pressurization although the OD significantly less than the ID of the hose.

Overall, this indicates that all hydraulic pilot hoses will “accept” more fluid when a pressure signal is initiated from the source and will duly expand in direct correlation with the VEC of the hose: dependent upon construction and materials. At pressure equilibrium (e.g., 3000 psi), the hydraulic pressure peak will transit the length of the hose at approximately the speed of sound.

2.3.4.3.3. Behavioral phenomena of thermoplastic hose

All hoses that are constructed of material that use a composition of polymers and fibers may be classed as thermoplastic hoses. When subjected to pressure changes internally and externally, they exhibit a viscoelastic time-related response. After an initial pressurization, the pressure decays over a period of time as the hose dilates (see the previous figure) [3].

The extent of the dilation is dependent upon a number of factors such as hose material, construction, age, environment, and so on. A similar effect occurs when the hose is depressurized, this being a time-related contraction effect.

Against logical intuition, hoses bundled together exhibit greater volumetric expansion (VE) than identical hoses pressurized in isolation. There is a mathematical proof for these phenomena but suffice it to say that the reason is simply because there are effects from adjacent bundled hoses which remain pressurized against those vented to zero gauge.

It is known that aging in hoses reduces VE which acts in our favor (in drilling BOP controls) but is considered detrimental in production control systems.

Minimizing the effects of VE promotes faster response times in hydraulically piloted BOP control systems since the pod-mounted relay valves will not “fire” until they have sufficient pressure in the hydraulic pilot signal: normally around 500–700 psi.

The following figures illustrate some of the effects of the volumetric expansion characteristic in thermoplastic hoses (Figures 18, 19, 20).