MODELING AND ANALYSIS

On the rapid discharge of subsea accumulators: remarks on

the normed design method and proposal of improvement
Roberto Cipollone1, Fabio Fatigati1, Davide Di Battista1 , Paolo Allara2 & Nicola Carini2
1University of L’Aquila, via Giovanni Gronchi18, L’Aquila 67100, Italy
2Saipem S.p.a., Via Martiri di Cefalonia 67, San Donato Milanese (Milan) 20097, Italy

Keywords Abstract
API 16D, blow out preventer, energy storage,
hydraulic accumulators, subsea accumulator This study is focused on the design of the subsea accumulators currently used
design to deliver a pressurized fluid whose function is to actuate safety operations on
a subsea well. API 16D, whose specifications regulate the design method of a
bank of accumulators, was revised, discussed, and resumed in terms of a nomo-
Correspondence gram whose engineering value is useful for designers. An additional nomogram
Fabio Fatigati, University of L’Aquila, via was derived allowing to perform the verification process which is suggested by
Giovanni Gronchi18, L’Aquila, 67100, Italy. the norm, in order to verify the fulfillment of the actuations requested to the
E-mail: [email protected] accumulators. This verification phase caught the interest of the Authors: API
16D assumes for the design that the Functional Volume Requirement (FVR) –
Funding Information
the sum of the quantities of pressurized fluid to be delivered for different
Italian Economic Development Ministry.
sequenced-­in-­time functions – is delivered at the Minimum Operating Pressure
Received: 11 December 2017; Revised: 12 (MOP) requested by the last actuation. This hypothesis simplifies the design,
May 2018; Accepted: 15 May 2018 but the result cannot fulfill the previous functions because of the pressure lack
(energy content in the pressurized propelling gas) which can happen. The use
of the two nomograms proposed simplifies the application of the design proce-
dure and allow to discuss main sensitivities of the variables involved (FVRs &
doi: 10.1002/ese3.203
MOPs), evidencing which variables deserve more attention and accuracy. In
particular, the use of the second nomogram permits to verify that the right
volumes of fluid during the actuations are delivered above a minimum pressure
level, so guaranteeing the fulfillment of the function. It is based on a thermo-­
fluid-­
dynamic discharge modeling of the accumulators. The paper outlines a
design direction which minimizes the number of accumulators and opens the
way to a different design approach, based on a fully physical discharge process
description.

Introduction
times must be very short, discharge of the pressurized
A system of accumulators is a set of pressure vessels fluid must be almost instantaneously: in the severe en-
charged with inert gas and used to store pressurized hy- vironments represented by high water depths, these ac-
draulic fluid. Energy accumulation through pneumatic-­ cumulators are sited close to the well. From this position,
hydraulic devices is widely used in renewable energy they assist a system of valves (Blowout Preventer, BOP)
generation (in order to solve the discontinuity of the which definitively close the well in case of critical situa-
source) [1, 2] or in mobile applications (principally for tions for human safety and environmental concerns. Main
energy recovery) [3, 4]. Moreover, they can be used to limitation of these devices is the low-­energy storage density
actuate specific functions, mainly related to safety reasons. [5]. An additional concern is due to the dependence of
In the oil and gas sector, for instance, these devices are the fluid pressure to the quantity of the fluid delivered:
commonly used to fulfill most part of the actuations which during discharge, the pressure inside the accumulators
can be near or far away from the accumulators themselves. decreases and it can reach levels unsuitable for a correct
In the off-­shore deep-­water drilling, when the actuation actuation. So, the requirements of energy accumulated

© 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd. 1
This is an open access article under the terms of the Creative Commons Attribution License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited.
On the Rapid Discharge of Subsea Accumulators R. Cipollone et al.

and power delivered push the design toward heavier and actuated by rapid discharge accumulators: a pressurized
larger systems [6]. fluid delivered by them is used to accomplish the func-
In the oil and gas sector, the use of pneumatic-­hydraulic tions which require given quantities of fluid at a specified
accumulators represents the only reliable technological pressure level which insure the right energy content.
option for valve management and safety ultimate actua- Actuating time must be lower than 3 minutes as required
tions. In spite of this, the specific literature is limited by API 16D [15]. In order to be sure about the actua-
and the sector demonstrates a certain degree of techno- tions, accumulators are mounted in a subsea position as
logical inertia. In most severe situations, BOP management backup power fluid supply source.
produces an irreversible situation represented by the cut All the energy accumulators dedicated for these specific
of the drill pipe and the sealing of the well. functions (AS, DM, EDS) must work properly, and in
Blowout is an uncontrolled jet of crude oil and/or case of AS and DM also independently from any other
natural gas which may be due to the high formation high-­ pressure fluid supply from topside: they represent,
pressure, lower mud density, short borehole mud slurry therefore, the last chance to prevent blow out and avoid
column height, or other improper measures [7, 8]. The unpredictable consequences.
procedure of a blowout preventer is done through a shut The severity and the enormous worldwide relevance of
in of the well, which, however, has some concerns related the blow out events invite to a design of the actuation
to over pressures and actuation times [9–11]. If the BOP system scientifically based, without ambiguities, reliable
system fails, the consequence could be devastating as hap- and consistent with experimental data [17–19]. This paper
pened in the Gulf of Mexico on 20 April 2010 during offers a contribution in this direction, considering the
the Deepwater Horizon accident, the worst Environmental severity of the events which could happen if a failure
Disaster in history of the United States [12, 13]. This occurs to one of the three actuations.
event increased the international scientific and technical The correct design of the bank of accumulators which
interest toward the reliability issues of the BOP, even deliver a pressurized fluid to accomplish AS, DMAS, or
though main cause of the disaster was also referable to EDS must comply the requirements reported by API 16D-­
other circumstances [14]. Method C [15]; some additional references can be found
In any case, in order to prevent definitively blow out in API Standard 53 [16]. The cited norm gives proper
when the evidence of a disaster is approaching, some rules to size and define the pressure conditions inside
emergency actions have been considered as last operations, the accumulators. Several operating parameters must be
most of them are irreversible. They are recognized by specified in order to manage the processes which influ-
the international standards API Standard 16D [15], and ence the quantity of the pressurized fluid that can be
by the API 53 [16] and are referred as Autoshear System delivered: AS, DM, and EDS actuations are considered
(AS), Deadman (DM) and Deadman/Autoshear (DMAS). as the most important safety sequence.
Autoshear System (AS) is defined as “safety system that The effect of the surface and subsea temperatures and
is designed to automatically shut in the well-­bore in the precharge pressure uncertainties are not to be neglected:
event of a disconnect of the LMRP. When the autoshear these uncertainties, particularly relevant during typical
is armed, a disconnect of the LMRP closes the shear off-­shore severe operating conditions, should be included
rams and this is considered a rapid discharge system.” in the design sizing [20]. As the water depth increases,
Furthermore, the Deadman (DM) is defined as “a system it is known that the delivery of a pressurized fluid become
designed to automatically shut-­in the well-­bore in event more difficult due to the effect of the hydrostatic head
of simultaneous absence of hydraulic supply and control and pressure increase inside the accumulators: the devel-
of both subsea control pods,” and it is considered also opment of the accumulator technology (to reduce the
a rapid discharge system. The Deadman/Autoshear System negative effect of the water depth on the volumetric ef-
(DMAS) is designed so that the function for the High ficiency) has also been analyzed [21]. In order to com-
Pressure Shear Rams is the same for the Autoshear and pensate the effect of the hydrostatic pressure on the usable
Deadman as reported in the API Specification 16D [15]. fluid, the increase of the precharge pressure can be handled,
An additional and optional system is the Emergency but this is limited to the rated pressure of the accumula-
Disconnect Sequenced System (EDS), which is defined in tor. Moreover, the liquid volume (when the precharge
the study by Han and Zhang [10] as a system that “shall pressure increases) could be too low, do not insuring the
be provided for a deep-­ water floating drilling rig when actuation requirements in terms of fluid delivered. An
there is a requirement to rapidly disconnect the riser in interesting technological alternative is offered by the so-­
the event of inability to maintain rig position within a called Constant Differential Accumulators [22] which
prescribed watch circle.” EDS is classified as a rapid dis- ensure to satisfy the functional requirement (fluid deliv-
charge system as well. AS, DMAS, and EDS need to be ered) without additional propelling gas pressurization.

2 © 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd.
R. Cipollone et al. On the Rapid Discharge of Subsea Accumulators

Other technologies have been presented in order to sustain of the accumulators discharging process when they are
the fluid pressure inside the accumulators when the water sited close to subsea floor. The nomogram allows easily
depth increases, replacing the pressurized gas inside the the verification process requested by the normed method
accumulators with springs and heavy weight [23]. Subsea and explains why, eventually, the design fails in terms of
operations can be enhanced providing greater shearing unsufficient pressure when specific quantity of fluid (FVRi)
pressure and delivering greater fluid volumes when high is extracted. Being the verification a very important pro-
strength casing with thicker walls must be cut as it hap- cess, the thermodynamic properties of the fluids involved
pens in deep-­water wells [24]. Also different fluids and (propelling gas and actuating fluid) were evaluated through
materials have been investigated for this purpose [25]. NIST™ database without any simplification concerning the
This paper analyzes some relevant aspects of the pro- nature of the propelling gas.
cedure standard reported by API 16D-­Method C, and it Finally, the paper discusses how to manage the situa-
reorganizes the design phase in one original nomogram tion in which the verification process is not fulfilled.
in which the role of all the operating conditions is easily Indeed, API 16D Method C leaves to the designer the
represented: this will give confidence about the importance choice to overcome this occurrence, that is, increasing
of the different terms requested by API Specification 16D the number of accumulators or changing the precharge
[15], outlining most relevant parameters and operating pressure with respect to the optimum value. In order to
variables which are more sensible to the number of ac- understand and to choose the best way to deal when the
cumulators needed for the actuation sequence. Considering verification process is not fulfilled, an analysis using the
that the nomogram puts in evidence the role of all the second nomogram is proposed.
variables involved in a design (according to the standard),
it has an interesting engineering value and adds to the
Materials and Methods
literature a further knowledge about the sensitivity played
by different variables. API 16D Method C [15] gives the guidelines to design
It is the aim of this paper, also, to highlight the “weak- rapid discharge accumulators used to actuate different
ness” of the standard procedure which proposes a design functions mainly related to safety operations. Indeed, these
simplification which does not inherently insure the success are the last resorts in case of a blowout occurrence and,
of the actuations for which the accumulators were de- as prescribed by the regulation, all the function must be
signed. The design nomogram presented allows to have actuated in <3 min in order to prevent blow out and
a graphical outlook of the design and to put in evidence fire occurrence. Once the precharge pressure was set at
the role of different variables referred to the environmental surface, the regulation design procedure allows to predict
conditions as well as the overall quantity of actuating the number of accumulators which ensures that the pro-
fluid to be delivered and the pressure at which the last pelling gas is at its minimum pressure after the discharge
actuation is requested. More recently, computational meth- of the entire volume of control fluid required by all the
ods have been presented [26] to support the design and actuations. Therefore, the regulation requires to fix
the Authors are also developing a fully physically consistent the pressure for the last actuation: it can be fixed by the
mathematical model. Nevertheless, a graphical approach requirements of the device or by the hydrostatic pressure
allows to observe tendencies and most important param- if this is higher than the previous value. So, the norm
eters concerning the design. The graphical treatment allows provides the optimal value of the precharge pressure in
to understand why the norm sometime fails (lack of pres- order to deliver the whole volume required (Functional
sure, i.e., of energy) understanding the “design distance” Volume Required, FVRtot, as sum of the all actuations),
to a failure. In fact, the most important aspect when minimizing the number of accumulators needed. In gen-
using the norm is that, for sake of simplicity, it simplifies eral, this approach cannot insure the requirements of the
the design phase considering the full delivery of the fluid intermediate actuations: because of this, the API 16D
(the sum of all the fluid quantities requested by the vari- Method C [15] advises to accomplish a further verifica-
ous actuations) at the Minimum Operating Pressure (MOP) tion step to check about the success of all the actuations.
which characterizes the last actuation (the one at the In the following, in order to clarify the use the regulation
lowest operative pressure). So, the normed procedure does and analyze the effects of the operational and environ-
not consider in any case the pressure levels which char- mental quantities on the design, the normed procedure
acterize the previous actuations. Due to this, the norm was reorganized and implemented in an algorithm which
requires a verification process which will put in evidence produced a design nomogram (Fig. 4) particularly easy
that all the actuations cannot be fulfilled. Following the to be used. A further innovative nomogram (Fig. 5) has
first graphical approach, a second original nomogram has been presented suitable to easily proceed with a verifica-
been derived in the paper through a thermodynamic model tion phase: it is based on a thermodynamic model of the

© 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd. 3
On the Rapid Discharge of Subsea Accumulators R. Cipollone et al.

accumulators discharge when they are sited close to

wellhead.

An algorithm for the bank of accumulators

design
The calculation method described in API Specification
16D [15] has been implemented in a MS Excel™ script,
coupled to NIST™ fluid database, to evaluate all relevant
properties of the propellant gas during the charging and
discharging phases with the control fluid. The initial phase
is the precharge of the accumulator: this precharge pres-
sure is related to the initial charge of inert gas in the Figure 1. Conclusion of the precharge phase.
accumulators when the accumulators are at a certain
surface temperature Tsurf (off-­shore platform level). After fluid PHydro,CF (eq. 4.1) for not regulated circuit or the
the precharge phase has been concluded (Fig. 1) and the hydrostatic pressure of the sea water PHydro for regulated
accumulators have been moved down to the wellhead, circuit (eq. 4.2), as specified in [16].
the control fluid is pumped in the bottles from the rig. P1 = PHPU + PHydro,CF (4.1)
Therefore, the inert gas inside the accumulators is com-
pressed ensuring to store potential energy. The regulation P1 = PHPU + PHydro (4.2)
suggests an optimum value of precharge pressure Popt and
explains how to calculate it. The optimum precharge pres-
In API Specification 16D [15], it was stated that the
sure allows to maximize the Volumetric Efficiency of the
value of PHPU is the at “pump stop pressure” if the
accumulator VE, which depends on the precharge density
accumulators are isolated with a check valve or at “pump
D0 value already calculated; this will minimize the number
start pressure” if the accumulator pressure fluctuates with
of accumulators to be submerged. VE, in fact, is the ratio
pump pressures as the main accumulators normally do.
between the usable volume fluid for the function i and
From this knowledge, the entropy S1 of the gas when
the total amount of gas inside the accumulators:
the accumulator is placed on the subsea floor and when
(Vi − V1 ) it is ready for operation can be evaluated. This quantity
VEi = . (1)
V0 ⋅ Fi is fundamental to evaluate the thermodynamic state dur-
Being the gas volume equal to the ratio between the mass ing the discharge phase because this process is modeled
of gas and the density at the condition considered as: as an adiabatic (isentropic) transformation;
m According to API Specification 16D [15], Di is the density
Vi = . (2) of the inert gas at withdrawal condition of interest (Fig. 3),
Di
such as minimum operating pressure (D2) or total discharge
Equation (1) gives VEi as follows: condition (D3). Concerning the real gas state at minimum
(
m m
) (
D0 D0
) operating pressure, this condition is defined by the MOP
Di
− D1 Di
− D1 (the lowest delivery pressure of the sequence corresponding
VEi = m = . (3)
⋅ F i Fi
to the last function) or by the sea water hydrostatic pressure
D 0
PHydro if it is higher than MOP, and the S1 being the dis-
Equation (3) shows the volumetric efficiency at charge phase modeled as an adiabatic isentropic transforma-
i-­
discharging condition; it depends also on two other tion. Otherwise, in the total discharge case which represents
thermodynamic quantities: D1 and Di. It applies that: D1 the situation of a full discharge of the fluid, if the gas pres-
is the density of the inert gas when the accumulators sure in the accumulators P3 is higher than Hydrostatic Pressure
have been moved down at wellhead and have been charged PHydro, the fluid in the accumulators will be discharged
with the control fluid (Fig. 2). According to API 16D in entirely and the thermodynamic state of the inert gas is
charged conditions, the thermodynamic state of the inert defined by P3 and S1. Nevertheless, if P3 is lower than PHydro,
gas is at Tsub being the charging process enough long to a part of the control fluid will remain in the accumulator
assume the gas temperature equal ambient temperature. after discharge and P3 will be called sea-­ head limited: the
The pressure P1 is given by the sum of the pressure of thermodynamic state of the gas is defined by PHydro and
the Hydraulic Power Unit PHPU (used to pump the fluid S1. The regulation disposes that the Volumetric Efficiency
inside) and the hydrostatic pressure of column of control is equal to the minimum between Volumetric Efficiency for

4 © 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd.
R. Cipollone et al. On the Rapid Discharge of Subsea Accumulators

MOP is lower than PHydro, D2 is the density of the gas

at the thermodynamic state defined by PHydro and S1.
According to the regulation, there is an optimum value
of precharge pressure: this happens when VEv is equal
to VEp. Therefore, this optimum precharge pressure could
be obtained as in the following:
( ) ( )
D0
D2
− DD0 D0
D3
− DD0 (7)
1 1
=
FP FV

API 16D introduces FP and FV in order to provide a more

conservative design of accumulators. These coefficients are
assumed equal when the API 16D Method C is used; the
effect is a reduction of the Volumetric Efficiency (VEp, VEv)
by 10%. Thus, equation (7) can be rearranged as follows:
D0 D0
= (7.1)
D2 D3
Figure 2. Accumulators at charge condition.
Equation (7) implies that D2 = D3. The optimum value
pressure-­limited case VEp (eq. 5), corresponding to the MOP for the precharge density is given by the condition D2 = D0
condition (the lowest among the MOPi of the functions in which insures that the accumulators will be empty in
the sequence or PHydro if this value is higher than MOP) terms of actuating fluid at the end of the last discharge.
and Volumetric efficiency for Volume limited case VEv (eq. 6) Considering that the temperature at the surface is known
corresponding to the total discharge condition only if P3 is (during the precharge phase), the optimum precharge
greater than PHydro, otherwise P3 is sea-­head limited. pressure is univocally defined.
Being the Functional Volume Requirement given by
the sum of all the specific volumes required by the single
( )
D0 D0
−
D2 D1 (5) actions defined as
VEP =
FP nf
∑
( ) FVR = FVRi (8)
D0
− DD0 i=1

VEV =
D3 1
(6)
FV the Volumetric Capacity (VC) required for the accu-
mulator system is:
In equation (5), D0 must be lower than D2. Moreover, FVR
VC = (9)
as reported in API Specification 16D [15], if the minimum VE

Figure 3. Accumulator state at withdrawal condition of interest (after the actuation of i-­th function).

© 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd. 5
On the Rapid Discharge of Subsea Accumulators R. Cipollone et al.

The number of accumulators na comes by dividing the Finally, the Volume of working fluid VL1 stored in ac-
entire VC by the volume of a single accumulator SVA: cumulators is outlined in Figure 4(D) as x-value, by
knowing the VC (y-value of Fig. 4C) and the density of
VC
na = (10) the propelling gas at precharge (y-value of Fig. 4A).
SVA
The analysis of the nomogram in Figure 4 allows to
Being VC known as well as the precharge density of understand the criticity of some operations, in particular
the gas, the volume of the propelling gas and that of the for very influencing parameters, like the volumetric ef-
actuating fluid can be obtained according to equations (11) ficiency. Some interesting observation on the operating
and (12), respectively: quantities can be summarized:
( ) 1. The density of the propelling gas D0 inside the accu-
D0
V1 = V0 (11) mulator (y-value of Fig. 4A) once the precharge phase
D1 has been concluded should be selected in order to obtain
( ) ( ( )) the maximum Volumetric Efficiency (VE) (x-value of
D0 D0 Fig. 4B) when the MOP of the last actuation has been
VL1 = VC − V 0 = V0 1 − . (12)
D1 D1 fixed. Therefore, the optimal gas precharge density D0opt
The design specified by the norm has been synthesized for a certain final MOP is that value which corresponds
in the nomogram reported in Figure 4. to the maximum of the respective VE curve in the (B)
It has been conceived following a sequence of steps, hav- part of Figure 4. As can be observed, the maximum of
ing chosen specific values for the horizontal and vertical VE shifts upward as the MOP of the last function in-
axes and suitable parameters which specify the different curves. creases, thus the D0opt becomes higher.
The steps reported in Figure 4 have been referred to 2. The precharge pressure P0 which ensure to obtain a
the data reported in Tables 1 and 2 which makes refer- given density of the propelling gas D0 increases as the
ence to a typical sequence of actuations and operating Surface Temperature grows, as can be observed from
conditions. the Figure 4(A). Therefore, the precharge phase should
To apply the nomogram, the following data must be be performed when the surface temperature is lower
specified: because the same density of the gas can be obtained at
lower pressure compared with the case when this tem-
1. water depth; perature is higher. Nevertheless, in this case, the ac-
2. HPU pressure; cumulators should be submerged as soon as possible
3. density of the seawater; because if the surface temperature grows, the pressure
4. density of the fluid; in the accumulators increases and may become higher
5. subsea temperature at the well. than its maximum pressure rating.
3. The MOP of the last function should be the lowest
Once the surface temperature is known on the rig (­ x-­value possible value in accordance with the functional require-
in Fig. 4A), for a specified precharge pressure selected ment, in fact the trend of the VE curves (Fig. 4B) cor-
for the gas inside the accumulator, the precharge density responding to higher values of the last MOP tends to
D0 could be obtained (y-value in Fig. 4A), according to be quite vertical. Thus, considering that the slope of
a real gas state equation. Indeed, in Figure 4(A), each the Functional Volume Requirement (FVR) curves
line represents a specific precharge pressure. As shown strongly increases as the VE decreases the Volume
by the graph, when the pressure is kept constant, the Capacity (VC) (y-value of Fig. 4C) needed to perform
gas density decreases if the temperature increases. All the considered actuation becomes larger.
the curves require an equation of state for real proper- 4. For a given precharge pressure P0, gas density D0 at
ties of the gas (Nitrogen). The choice of precharge pres- precharge is slightly linear with respect to surface tem-
sure radically influences the accumulator performance perature; the linearity remains more or less constant
in terms of number of accumulators and fluid regardless to P0. Similar linearity characterizes gas density
deliverable. D0 at precharge with respect to volumetric efficiency;
Then, introducing the precharge density D0 in in this case, the linearity (slope) depends on minimum
Figure 4(B) as y-value, it is possible to define the Volumetric operating pressure (MOP3). As much as MOP3 increases,
Efficiency VE (x-value of Fig. 4B) for the MOP considered. the slope increases (as absolute values): this means that
So, inserting in Figure 4(C) the VE as x-value fixing the volumetric efficiency changes as much as MOP3 decreases,
whole FVR, the Volumetric Capacity (VC) needed to fulfill requiring in this case more precision in the original
all the actuations can be defined (y-value of Fig. 4C). surface temperature. At higher MOP3, the volumetric

6 © 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd.
R. Cipollone et al. On the Rapid Discharge of Subsea Accumulators

Figure 4. Design chart in a nomogram form (water depth = 3657.6 [m] (12,000 [ft]); HPU = 345.77 [bara] (5015 [psia]); subsea temperature = 1.67[°C],
([35°F]); Seawater density = 1025.088 [kg/m3], 8.556 [lb/US gal], Control Fluid Density = 997.934 [kg/m3], (8.33 [lb/US gal])). PHydro is equal to: 369.14
[bara], 5353.944 [psia]).

Table 1. Sequence of three functions test case. Table 2. Environmental and operating conditions.

Function MOPi FVRi [m3], FVR∗ci [m3], FVRci [m3], Pump start pressure 713.87 (10353.9) Bara (psia)
[bara], ([US gal]) ([US gal]) ([US gal]) (absolute)
([psia]) Gas volume per bottle 0.052 (13.8) m3 (US gals)
Water depth 3657.6 (12,000) m (ft)
1 551.6 0.189 (50) 0.189 (50) 0.208 (55) Surface temperature at 28.7 (83.7) °C (°F)
(8000) precharge
2 510.2 0.114 (30) 0.303 (80) 0.333 (88) Subsea (mudline) water 1.7 (35) °C (°F)
(7400) temperature
3 482.6 0.076 (20) 0.379 (100) 0.417 Sea water density 1025.2 (8.556) m3/kg (lb/US
(7000) (110) gal)
Sea water head pressure 369.1 (5353.9) bara (psia)

efficiency remains almost constant and as a function of

surface temperature and gas density at precharge condi- requirement decreases. Volume capacity and liquid stored
tion P0. in the accumulators change almost linearly: variation
5. Volumetric efficiency versus volume capacity is not linear, for the liquid stored is greater when gas density at pre-
for a specific overall functional volume required: when charge pressure P0 decreases.
volumetric efficiency increases, volume capacity decreases
as more as the functional volume requirement decreases:
Improvement on the normed design to
this produces a significant variation in the volume ca-
verify all the actuations
pacity at higher volumetric efficiencies: small surface
temperature variations can produce higher volume ca- After the design procedure, the regulation disposes to
pacity changes, as much as the functional volume check if the design of the accumulators allows to fulfill

© 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd. 7
On the Rapid Discharge of Subsea Accumulators R. Cipollone et al.

all the functions in terms of volume and pressure re- This verification phase has been implemented in a new
quirements. This verification is mandatory because in nomogram conceiving a sequence of steps represented in
the design procedure according to equations (5) and Figure 5. Specific values for the horizontal and vertical
(6), VE is evaluated according to the final MOP condi- axes and suitable parameters which characterize the curves
tion corresponding to the last function. For this reason, have been suitably chosen: they correspond to main vari-
the regulation does not consider the volumes of control ables of the discharge phases.
fluid required by the other functions at their respective Figure 5 has been produced considering the same design
MOPi. data reported in Tables 1 and 2. Verification is insured
The verification can be done according to a mathemati- if the actual pressure Pai of the propelling gas, after the
cal model of the discharging process based on the mass i-­discharge, is greater than the maximum value between
conservation of the propellant gas from the charged con- its respective Minimum Operating Pressure (MOPi) or
dition to the pressure inside the accumulators when the Hydrostatic Pressure PHydro (eq. 17).
i-­specific is done (Pai). The use of the nomogram in Figure 5 can be specified
When the charging process of the accumulators is com- as follows.
pleted, the volume of the propellant gas in the accumula- In Figure 5A, the input (x-­ value) is the Volumetric
tors is given by equation (11). If a generic function is Capacity (VC) of the accumulators (see y-­values of Fig. 4C).
correctly actuated, starting form charged condition and For a curve referred to a specific cumulated Functional
being the control fluid incompressible, the amount of Volume Required FVR*ci:
control fluid delivered, is equal to FVRci (eq. 13). i
∑
Vi = V1 + FVRci i = 1,2 … nf (13) FVR∗ci = FVRi for i = 1,2 … nf (18)
k=1

as the difference between the Vi (the volume occupied

evaluated without multiplying each FVRi by the Fv,
by the propellant gas in the accumulators at the actuation
rV can be calculated: it has been reported in the y-­axis
of the generic i-­th function) and the volume V1.
of Figure 5A.
The density of the propellant gas after the delivery of
Moving to Figure 5B and for a given D0 (y-­value of
a specific quantity of actuating fluid FVRci can be calcu-
Fig. 4A), Pai is evaluated and reported in the x-­axis. On
lated as follows:
the same x-­axis, Figure 5B reports the MOPi as reference
D0 values. If, for every discharge, the Pai fulfills the condition
Dai = ( ) (14) expressed by the equation (17), the accumulators allow
rV + DD0 to meet the functional requests.
1

The volumetric ratio rV is defined as follows: As it can be observed in Figure 5, all the pressure
levels respect the condition expressed by the equation (17)
FVRci ⋅ FV as reported in equation (19):
rV = (15)
VC Pa1 > max (MOP1 ,Phydro )
Being FVRci given by Pa2 > max (MOP2 ,Phydro ) (19)
i Pa3 > max (MOP3 ,Phydro )
∑
FVRci = FVRi ⋅ FV . (16) Therefore, the accumulators allow to deliver the volume
k=1
of control fluid required. If at least one of the condition
expressed in the equation (19) is not satisfied, the design
The already defined Pai value (pressure of the propellant
process must be repeated.
gas Pai at the thermodynamic state defined by Dai) can
The complete fulfillment of the actuations is further
be immediately evaluated considering that the entropy of
demonstrated in Figure 6A, where each Pai is reported
the propelling gas remains constant during expansions.
as a function of the precharge pressure. N2 is considered
If Pressure Pai is higher than the maximum value between
as propelling gas.
the Minimum Operating Pressure MOPi and the Hydrostatic
When the precharge pressure increases, Pai remains
Pressure PHydro, the accumulators allow to deliver a quantity
constant till an optimum value (for the precharge pres-
of control fluid equal to FVRci. Otherwise, the accumulators
sure) to which it corresponds a fully voided accumulator.
are unable to deliver the right quantity of actuating fluid.
It is evident that to this precharge value, a minimum
Therefore, the design meets the functional request if:
number of accumulators corresponds (Figure 6B). After
Pai > max (PMOPi ,PHydro ), for i = 1,2 … nf. (17) this precharge pressure, a further precharge pressure

8 © 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd.
R. Cipollone et al. On the Rapid Discharge of Subsea Accumulators

Figure 5. Design verification process in a nomogram form (Water depth = 3657.6 [m] (12,000 [ft]); HPU = 345.77 [bara] (5015 [psig]); subsea
temperature = 1.67[°C], ([35 °F]); seawater density = 1025.088 [kg/m3], 8.556 [lb/US gal], Control Fluid Density = 997.934 (8.33 [lb/US gal]).
PHydro = 369.14 [bara], (5353.944 [psia]).

increase produces a reduction in the volumetric efficiency, • initial precharge pressure variation of the propelling gas;
realizing an increase in the number of the accumulators • increase of the number of the accumulators
and a residual pressure after each discharge higher than
till to the fulfillment of the function(s) previously failed,
the minimum requested. Figure 6A shows the success of
without indicating a specific choice. In any case, the
the design.
final design is far from being optimized in terms of
number of accumulators which was the first goal of
the norm.
Remarks on the Design of the Bank of
The failure happens because the energy stored in the
Accumulators
propelling gas is unable to deliver the requested quantity
If the fulfillment of all the actuations is not reached, the of the actuating fluid above a specific minimum operat-
regulation leaves to the designer the following choices: ing pressure. When the lowest pressure of the last

© 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd. 9
On the Rapid Discharge of Subsea Accumulators R. Cipollone et al.

Figure 6. (A) Pressure inside the accumulators when FVRi are extracted as a function of the precharge pressure. (B) Number of minimum accumulators
required varying the gas precharge pressure value. The design has been done according to the norm.

function MOP3 is assumed to be inside the accumulators Figure 7A shows the Pai values after each discharge:
after the full discharge of the actuating fluid (as the it is evident that the second actuation is not fulfilled
norm assumes), this hypothesis could not insure the being the MOP2 greater that the Pa2. This happens also
fulfillment of the intermediate actuations. To overcome beyond the optimum precharge pressure value, as defined
this situations, if this pressure is assumed equal to the by the norm (660.76 [bara]-­9583.6 [psia]). Only if the
MOP1 (first actuation), all the actuations are fulfilled precharge pressure is greater than 676.7 [barg] (9815
but a severe overdesign is produced (increase of the [psia]), the second actuation is fulfilled but with a num-
number of accumulators). ber of accumulators (91) greater than the minimum one
In order to put in evidence this situation, a new set (82).
of data concerning the accumulator design is reported in For sake of completeness, if MOP1 is assumed after
Table 3. The sequence of BOP actuations of this second the full fluid discharge, the optimal gas precharge pressure
test case has the same functions requirements of the first would be 712.4 [bara] (10,332.5 [psia]) and the number
sequence reported in Table 1, except for the second ac-
tuation which has an higher MOP, while the operating Table 3. Functions requirements.
conditions are the same (Table 2).
In this new situation, the overall cumulative FVR is Function MOPi FVRi [m3], FVR∗ci [m3], FVRci [m3],
equal to 0.379 [m3] (100 [US gals]) and MOP (third [bara], ([US gal]) ([US gal]) ([US gal])
([psia])
action) is 482.6 bara (7000 psia); so, according to
the API 16D design procedure, the optimal precharge 1 551.6 0.189 (50) 0.189 (50) 0.208 (55)
pressure, the minimum MOPi, and the FVR for the (8000)
entire sequence of the functions do not change with 2 537.8 0.114 (30) 0.303 (80) 0.333 (88)
respect to the previous case (Table 1). Thus, the (7800)
3 482.6 0.076 (20) 0.379 (100) 0.417
number of accumulator prescribed by the regulation is
(7000) (110)
the same.

10 © 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd.
R. Cipollone et al. On the Rapid Discharge of Subsea Accumulators

Figure 7. (A) Pressure inside the accumulators when FVRi are extracted as a function of the precharge pressure; (B) number of minimum accumulators
required varying the gas precharge pressure value.

of the accumulators would be 123. In Figure 8, this design API 16D Method C is used. The procedure developed
is represented in terms of pressure levels inside the ac- shows that failure happens due to the insufficient pressure
cumulators after each discharge: it is evident the overdesign (energy) inside the propelling gas. In reality, for sake of
produced by the norm. completeness, after the design, the norm invites to verify
the fulfillment of all the actuations, but it does not assist
the designer on how to modify the design if an actuation
Conclusions
in not insured.
In this paper, the API 16D Method C design procedure This aspect represents a critical issue of the norm
for rapid discharge subsea accumulators was implemented which should be defined in a unique way and in all
in a software platform. A result was to provide an origi- parts in order to avoid uncertainties, difficulties when
nal nomogram which allows the sizing of the accumula- a comparison is made among different choices, intro-
tors in a graphical form immediately evidencing the role ducing a degree of freedom which should not be allowed
of all the operating parameters involved and the sensitivity when a standard is applied. In case of failure of one
of these variables in the final design. Due to this, the actuation, the norm invites the designer to: (a) produce
procedure developed makes the design easier (within the a new design modifying the precharge pressure and
respect for the API 16D), transforming the normed pro- repeating the normed method; (b) increase directly the
cedure into the application of a nomogram. number of accumulators keeping the same optimal pre-
Moreover, in order to verify if the design obtained charge pressure. In any case, the number of accumulators
ensures the fulfillment of the requirements of all the ac- is anymore optimized.
tuations, a verification nomogram has been developed. A It is opinion of the authors that a novel design pro-
result was to provide a second innovative nomogram cedure of the accumulators scientifically based is needed,
whose theoretical base has been a thermodynamic model so definitively improving the design and removing the
of the discharge of the accumulators. hypotheses usually assumed. The procedure should allow
The verification nomogram allows to understand the a physical representation of the discharge process relating
reason of an eventual failure which can happen if the fluid deliveries and pressure levels.

© 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd. 11
On the Rapid Discharge of Subsea Accumulators R. Cipollone et al.

Figure 8. (A) Pressure inside the accumulators when FVRi are extracted as a function of the precharge pressure: the MOP1 is assumed after the full
discharge; (B) number of accumulators required varying the gas precharge pressure value.

Acknowledgments D2 gas density at MOP condition [kg/m3], [lb/ft3]

Di gas density at intermediate MOPi condition [kg/
The authors are grateful to Saipem S.p.A. for having sup- m3], [lb/ft3]
ported a specific PhD program on drilling technologies. D3 gas density at total discharge condition [kg/m3],
Dr. Franco Terlizzese, General Manager of DSG UNIMIG 3
[lb/ft ]
(“Direzione Generale Per la Sicurezza anche Ambientale Dopt gas density at optimum discharge condition and
delle Attività Minerarie ed Energetiche, Ufficio Nazionale surface temperature [kg/m3], [lb/ft3]
Minerario per gli Idrocarburi e le Georisorse”), Italian EDS emergency disconnection sequenced system
Economic Development Ministry is particularly acknowl- Fp design factor at pressure limited condition
edged for financial funding. Fv design factor at volume limited condition
FVRi functional volume required for i-function [m3],
Conflict of Interest [US gals]
FVR functional volume required for the entire sequence
None declared. of the functions [m3], [US gals]
FVRci cumulated functional volume required with vol-
ume factor Fv
Nomenclature
FVR*ci cumulated functional volume required at i-th
function actuation without volume factor Fv
AG air gap [m], [ft] LMRP lower marine raiser package
D0 gas density at precharge condition [kg/m3], [lb/ MOPi minimum operating pressure for an intermediate
ft3] function, [bara], [psia]
D1 gas density at charged condition [kg/m3], [lb/ MOP lowest minimum operating pressure of the last
ft3] function [bara], [psia]

12 © 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd.
R. Cipollone et al. On the Rapid Discharge of Subsea Accumulators

m mass of inert gas stored in bottle after the pre- pneumatic strain energy accumulator. Int. J. Fluid
charge phase [kg], [lb] Power 18(3):167–180. https://doi.org/10.1080/14399776
nf number of function .2017.1335141.
na number of accumulators 4. Chen, J.-S. 2015. Energy efficiency comparison between
P0 gas precharge pressure (gauge) [barg], [psig] hydraulic hybrid and hybrid electric vehicles. Energies
P1 gas pressure at charged condition [bara], [psia] 8:4697–4723.
P3 pressure at total discharge condition [bara], [psia] 5. Fan, Y., A. Mu, and T. Ma. 2013. Modeling and
PHPU hydraulic power unit delivery pressure [barg], control of a hybrid wind-­tidal turbine with hydraulic
[psig] accumulator. Energy 112:188–199.
PHydro hydrostatic pressure [bara], [psia] 6. Puddu, P., and M. Paderi. 2013. Hydro-­pneumatic
Phydro,CF hydrostatic pressure of column of con- accumulators for vehicles kinetic energy storage:
trol fluid [bara], [psia] influence of gas compressibility and thermal losses on
POpt optimum precharge pressure [bara], [psia] storage capability. Energy 57:326–335.
Pai actual pressure of propellant gas in the accumu- 7. Pourmovahed, A. 1993. Sizing energy storage units for
lator after discharge FVRci [bara], [psia] hydraulic hybrid vehicle applications. Am. Soc. Mech.
rv volume ratio Eng. Dyn. Syst. Control Div. New Orleans, LA
52:231–246.
S1 entropy of the gas at charged condition [kJ/kgK],
8. Van de Ven, J. D. 2013. Constant pressure hydraulic
[BTU/lb°F]
energy storage through a variable area piston hydraulic
SAV single accumulator volume [m3], [US gals]
accumulator. Appl. Energy 105:262–270.
Tsub subsea temperature at wellhead [°C], [°F]
9. Han, C., X. Yang, J. Zhang, and X. Huan. 2015. Study
Tsurf Air temperature on the rig [°C], [°F]
of the damage and failure of the shear ram of the
VC accumulators volumetric capacity [m3], [US gals]
blowout preventer in the shearing process. Eng. Fail.
V0 gas volume at precharge condition [m3], [US
Anal. 58:83–95.
gals]
10. Han, C., and J. Zhang. 2013. Study on well hard
V1 gas volume at charged condition [m3], [US gals]
shut-­in experiment based on similarity principle and
Vi gas volume at MOPi condition [m3], [US gals]
erosion of ram rubber. Eng. Fail. Anal. 32:
V2 gas volume at MOP condition [m3], [US gals] 202–208.
V3 gas volume at total discharge condition [m3], 11. Jardine, S. I., A. B. Johnson, D. B. White, and W.
[US gals] Stibbs. 1993. Hard or soft shut-in: which is the best
VL1 liquid volume at charged condition [m3], [US gals] approach? Drilling Conference – Proceedings. Pp.
VLi liquid volume at MOPi condition [m3], [US gals] 359–370. Available at https://www.scopus.com/inward/
VE volumetric efficiency record.uri?eid=2-s2.0-0027261592&partnerID=40&md5=d
VEp volumetric efficiency at pressure limited cdac742aefdd4e219a5447bef092157
condition 12. Harlow, W. F., B. C. Brantley, and R. M. Harlow.
VEv volumetric efficiency at volume limited 2011. BP initial image repair strategies after the
condition deepwater horizon spill. Publ. Relat. Rev. 37:80–83.
VEi volumetric efficiency at generic i-condition 13. Skogdalen, J. E., I. B. Utne, and J. E. Vinnem. 2011.
VF volume factor Developing safety indicators for preventing offshore oil
WD water depth [m], [ft] and gas deepwater drilling blowouts. Saf. Sci.
49:1187–1199.
14. The Bureau of Ocean Energy Management Regulation
References
and Enforcement (BOEMRE). 2011. Report regarding
1. Fan, Y. J., A. L. Mu, and T. Ma. 2016. Study on the the causes of the April 20, 2010. The Bureau of Ocean
application of energy storage system in offshore wind Energy Management Regulation and Enforcement,
turbine with hydraulic transmission. Energy Convers. September 14, 2011.
Manag. 110:338e46. 15. API Specification 16D, Second Edition, July 2004.
2. Lin, T., Q. Chen, H. Ren, Y. Zhao, C. Miao, S. Fu Specification for Control System for Drilling Well
et al. 2017. Energy regeneration hydraulic system via a Control Equipment and Control Systems for diverter.
relief valve with energy regeneration unit. Appl. Sci. 16. API Standard 53, Fourth Edition, November 2012.
7:613. Blowout Prevention Equipment Systems for drilling well.
3. Cummins, J. J., S. Thomas, C. J. Nash, S. 17. Lavasani, S. M., N. Ramzali, F. Sabzalipour, and E.
Mahadevan, D. E. Adams, and E. J. Barth. 2017. Akyuz. 2015. Utilisation of Fuzzy Fault Tree Analysis
Experimental evaluation of the efficiency of a (FFTA) for quantified risk analysis of leakage in

© 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd. 13
On the Rapid Discharge of Subsea Accumulators R. Cipollone et al.

abandoned oil and natural-­gas wells. Ocean Eng. of Petroleum Engineers. https://doi.
108:729–737. org/10.2118/173167-ms
18. Dong, G., and P. Chen. 2017. A review of the 23. McCurdy, P. J. A. 2009. Developments in accumulator
evaluation methods and control technologies for trapped technology: A review of fluid power options in subsea
annular pressure in deepwater oil and gas wells. J. Nat. BOP control systems. Society of Petroleum Engineers.
Gas Sci. Eng. 37:85–105. https://doi.org/10.2118/118415-ms
19. Chung, S., S. Kim, and Y. Yang. 2016. Use of 24. Amani, M., M. Mir-Rajabi, H. C. Juvkam-Wold, and J.
hazardous event frequency to evaluate safety integrity J. Schubert. 2006. Possible alternatives for gas-charged
level of subsea blowout preventer. Int. J. Nav. Arch. accumulators in deep water. Society of Petroleum
Ocean Eng. 8:262–276. Engineers. https://doi.org/10.2118/100305-ms
20. Shanks, E. F., W. Pfeifer, S. Savage, and A. Jain. 2012. 25. Zhao, X., Z. Qiu, Y. Zhang, H. Zhong, W. Huang, and
Enhanced subsea safety critical systems. Offshore Z. Tang. 2017. Zwitterionic polymer P(AM-­DMC-­
Technology Conference. https://doi.org/10.4043/23480-ms. AMPS) as a low-­molecular-­weight encapsulator in
21. Good, C. A., and J. P. McAdams. 2001. Mathematical deepwater drilling fluid. Appl. Sci. 7:594.
prediction and experimental verification of deep water 26. Russo, M., A. Zakeri, S. Kuzmichev, G. Grytøyr, E.
accumulator capacity. Offshore Technology Conference. Clukey, and E. B. Kebadze. 2016. Integrity assessment of
https://doi.org/10.4043/13234-ms. offshore subsea wells: evaluation of wellhead finite element
22. Cole, E. H. (Ted). 2015. Deadman/autoshear: Managing model against monitoring data using different soil models.
precharge pressure and temperature uncertainty. Society J. Offshore Mech. Arct. Eng. 138:061301.

14 © 2018 The Authors. Energy Science & Engineering published by the Society of Chemical Industry and John Wiley & Sons Ltd.