SPE-188598-MS

Implementation of Dispersed Gas Model for Kick Tolerance Analysis of an

HPHT Exploration Well in Norway

K. Karahasan and J. S. Oyovwevotu, Maersk Oil; W. E. Szemat Vielma, Schlumberger

Copyright 2017, Society of Petroleum Engineers

This paper was prepared for presentation at the Abu Dhabi International Petroleum Exhibition & Conference held in Abu Dhabi, UAE, 13-16 November 2017.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Minimum requirements for kick tolerance (KT) in Maersk Oil corporate standards drove implementation
of analysis using a dispersed model instead of a single-bubble model to achieve more realistic KT values
for a planned high pressure and high temperature (HPHT) exploration well with oil-based drilling fluids.
The large uncertainty bands of predicted pore and fracture pressures in an HPHT exploration well offshore
Norway had driven the casing design, which had very narrow margins and little flexibility in setting depths
for the casing strings. KT analysis using a single-bubble model indicated allowable kick volumes of less
than the required 50bbl minimum. To comply fully with the 50bbl KT, additional casing strings would be
required, with several liners and/or unconventional casing sizes, adding further complexity, cost, and risk
to the well and the operations. The development of a well control bridging document, including HPHT
procedures, competency assurance of the crew, and use of real-time pore pressure prediction service was
considered inadequate to fully manage the associated well control risk.
The benefit of accepting the less than 50bbl KT would be to enable drilling of the HPHT exploration
well with the planned five-string casing design, which was already more robust than in most of the offset
wells, which utilized a four-string design. With the five-casing-string approach, there is little room for a
contingency casing/liner string without possible compromise of some of the well objectives.
An internal dispensation process was initiated due to the noncompliance with Maersk Oil corporate
standards. To secure dispensation from such corporate standards, transient drilling modeling software was
used to calculate updated KT on the basis of the dispersed kick (gas) model. In addition, a plan was put in
place to update the well model and recalculate KT using real-time well information.
This case history presents the story of KT analysis of the subject well, including transient drilling
modeling to investigate the sensitivity of input parameters. Furthermore, a description of the software's use
to assure safe drilling operations is discussed. Lessons learnt from use of this software during planning and
well construction will also be discussed.

Introduction
The subject exploration well was drilled on a high pressure and high temperature (HPHT) prospect situated
on the Haltenbanken area of the Norwegian Sea (Fig. 1) in an established Jurassic play. The objective was to
2 SPE-188598-MS

prove the geological concept in the block by drilling a representative section of five potentially prospective
reservoirs that, according to the geological model, could be bearing dry gas or condensate.

Figure 1—Exploration block area.

In line with the established pore prediction prognosis methodology on exploration wells, three different
cases, respectively low, most likely, and high, were generated. This process involved an analysis of offset
wells, which are in similar geological settings and target the same play, using both drilling data and log-based
analysis. Furthermore, we performed an analysis of the likelihood of connectivity to offsets at any depth.
As in the other wells on the Haltenbanken area, the water depth of the wellsite required the use of a
semisubmersible rig, which in this case was a fifth-generation floater.

The Case
The casing design and shoe depth selection for the subject well was driven by the challenging pore pressure/
fracture gradient (PP/FG) profile as well as the wellbore stability issues observed in offset wells (Fig. 2).
Quite differently from most North Sea HPHT plays, the pressure profile was characterized by two distinct
pressure ramps; the first one was before the 13 5/8-in. casing planned setting depth, and the second was
before the 9 7/8-in. production casing planned setting depth at 4000m.
SPE-188598-MS 3

Figure 2—Well architecture and PP/FG.

In particular, the setting depth of the 16-in. liner was driven by both a minimum and maximum value
of the fracture gradient. A minimum fracture gradient was required to enable a manageable kick tolerance
(KT) for the most likely case in the 14 ½-in. × 17 ½-in. section, and a maximum fracture gradient was
in fact driven by the burst rating of the 20-in. casing for the load case full displacement to gas, which is
mandatory for exploration wells.
On this non-HPHT section of the well, a very narrow mud window was expected at the bottom of the
first pressure ramp. Consequently, the aim was to set the 13 5/8-in. shoe at a sufficient depth to cover
potential hydrocarbon-bearing and/or unstable formations, thereby maximizing the probability of the 12 ¼-
in. bottomhole assembly (BHA) reaching planned section total depth (TD).
In fact, on the same section of one of the offset wells, another operator had experienced three packoff
events before taking a severe kick toward the section TD. After a succession of wrong judgment calls during
the kick control event, the situation escalated beyond control, with loss of circulation; the BHA was left
in hole; and the section was plugged back. As a consequence, a 16-in. contingency liner was set above the
4 SPE-188598-MS

high-pressure zone. On the basis of this and other offset well experience, the 16-in. liner was part of the
base case plan for the subject HPHT exploration well.
The objective of the 12 ¼-in. section, on the other hand, was to cover formations that were troublesome
from a geomechanics perspective in order to set the 9 7/8-in. casing shoe above the first potential reservoir
and, thereby, enabled the reservoir to be drilled in an 8 ½-in. hole. A very narrow mud weight window was
expected to be available near the planned TD of the 12 ¼-in. section, where particular caution would be
required to minimize any potential influx volume.

Conventional Kick Tolerance Approach

At a very fundamental level, KT is defined as the maximum volume of influx which can be shut in and
circulated out of the well without breaking down the weakest point of the well exposed to the circulation
event. Quite often, but not always, that weakest point will be the casing shoe.
At the design stage, the selection of a safe casing seat is, in part, determined by the calculation of KT
for a gas kick. It can therefore be said that the KT is one of the driving factors of the well architecture, in
conjunction with geological constraints with respect to competence of the formations, for setting the casing
shoe.
The intensity of the kick is arguably one of the poorly defined and understood points of the KT calculation.
The kick intensity (KI) used in the calculations is, broadly speaking, supposed to reflect the degree of
confidence in the wellbore pressure prognosis. In the usual application, KI is added as an additional pressure,
in terms of pounds per gallon (ppg) or specific gravity (SG), to the prognosed pressure. Commonly, the
values used in development or exploration settings range between 0ppg and 0.5ppg (0.06SG).
KT should be calculated at the design stage and then evaluated following a leakoff test (LOT) in
accordance with the proposed mud weight program. Clearly, any change in either the estimated formation
pore pressure or the mud weight will impact the KT and should be taken into account.
For the calculation of conventional single-bubble KT, two scenarios must be considered; the final KT
volume is the smaller quantity of the two scenarios:
1. When the influx is at the hole bottom under the initial shut-in condition
2. When the influx top is displaced to the openhole weakpoint with the original mud weight.
With a relatively short openhole section, the calculated volume may extend from bit depth to above the
weakpoint (generally the casing shoe), which implies that the well can tolerate an unlimited volume of kick
without breaking down the weakpoint of the formation. If this occurs in a high-angle or horizontal hole
section where the potential kick volume can be high, it is important to check the maximum allowable gas
volume for the relevant load cases on the casing, the pressure ratings of the blowout preventer (BOP) stack
and choke manifold, and the gas handling capacity of the mud-gas separator (MGS) at a realistic well kill
rate.
The conventional single-bubble KT calculation proceeds as per the following seven steps:
1. Estimate the safety margin to be applied to the leakoff pressure at the openhole weakpoint. This should
include a suitable safety margin to account for (amongst other factors)

◦ Choke operator error

◦ Choke line friction (especially important in subsea applications)

◦ Annulus friction.
2. Calculate the maximum allowable static weakpoint pressure. The maximum allowable static
weakpoint pressure is given by
SPE-188598-MS 5

(1)
 where
Pmax = maximum allowable static weakpoint pressure [psi]
Plo = leakoff pressure at openhole weakpoint [psi]
SF = safety factor [psi].
3. Calculate the maximum allowable height of influx in the openhole section. The maximum height of
influx that can be taken without exceeding Pmax is given by

(2)

 where
H = height of influx [ft]
Pmax = maximum allowable static weakpoint pressure [psi]
MW = mud weight in the hole [ppg]
GG = gas gradient [psi/ft]
Dbit = bit depth [TVD ft]
Pf = formation pressure at depth considered [psi]
Dwp = depth of openhole weakpoint [TVD ft].
It is important to realize that the single-bubble theory has been in use since before the advent of
software with pressure, volume and temperature (PVT) correction capabilities. For this reason, the
gas gradient is considered constant and independent of the gas/wellbore pressure.
As established in the definition of KI, the formation pressure at calculation depth is an assumed
pressure:

(3)

 where
KI = chosen kick intensity [ppg]
Dbit = bit depth [TVD ft].
4. Calculate the volume V1 that corresponds to this height (as calculated in Step 2) at initial shut-in
condition:
(4)
 where
V1 = KT for initial shut-in [bbl]
C1 = annular capacity at initial shut-in condition [bbl/ft].
5. Calculate the volume Vwp that corresponds to the situation when the influx has been circulated to the
weakpoint:
(5)
 where
Vwp = volume of initial influx at weakpoint [bbl]
C2 = annular capacity below weakpoint [bbl/ft].
6. Calculate V2, which is what the weakpoint volume (as calculated in Step 5) would be at initial shut-
in. Using Boyle's law to convert Vwp to the volume at shut-in, we obtain
6 SPE-188598-MS

(6)

 where
V2 = Vwp at initial shut-in
Pf = pore pressure at calculation point
Pmax = maximum allowable weakpoint pressure.
7. The KT will be the lower quantity of V1 and V2.

Shortcomings of Conventional Kick Tolerance Approach

In a single-bubble model, the kick is considered to have entered the wellbore as a single, continuous phase,
displacing the drilling fluid, and with a clean influx (gas) and mud interface. The single-bubble model
does not account for influx distribution, gas dissolution in the drilling fluid, and multiphase behavior based
on pressure and temperature of the influx and the mud system (Szemat Vielma and Mosti 2014). These
simplified assumptions can lead to an erroneous calculated KT volume, with significant implications for
safe well operations.
This conventional approach has two main shortcomings. First of all, the assumption of a gas bubble
entering the wellbore in such an "instantaneous" fashion and displacing the drilling fluid entirely from a
whole section of annular space is not in line with diffusion mechanisms, as per the fluid dynamics governing
the reservoir-wellbore interface. As such, this approach does not reflect reality and disregards such important
parameters as the solubility of the gas in the mud, depending on the choice of an oil-based or water-based
system; PVT properties of the influx effluent; and reservoir delivery parameters, such as permeability and
porosity. Time-dependent evolution of pit gains and pressures at crucial observation points, such as the last
casing shoe or the section TD, and any sensitivity on choke operator errors and blowout preventer (BOP)
closure times cannot be simulated in this approach.
The second shortcoming is directly related to the consequences of having an unrealistic density profile of
the fluid composition from influx depth to the openhole weakpoint. Because the single-bubble gas density is
considerably lower compared to the density of the drilling fluid containing dispersed-gas bubbles, a single-
bubble model leads to a higher pressure at the casing shoe because the influx pressure is transmitted to
the weakpoint after a small reduction by the low hydrostatic pressure of the gas bubble, and then a rather
moderate reduction occurs, depending on the height of the remaining mud column before the weakpoint.
In other words, the single bubble is considered conservative versus the dispersed bubble, and this is the
major reason why it has been maintained regardless that is not always correct (Morrell et al 2017). The
comparative evolution profiles of both models can be observed in Fig. 3.
SPE-188598-MS 7

Figure 3—Comparative pressure evolution profiles of both KT models.

However, as explained in the preceding text, the kick will diffuse into the drilling fluid, producing reduced
mud density as a result of the mixture of drilling fluid and dispersed influx. This more accurate modeling
of fluid interaction and behavior on fluid density and wellbore pressures will result in lower casing shoe
pressure in comparison to the single-bubble approach.
Tables 1 and 2 summarize the KT calculated with a single-bubble model for a fixed mud weight at the
section TD of 4000m for the subject HPHT exploration well. The PP-FG most likely case result with 0.5ppg
KI is circled in blue (corresponding to operator requirement). Kick volume > 50bbl is highlighted in green
(complies with standard). Kick volume > 15bbl while < 50bbl is highlighted in amber (manageable). Kick
volume < 15bbl is highlighted in red (stop drilling).

Table 1—12 ¼-in. KT (bbl) with 1.86SG mud.

8 SPE-188598-MS

Table 2—12 ¼-in. KT (bbl) with 1.89SG mud.

If most likely case PP-FG conditions are encountered, it can be seen that only 5.7bbl of KT with 0.5ppg
(0.06SG) intensity is available if the planned mud weight of 1.86SG is used, whereas the operator standard
requires 50bbl. Furthermore, there is no KT if the high case PPs or lower FGs are encountered.
With the use of early-kick-detection systems and other appropriate risk-reduction measures, it would
be reasonable to accept a KT lower than 50bbl and dispensate correspondingly, but arguably a volume
below 10bbl becomes too small to be considered as within the as-low-as-reasonably-practicable (ALARP)
region. In any case, a KT of 5.7bbl was outside the risk-appetite boundaries of the operator for this HPHT
exploration well.
Facing this situation, an option is to increase the mud weight; increasing the mud weight will reduce the
amount of potential underbalance, thereby increasing available KT volume. This is demonstrated in Table
2 by increasing planned mud weight from 1.86SG to 1.89SG. For the most likely case, the underbalance is
thereby reduced from 0.06SG to 0.02SG and KT increases from 5.7bbl to12bbl.
However, mud weight definition, like any other well engineering decision, should adopt a holistic
approach that considers all relevant factors, rather than being an exercise in number crunching. Mud weight
increase for the sole purpose of improving theoretical KT values increases the risk of inducing losses. In
fact, on the subject well, even the use of a 1.86SG mud weight resulted first in heavy wellbore breathing
and then losses during the 9 7/8-in. casing run, which then led to considerable difficulties during plug and
abandon (P&A) operations. Therefore, it seems reasonable to assume that if a heavier mud weight had been
used for the sole purpose of improving a theoretical KT, these losses would have been even more difficult
to manage and the consequences for P&A would be even more dramatic.

Dispersed-Bubble Model Background

Transient simulation models are important for design of drilling operations and for decision support to ensure
selecting the best possible option in a critical operational situation. Including these models will ensure better
understanding of the ongoing processes, while having high accuracy in the results. In addition, it will be
possible to evaluate different operational strategies and understand how these will impact the downhole
well conditions. This will allow the drilling engineer to plan a diverse set of strategies to operate the well
without exceeding the available drilling window.
The physical problem is complex and is normally solved by the use of detailed numerical models. The
core of the solution is the conservation of mass and momentum. The well is then divided into a finite system
of grid boxes, from the inlet to the outlet, that are representative of the actual geometry of the well.
The overall conservation of mass can be expressed as

(7)

 where
SPE-188598-MS 9

ρ = density
ν = velocity.
The conservation of momentum can be expressed as

(8)

 where
P = pressure
Fw = wall friction
 g = gravitational constant
ρ = density
ν = velocity.
An accurate description of the interaction between the different phases in a kick scenario requires separate
consideration of mud mass, mass of free gas, mass of dissolved gas, and mass of formation oil. This
constitutes a total of five governing equations. To close the system, a set of accurate submodels is needed
to describe mud density, influx density, liquid PVT and phase properties, rate of gas dissolution, free gas
velocity, non-Newtonian rheology, two-phase friction losses, temperature conditions, and the interaction
with the formation.
The boundary conditions are defined, depending on the operational conditions of the well control
scenario. Of particular importance is the transient behavior of reservoir fluid properties, influx distribution
and interaction with the drilling fluid during the dynamic development of a kick event, and the subsequent
well control operations. This is treated on a compositional level using equation of state for accurate modeling
of the mud/reservoir fluid behavior, with changing temperature and pressure conditions during the kick
circulation. The composition of the reservoir fluid and the drilling fluid is considered and will give a realistic
representation for influx types ranging from heavy oil to gas.
The simulation model used in this case and presented in this paper has been developed over the past 30
years. Different submodels have been refined, improved, and verified by extensive laboratory work and
field testing. Particular care has been given to typical, narrow-margin conditions (e.g., HPHT and deepwater
wells), as these are extreme compared to several of the oversimplified assumptions used in standard models.
This model has performed very well in comparison with historical data for real wells.

Dispersed-Bubble Kick Tolerance Approach

By using the dispersed-bubble KT approach on the subject well, the full set of conditions and scenarios in a
kick control situation could be simulated in the transient modeling software, allowing meaningful sensitivity
analyses to be carried out, among others with respect to kick volumes, circulating rates, pit gain, crew
reaction time, formation production potential, and choke operator error.
Using the same parameters as in the most likely case figures that led to 5.7bbl on the single-bubble model
(Table 1), a realistic KT value of 15.4bbl was obtained on the dispersed-bubble model. Furthermore, as
opposed to the numerical result of a spreadsheet, this KT value indeed took into account the rig-specific
equipment peculiarities and the crew reaction times that had been observed during the drills, as the value
was set to represent the pit alarm level. As a consequence, the real volume of gas allowed to enter the
wellbore was even higher in terms of final pit gain than 15.4bbl, as can be observed on Fig. 4.
10 SPE-188598-MS

Figure 4—Simulation of the real volume of gas allowed to enter the wellbore.

The evolution of the pressure at the casing shoe during the circulation of the kick shows that the kick can
be circulated out safely, as the pressure profile stays below the leakoff test value. It can be observed that the
stabilized pressure at the casing shoe is about 7bar (100psi) below the shoe fracture pressure as per chosen
choke operator error, with the resulting dynamic safety margin peak recognizable and stopping just below
the leakoff pressure, which can be observed in Fig. 5. It is important to highlight that the dynamic safety
margin is set as per company-specific policies (artificial value) and does not change the well physics, but
ensures the limits are constrained and involved variables are controlled.
SPE-188598-MS 11

Figure 5—Simulation result of downhole pressure profile with respect to the casing shoe.

A particularly useful capability of the transient well control software for subsea and, especially, deepwater
operations is the simulation of the dissolved gas in mass fraction versus the free gas in volume fraction of
the mud. Fig. 6 shows that the dissolved gas in the mud breaks out closer to the surface when the wellbore
pressure reaches the bubblepoint pressure. The safety implications of an undetected influx that mixes with
drilling fluid and is circulated up the well in a system open to the atmosphere is that gas breaks out close to
surface with little reaction time to close-in the well. As a consequence, the marine riser may be unloaded
and, due to the resulting underbalance, further kicks can enter the wellbore.

Figure 6—Simulation of the dissolved gas vs. the free gas, both in volume fraction of the mud.
12 SPE-188598-MS

Further Risk-Reduction Measures

A risk management approach that focused on people, equipment, and process was devised to support the
process of waiver from corporate standards and was implanted during the well construction phase. These
measures were related to both human and technical factors to ensure that there was the required level of
vigilance, competence, and technical means on the rig and within the onshore organization to deal with a
reduced KT and to have the capacity to manage a well control incident by shutting in the well on a kick
below the required maximum size and thereafter circulating it out in a safe manner without breaking at the
formation's weakpoint and leading to an escalated situation.
Human factors were addressed to strengthen risk recognition and management capability through the
following actions:

• The drilling crew on the rig went through specific HPHT training prior to starting operations.

• This training was simulator-based, mimicking exactly the conditions anticipated on this well.

• Narrow margin drilling/equivalent circulation density (ECD) management training was carried out
in a separate session for the rig crew.
• Wellsite and onshore drilling engineers were trained in the use of the transient software and
constantly updated the model on the basis of real-time well data to carry out further KT analysis
relevant to actual well conditions.
• Kick drills and choke drills were conducted at regular intervals on the rig.

• Risks at their associated, anticipated geological depths were clearly identified to the entire team
onshore during the drilling-well-on-paper and prespud meetings onboard the rig.
• A well control bridging document, including HPHT procedures, was prepared and approved by
the operator and the drilling contractor. The objective of this document was to ensure that a single
policy and operating philosophy was followed during operations with respect to proper drilling
and tripping practices and well control procedures.
• All relevant personnel held valid International Well Control Forum (IWCF) certifications.

Technical factors were addressed through the following measures:

• Real-time PP prediction was performed by a specialist service company during the drilling
operations and results were compared with the predicted PP-FG values.
• A mud weight selection and communication protocol was established and strictly adhered to while
drilling the critical sections.
• Well-specific, narrow-margin drilling procedures were established as per operator's HPHT drilling
operations guideline.
• A 24/7 well operations parameter monitoring system from a third-party provider was used as an
additional early warning system for kicks and/or well stability issues. This real-time evaluation
system, together with real-time PP prediction, contributed to an informed decision-making process
during operations.
• In the intervals where risks of kick have been identified (offsets and geological prediction),
operations were conducted by controlling the drilling and tripping parameters (i.e., limiting rate of
penetration and controlling tripping speed to limit swab and surge effects). Swab calculations were
carried out using the transient modeling software to prevent swabbed kick events while tripping
out of hole.
• The well design included a contingency plan in case of significant deviations from expected PPs
and FGs, taking into consideration scenarios likely to result in drilling the reservoir partly or
entirely as a 6-in. hole.
SPE-188598-MS 13

• Well-specific leakoff test procedures were developed with input from the geomechanics team and
were included in the program.

Conclusions
• Use of the dispersed-bubble model with a transient multiphase simulator for KT calculations
provides more realistic assessment of KT for HPHT well planning and operations.
• Early recognition of well control risks, supported with the use of a transient simulator to understand
the interrelationship between actual downhole parameters and their boundaries, was crucial to the
success of planning and operations of this well with narrow-margin and low KT.
• Development of a coherent plan, addressing technical and human factors, allowed the team
to manage the drilling of this HPHT well without any well control incident during the well
construction phase.

Acknowledgments
The authors would like to thank the Maersk Oil Corporate Wells team and its leadership for the support
during the planning and well construction stages of the well. We are especially grateful to the management
of Maersk Oil for the permission to publish this paper and are grateful for the support provided by
Schlumberger.

References
Isambourg, P., Anfinsen, B.T. and Marken, C. 1996. Volumetric Behavior of Drilling Muds a High Pressure and High
Temperatures. SPE-36830-MS. http://dx.doi.org/10.2118/36830-MS.
Morrell, D., Davis, M. et al. 2017. Reduce Risk and Save Cost by Better Understanding of the Physics of Well Control.
Presented at the AADE National Technical Conference and Exhibition, Houston, USA, 11–12 April. AADE-17-
NTCE-036. http://www.aade.org/app/download/7246175889/AADE-17-NTCE-036+-+Morrell.pdf.
Mosti, I., Morrell, D., Anfinsen, B. T. et al. 2017. Kick Tolerance and Frictional Pressure Losses, Added Safety or Added
Risk? Presented at the SPE/IADC Drilling Conference and Exhibition, The Hague, The Netherlands, 14–16 March.
SPE-184655-MS. http://dx.doi.org/10.2118/184655-MS.
Rezmer-Cooper, I.M., James, J., Fitzgerald, P. et al. 1996. Complex Well Control Events Accurately Represented by an
Advanced Kick Simulator. Presented at the European Petroleum Conference, Milan, Italy, 22–24 October. SPE-36829-
MS. http://dx.doi.org/10.2118/36829-MS.
Rommetveit, R., Blyberg, A., and Lie Olsen, T. 1989. The Effects of Operating Conditions, Reservoir Characteristics
and Control Methods on Gas Kicks in Oil Based Drilling Muds. Presented at Offshore Europe, Aberdeen, UK, 5–8
September. SPE-19246-MS. http://dx.doi.org/10.2118/19246-MS.
Rommetveit, R. 1992. Proper use of PVT data improves well control during kicks in oil based drilling fluids.
Szemat Vielma, W. E. and Mosti, I. 2014, Dynamic Modelling for Well Design, Increasing Operational Margins in
Challenging Fields. Presented at the Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, UAE,
10–13 November. SPE-172016-MS. http://dx.doi.org/10.2118/172016-MS.
Umar, L., Thiam, Y. Y. B A., Murad, M S. et al. 2014. Dynamic Multiphase Kick Tolerance Allows Safe Drilling Which
Led to Huge Gas Discovery in a HPHT Exploration Well in Malaysia. Presented at the International Petroleum
Technology Conference, Doha, Qatar, 19–22 January. IPTC-17460-MS. http://dx.doi.org/10.2523/IPTC-17460-MS.