Sand Management

RESTRICTED-COMMERCIAL
SWR-G-255, AEAT-RHAD/RD01762/12/1 v3
Shearwater Sand
Management Manual
A report produced by AEA Technology for

Shell UK Exploration & Production
T. J. Lockett, J. Krol, J. Lilley
September 2000
RESTRICTED - COMMERCIAL SWR-G-255, AEAT-RHAD/RD01762/12/1 v3
Title Shearwater Sand Management Manual
Customer Shell UK Exploration & Production
Customer reference SWR-G-255
Confidentiality, Restricted - Commercial

copyright and
reproduction This document has been prepared by AEA Technology plc in
connection with a contract to supply goods and/or services and is
submitted only on the basis of strict confidentiality. The contents
must not be disclosed to third parties other than in accordance with
the terms of the contract.
File reference AEAT-RHAD/RD01762/12/1
Report number AEAT-RHAD/RD01762/12/1
Report status version 3 (final)
E1 Culham
Abingdon
Oxfordshire
OX14 3ED
Telephone +44 (0)1235 464088
Facsimile +44 (0)1235 463799
AEA Technology is the trading name of AEA Technology plc

AEA Technology is certificated to BS EN ISO9001:(1994)
Name Signature Date
Author T. J. Lockett Timothy Lockett 28/9/2000
Reviewed by J. Krol James Krol 28/9/2000
Approved by G. Eckold Geoff Eckold 28/9/2000
RESTRICTED - COMMERCIAL AEA Technology 2

Executive Summary
These guidelines and procedures were produced by AEA Technology working on behalf of
Shell UK exploration & Production’s Shearwater Team under contract C/49475/97/SC2:
2300019267.
Sand production is one of the biggest unknown quantities in exploiting the Shearwater field. In
generating and adopting these guidelines and procedures, the Shearwater production team
recognises the additional risk that occurs when sand is produced. Through good design,
selection of monitoring equipment, appropriate inspection, training and staff awareness, the risk
from sand production can be minimised.
Following these guidelines and procedures will ensure that:
• the Shearwater platform follows industry best practice in managing sand

• best use is made of the sand detector system that has been installed
• problems caused by sand are identified early and diagnosed rapidly
• uptime is maximised
• the Shearwater platform is maintained as a safe working environment (with regard to sand)
• the practical implementation of sand management can be demonstrated as/when required
(e.g. to the HSE)
Take a moment to look through the over-view to identify how sand may impact upon your
area of responsibility.

Contents
1 Over-view 6
1.1 RESPONSIBILITY 6
1.2 WHY MANAGE SAND ON SHEARWATER? 7
1.3 CONTROLLING SAND PRODUCTION 7
1.3.1 Looking for sand . . . 7
1.3.2 Deciding . . . the acceptable level 8
1.3.3 Acting to reduce or eliminate sand, and its effects 8
1.4 THE PROCESS FLOW PATH FOR SAND 9
1.5 SUMMARY 11
1.6 WHERE TO GET MORE HELP AND ADVICE 12
2 Sand detectors 13
2.1 BACKGROUND: OPERATING PRINCIPLES 13
2.2 ALARMS 13
2.2.1 Sand detector Alarm Example 1 15
2.2.2 Sand detector Alarm Example 2 16
2.3 OFFSHORE SIGNAL INTERPRETATION 17
2.3.1 Spurious signals and alarms 17
2.3.2 Reducing sand production 17
2.3.3 Which sand detector? 18
2.4 ONSHORE DATA REVIEW 20
2.4.1 How does it calculate sand flowrate? 20
2.4.2 Do the alarm set-points need to be changed? 23
2.4.3 Is the flow velocity appropriate? 24
2.4.4 Are the set values of ZERO appropriate? 26
2.4.5 Should additional maintenance / inspection be scheduled? 27
2.5 PERIODIC FUNCTIONALITY CHECKS 28
2.6 INSTALLATION CHECKS 29
2.6.1 Attachment 29
3 Minimising sand production and its effects 31

3.1 THE CAUSES OF SAND PRODUCTION 31
3.2 OPTIONS FOR DOWNHOLE EXCLUSION OF SAND 32
3.3 WELLHEAD REMOVAL OF SAND 33
4 Other indicators of sand production 34

4.1 CHOKE VALVES 34
4.2 EROSION OF FLOWLINES 35
4.3 SEPARATORS 35

4.3.1 Accumulation of sand in the test and first stage separators 35

4.3.2 Flushing the separators 36
4.3.3 Level indicator in the first stage separator (V1010) 37
4.3.4 Level indicator in the test separator (V1110) 37
4.4 LEVEL CONTROL VALVES 38
5 Wall thickness surveys 39

5.1 INTRODUCTION 39
5.1.1 Definitions and context 39
5.1.2 Risk-based Inspection Programme 39
5.1.3 Applicable inspection techniques 41
5.1.4 Practical aspects of ultrasonic inspection 42
5.2 IDENTIFICTAION OF VULNERABLE COMPONENTS 43
5.3 INSPECTION OF CHOKE VALVES 43
5.3.1 Inspection of the gallery region 43
5.3.2 Inspection of the outlet region 45
5.3.3 Choke valve inspection frequency 45
5.4 INPSECTION OF THE FLOWLINES IMMEDIATELY DOWNSTREAM OF THE
CHOKE VALVES 45
5.4.1 Pentland flowline – target tees 46
5.4.2 Pentland flowline – straight pipe downstream of the choke 47
5.5 INSPECTION OF THE REMAINDER OF THE FLOWLINES 47
5.5.1 90 degree bends 47
5.5.2 Target tees 48
5.6 INSPECTION OF THE PRODUCTION MANIFOLD AND TEST MANIFOLD 49
5.6.1 Production Manifold 49
5.6.2 Test Manifold 50
5.7 INSPECTION OF THE PRODUCTION HEADER AND TEST HEADER 50
5.7.1 Production Header 50
5.7.2 Test Header 51
5.8 INSPECTION IN THE EVENT THAT SAND IS DETECTED 52
5.9 INSPECTION IN THE EVENT THAT EROSION DAMAGE IS EVIDENT 52
5.10 SUMMARY OF THE INSPECTION SCHEME 53
6 Procedures 58
7 References 59
8 Index 60
Appendices
APPENDIX 1 SHEARWATER SAND MANAGEMENT PROCEDURES
APPENDIX 2 NON-INTRUSIVE INSPECTION TECHNIQUES

1 Over-view
1.1 RESPONSIBILITY
Sand presents a serious risk to the safe and efficient operation of the Shearwater platform. The
ultimate responsibility for sand management therefore rests with the Shearwater Asset Manager.
The process of sand management is implemented through the personnel (and equipment) that
together produce the Shearwater field. The following personnel and responsibilities are
specifically identified:
Personnel Responsibility with regard to sand production Relevant

procedures *
Asset Manager Ensuring that sand production is managed in a Compliance checklist
manner acceptable to the Shearwater co-venturers,
consistent with Health Safety and Environmental
legislation, and industry best practice.
OIM Maintaining a broad regard for the effects of sand Procedures 6, 7, 8, 9
production on the control and separation systems and 11
across the platform. Directing activities outwith
normal process operations.
Control room Acting on sand detector alarm conditions. Procedure 2
operator Observing the sand detector trend during well Procedure 4
bean-up.
Offshore Periodic check of sand detector system Procedure 1
instrument functionality.
technician Working with control room operators to find the Procedure 5
maximum sand-free rate for a well (if sand is
evident at the intended production rate).
Well Monthly (or more frequent as requested) review of
Programming sand detector output
(onshore)
Production Establishing new values of coefficients, and alarm Procedure 3
Technology thresholds.
(onshore)
Identifying unusual behaviour and seeking advice.
Contractors Installation (and maintenance if needed) of the sand
detector system (ClampOn).
Interpretation of unusual sand detector trends or

readings (AEA Technology, ClampOn).
Pipework and choke valve inspection (OIS). Procedure 10

Choke valve maintenance, repair or replacement.
* Procedures are included in Appendix 1.

1.2 WHY MANAGE SAND ON SHEARWATER?

On Shearwater, sand production could lead to:
• process upset (inadequate separation, and poor level control)

• loss of adequate well control (choke valve damage)
• loss of mechanical access to the well (liner buckling)
• loss of a well (sanding up)
• loss of containment due to erosion
Other fields are exposed to the same threats, but the high pressure of the Shearwater reservoir
means that the risks are greater. So, as far as is reasonably practicable, sand production cannot be
permitted to occur unchecked or unknown.
Management is the action of controlling a process or event to achieve an objective.
In this case, the objective is to deliver the target production from the Shearwater field without
undue risk to personnel, equipment, or the environment caused by sand. This document
describes the equipment and actions which will be used to control the process of sand
production in meeting this objective.
1.3 CONTROLLING SAND PRODUCTION

The key steps in controlling sand production are:
• Looking: identifying that sand is being produced

• Deciding: deciding whether or not the rate of sand production is acceptable
• Acting: taking action to reduce or eliminate sand production or its effects
1.3.1 Looking for sand . . .
Shearwater is equipped with sand detectors supplied by ClampOn. They are in use on other
Shell UK fields (e.g. Brent), are used by Shell in other locations (e.g. Gulf of Mexico) and by
other operators. As part of the selection process they were independently tested under
conditions representing those in the Shearwater flowlines. Their strengths and weaknesses are
largely known, and are discussed in section 2.
However, it should be understood that:
• the detectors work by listening for sand impacts with the inside wall of the pipework
(flowline or header) at a bend point
• detecting low levels of sand against the noise of a multiphase flow is not easy
• sand at a low flowrate can pass the detector without being indicated

• alarm levels will inevitably be set at a somewhat higher level, so even if sand is detected it
may not be sufficient to trigger an alarm.
So, while the sand detectors are the primary indicators of sand production, other indications of
(low levels of) sand production are equally important, such as:
• nucleonic sand build-up sensors will be fitted to the first and second stage separators, and to
the test separator
• choke valve trim erosion, will result in a gradual closing of the choke at fixed tubing head
flowing pressure (THFP), providing that choke erosion occurs more rapidly than the natural
decline in THFP as the reservoir is depleted.
1.3.2 Deciding . . . the acceptable level
As stated in section 1.1, the person ultimately responsible for sand production is the Shearwater
Asset Manager. The Asset Manager therefore defines the acceptable level of sand production for
the Shearwater field. At the present time, it has been defined to be:
‘Zero detectable sand’
If sand is detected, then the agreed action to eliminate it must be taken.
By definition, sand production that cannot be detected cannot be acted upon. The Asset
Manager is satisfied that the low level of sand production that might occur undetected poses an
acceptable risk.
Depending on the economic and technical risks (subsurface, safety, control, disposal), the Asset
Manager may establish a higher acceptable sand production rate in the future.
1.3.3 Acting to reduce or eliminate sand, and its effects
Sand production is caused by the reservoir in the near well-bore region breaking down into
grains. This process depends on the force that is applied to the rock. Sand production is
promoted by:
• Shocking the well (abrupt change in operation).
• Flowing at a high production rate (greater draw-down).
• Increasing water cut (dissolution of soluble salts).
• Increasing ability of the flow to transport sand (water has a higher density than gas or oil)
Conversely, sand production may be reduced or eliminated by:

• Bringing a well on in a gradual manner.
• Reducing the flowrate from a well.
• Fitting sand screens. During design it was established that down-hole sand screens can only
be fitted economically once the reservoir pressure declines to a maximum of 10000 psi.

The effects of sand on the topsides process may be reduced or eliminated by:
• Good design and materials selection (already implemented).
• Verifying continued fitness for purpose through inspection.
• Removing the sand early in the separation system. Sand will normally accumulate in the test
and first stage separators, where it can be periodically removed. However, sand may also
proceed through the produced water system (see Section 1.4).
• Removing sand at the wellhead. Weight provision (and space via a cantilever) has been
identified for fitting wellhead desanders if required (although none have yet been
designed/constructed for use on a 15000 psi system). Implementation of wellhead removal is
considered to be very unlikely since this option does not address all of the sand production
concerns (e.g. subsurface).
1.4 THE PROCESS FLOW PATH FOR SAND

The path taken by sand through the process is illustrated below:
V1110 Gas
Test separator V1010
st
Gas 1 Stage
Condensate to
Sand wash second stage separator
separator
Sand wash
PW level
Manual sand control
flush
Interconnecting Condensate to
Manual sand flush
bridge second stage
separator
Produced water to
clean-up package
Well
(1 of 6) ZA-4010:
HP & LP
hydrocyclones,
degasser
Produced water discharge
(to Casson V4010)

The path of the sand is as follows:
• up the well
• though the X-mas tree
• through the choke
• along the flowline and into the manifold (production or test)
• through the production or test header
• into the first stage or test separator
• where it will accumulate until:
• being flushed by manual operation
• or, if there is a significant produced water flowrate, sand will exit with the produced water
stream, and pass into the water cleanup system, before being discharged with the produced
water to the marine environment. The level control valve on the first stage separator
produced water outlet (part of the water cleanup package), and also the produced water
degasser level control valve, are specifically put at risk from this routing. Some sand may also
accumulate in the produced water degasser.
• or, if there is insufficient produced water and the separator is not flushed, sand will
eventually accumulate to such an extent that it is carried over the weir and into the
condensate stream to the second stage separator …. where it will accumulate until being
removed. The condensate level control valves are specifically put at risk by passing sand.
The progression of damage through the separation system can be minimised by producing as
little sand as possible to surface (through good well control), and by removing the sand from the
process as soon as possible (first stage or test separator).
If sand production comes from only one well, disruption to production from the platform as a
whole may be minimised by routing this well to the test separator.
Flushing the separator is a relatively involved process which includes setting up a temporary
desanding loop drawing water+sand from the separator, extracting the sand, and returning the
produced water. It is envisaged that this process can be conducted while the separator remains in
use.

1.5 SUMMARY
The threats posed to Shearwater by sand production, and the equipment and processes that have
been provided are summarised in the following two illustrations:
Flowlines and headers

experience wall loss due to
erosion (periodic inspection)
Separators
accumulate Choke valve erosion:
sand - periodic Trim -> loss of control
removal Body -> loss of containment
(inspection, repair/replacement)
Level control valves

wear - poor level
control (low level
trips)
Potential liner
buckling and loss of
access to the well
Disposal to sea (clean-
up of sand from oil
contamination)
The effects of sand production on Shearwater
Flowlines and headers

designed to tolerate sand
production (verified through
periodic inspection)
Sand detectors fitted to

flowlines and headers
Separators
fitted with Choke valves:
nucleonic sand Best available materials selection
level detectors Periodic inspection of the body
Trim damage gauged from
change in pressure drop
Provision for
periodic separator
flushing
Provision for fitting Gentle bean-up to

wellhead desanders minimise shocks to the
reservoir, establish
maximum sand-free rate
Provision for downhole
exclusion, once reservoir
pressure falls below 10000 psi
Managing sand production on Shearwater

1.6 WHERE TO GET MORE HELP AND ADVICE
Contact details Background

ClampOn mail@clampon.no Supplier and installer of the sand detector system.
00 47 55 94 88 50
AEA Tim Lockett Author of these guidelines and procedures.
Technology tim.lockett@aeat.co.uk AEA Technology conducted the erosion
01235 464088 assessment during design of Shearwater, and ran
the flow loop trials used during the selection of
the sand detection system.
Other Shell UK Brent: Brent now have over 100 ClampOn sand
Expro assets Graham Wilson detectors installed and operational experience
since 1997.
SIEP Shell Oil: During 1999 Shell Oil conducted over 100 sand
Lillian Skogsberg injection trials using ClampOn detectors on 19
platforms in the Gulf of Mexico

2 Sand detectors
2.1 BACKGROUND: OPERATING PRINCIPLES

The ClampOn sand detectors installed on Shearwater are non-intrusive acoustic devices, which
detect the noise made by sand grains as they hit the pipe wall at a bend point in the system. The
device is mounted on the outer wall of the pipe immediately downstream of a bend point, and
the noise made by the sand is detected and converted to an electrical signal. The signal is
conveyed to a computer in the control room via cables, where custom software converts the
signal into sand flowrate (grammes/second, g/s). The data is accessible through the Platform
Management System (PCS) and is echoed onshore through RTMS. Data may be displayed
graphically to highlight sand production trends, and the system is set to raise alarms if sand
production exceeds set limits. RTMS is also being configured to maintain a running total of the
sand production recorded by each detector.
A diagram of a typical system is illustrated below:
Alarms
Sand output (g/s)
Platform
PCS
Raw signal
Flow velocity
(Choke opening,
Wellhead flowing
pressure, temperature)
ClampOn PC
The usefulness of the system depends entirely on being able to interpret the output, either
manually or through automatic analysis, and raising alarms.
2.2 ALARMS
The action to be taken in the event of an alarm is given in Procedure 2. In addition to raising an
alarm if the unit considers itself to be faulty, two sand detection alarms are configured for the
ClampOn system on Shearwater:
• Alarm is raised if the instantaneous sand flowrate exceeds the set level, L1(g/s)
• Alarm is raised if the sand quantity in an individual batch exceeds the set level, M(kg)
When either of the alarms is activated or cancelled, an entry is made in the log file (held on the
ClampOn system computer). Both of these alarms (and the malfunction alarm) are mimicked to
the control room console and latch until cancelled by an operator.

Since it reflects an instantaneous condition, the L1 alarm is prone to being triggered by changes
in flowrate (choke valve positioning). Under these circumstances, if the L1 alarm is reset it will
not be repeatedly triggered. However, if the L1 alarm is triggered by production of sand, it is
highly likely that it will be repeatedly triggered once reset.
The M alarm depends on the mass of sand that has been observed by the detector over a time
interval defined by a starting and a stopping condition. The system starts to calculate the
quantity of sand whenever the instantaneous sand flowrate exceeds the L2 level, and is reset to
zero if the instantaneous sand flowrate falls below the level set as L3. L2 and L3 are set values
only and are not alarms in themselves. A mass alarm is only raised if the quantity recorded
exceeds the set level before the reset condition occurs. The M alarm is not prone to being
triggered by changes in flowing condition.
The following alarm levels have been configured on Shearwater:
Alarm Configuration on Shearwater

L1 Alarm Triggered if the instantaneous sand flowrate exceeds 3 g/s
M Alarm Triggered if the mass of sand exceeds 1 kg.
Counting is started if the instantaneous sand flowrate exceeds, L2, 0.5 g/s.
The counter is reset if the instantaneous sand flowrate falls below, L3, 0.1 g/s.
In view of the configuration of these alarms it is quite possible that:
• an L1 alarm may be triggered without sand being present (usually directly associated with
changes in flowrate)
• the L1 alarm may not be triggered when sand is present (at low levels)
• the M alarm may not be triggered when sand is present intermittently
These characteristics lead to the conclusions that . . .
The output from the sand detector should be periodically reviewed onshore to check for sand
production that was recorded, but was not sufficient to trigger an alarm condition.
Two examples are given in the sections below.

2.2.1 Sand detector Alarm Example 1
This example is taken from the flow loop tests used as part of the selection of the sand detector.
During this test 23 kg of sand passed the detector in two continuous batches, each lasting
approximately 3 hours. The recorded sand flowrate is shown on the left hand axis while the
condition of the alarm is shown on the right. This example is to illustrate behaviour, and the
alarm settings differ from those set on Shearwater.
Sand detector alarms - example 1

2.5 3
L1 set at 1 g/s
L2 set at 0.2 g/s
L3 set at 0.05 g/s
2
M set at 10 kg
Alarm status indicator

'M' alarm
Sand flow rate / (g/s)
active 2
1.5
1
'L1' alarm 1
active
0.5
0 0
08:00
09:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
Time
During this example, sand production began at 09:00. The L1 alarm was triggered repeatedly
from 09:00 until 09:30 indicating that the sand flowrate was exceeding the set level of 1 g/s.
The L1 alarm was then triggered intermittently until 10:10 as the sand flowrate intermittently
exceeded the set level. No further L1 alarm was triggered until 12:30 when the sand flowrate
again exceeded the set level of 1 g/s. After 15:00 the sand flowrate had fallen below the set level
of 1 g/s and the L1 alarm was not triggered again.
It will be appreciated that the L1 alarm would not have been triggered during the first period of
sand production if the level had been set to 1.5 g/s. Furthermore, the L1 alarm would not have
been raised at all if the level had been set to 2.5 g/s.
In contrast, the first time that the M1 alarm was triggered was at 14:05, once 10 kg of sand had
been recorded during the second period of sand production in the day. The M alarm was not
triggered during the period 09:00 – 12:00 because although the quantity of sand was being
monitored once the sand flowrate exceeded the L2 level of 0.2 g/s, the quantity of sand
recorded did not exceed the set level of 10 kg. At 12:00 the sand flowrate fell below the L3 set
level of 0.05 g/s, and consequently the mass of sand recorded up until that time was reset to
zero.

2.2.2 Sand detector Alarm Example 2
This example is to illustrate behaviour, and the alarm settings differ from those set on
Shearwater. The sand detector output is shown below for a period of 1 month:
Example 2: Sand detector output and alarm set-points

7
L1 = 5 g/s
6 L2 = 2.5 g/s
Sand flowrate / (g/s)
L3 = 1 g/s
5
M = 10 kg
4
2
1
0
25-Dec-99
30-Dec-99
4-Jan-00
9-Jan-00
3-Feb-00
14-Jan-00
19-Jan-00
24-Jan-00
29-Jan-00
The alarm status which corresponds to this trace is shown below:
Example 2: Alarm status

7 3
L1 = 5 g/s
6 L2 = 2.5 g/s
L3 = 1 g/s M1 alarm
5
Alarm status
M = 10 kg 2
4
3 L1 alarm
1
2
0 0
4-Jan-00
9-Jan-00
25-Dec-99
30-Dec-99
14-Jan-00
19-Jan-00
24-Jan-00
29-Jan-00
3-Feb-00
In this example, the sand is produced is discrete bursts. The L1 alarm set point has been set at 5
g/s to avoid being triggered by changes in flowrate. During the period shown, the L1 alarm is
triggered only on 1 occasion. In contrast the M alarm is triggered repeatedly, and is active for
10% of the elapsed time. Nevertheless, periods of sand production are clearly visible which were
not sufficient to trigger either alarm.

2.3 OFFSHORE SIGNAL INTERPRETATION

The action to be taken in the event of an alarm is given in Procedure 2.
The output trace from the relevant sand detector(s) should be periodically observed by the
control room operator, over a period of 1 – 6 hours, under the following circumstances:
• if a sand detector alarm is activated

• if a well is being brought on for the first time
• if a well is being brought on after some form of work-over
• if a well is being brought on after having been closed in for a period in excess of 1 week
• during a well-test
The key questions that must be answered offshore are:
• Does the signal suggest that sand is being produced? or is the signal spurious?
• If sand is being produced, how can it be minimised? Or eliminated?
These two issues are considered in the following two sections.
2.3.1 Spurious signals and alarms
If the signal has both of the following characteristics, it is a spurious signal and no action need be
taken:
• High sand detector level which only occurs when the flowrate is changed (choke position)
• Sand detector signal which occurs for a short duration (e.g. <2 minutes)
A spurious signal is unlikely to lead to a Mass alarm being triggered, but may lead to an L1 alarm
being triggered (see section 2.2). If an L1 alarm genuinely indicates a high flowrate of sand, then
it will repeatedly re-activate once it is cancelled.
2.3.2 Reducing sand production
Sand production can be minimised, and may be eliminated, by:
• Treating the well more gently (opening and closing the choke more gradually)
• Minimising the number of changes to well flowrate
• Reducing the flowrate from the well (finding the maximum sand-free production rate)
Procedure 4 addresses minimising sand production by gradual bean-up of a well.
An example of using the sand detector to determine the maximum sand-free flowrate is shown
overleaf.

Example of sand production, which stabilises with time, triggered by

changing the choke position
1.2 40
Delayed response corresponds to
bottoms-up time (sand-face to
35
1
Choke opening position / %

surface)
30
Sand flowrate / (g/s )
0.8
25
Functionality check -
0.6 see Procedure 1 20
15
0.4
10
0.2
5
0 0
11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00
Time
The choke was opened in a series of steps, and in this example, sand was produced to surface
following each increase in choke opening position. The operator judged that the sand decayed
away to an acceptable level at choke settings of less than 30%. At 35% the sand signal was judged
to be too strong, although it may have decayed away given more time.
Procedure 5 should be followed to find the maximum sand free production rate.
2.3.3 Which sand detector?
The sand detectors and their alarms have the following tag numbers:
Well, Flowline Sand Sand High sand Mass Alarm Malfunction

detector flowrate signal, L1 (M alarm) Alarm
(P&ID) (g/s) alarm
Well 1, flowline 1 XT-01111 XIA01111 XAH01111 XQAH01111 XMA01111
Pentland, flowline 7 XT-01171 XIA01171 XAH01171 XQAH01171 XMA01171
Production Header XT-01201 XIA01201 XAH01201 XQAH01201 XMA01201
Test Header XT-11001 XIA11001 XAH11001 XQAH11001 XMA11001
Spare N/A XIA01181 XAH01181 XQAH01181 XMA01181

The flow velocity at the sand detector is tag number FY-nnnn and is useful for checking that the
well is currently flowing, and for assessing whether changes have been made to the well flowrate
in the recent past.
The flow from any well passes two detectors, one on the flowline and one on the header (either
test header or production header depending on routing). Since detection is influenced by
flowrate and pipe diameter, different results will be obtained from the two detectors. Action
should be taken if either detector indicates that sand is present at a rate higher than is tolerable
(see section 1.3.2).
A trend may be displayed using the Tag numbers listed above. A suitable trend group for a well
routed to the test separator would be XIA01111, FY01111, XIA11001, FY11001. The state of
the sand detector is summarised on the wellhead graphic for each well, illustrated below.
The relevant area is entitled “SAND PRODUCTION” (top centre of the screen).The current
value of the sand flowrate is displayed (x g/s). The instantaneous sand flowrate alarm (L1 alarm)
results in the words “HIGH RATE” being displayed. The mass alarms (M alarm) results in the
words “HIGH ACCUMULATION” being displayed.
0.01 7.32

2.4 ONSHORE DATA REVIEW

Onshore data review should be conducted on a routine basis (recommended monthly) and also
under the following circumstances:
• Following the start-up of a new well

• Following a well test
• Following a work-over of a well
• Following re-start of a well which has been shut-in for more than 1 week
The onshore data review is used to consider the following:

1. Has sand been produced and detected, but was not sufficient to raise an alarm condition? If
so, do the alarm set points need to be changed?
2. Is the flow velocity that is supplied to the sand detector still appropriate?
3. Do the calibration coefficients used by the ClampOn software need updating?
4. Does additional inspection or maintenance need to be scheduled?
Before examining these in more detail, it is important to appreciate that the sand flowrate (g/s)
that is the primary output from the unit has already been processed according to calibration
coefficients that are set in the software. This process is discussed in section 2.4.1.
2.4.1 How does it calculate sand flowrate?
The raw output from each detector is automatically interpreted as a sand flowrate (g/s) by:
• Deducting the signal that occurs when sand is not present in the flow (ZERO), to obtain
the excess signal
• Dividing the excess signal by a scaling factor (STEP), to obtain sand flowrate (g/s)
It should be appreciated that:
• No sand will be indicated until the raw signal exceeds the set value of ZERO. If the set
value of ZERO is too high, then sand that is present in the flow will not be indicated. If the
set value of ZERO is too low, some background noise from the flow will be interpreted as
sand, which may result in unnecessary alarms being raised.
• A low value of STEP will result in the excess signal being reported as a large sand flowrate.
A high value of STEP will result in the excess signal being reported as a small sand flowrate.
• The values of both ZERO and STEP are functions of flow velocity

The following table of ZERO and STEP are recommended for use on Shearwater at start-up:
Gas + liquid velocity at ZERO STEP STEP STEP

flowline conditions (m/s) (Flowlines) (Production (Test
header) header)
0 2000 130* 14* 66*
1 2200 130 14 47
2 2400 330 38 120
5 3100 1170 130 430
10 4700 3060 350 1130
15 7150 5350 600 2000
20 11000 7960 900 2900
25 16800 10800 1200 4000
30 25700 13900 1600 5100
35 39300 17200 2000 6400
40 60100 20700 2400 7600
45 92000 24400 2800 9000
50 141000 28200 3200 10400
(equivalent to 147 MMscf/day
at wellhead conditions)
* included in order to prevent large sand flowrates being indicated when no flow is present.
Choosing appropriate values of ZERO is discussed in section 2.4.4. Advice concerning the
setting of ZERO and STEP levels may be sought from ClampOn or AEA Technology (see
section 1.6 for contact details). These values are held in the ‘Flow Table’ within ClampOn’s
Multitrend software [ref 5] running on the ClampOn computer. This screen is password
protected to prevent values being changed inadvertently. Procedure 3 describes how to change
these coefficients.
These values are consistent with the flow loop work conducted as part of the selection process
for the sand detectors, but may need to be changed (Procedure 3). The following guidelines are
given regarding changing these values:
• The ZERO values are known to be influenced by factors specific to the flowline
configuration and gas-liquid ratio. These values should therefore be reviewed following
start-up of the platform, and at each well-test. Section 2.4.4 discusses setting alternative
values for ZERO.
• The STEP values used on Shearwater can only be updated in the light of new data or new
understanding of the response of ClampOn sand detectors to sand under conditions
representative of the Shearwater flowlines. This can only be achieved from sand injection
work conducted either on Shearwater itself (not advised while the wellhead pressure is
high), sand production experience on Shearwater (not envisaged), or work conducted on
other fields which have similar gas and liquid flowing velocities, or on flow loops. Current
understanding indicates that increasing the pipe diameter reduces STEP at a given velocity.
Hence different values are given for the detectors on the flowlines, from those on the test
and production headers. If sand is produced it will be possible to review the consistency of
quantities given by the detectors on the flowline, the test header and the production header.

The ZERO and STEP values tabulated above are based on the results of the flow loop tests
conducted during the selection of the sand detectors for Shearwater, which were conducted at
gas-liquid velocities representative of the Shearwater flowlines [ref 1]. These values are also
broadly consistent with the results of sand injection work conducted by Shell Oil in the Gulf of
Mexico, and by other operators on other fields [ref 2], as illustrated in the following two figures:
Effect of velocity on background flow noise

100000
GLR 1-20
90000 m3/m3
80000
70000 Shell GoMex
(GLR = 80 -
60000 130 m3/m3)
Unusually
ZERO
50000 high level of

background Flow loops
40000 noise (GLR 10-20
m3/m3)
30000
20000 Exponential
based on
10000 flow loops
0
0 10 20 30 40 50
Mixture velocity (Gas+Liquid) / (m/s)
Effect of velocity on STEP value

1000000
120-150 micron
STEP (Shearwater flowlines)
100000
Shell 120 micron
Pow er fit to
10000 flow loop data
Flow loop 120

micron
1000
Flow loop 275
micron
100
0 10 20 30 40 50
Gas + Liquid velocity / (m/s)

2.4.2 Do the alarm set-points need to be changed?
As described in section 2.2, it is possible that sand may be produced but may be insufficient to
raise an alarm condition. Conversely, sand may also be indicated when the flowrates are
changed, even though none is present. These conditions may be diagnosed by comparing the
sand flowrate record with the alarm set points, and taking account of the flow velocity record.
Two contrasting examples are given below.
Example 1:
Example sand detector record 1

10 30
Sand detector
9
FIow velocity 25
Flow velocity / (m/s)

8
7
20
6
5 15
4
10
3
2
5
1
0 0
1-Jan-00
6-Jan-00
11-Jan-00
16-Jan-00
21-Jan-00
26-Jan-00
31-Jan-00
In the example shown above, the sand detector output is solely associated with bringing a well
back on after a period of shut-down. This trace is characteristic of a well which does not
produce sand.
The alarm levels should be set such that behaviour of this type does NOT trigger an alarm. In
this example, the L1 alarm was set at a value of 5 g/s to avoid being triggered by this behaviour.
Procedure 3 describes how to change alarm set points.
Example 2:
In the example given overleaf, a burst of sand is produced from the reservoir whenever the
choke is opened. The sand then decays away quite rapidly, within an hour or so. Sustained sand
production appears to be triggered by opening the choke to 35%. In view of the requirement
to highlight sand production as soon as it is detected on Shearwater (see section 1.3.2), the alarm
levels should be set such that sand production of this type triggers an alarm.

Example 2: Sand detector response

1.2 40
35
1
Choke opening position %

30
0.8
25
0.6 20
15
0.4
decay 10
0.2
5
0 0
09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00
Time
When considering whether or not sand has been produced, refer also to section 2.4.4.
2.4.3 Is the flow velocity appropriate?
The ClampOn software uses the flow velocity to determine the values of ZERO and STEP that
should be used. If the velocity that is supplied is not appropriate then the resulting sand detector
output will be unreliable. The velocity that is supplied to the sand detector is stored as FYnnnn .
For each sand detector on the flowlines, the flow velocity is calculated from the flowing tubing
head pressure (FTHP) using a well performance curve for the well. For the detectors on the test
header and production header, the flow velocity is calculated from the individual well flowrates
according to the position of the routing valves.
The only time that the supplied velocity may be checked is during a well test. The velocity of
the flow in the flowline at the sand detector should be calculated from the flowrates recorded
during the well test, and compared with the velocity passed to the ClampOn system at that
time. Once the system has been configured correctly, discrepancies should merely reflect the
changing deliverability of the wells (particularly increases in water cut).
The graph overleaf shows the well performance curves (FTHP vs Wellhead flowrate) which was
implemented at the start of production. Also shown is the equivalent flow velocity curve for the
sand detectors in the individual well flowlines (internal pipe diameter 139 mm).

Initial well performance curves Initial flow velocity curves
900 900
800 800
700 SW01 SW01
700
SW07
FTHP / barg
600 600 SW07
FTHP / barg
500 SW05 500 SW05
400 SW08 400 SW08
300
SW04 300 SW04
200
SW09 200
SW09
100
100
0
0 50 100 150 200 0
0 10 20 30 40 50
Well flowrate (wet gas at wellhead) /
(MMscf/day) Flowline velocity / (m/s at 90 barg , 138o C)
The flow velocity, V (tag FY-nnnn) for each well is calculated as follows:
 π D2 

V = Q  3600 
 4 
where: Q is the volumetric flowrate of gas in the flowline in Actual m3/hr (tag FX-nnnn), and D
is the internal diameter of the section of line on which the sand detector is mounted, in m.
Q is determined by PCS according to the flowing tubing head pressure using a GENLIN block
look-up table whose values were supplied by (and will be updated by) Well Programming (refs
FX011n5). Q is supplied to the ClampOn system, V is calculated within the ClampOn system,
and V is then returned to PCS along with the sand flowrate.
Q is related to the gas flowrate at wellhead conditions as follows:
Qwh . 1000000  ρ STD β 

Q= α + 
35.3 . 24  ρ FL 1000 
where: Qwh is the wet gas flowrate at the wellhead in MMscf/day, ρSTD is the gas density at
standard conditions (15°C, 1 bara), ρFL is the gas density at flowline conditions (90 barg, 138°C),
α is the gas yield factor in Sm3 of gas at flowline conditions per Sm3 of gas at wellhead conditions, and β
is the liquid yield factor in m3 of liquid per 1000Sm3 of gas at wellhead conditions.

2.4.4 Are the set values of ZERO appropriate?
As described in section 2.4.1, if the value of ZERO is set too high, then this may mask sand
production or other information that is useful. In contrast, if the value of ZERO is set too low,
then sand will be indicated unnecessarily, which may cause the Mass alarm to be triggered. As
discussed in section 2.4.1, the correct value of ZERO is a function of flow velocity.
Until it is made available within RTMS, the raw signal (i.e. unprocessed) may only be obtained
from data downloaded directly from the ClampOn PC. A number of important characteristics
can be determined from the raw signal:
Aspect Inference
Sand is unlikely to be produced at a constant level.
1 How stable is the signal over
Substantial variations are often seen over a time frame of 1
a time frame of 1-6 hours at
hour as the near well-bore region collapses and is re-
constant flowing conditions?
stabilised. Repeated collapse is often associated with
changes in production rate (as the reservoir is shocked). A
constant level is more likely to be noise from the
multiphase flow and/or the choke valve
Depletion of the reservoir, increased draw-down, and
2 How does the signal today
water production are factors which increase the likelihood
compare with the signal last
of sand being produced. The level of background noise
week, month, or year?
may be influenced by the pressure drop across the choke
valve.
The near well-bore region may shed some sand following
3 If the signal increases
a change in flowing condition, before being re-stabilised.
following a change to the
Shed sand will be carried to the surface in the multiphase
flowing conditions, does the
flow. The time delay between opening the well up and
increase reflect the
transportation process? Does seeing sand at the surface should reflect the velocity up the
the signal then decay with tubing and the length of tubing.
time, or remain at the new
level?
It is difficult to envisage a situation where this behaviour
4 High signal at low
genuinely reflects the sand flowrate. Such behaviour is
production rate. Low signal at
more likely to be caused by noise generated by the choke.
high production rate.
At low production rate, the pressure drop is high, while at
high production rate, the pressure drop will be lower.
ClampOn should be asked to survey other locations to
determine whether another bend or a target tee can be
used more successfully.
The value of ZERO that is set should:
• allow unwanted noise to be masked out (typically 1.1 times the minimum signal seen in a
period of stable flow and stable signal)
• allow genuine sand production events to be seen
• be broadly consistent with the values given in the table in section 2.4.1

• be broadly consistent across the seven sand detectors on Shearwater

• allow a consistent view to be obtained between the sand detector on the flowline and the
sand detector on the test header
Advice concerning the setting of Zero levels may be sought from ClampOn or AEA
Technology (see section 1.6 for contact details).
2.4.5 Should additional maintenance / inspection be scheduled?
The topsides equipment is relatively robust, but will become damaged if substantial quantities of
sand are produced. Section 5.8 requires that additional inspection work is undertaken if 500 kg of
sand is produced since the last inspection.
In order to instigate this work, it is essential that the results of this onshore review (conducted
within the remit of Well Programming) are passed to the maintenance / inspection team
(including the integrity management subcontractor) on a regular basis (e.g. monthly).
The ClampOn sand detector system does not maintain a running total of the sand which has
been recorded by a detector. This function will therefore be provided by RTMS.

2.5 PERIODIC FUNCTIONALITY CHECKS

The operation of the unit may be checked as follows:
• Offshore: The unit will give a substantial increase in signal when the pipe wall is rubbed
fairly vigorously with a wire brush in the vicinity of the detector (Procedure 1). For best
results, the wire brushing should be applied for a minimum of 1 minute as the output is
usually averaged over a time interval of 1 minute before being stored or logged. Since this
process results in a high output signal being generated, the L1 sand detection alarm may be
triggered – providing a further check that the system will function as required if/when sand
is detected.
• Onshore: When flowing, the output may be compared with a historical trace at a similar
Flowing Tubing Head Pressure. If the unit continuously indicates that there is no sand in
the flow (i.e. records a value of 0 g/s) then the raw signal must be used for this check. Raw
signal is discussed in section 2.4.4.
• Onshore: In the absence of flow, the unit gives out a constant raw signal of approximately
2000. Raw signal is discussed in section 2.4.4.
Until it is made available within RTMS, the raw signal (i.e. unprocessed) may only be obtained
from data downloaded directly from the ClampOn PC.
If the unit does not give the expected signal (i.e. the noise level is substantially lower than
expected at a given flow velocity), it may need to be removed and re-attached to the pipe.
Alternatively, the electronics in the unit may have failed (rare). Refer to section 2.6.
Later versions of the ClampOn sensors (DSP unit) are also able to conduct a self-check to verify
adequate attachment to the pipe wall. The units installed on Shearwater are ClampOn P2000
units and do not have this capability. Replacement units will be of the DSP specification since
the P2000 is no longer supplied. An upgrade to the software will be needed to enable this
feature of the new sensors.

2.6 INSTALLATION CHECKS

The sensing head is relatively robust, but contains electronics and should be handled
accordingly. It must be installed on the downstream side of a bend point (bend, elbow, tee) in
the line, and located on the outside of the corner. Examples of correct and incorrect placement
are shown below:
Choke
The proximity of other sources of ultrasonic noise, such as the choke valve, can have an adverse
effect on detection since it increases the background noise level against which the noise from
the sand impacts is set. Placing the detector close to a choke valve (upstream or downstream) is
therefore not recommended. Gas duty, and a large pressure drop across the choke promote the
generation of noise.
On the Shearwater flowlines, the sand detectors are located on the downstream side of
5-D 90° bend furthest from the choke valve.
Shearwater also has sand detectors located on the production header (second 1.5-D 90° bend off
the manifold) and test header (first 1.5-D 90° bend off the manifold).
2.6.1 Attachment
The sand detector sensing head is attached to the pipe by screwing it into a base unit which has
been clamped to the pipe using metal straps. Once installed, the base unit should not need to be
removed. The sensing head may need to be replaced if the electronics fail, or may need to be
removed and re-attached if the signal becomes weak/unreliable. The most likely cause of an
unusual decline in signal is poor contact. A silicone grease is used to ensure good ultrasonic

contact between the sensing head and the surface of the pipe. The effectiveness of the grease
may decline with age, particularly at high temperature.
Although the units are only low voltage power, the power to the sensing head should be
disconnected before it is removed or attached to the pipe. This is to protect the sensor itself
from the effects of rough treatment and to protect the electronics from very large signal spikes
and to prevent alarms being raised unnecessarily.
Refer to the ClampOn sand detector user guide [ref 5] for the process of removal and re-
attachment of a sensing head.

3 Minimising sand production and

its effects
3.1 THE CAUSES OF SAND PRODUCTION

Sand production is caused by the reservoir in the near well-bore region breaking down into
grains. This process depends on the forces that are applied to the reservoir (static and dynamic).
The major forces stem from the weight of the rock above, and the draw-down (that is the
difference between the closed in reservoir pressure and the bottom hole pressure when flowing).
Drag by the fluid provides another force on individual sand grains. Sand grains may also be held
together by precipitated salts which are stable while the well is dry, but dissolve if water breaks
through.
When the flowing conditions are changed, a transient pressure pulse occurs as the flow
accelerates or decelerates.
Sand production is therefore promoted by:
• Shocking the well (abrupt change in operation).

• Flowing at a high production rate (greater draw-down, more drag).
• Increasing water cut (dissolution of soluble salts).
• Increasing ability of the flow to transport sand (water has a higher density than gas or oil)
Conversely, sand production may be reduced or eliminated by using the wells differently:
• Bringing a well on in a gradual manner (See Procedure 4).

• Reducing the flowrate from a well.
• For some wells, a maximum sand-free flowrate can be determined (see Procedure 5).
The process of bringing a well on in a step-wise manner to find the maximum sand-free
condition is illustrated in section 2.4.2 (example 2).

3.2 OPTIONS FOR DOWNHOLE EXCLUSION OF SAND

During the well design, options for excluding sand were considered [ref 3] and the decision was
taken to defer their implementation until such times as sand production is known to occur and
is generating a problem that cannot be tolerated (either down-hole or at surface). Options
include deploying down-hole sand screens, and incorporating sand stabilisation within a
FRAC/PAC operation.
The key aspects in this decision were:

• The initial reservoir pressure of 15000 psi makes formulating a kill fluid expensive. This is
substantially reduced once the reservoir pressure falls to less than 10000 psi.
• There is a risk of damaging the well during the sand screen deployment or FRAC/PAC.
• There is a risk that the sand screen may become damaged during deployment and would not
be successful in preventing sand production.
• Even if installed successfully, sand screens may become damaged in time.
• The production from the well will almost certainly be impaired by installing a sand screen.
• Sand production may not become a problem on all of the Shearwater wells.
• Sand production is not expected in the early years of field life. Production during the early
years of field life are also very important in the economics of the project.
• If sand is produced primarily from one well, then the remaining wells may have the
capability to deliver additional flow while this well is operated at reduced flow (sand-free)
while a work-over is arranged. This minimises deferment.
Nevertheless, the option to install one or more sand screens remains available should the
situation change. Sand screen design and deployment is a matter of on-going development by
the various suppliers, and specific screens will not be discussed here.

3.3 WELLHEAD REMOVAL OF SAND

Under some circumstances a cyclone can be installed at the wellhead to remove sand from the
flow and thereby protect downstream equipment. Desanding cyclones are usually operated
manually and are therefore not suited to use over a prolonged period of time. However, once a
well has been operated through the desander for a period of time at a high flowrate, it is likely
that the well can be operated at a lower flowrate without producing sand. Brent (Delta) were
supplied with a desanding cyclone for use as part of well clean-up [ref 4].
Within the Shearwater design, installation of (one or more) desanding cyclone was considered.
Provision was made to accommodate the weight of the units, and space was identified as being
available on a cantilever deck should this be required. The desander would have to be installed
upstream of the choke valve in order to protect the choke itself, and in order to operate the
desander with a single dense phase flow (and not a two-phase flow).
It was envisaged that temporary pipework would be installed to route the flow to the desander
from the kill wing valve of the well to be desanded. The ‘cleaned’ flow would be returned to
the normal process through the kill wing valve of a neighbouring well (which would be shut-in
for the duration of this operation). The flow would then proceed through the choke valve and
flowline of the second well. The sand would be collected from the desander by an accumulator
and emptied off-line by flushing with water.
It should be noted that a desanding cyclone has yet to be specifically designed / constructed that
is rated to 15000 psi. Hence the lead time for such a unit is likely to be substantial.

4 Other indicators of sand

production
4.1 CHOKE VALVES

Choke valves regulate the flow by the size and number of ports that the flow is permitted to pass
through. The ports are set into a flow cage. A sleeve on the outside of the cage is moved by the
actuator to cover or uncover more ports.
The cage and sleeve together form the flow control trim, and the key components are
manufactured from tungsten carbide to minimise the effects of erosion. However, erosion
cannot be eliminated and the trim will become eroded if sand is produced. The process of
erosion of the trim is characterised by a change in the opening characteristics of the valve with
time, that is, the flowrate at a given opening position increases with time, or alternatively, the
opening position that is set to obtain a given flowrate reduces with time.
The trim may also become damaged by brittle fracture, resulting in an abrupt increase in the
flow at a given opening position, or alternatively, an abrupt reduction in the choke opening
position for a given flowrate. Brittle fracture will result in a substantial quantity of debris being
carried through the flowline in a short period of time, which may appear on the sand detector
trace. If this occurs, the non-return valve on the flowline should also be considered to have
been damaged.
These two processes may also combine, such that erosion of the trim results in sections failing by
brittle fracture, which greatly accelerates the degradation process.
The status of the choke valve internals may be gauged by:
• Recording the flowing tubing head pressure (FTHP) as a function of choke opening
position during each well test and using it as a reference guide during normal operation.
• Leak testing the choke at the start of each well test.
Procedure 6 describes these activities.
The process implications from progressive damage to the choke trim are minimal provided that
the choke is automatically controlled to achieve a set FTHP.
However, if the choke is operated without being automatically linked to the control of FTHP,
and the internals become damaged abruptly, then the flow from the well will be abruptly
increased (depending on well productivity). The process implications of this will be as follows:
• If flowing to the test separator and the gas outlet is not able to accommodate the increased
flow, an ESD will be triggered by the pressure sensor PXHH-11026. If the liquid outlet is
not able to accommodate the additional flow, an ESD will be triggered by LXHH-11028.

• If flowing to the first stage separator and the gas outlet is not able to accommodate the
increased flow, an ESD will be triggered by the pressure sensor PXHH-10022. If the liquid
outlet is not able to accommodate the additional flow, an ESD will be triggered by LXHH-
10028.
Procedure 7 describes the action to be taken if damage to the choke valve is suspected or
confirmed.
Inspection of the choke valve body is discussed in section 5.
4.2 EROSION OF FLOWLINES

Inspection of the flowlines is discussed in section 5.
4.3 SEPARATORS
4.3.1 Accumulation of sand in the test and first stage separators
As described in Section1.4, sand which enters either of these vessels will accumulate until the
sand layer is sufficient for sand to be carried out of the vessel in the produced water stream.
Since the produced water outlet stands 150 mm off the internal vessel wall, the maximum
volume of sand that can be accommodated is estimated as being equal to 50% of the volume of
an even layer of sand 150 mm thick in the bottom of the vessel. For the first stage separator, this
corresponds to a volume of approximately 1 m3, which is equivalent to 1.5 tonnes of dry sand.
For the test separator, the volume of sand that can be accommodated is approximately 0.5 m3
(750 kg).
If sand is produced from the reservoir while the produced water flowrate is zero, then sand will
accumulate further, until it is carried over the weir and into the condensate stream. However,
since sand production is often associated with the onset of water production, this occurrence is
less likely.
Neither of the separators (test or first stage) is currently equipped with a sensor to specifically
detect accumulations of sand in the bottom of the vessel (although such sensors are intended to
be fitted). The non-intrusive techniques that can be mobilised to give an indication of
accumulations of sand in vessels are:
• Thermal imaging (the areas of vessel wall covered by sand show up as being colder)
• X-ray (or gamma) transmission (sand is dense and reduces transmission)
Thermal imaging would be difficult to deploy on Shearwater as the vessels are lagged and the
bottom of the vessels are close to deck level.
Resolution of the potential sand layer is the key factor when considering a transmission-based
detection system. This technique can be difficult to apply since transmission is affected by the
relative thickness of steel, lagging, water, oil, gas and sand layers. Conducting a baseline scan of
the vessel (sand-free) and recording the oil and water levels gives greatest potential for success.

4.3.2 Flushing the separators
The separators may be flushed while on-line by utilising the sand wash sparge nozzles and sand
wash outlet connections on each separator. This process is described in Procedure 8.
A sand removal cyclone and pump will be connected to the sand wash outlet from the separator
using temporary pipework. Produced water is drawn from the separator, through the sand
cyclone, and is sprayed into the separator through the sand wash nozzles. The sand exits the
sand cyclone as a concentrated slurry and falls into an accumulator. The process is illustrated
below:
Gas outlet
Produced fluid
inlet
V1010 or V1110
To sand wash
nozzles
Sand wash drain Produced

Sand removal water outlet
cyclone
Condensate
outlet
Sand accumulator
The supply pressure of the sand wash should only be approximately 0.7 bar greater than the
pressure in the vessel to minimise the risk of damaging the vessel in the vicinity of the sand wash
nozzles. Best sand washing results are obtained when the wash is performed frequently, as this
minimises the chance for sand to become compacted. Clearly, the maximum interval between
sand wash operations depends on the volume that can be accommodated (see 4.3.1) and the rate
at which sand is flowing into the separator.
The produced water level control valve should be closed during this process in order to protect
it (and downstream equipment) from sand which will become fluidised by this process. Since
produced water will continue to flow into the vessel, the level will increase during this process.
The produced water level should therefore be close to its minimum at commencement, and the
duration of this process will be limited by the rate of produced water inflow.
For the test separator, the time available for desanding may be maximised by performing this
process off-line, or while processing the well which has the lowest water cut. For the first stage
separator, the time available for the sand flush may be maximised by minimising production
from the well with the highest water cut, and by routing a well to the test separator and using

the difference in low-high produced water levels to provide an additional buffer zone. The
practicality of this process should be reviewed prior to implementation.
The sand which has been collected in the accumulator may be quantified, sampled, and then
discharged as a separate process when cooled sufficiently to prevent flashing of the water.
Quantification is required for discharge inventory tracking, but is also useful for assessing the
quantity indicated by the sand detectors, and to assess the effectiveness of the sand wash process.
A sand sample should be sent for size analysis. The sand detector STEP values may then need to
be updated once the sand size is known.
4.3.3 Level indicator in the first stage separator (V1010)
The produced water level control on the first stage separator is achieved by a level sensor
mounted in a bridle (LICA-10021), which has a 3" nominal diameter off-take at the bottom of
the vessel. As sand accumulates in the vessel, some sand will inevitably fall into the off-take.
When enough sand has accumulated in the off-take, the level control will become sluggish, that
is, the level in the bridle will lag behind the level in the vessel. The ultimate consequence of this
change in response is that the PW level in the vessel may rise or fall beyond its normal operating
range.
An undue fall in the PW level will be detected by one of the two independent PW level
sensors, which trigger an ESD if the produced water level falls too low (LXLL-10023 and
LXLL-10024). This condition is identical to that which arises if the PW level control valve leaks
when closed (e.g. due to internal erosion damage).
An undue rise in the PW level does not trigger an ESD, but causes PW to overflow the weir
and exit with the condensate stream to the second stage separator. This behaviour therefore
increases the PW duty on the second stage separator. Depending on the overall performance of
the separation system, this behaviour may result in an increased oil-in-water discharge, or failure
to meet water-in-oil for export. This condition may be diagnosed by observing that the PW
discharge from the first stage separator is nil over an extended period of time (i.e. flow to HP
hydrocyclones is zero) while PW is being discharged from the second stage separator (i.e. flow
to the LP hydrocyclones is high). If the total liquid flowrate cannot be accommodated by the
line to the second stage separator, the level in the first stage separator will rise until an ESD from
LXHH-10028 is triggered.
4.3.4 Level indicator in the test separator (V1110)
Level control in the test separator is achieved by way of a nucleonic sensor. The operation of
this type of sensor is not affected by accumulation of sand in the bottom of the vessel. The test
separator is therefore better equipped to deal with the effects of sand production than the first
stage separator. Hence, production from a well, which is thought to produce sand, should be
routed through the test separator.

4.4 LEVEL CONTROL VALVES

The produced water level in the first stage separator (V1010) is controlled by a valve at the inlet
to the HP hydrocyclones (FCV-40152, V-4015).
The level of produced water in the test separator (V1110) is controlled by LCV-11021.
If sand is produced, the operation of these two valves may become compromised over time due
to internal erosion damage of the control ports and seat. If either of these valves leaks
substantially when closed (due to internal damage) then the level in the separator will fall outside
of the normal range and an ESD will be triggered. The relevant first stage separator ESDs are
LXLL-10023 and LXLL-10024, while the relevant test separator ESDs are LXLL-11023 and
LXLL-11027.
Procedure 9 describes the action to be taken in the event of a repeated LL trip/alarm.
Erosion damage to these valves may be minimised by ensuring that they are equipped with a
tungsten carbide trim (10% binder) to maximise resistance of the material to erosion. Erosion
damage may also be minimised by ensuring that the size of the trim is appropriate for the
flowrate duty to avoid continually operating valves over a small opening range (e.g. 0 – 20%).

5 Wall thickness surveys
5.1 INTRODUCTION
This section establishes the legal requirement to maintain integrity of the Shearwater flowline
system and develops a Risk-Based Inspection (RBI) programme for the items which are
vulnerable to erosion.
This manual requires that areas of the production system are scanned using a manual ultrasonic
technique, on a time periodic basis if sand is not produced. Work is to be undertaken while the
plant is operational (i.e. flowing) unless practical constraints, safety concerns, or pipework
vibration make this impractical.
Further details of the areas to be scanned are given in section 5.3 – 5.7. The requirement for
additional inspection work to be completed in the event that sand is produced is discussed in
section 5.8. The requirement for additional inspection work to be undertaken if erosion damage
is found is discussed in section 5.9. A summary of the inspection scheme is given in section 5.10.
5.1.1 Definitions and context
It is assumed that the hydrocarbon containment system (including the chokes, flowlines and
manifold lines up to the first stage and test separators) will be regarded as a safety critical
element/system and will, therefore, need to be addressed by the Design and Construction
Regulations (DCR). The failure of this safety critical system, by whatever means, could ‘cause
or contribute substantially to’ a major accident.
The Shearwater Sand Management Manual addresses both the management of sand production
and its potential effects on the topsides plant. The integrity management requirements necessary
to address DCR need to consider all of the credible damage mechanisms. Within this manual,
only the damage mechanism of erosion due to produced sand is addressed.
The Verification Body and the Integrity Management Contractor will need to incorporate the
requirements contained within the Shearwater Sand Management Manual into the respective
Verification Scheme/Inspection Scheme (along with other requirements to address other
credible damage mechanisms).
5.1.2 Risk-based Inspection Programme
The consequences of failure of the system, together with the probability of failure, are likely to
result in the system being regarded as a high risk system. The risk-based inspection programme
for Shearwater must, therefore, require concentrated integrity management attention.
The provision of sand detectors on the individual flowlines and the production and test headers
should provide the Integrity Management Contractor with a ‘monitoring’ element within the
greater definition of managing integrity (inspection, maintenance, monitoring, testing etc.).

This is akin to the use of corrosion monitoring devices e.g. corrosion probes and weight-loss
coupons within a corrosion management scheme.
Such devices must be regarded as being only part of the data supply/capture necessary to
indicate the current condition of pipework/equipment. Over (and sometimes complete)
reliance on such monitoring devices has led to numerous loss/threatened loss of containment
incidents within the Industry.
Therefore, this section provides a ‘back-stop’ inspection programme to check for erosion
damage and which is independent of the output from the sand detectors (and, if fitted, the sand
level indicator within the first stage separator) and which reflects the potential for erosion to
occur whilst sand is not being detected by the sand detectors.
The RBI programme identifies:

• a series of locations that are to be inspected (the most vulnerable locations).
• the frequency on which they should be inspected.
• the most appropriate inspection technique is also identified (although a range of inspection
techniques could be used).
In the event that sand production is indicated by any of the sand detectors this ‘back-stop’
erosion inspection programme will be over-ridden by a more stringent erosion inspection
programme. However, in practice it is the frequency of inspection that is increased since the
vulnerable components and inspection techniques remain unchanged.
In the event that erosion damage is found, it is also appropriate to increase the frequency of
inspection. It may also be appropriate to change the inspection technique in favour of a method
that gives greater coverage or assurance.
The inspection programme is summarised in section 5.10.

5.1.3 Applicable inspection techniques
The techniques that are discussed here are all non-intrusive (ultrasonic), and the various
methods of using the technique are discussed in Appendix 2.
Ultrasonic inspection is achieved by placing a sensing head (transducer) onto the outside surface
of the pipework. The transducer sends out an ultrasonic signal which is reflected by the inside
surface. A unit records the time taken for the pulse to arrive and calculates the wall thickness of
the material. A block of known thickness is used to calibrate the unit. A fluid (couplant) is used
to ensure good ultrasonic contact between the transducer and the pipe wall.
The main issues in ultrasonic inspection are:
Issue Aspect Difficulty Aid

Coverage Series of points Can the same point be Guide bosses and
(% area examined) revisited? permanent mounting of
probes.
Are the points in the right RBI planning and

places? experience of damage
on this particular plant.
Area scanning Has the whole area been Good practice, or use
scanned? Recording the data of automated systems.
on a suitable spatial resolution. Coverage by automated
Proving coverage. systems is auditable.
Are the areas in the right RBI planning

places? Are the areas big
enough to capture expected
damage sites?
Service Inspection High temperatures Personnel protection,
interruption while the plant equipment cooling,
is running hot calibration block.
Movement of the line Shut down the line.

5.1.4 Practical aspects of ultrasonic inspection
The ultrasonic velocity (speed through the material), and attenuation (signal dispersion) are both
influenced by temperature. Both will increase at elevated temperatures, but the change is not
expected to be significant over the temperature and wall thickness ranges of concern to
Shearwater. However, it is good practice to attach calibration blocks of the same material to the
component surface such that the block attains a similar temperature to the pipe itself.
• The taking of spot thickness measurements at temperatures up to 150°C is relatively
straightforward provided reasonable access is available.
• Access appears from drawings to be particularly difficult, especially to the underside of the
inlet header and this is likely to be extremely uncomfortable at operational temperature,
possibly unsafe. The impact of this can only be accurately assessed through site trial.
• Scanning an area becomes more difficult as the viscosity of high temperature couplants vary
with temperature and they dry off quickly.
• Overhead scanning creates safety concerns of hot couplant falling onto the technician.
• Scanning aids are available for use by technicians during inspection of hot surfaces. A water-
cooled probe jacket is recommended for use since it is easier to hold, as well as protecting
the probe.
• Suitable protective clothing must be worn by the technician when working on/near surfaces
which are hot. Air cooled protective suits may be appropriate in some high temperature
situations.

5.2 IDENTIFICTAION OF VULNERABLE COMPONENTS

The components of the hydrocarbon containment system (between the choke valves and the
inlet nozzles of the first stage separator and the test separator) deemed to be most susceptible to
erosion damage by produced sand are the:
• the choke valves.

• the section of flowline immediately downstream of each choke valve.
• the target tees and 90o bends in the remainder of the flowline (up to the manifold).
• the test and production manifolds (flowlines to manifolds tees).
• the 90o bends in test and production headers (manifold to respective separator).
These components have been listed above in order of the magnitude of potential erosion
damage. However, of greater importance is the ability of the component to operate safely with
the level of damage.
5.3 INSPECTION OF CHOKE VALVES

The choke valve may be considered to comprise the body and the control trim. The control
trim is not a pressure retaining component and cannot be inspected by non-intrusive means.
Damage to the trim can be diagnosed by leak testing or from records of well performance (see
Procedure 6).
The body is further split into two sections: the ‘gallery’ region, which is in line with the inlet,
and the ‘outlet’ region, where the flow expands through a cone from the trim to the bore of the
flowline pipework. The outlet region is fitted with a tungsten carbide wear sleeve which is
resistant to erosion damage. The gallery region cannot be protected by a tungsten carbide
sleeve.
The following sections describe the use of non-intrusive inspection techniques. In addition, one
valve is to be removed within 12 months of platform start-up, on an opportunistic basis, and
examined for erosion damage.
5.3.1 Inspection of the gallery region
The geometrical configuration of the choke valve (axially-bored block) restricts inspection
coverage by simple 0°compression wave ultrasonic probes to the area adjacent to the three sides
of the gallery region as shown overleaf.

Cross-section
Section through the
through the Choke
control trim
Body in line with the
inlet
Zones of expected
damage in the
gallery region
Inspection of Gallery Region of Choke Valves using 0°° compression probe
Manual scanning axially along the face(s) of the choke body will be used to measure the wall
thickness in this region.
However, experience suggests that the sites of the major erosion damage to the gallery region
are expected to be located either side of this ‘tangent’ region. A time-of flight-diffraction
(TOFD) inspection technique will therefore also be used. The detail of this inspection set-up is
shown below:
Cross-section
through the Choke
Body in line with the
inlet
B
Receiver A Zones of expected
damage in the
gallery region
Ligament
Transmitter
Inspection of Gallery Region of Choke Valves using TOFD

TOFD inspection as shown will determine the ligament of material remaining between the
bore of the gallery (Signal B) and the chord between the ultrasonic probes (Signal A). Grey-
scale ultrasonic imaging techniques will be used to display the data as the probes are scanned
along the length of the choke valve body and remaining ligament information can be calculated
using data processing routines. It is recommended that the normal response of this technique is
established onshore using a new (unused) choke valve of the same type as installed on
Shearwater, before using this technique on the Shearwater platform chokes.
5.3.2 Inspection of the outlet region
It is also understood that each choke valve is furnished with a tungsten carbide insert in the
outlet section. The tungsten carbide sleeve will experience some erosion damage (but at a
slower rate than steel in this duty). The choke body will therefore be protected, but only while
the tungsten carbide sleeve remains intact.
The body of the choke valve will therefore be inspected in the outlet region to check that wall
thickness loss from the valve body has not occurred. Wall thickness loss would be indicative of
the insert being either completely missing or locally missing. The inspection necessary to
monitor for this type of damage is identical to that used to inspect the gallery region i.e.
compression wave inspection of the tangent regions and TOFD inspection of the regions in
between.
The inspection of the insert itself will not be possible by ultrasonics since the interface between
the insert and the choke body will not transmit ultrasound.
5.3.3 Choke valve inspection frequency
A baseline survey is to be conducted prior to start-up. Inspection methods may be developed /

verified onshore using a ‘spare’ choke valve if available, with reference to drawings detailing the
wall profile and thickness. Since choke valves are manufactured to close tolerances, a baseline
survey will then be completed on 3 out of the 6 valves on Shearwater. Baseline measurements
on the remaining valves will only be undertaken if there is significant discrepancy (1mm) in the
measurements across the baseline readings of the three valves that are inspected.
Each choke will then be inspected monthly for the first 3 months (to give extra protection
against the effects of debris and drilling mud brought up early in well life). Thereafter the choke
valves will be inspected every 3 months – assuming that there are no indications of sand being
produced.
5.4 INPSECTION OF THE FLOWLINES IMMEDIATELY

DOWNSTREAM OF THE CHOKE VALVES
A suitable length of the flowline immediately downstream of each choke valve is to be inspected
for evidence of erosion damage. This region is considered to be vulnerable because the flow out
of the choke valve may be subject to swirling effects and may not be uniformly distributed
across the full area of the pipe.

The inspection frequency of all the components described in section 5.4 will match that of the
chokes (5.3.3), i.e. Baseline 3 out of 6, Inspected monthly for the first 3 months, then every 3
months thereafter – assuming that there are no indications of sand being produced.
The flowline/pipe length to be covered is approximately 5 x the internal diameter (750mm) or

to the first bend/target tee – whichever is the shortest length. Because these sections of the
flowlines all lie in the horizontal or near horizontal plane and will encounter unpredictable flow
patterns as the wellstream fluids exit the choke valves, full circumferential coverage will be
needed. This length of pipe will start as close to the flange/hub clamp as is practicable. The
presence of a pipe support within this section of flowline may preclude the inspection of a part
of this section of the flowline.
Initial on-line inspection will consist of grid-based area scanning using a hand-held ultrasonic
probe to scan the flowline in a grid pattern and manually recording the results onto a
‘developed’ 2-dimensional sketch of the flowline section. It is recommended that the grid
pattern is based on an approximate grid of 150mm x 150mm – 150mm linear length down the
flowline and 4 x circumferential regions (12 –3 o’clock, 3 –6, 6-9 & 9-12 in relation to the
direction of flow). This would result in 20 grid boxes. Each grid box should be scanned to record
the minimum wall thickness within the box as well as the wall thickness at each corner.
Flowline 4 (well SW07) has a 30o bend, in the horizontal plane, immediately downstream of the
choke valve. This 30o bend is immediately upstream of a straight section of pipe. Because of
the small angle of this bend, and the fact that the wellstream fluids will exit the choke valve in an
unpredictable way, this bend is to be subject to 360 o manual scanning between the hub clamp of
the choke valve and the circumferential butt weld with the straight pipe section. The grid box
will be based on 150mm increments marked on the outside surface of the bend – this will result
in narrower box widths on the inside on the bend.
The flowline from the Pentland well (SW05) has two tees within the length immediately
downstream of the choke and is discussed in the following sections.
5.4.1 Pentland flowline – target tees
Flowline 7 (Pentland well SW05) differs from all of the other flowlines in that a target tee is
located immediately downstream of the choke valve – the downstream hub of the hub clamp
assembly is welded directly to an 8″ equal tee without a ‘pup-piece’ between these components.
A second target tee is then located immediately after a straight section of pipe (straight section of
pipe is ~1200mm long).
The end cap of the target tees will be manually scanned using a grid pattern as follows:
• straight section of end cap (upstream of weld to equal tee – to the start of the radiused
section) sectioned into 4 quadrants (12–3 o’clock, 3–6, 6–9 & 9–12) as viewed from the end
cap,
• a single circular section approximately 120mm in diameter centred on the centre point of
the end cap’s end,
• the section between the latter circular section and the former straight sections and including
the radiused section of the end cap, sectioned into the same 4 quadrants (12–3 o’clock, 3–6,
6–9 & 9–12).

In addition to the above, the shoulder of the tee itself will be scanned on the outlet leg up until
the attachment weld to the downstream pipe. The region will be split into 4 quadrants (grid
boxes marked as 1, 2, 3 & 4) as shown overleaf.
Each of the respective grid boxes (rectangles, circle and ‘other’ sections) – 13 in total – should
be scanned to record the minimum wall thickness within the ‘box’ as well as the wall thickness
at each corner. This will produce 20 point wall thickness readings and 11 minimum ‘scanned’
wall thickness readings.
Outlet leg,
End-cap side
Grid Box Shapes on Target Tee End Cap
5.4.2 Pentland flowline – straight pipe downstream of the choke
The section of the Pentland flowline (flowline #7, SW05) pipe between the 2 x target tees
immediately downstream of the choke valve will also be manually scanned. This linear length
of flowline will start immediately downstream of the circumferential butt weld between the
pipe and the upstream target tee. This section of straight pipe will be inspected in a similar way
to the other flowlines i.e. a grid pattern covering 750mm linear length and 360o of the
circumference of the pipe.
5.5 INSPECTION OF THE REMAINDER OF THE FLOWLINES

5.5.1 90 degree bends
The 1st 90o bend downstream of the choke valve on each flowline will be subject to manual
scanning, with the exception of the Pentland flowline (#7) since the two target tees will be
inspected. All of these bends are associated with a change of flow from the horizontal to vertical
downflow – considered to be more susceptible to erosion damage than bends in the horizontal
plane. Scanning will be confined to the outside of the bends for a strip that is 60mm wide and
which extends for the full length of the outside of the bend – up to each circumferential butt

weld. The full length of the outside of the bend is ~1000mm and, therefore, grid boxes of
100mm length x 60mm width are suggested. This would result in 10 grid boxes. Each grid box
should be scanned to record the minimum wall thickness within the box as well as the wall
thickness at each corner.
The presence of a part-encirclement locating lug on the outside of the bend on flowline #1 will
prevent full coverage of this bend.
Flowlines 5 & 6 (wells SW08 and SW09) have 30o bends, in the horizontal plane, immediately
upstream of the 90o bends that have been selected for inspection. This will tend to bias any
erosion damage to the 90 o bend off-centre in the direction of the outside bend of the 30o bend.
Because of this, the grid boxes for manual scanning of these bends will be extended to a total
width of 150mm (30mm/120mm from centreline of outside of bend, biased according to the
orientation of the upstream 30o bend).
A baseline inspection will be completed on 3 out of the 5 bends. The inspection frequency (for
all 5) during production is set at 6 months - assuming that there are no indications of sand being
produced (including there being no indication of wall thinning of the choke valves and the
sections of straight pipe immediately downstream of the chokes).
5.5.2 Target tees
Upstream of the Non-return Valve and isolation/routing valves on each of the flowlines (route
to the test or production manifold) there are 2 x target tees (1 orientated horizontally and 1
orientated vertically). It is proposed to conduct a baseline inspection of these target tees on a
sample basis with 2 of the horizontally orientated target tees and 2 of the vertically orientated
target tees being inspected (one from each of four flowlines). Downstream of these isolation
valves there are a minimum of five target tees on each flowline. Two of the total number of
target tees will have a baseline inspection.
Review of the piping isometrics shows that the horizontally orientated target tees on flowlines
1, 2 & 4 (wells SW01, SW04 & SW07) have the longest sections of straight pipe immediately
upstream of the target tees and, therefore, these horizontally orientated target tees are selected
for manual scanning. The vertically orientated target tees of flowlines 5 & 6 (wells SW08 and
SW09) will also be manually scanned. All of the target tees will be sectioned into grid boxes in
an identical manner to that detailed previously (section 5.4.1), however, the vertically orientated
target tees will use a Platform North (N), S, W & E quadrant notation. Only 180 o of the outlet
leg will be inspected – the end cap side (grid boxes 1 & 2 on the Figure in section 5.4.1).
Apart from the baseline inspection, these target tees will not be subjected to any further
inspection unless significant quantities of sand are produced (see section 5.8), or in the event
that erosion damage is found on another component (section 5.9).
Note: The inspection of the target tees immediately downstream of the choke valve on the
Pentland well is discussed in section 5.4.1.

5.6 INSPECTION OF THE PRODUCTION MANIFOLD AND TEST

MANIFOLD
5.6.1 Production Manifold
Flow from all of the individual wells (except flowline 2, well SW04) enters the production
manifold in a vertically downwards manner via an 8″ x 18″ reducing tee – the manifold is 18″
nominal diameter. Piping ‘pup-pieces’ connect the individual tee pieces to form the manifold.
If sand is present within the wellstream fluids the bodies of the reducing tees opposite to the 8″
inlet will suffer impingement/erosion attack. The flowline configuration upstream of the
reducing tees will introduce a bias to this impingement location. This impingement location is
influenced further by the flow within the manifold (from the other wells) resulting in potential
erosion damage of the ‘pup-pieces’. Flow from flowline 2 (well SW04) enters the manifold
through an eccentric reducer at one end of the manifold and, therefore, is not considered to
produce ‘direct’ impingement attack of the manifold.
It is proposed to inspect the bottom sections of the manifold – opposite the flowline inlets –
manually scanning, using a grid box pattern with the following details:
• the start of the linear section of grid boxes is 175mm upstream of the vertical centreline of
the tee,
• 4 x 175mm linear sections, through the tee and into the downstream ‘pup-piece’,
• 9 o’clock position of the tee and ‘pup-piece’ to the 7:30 position, 7:30 to the 6 o’clock
position, and 6 o’clock position to the 4:30 position (each covering ~ 175/180mm of the
external surface) - as viewed in the direction of flow down the manifold.
Grid Boxes on Production Manifold

This would produce 12 grid boxes. Each grid box should be scanned to record the minimum
wall thickness within the box as well as the wall thickness at each corner.
A baseline inspection survey will be conducted on 3 out of the 5 flowline inlets, including the
two which are farthest from the manifold outlet (currently flowlines #5 and #6). Thereafter the
manifold will not be subject to any further inspection unless significant quantities of sand are
produced (see section 5.8) or erosion damage is evident on another component (see section
5.9).
5.6.2 Test Manifold
Inspection of the bottom sections of the test manifold – opposite the entry points of the
flowlines (entry is through 8″ x 12″ reducing tees) is to be carried out in an identical manner to
the production manifold. The only differences between the two inspection areas being
associated with the actual grid box dimensions and the bias of the potential erosion damage,
therefore, the grid box pattern will have the following details:
• the start of the linear section of grid boxes is 125mm upstream of the vertical centreline of
the tee,
• 4 x 125mm linear sections, through the tee and into the downstream ‘pup-piece’,
• 3 o’clock position of the tee and ‘pup-piece’ to the 4:30 position, 4:30 to the 6 o’clock
position, and 6 o’clock position to the 7:30 position (each covering ~ 125/130mm of the
external surface) - as viewed in the direction of flow down the manifold.
This would produce 12 grid boxes. Each grid box should be scanned to record the minimum
wall thickness within the box as well as the wall thickness at each corner.
A baseline inspection survey will be conducted on 3 out of the 6 flowline inlets. Thereafter the
manifold will not be subject to any further inspection unless significant quantities of sand are
produced (see section 5.8) or erosion damage is evident on another component (see section
5.9).
5.7 INSPECTION OF THE PRODUCTION HEADER AND TEST

HEADER
5.7.1 Production Header
The pipe run between the production manifold and the first stage production separator – the
production header – is mainly 18″ nominal diameter and contains 1.5D radius bends. The 90o
bends are furnished in 20″ nominal diameter with matching internal diameter to the 18″ line
pipe thus providing localised increased wall thickness (approximately 31mm nominal wall
thickness).
One of these 90o bends has been selected for periodic erosion inspection by using the rationale
of vertical to horizontal flow promoting greater potential erosion damage. The 90o bend
chosen for inspection is the 1st bend on the header off the manifold (i.e. the 90o bend shown on
Drawing Number 01-PI-4120, Rev. 2).

Previous work by AEA Technology identified that the greatest erosion damage sites associated
with annular multiphase flow is concentrated at the outside of the bend - at around 67o into the
bend in the direction of flow. In order to detect any damage obeying this rationale and
previous work, and to allow for variable flowrates and differing fluid properties, it is proposed
that the outside of the bend is inspected between ~50o and ~80o. This equates to a region
between 1060mm and 1420mm around the bend split into 2 x 180mm grid boxes each of
130mm width centred around the centreline of the outside of the bend.
Each of the grid boxes – 2 in total – should be scanned to record the minimum wall thickness
within the grid box as well as the wall thickness at each corner. This will produce 6 point wall
thickness readings and 2 minimum ‘scanned’ wall thickness readings.
A baseline inspection survey will be conducted. Thereafter the manifold will not be subject to
any further inspection unless significant quantities of sand are produced (see section 5.8) or
erosion damage is evident on another component (see section 5.9).
5.7.2 Test Header
The pipe run between the test manifold and the test separator – the test header – is mainly10″
nominal diameter and contains 1.5-D radius bends. The 90o bends are furnished in 12″ nominal
diameter with matching internal diameter to the 10″ line pipe thus providing localised increased
wall thickness (approximately ?mm nominal wall thickness).
One of these 90o bends has been selected for periodic erosion inspection by using the rationale
of vertical to horizontal flow promoting greater potential erosion damage. The 90o bend
chosen for inspection is the 2nd bend on the header off the manifold (i.e. the upper 90o bend
shown on Drawing Number 01-PI-41?3, Rev. 1). The actual inspection area of the bend has
been derived by considering the same rationale/previous work as the production header. It is
proposed that the outside of the bend is inspected between ~50o and ~80o. This equates to a
region between 640mm and 870mm around the bend split into 2 x 115mm grid boxes each of
90mm width centred around the centreline of the outside of the bend.
Each of the grid boxes – 2 in total – should be scanned to record the minimum wall thickness
within the grid box as well as the wall thickness at each corner. This will produce 6 point wall
thickness readings and 2 minimum ‘scanned’ wall thickness readings.
A baseline inspection survey will be conducted. Thereafter the manifold will not be subject to
any further inspection unless significant quantities of sand are produced (see section 5.8) or
erosion damage is evident on another component (see section 5.9).

5.8 INSPECTION IN THE EVENT THAT SAND IS DETECTED

As discussed above, the choke valves (and pipework immediately downstream of the chokes) are
the most vulnerable components to erosion damage and the routine inspection programme will
ensure that integrity is maintained if sand is produced but is NOT detected by the sand
monitoring system. This section concerns the escalation of the inspection scheme in the event
that sand is detected.
The topsides pipework was designed to be able to endure service with a substantial sand loading
on a continuous basis. The detection of sand does not therefore imply that integrity will become
compromised. However, clearly it is prudent to consider increasing the frequency of inspection
of the flowline pipework if substantial quantities of sand production are encountered, or if
erosion damage is found in the areas which are inspected more frequently.
Additional inspection work will therefore be conducted on the most vulnerable components
once 500 kg of sand has been produced (and detected) by any one well since the most recent
inspection of this flowline (whether this amount of sand is produced in 1 day or 2 months). The
need for inspection is deliberately linked to the mass of sand produced since this reflects the
ability to cause damage. In addition, the inspection programme will also extend further to check
for erosion of the other vulnerable components, including the entry of the flowline into the
manifold.
The bend on the production header and test header will similarly be inspected once 500 kg of
sand has been passed in total across the relevant wells, according to the routing of those wells.
A running total of the sand quantity indicated by each of the sand monitors will be stored within
RTMS and can be accessed through Well Programming. In addition, the accumulation of sand
in the first stage, second stage, and test separators will be monitored by nucleonic sand level
sensors (when fitted).
5.9 INSPECTION IN THE EVENT THAT EROSION DAMAGE IS

EVIDENT
In the event of wall thinning becoming evident from the manual scanning of any component,
the use a more refined inspection using an automated or semi-automated ‘corrosion mapping’
system is recommended to better define and track erosion damage to that component. The
exact system to be used (e.g. ‘Seescan’, ‘Mapscan’, ‘Andscan’) will depend upon the accessibility,
practicability and spatial resolution requirements. The system will be used to produce data of
sufficient quantity and quality, and at suitable periods, to enable the increase in erosion damage
to be monitored whilst still maintaining integrity.
In the event of wall thinning becoming evident from the manual scanning of any component,
equivalent components within the line (e.g. the other bends or elbows) shall also be inspected
using the same technique as for the component which has been found to be damaged.
In addition, since the flowlines are broadly similar in layout, the equivalent component on the
remaining flowlines shall also be inspected.

5.10 SUMMARY OF THE INSPECTION SCHEME

5.10.1 Baseline inspection
A baseline inspection is needed in order to establish the as built wall thickness of relevant
vulnerable components as a context for future wall thickness measurements taken during
service. Sampling across similar components will be used to minimise the workload / insulation
removal. The effect of sampling will need to be accounted for when interpreting future wall
thickness measurements of components on which a baseline measurement was not taken.
What Baseline sample How Plant condition

Choke body (gallery Baseline 3 out of 6*. TOFD Prior to start-up
and exit region)
0° compression probe
Pipework immediately Baseline 3 out of 6*. 0° compression probe Prior to start-up
downstream of the
choke, including 30° Bends: 4 x 5 Grid
bends where relevant, with area scan,
and including the first manually deployed.
two target tees in the
Pentland flowline Tees: 13 box grid
with area scan,
manually deployed.
First 90° bend Baseline 3 out of 5*. 0° compression probe Prior to start-up
downstream of the
choke valve (Pentland 1 x 10 Grid with area
is exempt due to scan, manually
inspection of target deployed.
tees near the choke
valve)
Flowline target tees Baseline four, one 0° compression probe Prior to start-up
upstream of the from each of four
isolation valves. flowlines. 11 box grid with area
scan, manually
deployed.
Flowline target tees Baseline two, from 0° compression probe Prior to start-up
downstream of the each of two
isolation valves. flowlines. 11 box grid with area
scan, manually
deployed.
Production manifold Baseline 3 out of 5*, 0° compression probe Prior to start-up
(opposite the flowline 2 of which should be
inlets) the farthest top entry 3 x 4 Grid with area
inlets from the scan, manually
manifold outlet. deployed.
* If the results are inconsistent (by more than 1 mm) then the remainder will also need a baseline inspection.

What Baseline sample How Plant condition

Test manifold Baseline 3 out of 6*. 0° compression probe Prior to start-up
(opposite the six
flowline inlets) 3 x 4 Grid with area
scan, manually
deployed.
Production header Baseline. 0° compression probe Prior to start-up
(1st 90° bend off the
manifold) 1 x 2 Grid with area
scan, manually
deployed.
Test header Baseline. 0° compression probe Prior to start-up
(2nd 90° bend off the
manifold) 1 x 2 Grid with area
scan, manually
deployed.
* If the results are inconsistent (by more than 1 mm) then the remainder will also need a baseline inspection.

5.10.2 On-going routine inspection
On-going routine inspection is needed in order to ensure that integrity is maintained (and can
be demonstrated) for the components which have the highest potential for erosion damage
because they are in a region where there is considerable uncertainty as to the nature /
uniformity of the flow. This routine inspection programme will also ensure that the sand
detection system is not relied upon to protect the integrity of the process.
What Frequency (if no How ** Plant condition

sand is detected) *
Choke body (gallery Monthly for the first TOFD While operating
and exit region) 3 months. unless practicality,
0° compression probe safety or movement
Every 3 months renders this
thereafter. unworkable.
Pipework immediately Monthly for the first 0° compression probe While operating
downstream of the 3 months. unless practicality,
choke, including 30° Bends: 4 x 5 Grid safety or movement
bends where relevant, Every 3 months with area scan, renders this
and including the first thereafter. manually deployed. unworkable.
two target tees in the
Pentland flowline Tees: 13 box grid
with area scan,
manually deployed.
First 90° bend Every 6 months. 0° compression probe While operating
downstream of the unless practicality,
choke valve (Pentland 1 x 10 Grid with area safety or movement
is exempt due to scan, manually renders this
inspection of target deployed. unworkable.
tees near the choke
valve)
* Each component to be re-inspected if 500kg of sand is produced since the last inspection of this component.
** If damage is found using manual inspection, use of an automated technique with greater data gathering and
handling capability is recommended.
This inspection scheme should be reviewed and may be altered by the Integrity Manager
(subject to approval by the Verification Body) in the light of a growing history of inspection
results and production records.

5.10.3 Additional inspection following sand production
The following components will be inspected following production of sand in addition to those
identified in section 5.10.2.:
What Frequency How * Plant condition

One of the flowline After each 500 kg of 0° compression probe While operating
target tees upstream of sand detected from unless practicality,
the isolation valves. one well. 11 box grid with area safety or movement
scan, manually renders this
deployed. unworkable.
Production manifold After each 500 kg of 0° compression probe While operating
(opposite the relevant sand detected from unless practicality,
flowline inlet) one well. 3 x 4 Grid with area safety or movement
Production header After each 500 kg of 0° compression probe While operating
(1st 90° bend off the sand detected in total unless practicality,
manifold) from relevant wells 1 x 2 Grid with area safety or movement
(depending on scan, manually renders this
routing). deployed. unworkable.
Test manifold After each 500 kg of 0° compression probe While operating

(opposite the six sand detected from unless practicality,
flowline inlets) one well. 3 x 4 Grid with area safety or movement
Test header After each 500 kg of 0° compression probe While operating
(2nd 90° bend off the sand detected from unless practicality,
manifold) relevant wells 1 x 2 Grid with area safety or movement
(depending on scan, manually renders this
routing). deployed. unworkable.
* If damage is found using manual inspection, use of an automated technique with greater data gathering and

5.10.4 Additional inspection if erosion damage is found
The following components will only be inspected if erosion damage is found on the
components identified above (5.10.2, 5.10.3).
What Frequency How * Plant condition

Other bends in the Only if erosion 0° compression probe While operating
flowline damage is evident on unless practicality,
one of the bends Grid with area scan, safety or movement
being inspected manually deployed. renders this
routinely unworkable.
Equivalent Only if erosion 0° compression probe While operating
components in other damage is evident on unless practicality,
flowlines one flowline Grid with area scan, safety or movement
manually deployed. renders this
unworkable.
* If damage is found using manual inspection, use of an automated technique with greater data gathering and

6 Procedures
Name What? When? Owner Section
Compliance Sand management over-view 6 monthly Asset

checklist Manager
Procedure 1 Verifying operation of the sand weekly Offshore Inst. 2.5

detectors: the wire brush test. Tech.
Procedure 2 Action(s) to be taken in the event on alarm Control 2.2
of a sand detector alarm. Room Op. 2.3
Procedure 3 Altering the operation of the sand as instructed by Offshore Inst. 2.4.1
detectors: Viewing and changing Prod. Tech.*2 Tech. 2.4.2
coefficients / alarms. 2.4.4
Procedure 4 Well bean-up to minimise sand as instructed by Control 3.1
production from a ‘problem well’ Well Programming*1 Room Op.
Procedure 5 Finding the maximum sand-free as instructed by Offshore Inst. 3.1
production rate. Well Programming *1 Tech.
Procedure 6 How to check for possible damage monthly, OIM 4.1
to choke valves. or at well test*2
Procedure 7 Action to be taken if choke valve as required OIM 4.1
damage is suspected or confirmed.
Procedure 8 Flushing the separator (Test as required OIM 4.3.2
and/or First Stage Separator).
Procedure 9 Action to be taken in the event of on alarm / trip OIM 4.4
repeated LL-level trip on Test or
First Stage Separator.
Procedure 10 Non-intrusive inspection. as per inspection Integ.
• choke valves schedule Mangmt. 5.10
• pipework immediately 5.3
downstream of choke valves 5.4
• the remainder of the flowlines 5.5
• production and test manifolds 5.6
• production and test headers 5.7
Procedure 11 Action to be taken if sand is as evident OIM
evident from the nucleonic sand
level sensor in the separator(s)
*1: Well programming (onshore) are responsible for the routine review of sand detector output on a
monthly basis or following specific events, such as if an alarm is triggered, when bringing a new well on,
or following a well work-over.
*2: Production Technology (onshore) are responsible for checking for possible choke damage evident
from well test results (higher flowrate than expected at given choke position), and for advising that new
sand alarms are to be set (based on the results obtained from Well Programming).

7 References
1. ‘Detection of sand in high GLR multiphase flows’, AEAT-2554, January 1998. (Work
conducted specifically for Shearwater).
2. Lockett T.J. & Worsley J., ‘Assessment of operating envelopes for non-intrusive sand
detectors’, AEAT-RHAD/RD01535003/12/2 v2 May 2000. (Copy supplied to SIEP)
3. ‘Shearwater sand control and treatment strategy for the SDA’, SWR/WE/083, March 1998.
4. ‘Wellhead desander delivers the goods’, Offshore Engineer, October 1996.
5. ClampOn user guide: ‘Technical Manual’, 62.310.0016.00, SDA PO.No: 32-I034-001,
SDDR Code 137, 10/5/00

8 Index
Alarm.......................................................................................................13, 14, 15, 16, 18, 64

L1 alarm ......................................................................................... 14, 15, 16, 17, 18, 19, 23
L2 ...............................................................................................................................14, 15
L3 ...............................................................................................................................14, 15
M alarm .................................................................................................... 14, 15, 16, 18, 19
Bean-up ............................................................................................................................... 65
ClampOn.................................................................. 6, 7, 12, 13, 20, 21, 24, 26, 27, 28, 30, 59
ESD .............................................................................................................. 34, 35, 37, 38, 77

LXHH-10028 .............................................................................................................35, 37
LXHH-11028 .................................................................................................................. 34
LXLL-10023 ...............................................................................................................37, 38
LXLL-10024 ...............................................................................................................37, 38
LXLL-11023 .................................................................................................................... 38
PXHH-10022 .................................................................................................................. 35
PXHH-11026 .................................................................................................................. 34
Flow velocity ..................................................................................... 19, 20, 23, 24, 25, 26, 28
Inspection
Area scanning .......................................................... 46, 52, 53, 54, 55, 56, 57, 84, 88, 91, 94
A-scan.............................................................................................................................. 91
Temperature effects ......................................................................................... 41, 42, 81, 92
TOFD (Time Of Flight Diffraction) ......................................... 44, 45, 53, 55, 76, 84, 91, 95
Ultrasonic....................................................................................... 44, 53, 54, 55, 56, 57, 76
Interpretation ......................................................................................................................... 6
Offshore ..................................................................................... 6, 17, 28, 58, 59, 64, 67, 92
Onshore .......................................................................................................... 20, 28, 64, 82
Level control............................................................................................................. 37, 38, 65

FCV-40152.................................................................................................................38, 80
LCV-11021...................................................................................................................... 38
LICA-10021 .................................................................................................................... 37
PCS ................................................................................................................................13, 76
Procedures ..................................................................................................................6, 58, 63

Procedure 1................................................................................................. 6, 28, 58, 63, 64
Procedure 10...........................................................................................................6, 58, 63
Procedure 11...............................................................................................................58, 63
Procedure 2...................................................................................... 6, 13, 17, 58, 63, 64, 76
Procedure 3................................................................................................. 6, 21, 23, 58, 63

Procedure 4................................................................................................. 6, 17, 31, 58, 63

Procedure 5................................................................................................. 6, 18, 31, 58, 63
Procedure 6..................................................................................................... 34, 43, 58, 63
Procedure 7........................................................................................................... 35, 58, 63
Procedure 8........................................................................................................... 36, 58, 63
Procedure 9........................................................................................................... 38, 58, 63
RBI (Risk-based inspection)...................................................................................... 39, 40, 41
Responsibilities .................................................................................................................6, 64
RTMS .......................................................................................................... 13, 26, 27, 28, 52
Sand detector
configuration............................................ 6, 14, 15, 16, 18, 20, 21, 43, 49, 58, 63, 67, 71, 72
functionality ..................................................................................... 6, 28, 64, 67, 68, 69, 70
output .............................................................................6, 16, 23, 24, 58, 76, 79, 80, 82, 83
spurious output................................................................................................................. 17
Step........................................................................................................... 20, 21, 22, 24, 37
Zero.......................................................................................................... 20, 21, 22, 24, 26
Sand exclusion ..................................................................................................................8, 32
Sand management .....................................................................................................3, 5, 6, 62
Sand removal
Separators................................................................................. 33, 36, 37, 64, 77, 78, 79, 80
Wellhead......................................................................................................... 10, 33, 36, 77
Total sand quantity .................................................................................................... 13, 27, 52

RESTRICTED: COMMERCIAL SWR-G-255, AEAT-RHAD/RD01622/12/1 v3
Appendices
CONTENTS
Appendix 1 Shearwater sand management manual procedures
Appendix 2 Non-Intrusive inspection techniques
AEA Technology 62
Appendix 1
Shearwater Sand
Management Procedures
CONTENTS
Compliance checklist
Procedures
Procedure 1 Verifying operation of the sand detectors: the wire brush test.
Procedure 2 Action(s) to be taken in the event of a sand detector alarm.
Procedure 3 Altering the operation of the sand detectors: Viewing and

changing coefficients / alarms.
Procedure 4 Well bean-up to minimise sand production from a

‘problem well’
Procedure 5 Finding the maximum sand-free production rate.
Procedure 6 How to check for possible damage to choke valves.
Procedure 7 Action to be taken if choke valve damage is suspected or

confirmed.
Procedure 8 Flushing the separator (Test and/or First Stage Separator).
Procedure 9 Action to be taken in the event of repeated LL-level trip on

Test or First Stage Separator.
Procedure 10 Non-intrusive inspection.
Procedure 11 Action to be taken if sand is evident from the nucleonic sand

level sensor in the separator(s)
AEA Technology 63
SAND MANAGEMENT REVIEW (COMPLIANCE CHECKLIST)

v1.02
Person Shearwater Asset Manager
Responsible
Purpose To ensure compliance with the Shearwater Sand Management
Manual, and ensure that compliance can be demonstrated as/when
required.
Scope To summarise/record the checks that have been completed,
and to document the actions taken/planned if sand is evident.
Input Information is needed from: OIM, Offshore Instrument Technician,
Onshore Production Technology
Aspect Check Result Compliance

Function Have the sand detectors Y/N Date of last check
been checked for (Procedure 1):
functionality?
If No:
which unit(s) are
not working?
Offshore Has a sand detection Y/N If Yes, date of

detection alarm been triggered since Alarm(s): _______
and the last check? (excluding Well(s): ________
reaction functionality tests)
If Yes: If Procedure 2 was
Was Procedure 2 not implemented,
implemented? is action being
Y/ N taken to address
this? Y/N
Offshore Is sand production Y/N Record the current
detection evident from the sand sensor readings:
level sensor in the test and If Yes: Test separator:
first stage separators? Does the separator
require flushing?
Y/N 1st Stage separator:
Onshore From the sand detector Y/ N

detection records, is sand evident at If Yes:
review a level below the alarm Is this condition Record the new
set-points? acceptable? Y / N tolerable sand
flowrate:
If No, have the
alarm set-points Record the new
been changed? alarm set-points:
Y/N
Page 1 of 2
AEA Technology 64
Choke Do well test results or Y/N If Yes: record the

trim leak test suggest that the Well(s): well and schedule
erosion choke trims are damaged? maintenance.
Level Have any low-low liquid Y /N If Yes and sand is

control level trips occurred on the evident, schedule
valves test or first stage separator maintenance of the
since the last review? valve
Inspection Has any inspection work Y/N If Yes: record the

been conducted since the inspection report
last review? reference/date:
______________
Separators Was either of the Y/N If Yes: record the
separators flushed since operation report
the last review? reference/date:
______________
Forward activity plan
On the basis of this status review:

Should additional inspection work be Y/N
scheduled? (required if significant Choke:
quantities of sand have been produced, Flowline:
or if damage has been found) Manifolds:
Headers:
Should well operation be changed? Y/N
(required if sand is evident) Well(s):
Bean-up more slowly:
Establish max sand free rate:
Route to test separator:
Should either separator be flushed? Y/N
Test separator:
First stage separator:
Is maintenance required? Y/N
Level control valves:
Broken sand detectors:
Choke valves:
Completed by:
Approved by: (Asset Manager)
Date:
By completing this review you have demonstrated that sand is being managed on
Shearwater. The next Sand Management Review is due: __________(6 monthly)
Page 2 of 2
AEA Technology 65
SSMM-001: Verifying the operation of the sand detectors

Purpose: To provide confidence that the sand detector sensors are functioning normally.
2. Read through this procedure to

1. START: assess the hazards, and take appropriate
action. Hazards are likely to include:
• access
• hot surfaces
• pipework vibration & noise
4.Display a trend graph for the 3. Identify the sand detector to be tested.
sand detector to be tested. • Testing may only be conducted when
the well is flowing (as 0 g/s is
enforced when the well is shut in).
• To avoid confusion of the results,
testing should only be carried out
5. Advise the control room during a period of steady flow.
operator of the sand detector to
be tested, and that the sand
detector alarms may be triggered
by this test. This alarm may be
cancelled and ignored.
6. Vigorously rub the pipe wall in the vicinity of

the sand detector with a wire brush for a period 9. Send the
of 30 – 60 seconds. Have all the completed form
detectors (SSMM-001a) to
• Brushing must occur on the pipe wall and been tested? the onshore Well
not the lagging. Programming.
• Lagging should not need to be removed
Advise relevant
since sufficient bare pipe should be available). parties of the
outcome (as given
on the form).
END
7. Advise the control room
operator that the sand detector 8. Record the result of
test of the detector has been the functionality test on
completed, and that future alarms form SSMM-001a.
should be treated as per
procedure 2.
AEA Technology 66
FORM SSMM-001a
Purpose: To document the functionality checking of sand detectors.
Test completed by: Name:

Position: Offshore Instrument Technician
Detector Date:Time Sand flowrate Maximum Return to Result *Note 2

(refer to indicated indicated previous
section before the sand rate level after ok / not ok
2.3.3 for test (if 0 g/s produced by test
location) write zero) the test *Note 1 completed?
(g/s) (Yes / No)
XIA-01111
XIA-01121
XIA-01141
XIA-01151
XIA-01161
XIA-01171
XIA-01201
XIA-11001
Notes:
1: If an alarm was triggered, record the alarm type: L1-alarm (XAH-nnnn) or M-alarm (XQAH-nnnn)
2: Result if ‘ok’ if the test produced an increase and the signal, then returned to previous level. Otherwise ‘not ok’.
Send the completed form to Well Programming (for record keeping).

Advise the Asset Manager that this test has been completed.
If the response from any sensor was ‘not ok’, the following additional actions must be taken:
6. If the failed sensor is on one of the flowlines, the well must only be flowed to the test separator
(i.e. via XT-11001).
7. Request Well Programming to review the raw signal at the time of the test to determine whether
the increase in signal caused by the wire brushing was masked out by inappropriate calibration
coefficients (Zero set too high).
8. Advise the Maintenance Manager so that sensor replacement can be initiated.
9. Advise the Asset Manager that Shearwater has a reduced capability to detect sand until the
problem is diagnosed and fixed.
AEA Technology 67
SSMM-002: Actions to be taken in the event of a sand alarm

Purpose: To enable the Shearwater control room to respond appropriately to sand detector alarms
(excluding alarms which may be triggered during work to test the functionality of the detectors).
1. START:
Is the alarm
from 2. Go To
XT-01201 Procedure 2a
?
Is it an L1
Is it an M 6. Malfunction
type alarm?
type alarm? alarm (XMA-nnnn)
(XAH-
(XQAH-
nnnn)
nnnn)
7.Until repaired, this well may

be operated ONLY by flowing
Has the
to the test separator (via
operation of
XT-11001).
this well been
• Advice Maintenance
changed in the
Manager.
last 20 minutes?
• Advise Well Programming
END
5. Sand production suspected.

3. Cancel the L1 • Route the well to the test separator
Alarm condition • Display a trend graph of the two sand detectors
(XIA-nnnn and XIA-11001)
• Bean-back the well until no sand is indicated, or if
sand continues shut-in the well. The minimum well
flowrate permitted is 30 MMscf/day, otherwise sand
Does it may accumulate in the well/tubing.
repeatedly • Advise Asset Manager of suspected sand production
re-activate? from this well.
• Request Well Programming to analyse data onshore,
and review alarm set-points.
END
4. Spurious alarm.
• Request Well Programming to analyse data
onshore, and review alarm set-points with
Production Technology. END
AEA Technology 68
SSMM-002a: Actions to be taken in the event of a sand alarm

(Production header sand detector – Part 1)
Purpose: To enable the Shearwater control room to respond appropriately to sand detector alarms
(excluding alarms which may be triggered during work to test the functionality of the detectors).
2. An alarm has been activated on

1. From the sand detector on the
Procedure 2 production header (XT-01201).
If an alarm comes up on an
individual flowline sand detector,
revert to Procedure 2.
Is it an L1
Is it an M 6. Malfunction alarm
type alarm?
type alarm? (XMA-01201)
(XAH-
(XQAH-
01201)
01201)
7. Until repaired, Shearwater has a

reduced capability to monitor for
Have wells
sand production.
been switched
• Advice Maintenance Manager.
to/from the
• Advise Prod. Tech.
production
END
manifold in
the last 20
minutes?
5. Sand production is suspected on one or more

wells. Continue on Procedure 2b.
3. Cancel the L1
Alarm condition
4. Spurious alarm.
• Record which wells were
Does it re- flowing to the production header
activate? at this time.
• Request production team to
analyse data onshore and review
alarm set-points.
END
AEA Technology 69
SSMM-002b: Actions to be taken in the event of a sand alarm

(Production header sand detector – Part 2)
Purpose: To enable the Shearwater control room to respond appropriately to sand detector
alarms(excluding alarms which may be triggered during work to test the functionality of the
detectors).
2. Sand production is suspected,

1. From based on an alarm on sand
Procedure 2a detector on the production
header (XT-01201).
If an alarm comes up on an
individual flowline sand detector,
revert to Procedure 2.
3. Call up a trend graph with

each of the sand detectors on the
flowlines which are currently
routed to the production
manifold. (XIA-nnnn)
4. Sand production suspected from one well.

• Route the well to the test separator
• Display a trend graph of the two sand detectors
Does any (XIA-nnnn and XIA-11001)
• Bean-back the well until no sand is indicated, or if
one well
have a sand sand continues shut-in the well. The minimum well
rate higher flowrate permitted is 30 MMscf/day, otherwise sand
than the may accumulate in the well/tubing.
• Advise Asset Manager of suspected sand production
other wells?
from this well.
• Request Well Programming to analyse data
onshore, and review alarm set-points.
END
5. Sand production suspected, unknown source. Possible configuration

mismatch between flowline and production header sand detector.
• Advise Asset Manager of suspected sand production, unknown source.
• Request URGENT onshore data review and review of alarm set-points.
• Production may continue with close monitoring of flowline sand
detector outputs and nucleonic sand level sensor in the production
separator (when/if fitted). If sand is seen in the separator, refer to
procedure 11.
END
AEA Technology 70
SSMM-003: Changing calibration coefficients and alarm settings

Purpose: To implement new values in the sand detector software (supplied from onshore data
review & already approved for implementation through software change control system).
1. START: 2: Display the ClampOn multi view

screen (software Multitrend running
on the ClampOn rack mounted PC)
3: Use the +/- screen buttons to select

the sand detector which needs to be
updated.
Do you want
to update the 4: Click on the Menu button and select ‘[F1] Alarm
alarm Settings’
settings?
5: A Password box will appear. Enter the Shearwater

Operator (single) password and click ‘ok’.
6: A box will appear containing the current alarm

settings for this detector (L1, L2, L3, M), and two
11. Continue on to
tick boxes for disabling alarms (which should be
Procedure SSMM-003a.
blank, to indicate that the alarms are active). The
sand detector sensor number is also shown.
7: In turn, click in the entry field for each of the L1,

L2, L3 and M settings, type in the new value, and
press the Enter key to register the new entry. Note
that a comma is used as a decimal point, so that
10.00 appears as 10,00
Are the alarm

settings all 8: Click on ‘[F10] Save’. You are also
correct? asked to confirm the save operation.
9. The new settings take immediate effect.

• To view the settings, select the graph trend
screen for this detector and select ‘Alarm
10: Either go to Step 7 to correct Settings’ from the ‘Select Statistics’ pull-down
the entry or Click on ‘[F9] Abort’ menu.
and Go back to Step 1 to start again. • Advise Production Technology that the
requested change has been implemented.
END
AEA Technology 71
SSMM-003a: Changing calibration coefficients and alarm settings

(Part 2)
Purpose: To implement new values in the sand detector software (supplied from onshore data
review & already approved for implementation through software change control system).
1. Continued from 2 The ClampOn multi-view screen is

SSMM-003 displayed and the detector to be updated
has been selected.
Do you want
to update the
3: Click on the Menu button and select ‘[F5] System
ZERO &
Setup’
STEP
settings?
4: A Password box will appear. Enter the Shearwater
Full access password and click ‘ok’.
5:A ‘General Settings’ box will appear. From this

box, other aspects of the configuration may be
altered. This procedure is only concerned with
11. This procedure does not
changes to the flow table. Click on ‘[F5] Flow Table’
apply, revert to Procedure
SSMM-003.
6: A box will appear containing the current flow
table for this detector (Columns of Zero, Step and
Exp next to a column of flow velocities), The sand
detector sensor number is also shown.
7: In turn, click in the entry field for each of the

Zero and Step settings, type in the new value, and
press the Enter key to register the new entry. Note
that a comma is used as a decimal point, so that
10.00 appears as 10,00. Values which are to remain
unchanged need not be re-entered.
Are the
settings all 8: Click on ‘[F10] Save’. You are also asked
correct?\ to confirm the save operation. Also click on
‘[F10] Save’ at the ‘General Settings’ box.
9. The new settings take immediate effect.

10: Either go to Step 7 to correct • Advise Production Technology that the
the entry or Click on ‘[F9] Abort’ requested change has been implemented.
and Go back to Step 1 to start again. END
AEA Technology 72
SSMM-004: Minimising sand production by gradual bean-up

Purpose: To establish whether sand production can be minimised by bringing a known sand
producing well on-line more gradually (i.e. slower bean-up rate).
Limited Applicability: This procedure only applies to wells which are known to be prone to sand
production. A test to establish the maximum sand-free rate (procedure 5) should have already been
conducted for this well. This procedure therefore assumes that the sand production is primarily
caused by the shocking of the reservoir rather than the steady state production rate.
1. START 2. Obtain a copy of the test to

(choke closed) determine the maximum sand-free rate
(SFR) for this well. Trend XIA-nnnn.
3. Open the isolation valves and open

the choke valve to just clear the dead
band (at a FTHP fall rate of 5bar/ 5min)
4. Wait for a 5. Open the choke to reduce the FTHP by an

minimum of 20 amount equal to one fifth of the difference
minutes between the CITHP and the SFR (at a FTHP fall
rate of 5bar/ 5min)
Is sand
indicated Has the SFR Is the choke
now? been 100% open?
reached?
6. Leave the choke

alone, monitor sand 7. Continue operation of this well at 9. Failure to agree
for 20 minutes. this rate. Minimise the number of with test of SFR.
changes to the operation of this
well.
END
Did the sand 8. Close the well in. 10. Continue flowing
decay away? Request advice from the well.
onshore Production Request advice from
Technology. Advise onshore Production
Asset Manager. END Technology.
END
AEA Technology 73
SSMM-005: Finding the maximum sand-free production rate

Purpose: To establish the minimum flowing tubing head pressure (maximum choke opening
position) at which a well can be operated while remaining sand-free.
1. START: Close the choke valve

Complete the following entries:
Well : ____ Sand detector: _________
Date: ___________ Time: _________
2. Record the closed-in tubing head

pressure (CITHP). Trend XIA-nnnn.
CITHP = ___________
3. Open the isolation valves and open

the choke valve to just clear the dead
band (at a FTHP fall rate of 5bar/ 5min)
5. Record the FTHP

4. Wait for a 6. Open the choke
FTHP = _______
minimum of 20 a further 5% (at a
FTHP = _______
minutes FTHP fall rate of 1
FTHP = _______
bar/min).
FTHP = _______
Has the
Is sand target
indicated flowrate Is the choke
now? (XIA- been 100% open?
nnnn) reached?
7. Leave the choke

alone, monitor sand 10. The well can be operated 11. Target flowrate
for 20 minutes. sand-free at this FTHP. The cannot be achieved
Sand Free Rate is equal to with this choke
the current FTHP. trim.
SFR = _______ barg
12. For now, the Sand
Did the sand Free Rate is equal to
9. The Sand Free Rate is the current FTHP.
decay away?
equal to the last FTHP SFR = _______ barg
recorded in Box 5.
SFR = _______ barg
8. The Sand Free Rate is Advise Production Technology of

equal to the current FTHP. the result.
SFR = _______ barg END
AEA Technology 74
SSMM-006: Checking for choke valve trim damage

Purpose: To establish whether there is evidence of erosion damage to the choke valve control trim
(plug/cage). (This is important because damage to the trim can promote damage to the valve body
exit region and to the pipework immediately downstream of the choke.)
Theory: drift in flowrate response

1. Collate the well test results from this well
2. Plot THFP (Tubing Head Flowing Pressure) vs choke opening position
3. Plot gas flowrate vs choke opening position
4. Plot gas flowrate vs THFP
5. Estimate the gas flowrate which would be expected to be obtained at the choke position used for
this test (based on previous results)
6. Expected trend: With time (successive well tests) there should be little change, or there may be a
decrease in gas flowrate at fixed choke position, due to reservoir decline.
7. Unexpected trend: If erosion damage has occurred, the gas flowrate from this test will be greater
than expected (from previous results). However, the gas flowrate from this test should be
consistent with the FTHP used for this test. (i.e. FTHP is correspondingly low)
Practical: leak test

1. Shut-in the well.
2. Route the well to the test separator.
3. With the Production Wing Valve closed, record the gas flowrate through the test separator over
a 10 minute period (to check for a leak back through the manifold valves, and to account for any
offset on the flowmeter).
4. With the choke valve closed (0%), open the isolation valves.
5. Record the gas flowrate through the test separator over a 10 minute period.
6. Compare the leak rate with a previous leak test for this well, or with leak tests on other chokes.
Notes:
1. This test provides an insight into the state of the choke valve seat rather than the control trim. Since erosion damage
when the well is flowing may not occur on the seat, this test does not confirm that the control trim is undamaged.
However, if the valve leaks excessively, then this suggests that the seat is damaged (eroded), which also indicates that
the control trim may also be damaged (eroded).
2. This test also depends on the reliability and accuracy of the gas flowrate measurement on the test separator.
3. Choke valves are not intended to provide a tight shut-off, and a degree of leaking is to be expected. This test is
looking for substantial leakage. As a guide:
• A leak rate of less than 1% of the normal gas flowrate from this well would not be regarded as concerning.
• A leak rate exceeding 5% of the normal gas flowrate from this well should initiate action to replace the valve trim
(refer to procedure 7).
• An early re-test should be scheduled if the leak rate is in the range 1 – 5% of normal gas flowrate from this well.
AEA Technology 75
SSMM-007: Action to be taken if damage to a choke valve is

suspected or confirmed
Purpose: To ensure that appropriate action is taken to minimise the risk to personnel and the
platform.
Damage to the valve body (non-intrusive inspection, TOFD or 0° compression probe):

1. Shut down the well for further inspection work under conditions which are more favourable to
inspection.
4. Recheck the choke valve body damage to confirm the original results.
5. Check for erosion damage of the pipework immediately downstream of the valve.
6. If the ‘apparent’ damage has gone from the choke valve, and there is no damage evident on the
pipework immediately downstream of the choke, resume production. Request onshore review
of sand detector historical data (Well Programming).
7. If erosion damage of the choke body remains evident (i.e. is confirmed), do not resume
production from this well. Refer the results to the certifying authority for permission to resume
production from this well while a replacement valve is obtained. It may be appropriate to use
finite element modelling to assess the probability of failure with the known level of damage. A
pressure test may also be used to gain confidence that the valve body will continue to withstand
the design pressure. The choke valve vendor may also provide useful input.
8. While the well is inoperable, it may be advantageous to seek additional confirmation of the
damage by removing the valve in order to assess erosion damage, in the outlet region, by visual
means. Damage in the gallery region may or may not be visible from the inlet. Bonnet removal
and dismantling of the valve may or may not be possible in the well-bay. Request onshore
review of sand detector historical data (Well Programming).
Damage to the control trim (from flowrate drift or leak test – see procedure 6):
Damage to the control trim does not threaten plant integrity. Nevertheless, damage to the control
trim can promote erosion of the valve body (outlet region) and erosion of the downstream
pipework. The following actions should be taken:
1. Ensure that the flowrate from this well will be continuously monitored by PCS and that the
valve position will be automatically corrected to maintain a set target condition (usually in
FTHP). This ensures that the well will not ‘run away’ if the valve trim fails abruptly.
2. Schedule non-intrusive inspection of the choke valve and the pipework immediately
downstream of the choke valve.
3. Display a trend of the sand detector output.
4. If possible, route the well to the test separator so that the flow passes two sand detectors and any
sand that is produced can be more easily removed from the process.
5. Resume operation from this well. Monitor for evidence of sand production on the flowline and
test header sand detectors. Action Procedure 2 if sand production is evident.
AEA Technology 76
SSMM-008: Flushing the first stage or test separator
Purpose: To outline the process of removing sand from the separators. (This is important because it
limits the progress of sand through the produced water system.)
Sand cyclone loop (on-line):
In view of the complexity of this operation, it is recommended that the test separator is cleaned in
this way before the first stage separator is attempted. The off-line version of this operation presents
fewer challenges and is documented on the following page.
1. Configure temporary pipework, pump, and the sand cyclone/accumulator to form a loop,
drawing produced water from the sand outlet and returning the cleaned water through the sand
wash nozzles. (see section 4.3.2)
2. Reduce the throughput of produced water to a minimum. Determine the maximum period of
time for which the desanding loop can be operated (based primarily on the produced water
flowrate into the vessel and the volume available between low and high conditions).
3. Bleed the sand cyclone loop.
4. Lower the produced water level to the minimum level (without tripping on a low level ESD).
5. Close the produced water outlet valve (to minimise sand drawn into the produced water outlet).
6. Start the sand cyclone loop pump and run for the agreed duration (set in step 2). The pressure of
the water pumped to the sand wash system should be no more than 0.7 bar greater than the
pressure in the vessel (to minimise the risk of erosion damage to the vessel wall near to the sand
wash nozzles). Stop the desanding loop pump.
7. Allow a time interval (set in Step 2) for sand which has become fluidised in the produced water
layer to settle.
8. Open the produced water outlet valve.
9. Repeat steps 4 to 8 as required to clean the separator. Once fitted, the nucleonic sand level
sensors in the first, second and test separators should provide some evidence of sand removal
success. Temporary sand detectors (installed on the sand cyclone inlet and produced water
return) would also assist in confirming success.
10. Allow the sand cyclone loop to cool (to below 70°C).
11. Depressurise the sand cyclone loop.
12. Estimate the quantity of sand obtained for audit/discharge purposes, and take a sample for sand
characterisation and size analysis.
13. Flush the sand from the accumulator (discharge in accordance with prevailing environmental
regulation / policy).
AEA Technology 77
Desanding cyclone loop (off-line):
1. Configure temporary pipework, pump, and sand cyclone/accumulator to form a loop, drawing
produced water from the sand outlet and returning the cleaned water through the sand wash
nozzles. (see section 4.3.2)
2. Determine the period of time for which the sand cyclone loop will be operated (based primarily
on the throughput of the sand cyclone/pump and the volume to be removed). Alternatively
determine the criterion to be met (e.g. minimal sand flowrate in the flow from the separator).
Determine the minimum time interval from completion of the sand removal operation before
the separator may be returned to service (to allow sand which has become fluidised in the
produced water layer to settle).
3. Bleed the sand cyclone loop.
4. Shut down operation of the separator (no flow in, no flow out, still pressurised and hot).
5. Start the sand cyclone loop pump and run for the agreed duration (set in step 2) or until the
agreed criterion has been met. Once fitted, the nucleonic sand level sensors in the first, second
and test separators should provide some evidence of sand removal success. Temporary sand
detectors (installed on the sand cyclone inlet and produced water return) would also assist in
confirming success. The pressure of the water pumped to the sand wash system should be no
more than 0.7 bar greater than the pressure in the vessel (to minimise the risk of erosion damage
to the vessel wall near to the sand wash nozzles).
6. Stop the sand cyclone loop pump and isolate it from the separator/process.
7. Allow a time interval (set in Step 2) for sand which has become fluidised in the produced water
layer to settle.
8. Resume process flow through the separator.
9. Allow the sand cyclone loop to cool (to below 70°C).
10. Depressurise the sand cyclone loop.
12. Flush the sand from the accumulator (discharge in accordance with prevailing environmental
regulation / policy).
Shut-down and remove by hand:
1. Vessel entry as per platform procedures.

2. Care should be exercised not to damage the sand washing pipework in the bottom of the vessel.
AEA Technology 78
SSMM-009: Action to be taken in the event of repeated Low

Liquid-level trip on Test or First Stage Separator
Purpose: To highlight this as potentially being a sand-related problem and assist in the diagnosis.
Test separator:
Since the level sensor is via a nucleonic sensor (and not a bridle), erosion of the produced water
level control valve trim (LCV11021) is the only link to sand production (valve passing when closed).
1. Check the nucleonic sand level sensor (when fitted) for indications of sand being present in the
separator. Request onshore review of sand detector output to check for sand which was not
sufficient to trigger an alarm.
2. Check records to determine whether this problem occurs when flowing from a particular well or
at a particular flowrate. If due to erosion of the PW level control valve trim, the low level trip is
most likely to occur when flowing a well with a low water cut and a low production rate. Check
records to estimate when this problem began. Erosion due to sand is unlikely to occur
immediately (for a new field).
3. If the other isolation valves (if any) in the produced water line are left open when the separator is
not in use, does the liquid level decline. If the trips are only associated with periods of shut-
down, use a manual valve in this line (with suitable management protocol to ensure that it is
opened when needed).
4. If the checks indicate that sand damage is a possibility, schedule the removal of the valve trim for
replacement. If trim erosion damage is found, schedule a sand wash of the separator (see
procedure 8). Consider upgrading the trim material to a tungsten carbide (erosion resistant
specification) if available for this valve.
AEA Technology 79
The produced water level indication is by way of a 3″ bridle whose off-take is at the bottom of the
separator. Sand accumulation in this line may cause the response to be delayed compared with the
level in the separator. Erosion of the trim in the produced water level control valve (HP
hydrocyclone inlet – FCV-40152, V-4015) could also explain this behaviour.
1. Check the nucleonic sand level sensor (when fitted) for indications of sand being present in the
separator. Request onshore review of sand detector output to check for sand which was not
sufficient to trigger an alarm.
2. Check records to estimate when this problem began. Problems due to sand accumulation or trim
erosion are unlikely to occur immediately (for a new field).
3. If the checks indicate that sand damage is a possibility, or if other possibilities are ruled out,
schedule the removal of the valve trim for replacement. If trim erosion damage is found,
schedule a sand wash of the separator (see procedure 8). Consider upgrading the trim material to
a tungsten carbide (erosion resistant specification) if available for this valve.
4. Consider the options for flushing the level control bridle (either flowing produced water to drain
or pumping water into the separator through the bridle off-take).
AEA Technology 80
SSMM-010: Non-intrusive inspection

Purpose: To highlight specific concerns in the practical application of non-intrusive inspection on
Shearwater. For details of the areas to be inspected, refer to the guidance notes (sections 5.3 – 5.7,
summarised in section 5.10).
Preparation:
1. Check with Well Programming (onshore) to ascertain whether sand production has been
evident.
2. Identify the components to be inspected, and refer to the relevant guidance notes which describe
the area(s) to be covered.
3. Identify the hazards specific to the inspection of the particular component and identify ways in
which the risks may be minimised. In general, the hazards are expected to include:
• Access (working at height),
• Working in close proximity to hot surfaces (lagging removed),
• Possible contact with hot couplant fluid (e.g. drips),
• Prolonged working in a relatively stressful and hot environment,
• Movement of the work-piece (e.g. due to slugging within the line),
• Noise (particularly when working on/next to the choke valves).
Risks may be minimised/eliminated by:

• Use of suitable scaffolding or steps
• Awareness of the hazards.
• Taking breaks to re-evaluate work progress.
• Wearing appropriate protective equipment.
• Use of water-cooled sensing heads.
• Provision of clean cold water nearby to treat scalds.
• Shutting the flowline down to eliminate excessive vibration.
• Shutting the line down for a prolonged period to allow it to cool.
• Noise survey
Inspection:
1. Remove the lagging to expose the area to be inspected
2. On the first occasion, mark up the grid area to be used (with a marker that is durable under the
prevailing conditions).
3. Locate the calibration block and position it such that it will attain a temperature similar to that of
the work-piece.
4. Use a couplant which is suitable for the surface temperature (150°C) and which is compatible
with 25Cr Duplex.
5. Scan 100% of each grid square, recording the wall thickness at the grid corners and the minimum
wall thickness at any location in the grid square.
AEA Technology 81
Completion:
1. Replace any lagging that has been removed and seal (tape) around the edges to prevent ingress of
water.
Reporting:
Results for the flowlines, test header and production header will be held within the Integrity
Management System for the Shearwater platform and this reporting system will be maintained by
the Integrity Management Contractor (OIS).
The primary requirement is that changes in wall thickness (from previous inspections) are
highlighted.
Where damage is evident, the following steps should be implemented (section 5.9):
• Increased monitoring (resolution, and frequency) of the component which had been found to be
damaged
• The ‘equivalent’ components in the other flowlines must also be checked.
• Onshore review of the sand detector output may also be helpful.
AEA Technology 82
SSMM-011: Action to be taken if sand is evident from the

nucleonic sand level sensor in the separator(s)
Purpose: To ensure that the appropriate actions are taken if sand becomes evident by way of
accumulation in the separators.
1. Record daily and advise asset manager.

2. Request onshore data review of the sand detector output. (Well Programming)
3. If no indication is given by the sand detectors, implement a programme of rotating the wells
through the test separator to better diagnose sand production.
4. Once the well has been identified, refer to asset manager for approval to flow this well to test
separator in order to confirm sand production, and then attempt to determine the maximum
production condition for sand-free operation.
Test separator:
1. Check sand level sensor records, and cross-check with onshore review of sand detector output
(Well Programming) to ascertain which well has been producing sand.
2. Once the well has been identified, refer to asset manager for approval to flow this well to test
separator in order to confirm sand production, and then attempt to determine the maximum
production condition for sand-free operation.
AEA Technology 83
Appendix 2
Non-intrusive inspection
CONTENTS
Introduction and Terminology
1 Spot measurements (manual)

2 Line scanning (manual)
3 Area and grid scanning (manual)
4 Automated area scanning (corrosion mapping)
5 Fixed probes
5 TOFD
6 Other considerations relevant to Shearwater
7 Other inspection techniques considered
8 Summary of non-intrusive inspection techniques
AEA Technology 84
INTRODUCTION AND TERMINOLOGY
This appendix describes the non-intrusive inspection techniques that are commonly
used in the industry, and may be used on Shearwater. In each case advantages and
disadvantages are listed. The following terminology is used:
On-line Production is continued during the inspection work. Hence, the line
will be hot (approx 150 °C), and may move depending on the flow
inside the pipe.
Off-line Production is stopped during the inspection work, and sufficient time
has elapsed to allow the pipework to cool to more acceptable
temperature (approx 70 °C). The pipework will not be moving.
PoD Probability of Detection: the likelihood that damage located
somewhere on the piece being inspected is found.
1. SPOT MEASUREMENT
Isolated point ultrasonic wall thickness measurements taken at specific locations

such as at the outsides of bends, at intervals along piping systems, at Tee
intersections, etc. Two types of instrumentation are possible, thickness meters (D-
meters) or ultrasonic flaw detectors.
Advantages
Readily available, low cost and fast. D-meters can be operated by non-inspection
qualified staff. Data loggers are available which store data for off-line retrieval and
processing by data management software.
Disadvantages
PoD is poor due to very low coverage. Localised attack could easily be missed.
Prone to human error. For long-term monitoring purposes it is extremely difficult
to re-locate the probe in the same position unless permanent marking or bosses are
used.
On/off-line
Lagging would need to be removed locally where measurements are to be taken.
Access could be made through specially constructed access ports with removable
AEA Technology 85
covers. Process vibration could make measurement difficult. Due to the small
number of measurement sites, locations could be selected which are more readily
accessible.
2. LINE SCANING
Ultrasonic wall thickness measurements taken by scanning a probe along a line at

strategic locations on each item of piping. This could involve, for example, a
complete circumferential scan around a pipe at each end, adjacent to welds or
fittings where the minimum observed thickness is recorded. Similar circumferential
scans would be carried out adjacent to welds on Tee’s and axial scans would
additionally be carried out on the external radii of bends and along manifolds/inlet
headers opposite inlet pipes.
Advantages
Readily available, low cost and fast. More likely than spot measurements to detect
localised damage if the locations have been well selected.
Disadvantages
D-meters not applicable to this form of scanning. PoD is still low due to relatively
low coverage. Prone to human error. For long-term monitoring purposes,
probability of re-locating the probe at the same position at successive inspections is
low.
On/off-line
More lagging would need to be removed locally where measurements are to be
taken. Access could be made through locally removable sections of insulation.
Process vibration could make measurement difficult. Dynamic scanning is more
difficult at elevated temperatures than spot measurements.
AEA Technology 86
3. AREA/GRID MEASUREMENT
d
d
Ultrasonic wall thickness measurement taken by marking the item into square grids.
The wall thickness at each grid intersection is recorded along with the minimum
wall thickness measured in each grid area. Grid areas are scanned 100%. The grid
dimensions (d) are usually equal to one quarter of the pipe circumference.
Some ultrasonic flaw detection equipment has ‘echo-dynamic’ capabilities which

effectively stores all ultrasonic responses within a scan so that the minimum wall
thickness recorded can be observed. This is a potentially useful feature for high
temperature inspection, or if a large area is to be scanned.
Advantages
Readily available and reasonable cost. More likely than key point measurements to
detect localised damage as it will detect off axis erosion.
Disadvantages
Slower than spot or line measurement, as much more scanning is involved along
with increased data recording/reporting. PoD improves due to greater coverage,
but is still prone to human error because there is no auditable record of the area
which has supposedly been scanned. For long-term monitoring purposes,
probability of re-locating the probe at the same position at successive inspections is
reasonable since the entire area is covered.
On/off-line
All lagging would need to be removed at areas requiring inspection. Process
vibration could make measurement difficult. Scanning is more difficult at elevated
temperatures than key point measurements and some locations could be very
difficult, if not impossible for extended periods of access (e.g. manifolds).
AEA Technology 87
4. AUTOMATED AREA MAPPING (ALSO KNOWN AS CORROSION

MAPPING)
Ultrasonic wall thickness measurements are taken by scanning an area with

simultaneous signal processing to convert wall thickness measurements into colours
to represent defined wall thickness intervals. The position-related colour-coded
wall thickness measurements are plotted using a PC to generate a two dimensional
topographical map of wall thickness or C-scan.
There are two approaches to tracking the probe over the work-piece, automated
and semi-automated. In a semi-automated case approach, the probe is moved by
hand but positional information is automatically recorded (either by mechanical
linkages or by an optical tracking device). In an automated approach the tracking of
the probe over the workpiece is also automated.
Most commercially available equipment allows the spatial resolution of the scan to
be adjusted, and is typically set at 4mm x 4mm for general corrosion assessment and
10mm x 10mm for erosion. Typical individual scan areas are 90° or 120° segments
by 300mm or 500mm of pipe length. Data from individual scan areas can be re-
displayed off line for analysis and some systems have the facility for compiling
composite images. In some cases, corrosion maps have been used within finite
element models to determine the maximum safe operating pressure.
Advantages
Very high PoD due to auditable evidence of coverage – good signals must be
obtained or the readings are not plotted, i.e. many of the potential human errors of
inspection are eliminated. Calibration scans can be carried out and recorded pre-
and post-scanning to demonstrate accuracy of measurement. Monitoring is
improved as images can be assembled together to form composite images. Erosion
behaviour can be more readily observed when successive images are observed side
by side and localised trends can also be recognised more easily.
AEA Technology 88
Disadvantages
Specialised, costly service. Low productivity, partly because good signals must be
obtained from every point scanned.
On/off-line
All lagging would need to be removed at areas requiring inspection. Process
vibration could make probe tracking difficult and may render optical tracking
unsuitable. Scanning is more difficult at elevated temperatures and some locations
could be very difficult, if not impossible for extended periods of access (e.g.
manifolds). Mechanical scanning devices may not cope with geometries requiring
inspection.
AEA Technology 89
5. FIXED PROBES
Strategically located probes can be permanently fixed in position using an

appropriate adhesive. These are positioned under the insulation and a flying lead
enables wall thickness measurements to be taken without removal of insulation
material. Probes can be held in place using a template for re-positioning purposes of
any probes which may fail or become dislodged with time.
Advantages
Many of the potential human errors are eliminated. One or more probes could be
permanently attached to calibration sample(s) attached to pipe. Calibration
measurements can be carried out and recorded pre- and post-measurement to
demonstrate accuracy. Long-term monitoring is very accurate as probes are always
at exactly the same location with identical coupling conditions. Rates of wall loss
can therefore be very accurately ascertained. Data could be collected by NDT
technicians with proprietary equipment or, if initial investment was made, with
smart ‘plug & play’ system with data logger for off-line analysis. Therefore, if
erosion rates are high, frequent measurements could be made without the need for
mobilising inspection crews.
Disadvantages
Relatively low PoD due to small surface area covered unless a high number of
probes are used. Even if probes are strategically targeted, there is a possibility that
erosion could occur at other locations. Initial high cost of installation, but low cost
thereafter. Commercially available probes are rated for continuous use at 260°C1,
but some failures must be expected with time. Probes could become dislodged or
physically damaged. Adhesive could also degrade with time. Vibrations could
cause connections to work loose over extended periods.
On/off-line
Lagging would not require removal. Process vibration would not affect
measurements.
1
The probes must be attached to the surface using a glue or mechanical device which is also suited
for continuous exposure to 260°C. Fleximat is currently only rated for exposure to a surface
temperature of 125°C.
AEA Technology 90
6. TOFD INSPECTION OF CHOKE VALVE BODY
B
Receiver A
Ligament
Transmitter
Erosion is expected to occur at the areas of the gallery as indicated above. TOFD
inspection as shown will determine the ligament of material remaining between the
bore of the gallery (Signal B) and the chord between the ultrasonic probes (Signal
A). Grey scale ultrasonic imaging techniques are used to display the data as the
probes are scanned along the length of the choke valve body and remaining
ligament information can be calculated using data processing routines. Ultrasonic
waveforms are digitised and stored for off-line retrieval, comparison, and analysis.
A less accurate derivative of this technique is to perform the scan using

conventional flaw detection equipment and comparator test samples with simulated
erosion at various depths. The technique could also be applied using the fixed
probe approach. An additional, compression probe technique could be applied
from the sides of the choke body to interrogate the tangent of the bore closest to the
outer surface. This technique could be applied by either manual scanning with a
conventional flaw detector or by corrosion mapping.
Advantages
Many of the potential human errors are eliminated if TOFD imaging system is used.
Data is analysed off-line, allowing the operator to concentrate on collecting quality
data rather than having to simultaneously interpret the images. Accuracy of
measurement is higher with an imaging system which leads to a corresponding
accuracy for wall thickness monitoring. Base-line TOFD inspection with
subsequent fixed probe option is expected to give high levels of reliability with the
low cost option of manual TOFD measurement during service without the need for
removal of insulation material. The manual UT measurements would be more
accurate if a digital flaw detector with A-scan storage facility and off-line processing
capability is available.
Disadvantages
Relatively high cost of inspection requiring specialised inspectors with advanced
TOFD equipment.
AEA Technology 91
On/off-line
Lagging would not require removal if fixed probes are used. Process vibration is not
expected to affect measurements.
6. OTHER CONSIDERATIONS RELEVANT TO SHEARWATER
Material
A study was carried out by the Harwell Offshore Inspection R&D Service (HOIS)
as part of a ‘New Materials Programme’ in 1997 (Ref HOIS(97)P4 Issue 0.1), and
concluded that duplex stainless steel parent material (as opposed to weld material
and heat affected zones) were no more difficult to inspect with the methods used
than ferritic steels of the same geometry. This is in line with AEA Technology’s
field service experience. There are however, anecdotal references in the above
mentioned report to difficulties with the detection of erosion and corrosion in
heavy section duplex stainless steel pipe. The use of calibration step wedges
manufactured from the same material as the pipes to be inspected will provide
assurance of wall thickness measurement integrity.
Temperature effects
Ultrasonic velocity and attenuation are influenced by temperature. Both will
increase at elevated temperatures, but are not likely to be significant over the
temperature and wall thickness ranges of concern. However, it is good practice to
attach calibration blocks of the same material to the component surface such that the
block attains a similar temperature to the pipe itself.
• The taking of spot thickness measurements at temperatures up to 150°C is
relatively straightforward provided reasonable access is available.
• Access appears from drawings to be particularly difficult, especially to the
underside of the inlet header and this is likely to be extremely uncomfortable at
operational temperature, possibly unsafe. The impact of this can only be
accurately assessed through site trial.
• Scanning an area becomes more difficult as the viscosity of high temperature
couplants vary with temperature and they dry off quickly.
• Overhead scanning creates safety concerns of hot couplant falling onto the
technician.
• Scanning aids are available for use by technicians during inspection of hot
surfaces. A water-cooled probe jacket is recommended for use since it is easier
to hold, as well as protecting the probe.
• Suitable protective clothing must be worn by the technician when working
on/near surfaces which are hot. Air cooled protective suits may be appropriate
in some high temperature situations.
AEA Technology 92
7. OTHER INSPECTION TECHNIQUES CONSIDERED
Flash radiography
This is intended for imaging the external profile of insulated material for the
detection of Under Insulation Corrosion (UIC) by taking tangential shots. This
technique is considered unsuitable for this application as it does not penetrate the
pipe wall thickness to image internal wall condition.
CHIME
This is a long range ultrasonic inspection technique which requires the material to
be flooded with ultrasonic waves using separate transmitter and receiver. The
waves can be propagated around the pipe circumference or axially with distances
between the probes of up to 1m. The best achievable resolution of the system is
10%, i.e. wall loss would need to be 4mm in a 40mm thick pipe section before it
would be detectable using this technique. This resolution is considered to be too
coarse for this application.
Pulsed eddy current

Two systems are available, one licensed from ARCO by the Dutch NDT company,
RTD called INCOTEST, and another Shell Global Services in-house
development. Both organisations have been contacted.
RTD has confirmed that the technique is not suitable for duplex stainless steels.
SGS are investigating the matter and there does seem to be some possibility that this
technique may work. The technique allows wall thickness measurements to be
made through insulation material at operating temperature. Measurement
accuracies of ±5% of wall thickness are typical. If pulsed eddy current does work in
this application it could prove to be a very cost effective on-line screening tool.
AEA Technology 93
8. SUMMARY OF NON-INTRUSIVE INSPECTION TECHNIQUES
Pipework and fittings
Lagging removal
Bends, tees
(hot/flowing, or
cool and static)
threshold (mm)
Monitoring of
Availability of
Probability of
detection (%)
Speed of data
and straight
erosion rate
Line status
pipe runs
Detection
technique
collection
(if fitted)
Cost
Spot readings On-line ***** **** **** **** * 2.0 20
Off-line ***** ***** ***** **** * 1.0 20
Line scans On-line ***** *** *** *** ** 2.0 50

Off-line ***** **** **** *** ** 1.0 50
Grid/area On-line ***** ** ** * *** 2.0 90

scans
(manual) Off-line ***** *** *** * *** 1.0 90
Area scans On-line * * * * **** 0.25 95
(automated) Off-line * ** ** * **** 0.25 95
Fixed probes On-line *** *** ***** ***** ***** 0.1 50
Off-line *** *** ***** * ***** 0.1 50
Clearly, corrosion mapping is preferable on the grounds of accuracy and reliability,

but the need for lagging removal, cost and speed are limitations. It is proposed that
this technique is held in reserve for the event that critical information is required.
Spot readings and line scans are attractive in terms of practicality, but are not
considered to deliver the required information on component condition in this
application. The off-line grid/area scanning survey offers the best all round
capabilities, but these are expected to be limited to an as yet, unknown degree when
performed on-line.
AEA Technology 94
Choke valve body
Lagging removal
Choke
(hot/flowing, or
cool and static)
threshold (mm)
Monitoring of
Availability of
Probability of
detection (%)
Speed of data
Valve Body
erosion rate
Line status
Detection
technique
collection
(if fitted)
Cost
Manual U T On-line ***** ** ** * *** 2.0 90
Off-line ***** *** *** * *** 1.0 90
Corrosion On-line * * * * **** 0.25 95

mapping
Off-line * ** ** * **** 0.25 95
Fixed probes On-line *** *** ***** ***** ***** 0.1 90

Off-line *** *** ***** * ***** 0.1 90
TOFD On-line * * ** * **** 0.3 95
imaging
Off-line * * *** * **** 0.3 95
TOFD manual On-line *** *** ** * ** 2.0 90
Off-line *** *** *** * *** 1.0 90
Corrosion mapping and TOFD with imaging capability are preferable in terms of
detection, but fixed probes are preferable on the grounds of cost and erosion rate
measurement.
AEA Technology 95
Appendix 3
Title
CONTENTS
AEA Technology 96
RESTRICTED - COMMERCIAL AEA Technology

Sand Management

Uploaded by

Copyright:

Available Formats

Sand Management

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sand Management

Uploaded by

Copyright:

Available Formats

What is the purpose of the document?

What is the purpose of the document?

What risks are associated with sand production?

What risks are associated with sand production?

RESTRICTED-COMMERCIAL

A report produced by AEA Technology for

T. J. Lockett, J. Krol, J. Lilley

Title Shearwater Sand Management Manual

Customer Shell UK Exploration & Production

Customer reference SWR-G-255

Confidentiality, Restricted - Commercial

File reference AEAT-RHAD/RD01762/12/1

Report number AEAT-RHAD/RD01762/12/1

Report status version 3 (final)

AEA Technology is the trading name of AEA Technology plc

Name Signature Date

Author T. J. Lockett Timothy Lockett 28/9/2000

Reviewed by J. Krol James Krol 28/9/2000

Approved by G. Eckold Geoff Eckold 28/9/2000

RESTRICTED - COMMERCIAL AEA Technology 2

Following these guidelines and procedures will ensure that:

• the Shearwater platform follows industry best practice in managing sand

RESTRICTED - COMMERCIAL AEA Technology 3

3 Minimising sand production and its effects 31

4 Other indicators of sand production 34

RESTRICTED - COMMERCIAL AEA Technology 4

4.3.1 Accumulation of sand in the test and first stage separators 35

5 Wall thickness surveys 39

RESTRICTED - COMMERCIAL AEA Technology 5

Personnel Responsibility with regard to sand production Relevant

Interpretation of unusual sand detector trends or

Pipework and choke valve inspection (OIS). Procedure 10

* Procedures are included in Appendix 1.

RESTRICTED - COMMERCIAL AEA Technology 6

1.2 WHY MANAGE SAND ON SHEARWATER?

• process upset (inadequate separation, and poor level control)

Management is the action of controlling a process or event to achieve an objective.

1.3 CONTROLLING SAND PRODUCTION

• Looking: identifying that sand is being produced

1.3.1 Looking for sand . . .

However, it should be understood that:

RESTRICTED - COMMERCIAL AEA Technology 7

1.3.2 Deciding . . . the acceptable level

‘Zero detectable sand’

If sand is detected, then the agreed action to eliminate it must be taken.

1.3.3 Acting to reduce or eliminate sand, and its effects

Conversely, sand production may be reduced or eliminated by:

RESTRICTED - COMMERCIAL AEA Technology 8

1.4 THE PROCESS FLOW PATH FOR SAND

RESTRICTED - COMMERCIAL AEA Technology 9

The path of the sand is as follows:

RESTRICTED - COMMERCIAL AEA Technology 10

Flowlines and headers

Level control valves

The effects of sand production on Shearwater

Flowlines and headers

Sand detectors fitted to

Provision for fitting Gentle bean-up to