TANK LEAK DETECTION AS A RISK MANAGEMENT TOOL

4 5

5 3

5 4

DEFINITION EVALUATION

EVOLUTION CERTIFICATION

DEDICATIONEMERSION

9.0122 10.811

10.811 6.941

10.811 9.0122

EDE

20

TRANSITION

40.078

T C20

TRACTION

40.078

TE D

5

EVOLUTION

10.811

E XECUTIVESUMMARY

Storage tanks are an integral part of theinfrastructure of the oil industry. Imaginean oil industry without storage; it wouldbe a very different place. All operationswould have to happen in real time.Refining could only take place whena ship was discharging crude oil andanother ship was receiving therefined products.

At a terminal, ship discharges could onlytake place when sufficient road or railcarswere available to receive the products.Then imagine the low discharge rate andthe time this would take.

The industry would operate at a verymuch slower pace and on a very muchsmaller scale without storage facilities,if indeed it could operate at all.

In times of global or regional instabilities,storage tanks also provide a means ofsecuring supply to countries. Termedstrategic petroleum reserves, it isestimated that this totals around 4.1billion barrels globally. This may be inthe form of crude oil or refined product.

15

OPTIMISATION

30.974

OLOCATI14

NAVIGATION

28.086

NWhile the importance of storage tanks is clear, they do need to be in the right place for them to be of use.

Easy access to shipping is vital, hence the majority of storage tank farms are built on the coast at a large port.

Furthermore, the terminal must connectseamlessly with the users of the refinedproducts either via pipelines (perhaps to a distribution terminal) or via loadinggantries for internal distribution by roadand rail. Such locations are relativelyunique. The increase in size of ocean-going tankers has further restricted the locations suitable for large storage terminals.

This indicates that there is a vital link inthe supply chain with limited potential forgrowth. This growth (i.e. building newtanks) is further limited by environmentaland planning restrictions; society simplydoes not want more tanks to be built.

The oil industry therefore has a scenariothat requires that the current stock ofstorage tanks must be used at maximumefficiency at all times. However, this isat odds with another requirement: that of safety.

HEALTH AND SAFETYREQUIREMENTS ANDENVIRONMENTAL/ECONOMICCONSIDERATIONSAll liquids associated with the oil industry, from crude oil to the plethoraof refined products, are hazardous tohealth and to the environment. Wheneven the smallest of releases into theenvironment is to be avoided, storingthem in large quantities only servesto magnify the issue. Storage tankintegrity is therefore critical from anenvironmental point of view. There arealso economic considerations: the liquidsare valuable, some more so than othersand therefore a leak is a financial loss. If a leak does occur then the cost of anyclean-up can be significant. On top of that,there may be punitive costs applied plusthe impact on the company’s reputation.But what can be done to minimise thepossibility of a leak from a storage tank?

MANAGING RISKThis starts with the design and buildof the storage tank. International codesare available, for example API 650,which give guidance on the matter.The following is an extract fromthat standard:

1.1.1 This standard covers material,design, fabrication, erection, and testingrequirements for vertical, cylindrical,aboveground, closed- and open-top,welded steel storage tanks in varioussizes and capacities for internal pressuresapproximating atmospheric pressure(internal pressures not exceeding theweight of the roof plates), but a higherinternal pressure is permitted whenadditional requirements are met (see1.1.10). This standard applies only to tanks whose entire bottom is uniformlysupported and to tanks in non-refrigerated service that have amaximum operating temperature of 90°C (200°F) (see 1.1.17).

1.1.2 This standard is designed to providethe petroleum industry with tanks ofadequate safety and reasonable economyfor use in the storage of petroleum,petroleum products, and other liquidproducts commonly handled and storedby the various branches of the industry.This standard does not present orestablish a fixed series of allowable tanksizes; instead, it is intended to permit thepurchaser to select whatever size tankmay best meet his needs. This standardis intended to help purchasers andmanufacturers in ordering, fabricating,and erecting tanks; it is not intended toprohibit purchasers and manufacturersfrom purchasing or fabricating tanks thatmeet specifications other than thosecontained in this standard.

While testing forms part of this standard,it refers to testing performed during theconstruction/fabrication process to ensurethat the tank is put into service in a fitstate, i.e. without leaks and structurallysound. There will be more on this later asmany integrity issues relate to tankswhen they are first put into servicefollowing construction or following majorrepair work.

Once the tank is in service, what stepscan be taken to give confidence that thetank will not leak? Again, standards areavailable that give guidance as to therecommended checks and tests to beperformed. Typically, codes are API 653and EEMUA 159.

The inspections performed can beclassified as follows:

Routine in-service inspection

• The external condition of the tank shallbe monitored by close visual inspectionfrom the ground on a routine basis.

• The interval of such inspections shallbe consistent with conditions at theparticular site, but shall not exceed one month.

• This work shall be performed bycompetent personnel, but notnecessarily an authorised inspector (asdefined by the standard).

External inspection

• Visual external inspection to beperformed by an authorised inspectorat least every 5 years or less if theshell plate corrosion rate dictates otherwise.

Ultrasonic thickness inspection

• External ultrasonic thickness testing ofthe tank shell used to determine rateof uniform general corrosion while thetank is in service.

• Inspection intervals can varydependent upon whether the corrosionrate is known or not. When it is not,the maximum interval should be 5years. When it is then the interval isdetermined by calculation based oncurrent shell thickness, minimumallowable shell thickness and theknown corrosion rate. If this figure isgreater than 15 years then a maximumof 15 years between inspectionsshould be adopted.

Cathodic protection surveys

• Where exterior tank bottom corrosionis controlled by a cathodic protectionsystem, periodic surveys of the systemshall be conducted in accordance withAPI RP 651. This work shall beperformed by a competent person.

Internal inspection

The purpose of the internal inspectionis as follows:

• Ensure that the bottom is not severelycorroded and leaking.

• Gather plate thickness data.

• Identify and evaluate tank bottomsettlement.

HOW OFTEN?

The frequency of internal inspections isdependent upon knowledge of platecorrosion rates. Where this informationis available, either through previousinternal inspections or is anticipatedbased on tanks in similar service, asimple calculation can be performed toensure a minimum plate thickness is notreached over a known period of time.Not withstanding this, the standardindicates that the internal inspectioninterval should not exceed 20 years.

When corrosion rates are not knownand similar service experience is notavailable an internal inspection shouldbe performed to determine bottomplate thickness. This should beperformed within 10 years of the tankbeing put into service.

In practice, it is this 10-year period thathas been adopted as the target forinternal inspection intervals by manyorganisations. In some countrieslegislation is in place that requires tanksto be taken out of service for an internalinspection and re-calibrated on theirtenth anniversary. In the UK suchlegislation is not in place, but an operatormust have in place a documentedmaintenance plan/policy that sets outwhat actions will be taken, and when,to ensure that the facility (which includesthe storage tanks) is operated in asafe manner and the potential for

contamination of the environment isminimised. This document can be subjectto review by the Environment Agencyand may be included in any licencegranted for the operation of the facility.

GOING INSIDE

Many operators write into theirmaintenance policy that tanks will besubject to an internal inspection every 10 years, thereby meeting requirementsas set out by the API document.

But what happens if the tank cannot betaken out of service at this time?

The act of tank entry raises a number ofissues: operational, financial, and healthand safety – all giving good reasons fornot entering a tank unless it is absolutelynecessary. For example, entering a tankthat has contained hydrocarbon liquids orindeed other hazardous liquids hassignificant health and safety implications.

The tank is an enclosed space and thestored products are hazardous to health.Strict controls are required to minimisepotentially life-threatening risks to thoseentering the tank. Therefore, manyoperators have looked for alternativemeans of ensuring the integrity oftheir tanks, without the need for entryunless absolutely necessary (i.e. if workneeds to be done) and have adoptedwhat is termed as an RBI (risk-basedinspection) approach.

Section 6.4.3 of API 653 outlines RBI.Operators use this to justify not enteringtheir tanks. However, this may beconsidered somewhat weak byregulators as it does not provide positiveproof that the tank is tight. In order todemonstrate the conclusions reachedconcerning the integrity of the tank,based on the various factors consideredin their risk-based analysis, operatorsshould ideally include a precision leaktest as a final check. The precision leaktest will document and confirm that theirtank is not leaking, supporting that theiractions have been sufficient.

EXTRACT FROM API 653

6.4.3 Alternative InternalInspection Interval

As an alternative to the procedures in6.4.2, an owner-operator may establishthe internal inspection interval using risk-based inspection (RBI) procedures.Combining the assessment of thelikelihood of tank leakage or failure andthe consequence of tank leakage orfailure is the essential element of RBI.A RBI assessment may increase ordecrease the internal inspection intervalsobtained using the procedures of 6.4.2.1.The RBI process may be used toestablish as acceptable the risk of aminimum bottom plate thickness at thenext inspection interval independent ofthe values in Table 6-1. The RBIassessment may also increase ordecrease the 20-year inspection intervaldescribed in 6.4.2.1. The initial RBIassessment shall be reviewed andapproved by an authorised inspector and an engineer(s), knowledgeable andexperienced in tank design (includingtank foundations) and corrosion. The RBI assessment shall besubsequently reviewed and approved byan authorised inspector and anengineer(s), knowledgeable andexperienced in tank design (includingtank foundations) and corrosion, atintervals not to exceed 10 years, or moreoften if warranted by changes in service.

After an effective RBI assessment isconducted, the results can be used toestablish a tank inspection strategy andbetter define the most appropriateinspection methods, appropriatefrequency for internal, external andon-stream inspections, and preventionand mitigation steps to reduce thelikelihood and consequence of a tankleak or failure.

Factors that should be considered intank RBI assessments include, but arenot limited to, the following:

Likelihood Factors:

• Original thickness, weld type, and ageof bottom plates.

• Analysis methods used to determinethe product-side, soil-side and externalcorrosion rates for both shell andbottom and the accuracy of themethods used.

• Inspection history, including tankfailure data.

• Soil resistivity.

• Type and quality of tank pad/cushion.

• Water drainage from bund area.

• Type/effectiveness of cathodicprotection system and maintenancehistory.

• Operating temperatures.

• Effects on internal corrosion rates dueto product in service.

• Internal coating/lining/liner type, ageand condition.

• Use of steam coils and waterdraw off details.

• Quality of tank maintenance, includingprevious repairs and alterations.

• Design codes and standards and thedetails utilised in the tank construction,repair and alteration (including tankbottoms).

• Materials of construction.

• Effectiveness of inspection methodsand quality of data.

• Functional failures, e.g. floating roofseals, roof drain systems, etc.

• Settlement data.

• Tank bottom details (single, double,Release Prevention Barrier (RPB,internal reinforced linings, etc.)

Consequence Factors:

• Product type and volume.

• Mode of failure, e.g. slow leak to theenvironment, tank bottom rupture ortank shell brittle facture.

• Dike containment capabilities (volumeand leak tightness).

• Identification of environmentalreceptors such as wetlands, surfacewaters, ground waters, drinking wateraquifers, and bedrock.

• Distance to environmental receptors.

• Effectiveness of leak detectionsystems and time to detection.

• Mobility of the product in theenvironment, including for releases to soil, product viscosity and soil permeability.

• Sensitivity characteristics of theenvironmental receptors to the product.

• Cost to remediate potentialcontamination.

• Cost to clean tank and repair.

• Loss of use.

• Impact on public safety and health.

More qualitative approaches may beapplicable that do not involve all of thefactors listed above. In these cases,conservative assumptions must be usedand conservative results should beexpected. A case study may benecessary to validate the approach.

The results of the RBI assessment are tobe used to establish a tank inspectionstrategy that defines the mostappropriate inspection methods,appropriate frequency for internal,external and on-stream inspections, andprevention and mitigation steps toreduce the likelihood and consequenceof tank leakage or failure.

By including a precision leak test, thiswould confirm that the procedures put inplace are working and that deferment ofthe internal inspection is justified.

The key is that the maintenance planprovides sufficient detail of the approachto be adopted.

The advantages to a risk-basedinspection approach are clear. By movingaway from a strict calendar-basedinspection programme this more flexibleapproach allows an operator to maximiseutilisation of the tank assets to meetoperational needs. It also encourages amore focused approach to maintenancespend, targeting those tanks that aremost in need of maintenance rather thanthose that have reached a certain age,but may be completely serviceable.

HOW TO

ESTEAK12 24.305

LLOCATION

20 40.079

TTRACTION

The next section considers various options for performing leak tests to verify the RBI findings. Obviously, if the

tank is fitted with a permanent leak detection system or has a double bottom with interstitial monitoring then

these should be capable of providing the verification required. However, many tanks have only a single bottom

and do not have any permanent leak monitor systems in place. This section considers what techniques are

available in such circumstances.

Clearly, for the test to provide thisconfirmation it must be suitable forthe application and ideally have had itsperformance verified by an independentthird party.

An immediate response is that theautomatic gauging system used for stockdetermination can be used during quietperiods to act as a leak detectionsystem. On the face of it this seems asensible approach, utilising equipmentalready installed on the tank andproviding constant feedback of levelmeasurement to a central control room.They are normally calibrated beforeinstallation and typically thereafter on aregular basis. Indeed, many systems aremarketed as having leak detection builtin to their functionality. However, oncloser examination, does it really providethe level of discrimination required toconfirm that a tank is tight?

Current level gauging systems claim abest measurement capability of±0.5mm. If we consider a tank ofdiameter 30m the surface area of thetank is 707m2. A level change of 0.5mmequates to a volume of 354 litres. It isonly once a volume change of thismagnitude or greater has occurred that aleak could be identified with someconfidence. Remember as well thatthroughout the test period the tankcontents and the shell will be subject tothermal gains and losses.

As the system measures level directly,changes in liquid temperature will impactdirectly on level, perhaps triggering falseleaks or masking real leaks. Based onthe above, it is suggested that the use ofexisting level gauges does not give the

confidence required to determinewhether a tank is leaking or not, exceptin situations when the leak is very large.

An alternative is the use of acousticemissions. This technology uses an arrayof transducers located at intervalsaround the tank that listen for the noisescreated by active corrosion andpotentially for the noise created byleaking liquid. Careful interpretation ofthe received signals against a databaseof known responses enables theoperator to map the complete tank floorand highlight the level of active corrosiontaking place. This is extremely useful asit provides the owner (and specificallythose charged with maintaining thetanks) with details of where corrosionactivity is high and logically where thereis a greater likelihood of a leak pathbeing present. For the test to take placethere must be product in the tank, but itmust be static. As the technologymonitors sound the test area must beacoustically quiet during the test period(this could be a number of hours)otherwise the low-level noise generatedby the active corrosion is swamped byextraneous noise potentially masking anactual occurrence.

While not strictly a leak test it doesprovide information that can be usedto verify the assumptions made fromthe RBI. For example, if RBI points to highrates of corrosion and this is supported byacoustic testing then the tank is a primecandidate for early internal inspection. IfRBI suggests little corrosion and this iscorroborated by acoustic testing, then thetank can remain in service withincreased confidence.

However, acoustic emission is aqualitative test; monitoring oneparameter (sound) and from that,predicting another parameter (corrosionactivity) in conjunction with a databaseof known responses. A better approachwould be to apply a quantitativeapproach to tank leak determinationthrough precision, mass-based testing.

A mass-based technique monitors thepressure head generated by the columnof liquid in the tank over a period of time.The basic equation that is applied by thetechnique is:

p = ρ * g * h

where,

• p is the pressure generated by thecolumn of liquid

• ρ is the density of the liquid in the tank

• g is the constant term, accelerationdue to gravity

• h is the height of the liquid column

Changes in liquid temperature result in achange in density and therefore a changein volume. If the liquid is constrainedwithin a tank then any change in volumewill manifest itself as a change in level,yet there is no physical loss or gain ofliquid by the system. With a leakdetection system based on levelmeasurement, even small temperaturefluctuations are sufficient to mask anyleaks that may be present. By monitoringpressure rather than level, the impact ofchanging liquid temperature is neatlyeliminated. Considering the aboveequation, as temperature changes thedensity of the liquid also changes.

Within a fixed container thecorresponding change in volume of theliquid will result in a change in the heightof the column. However, since thechanges in density and liquid height areinversely proportional the net result inpressure change is zero.

However, the shell of the tank containingthe liquid is not immune fromtemperature changes. A good example isthe effect of solar radiation on the tank:as the sun rises in the morning the effectis to warm the tank shell causing thesteel to expand and increase thecapacity of the tank. This causes theliquid level to fall resulting in thepressure falling (assuming there hasbeen no change in the density of theliquid). This will continue through the day,but as the sun begins to set and the heatgoes out of the day, the tank shell coolsand contracts.

The capacity of the tank reduces and thelevel rises. This rise in level results in acorresponding increase in pressure.The diurnal effects are marked in redon the data plot above.

In reality, what happens is that thereis a combination of the two effects inoperation throughout the testingprocess. Tank shell dynamics can bepartially corrected for, but much betterthat their impact is minimised and this isachieved, to a large degree, by onlyanalysing data gathered during thenight-time period, which should bemore stable.

But how is the pressure monitored? The system used by SGS utilises adifferential pressure (DP) transmitterlocated outside the tank in a portablecontrol unit. The head pressure of theliquid in the tank is transferred to the

high-pressure side of the DP transmitterusing a “nitrogen bubbler” system,similar to that previously used for tanklevel gauging. The low-pressure side ofthe DP transmitter connects to an openhose located in the vapour space abovethe liquid, the barometric pressure.

Temperature probes are used tomonitor both liquid and air temperaturethroughout the test. All data fromthe various transmitters are inputto a control unit for data loggingpurposes and subsequent analysis.This analysis is performed off-line usingbespoke software.

Schematic data plot showing diurnal effects – circled in red. Courtesy of Mass Technology Corporation.

The performance of the system hasbeen determined through independentevaluation in the USA by Ken WilcoxAssociates, who performed a series oftests on behalf of the US EnvironmentalProtection Agency. The evaluationprogramme determined the thresholdof detection for the system and thetesting protocol required to achieve thatlevel of performance.

The threshold of detection if expressedin terms of level change is constantregardless of tank diameter and equatesto approximately 0.004mm per hour.For a mass-based system, the detectionthreshold (expressed volumetrically)varies directly with tank diameter. For atank of 30m diameter, the threshold ofdetection is less than 3 litres per hour.

Recalling the example provided earlier ofa potential detection threshold for astandard level gauging system in asimilar size tank, the precision mass-based approach is two orders ofmagnitude better. Test duration is also

related directly to tank diameter withtanks up to 9m in diameter requiring a24-hour test, and up to 120 hours fortanks with diameters in excess of 67m.

A key requirement for this techniqueto provide a conclusive result is for alltank penetrations to be adequatelyisolated. Experience shows that singleisolation using the tank-side valve isnot adequate. Either the penetrationshould be physically blinded or a doubleblock and bleed arrangement needs tobe in place and monitored during thetest. The system will indicate if there isloss of liquid from the tank, but it cannotdifferentiate between liquid lost througha hole in the tank floor and liquid passingthrough a nominally closed valve.

External floating roofs also create apotential problem if rain occurs duringthe test. Rain collecting on the roofincreases its weight, which is registeredby the system as an increase in liquidhead pressure. Any real loss of liquid ismasked until the roof attains the sameoperating conditions as it was in at thestart of the test. For the test to beconclusive it may need to be extended in order to capture enough valid data for analysis.

Clearly, all techniques have theirlimitations. But provided thoselimitations are known and understooduseful information can be generated,whether from acoustic emissions or byprecision mass monitoring, which can beused in conjunction with data from theRBI process.

Schematic of the Compact Bubbler Leak Detection System. Courtesy of Mass Technology Corporation.

One option is for the operator to donothing, keep the tank in service andhope that the tank integrity will not becompromised until it can be taken out ofservice for an internal inspection.

Is this really a viable option? Withoutthe ongoing information gathered aspart of an RBI programme what levelof confidence does the operator havein making a decision to keep the tankin service?

Such a scenario, which one suspects isnot an isolated occurrence, would beimproved by performing a precision leakdetection test. While the operator is stillbeyond their own maintenance policy, itat least demonstrates that efforts havebeen made to assess the tank beforeextending its service life. Clearly, if thetank is shown to be leaking then its lifecannot be extended and it should betaken out of service immediately.

Even when the operator is able to meettheir own maintenance programme thereis a place for a precision leak test.Performing an internal inspection on astorage tank is not an easy task. Theenvironment is difficult and there can bea very large area to examine.

A source in the USA found that around7% of new/repaired tanks are identifiedas leaking when hydro-tested (seesection below). Therefore it is reasonableto conclude that a percentage of alltanks subject to internal inspection willbe returned to service with a leak paththat has been missed.

Below is an extract from API 653:5.2.1 In all reported incidents of tankfailure due to brittle fracture, failureoccurred either shortly after erectionduring hydrostatic testing or on the firstfilling in cold weather, after a change tolower temperature service, or after arepair/alteration.

Tanks could continue to leak for manymore months or years until the loss ofliquid is identified.

When a tank is entered the greatestpossible care and professionalism istaken when performing the inspection,but it is a difficult environment and if noleaks are found then it is generallyconcluded that the tank was fine beforeit was taken out of service. But that maynot be the case.

By performing a precision leak detectiontest prior to the tank being taken out ofservice the operator will know whetherthe tank is tight and they are notexpecting to find any potential leak pathsor that it is leaking at a rate of x litres perhour. Knowing that the tank is leakingbefore the internal inspectioncommences means that they areexpecting to find a leak path or paths.The work would continue until leak pathsare found.

As a colleague says: “Trying to finda leak in a tank is like looking for aneedle in a haystack. Precision leaktesting tells you that there is a needlein the haystack!”

NEW AND REPAIRED TANKS

Before a new or repaired tank is putinto service it is subject to a hydro-test.While this is primarily a test of thestructural integrity of the tank manyoperators also use it as a means ofconfirming that the tank is tight, beforereturning it to service.

But how valid is a hydro-test as a meansof detecting small leaks from a tank? Inmany instances the decision as towhether the tank is leaking or not isbased on level changes during the testperiod. How the level is monitored maybe as rudimentary as checking theposition of the floating roof (if one isfitted). A level gauging system may beused, but we have already seen that thismethod of leak detection will onlyidentify major leaks of several hundredlitres per hour. Some operators utilisefluorescent chemicals in the water andcheck around the base of the tank usingfluorescent lights. This is neitherscientific nor likely to identify a leakunless it occurs in a visible area.

Is a leak test strictly necessary as surelythe tank was inspected before it wasreturned to service? Data shows thatapproximately 7% of tanks subjected toa leak test leak while under hydro-test.How many of these would have beenidentified without a precision leakdetection test? Consider the financialand environmental impacts of returninga leaking tank to service.

By performing a precision leak test itgives the operator the confidence toreturn the tank to service; remember itcould be many years before the tank issubjected to an internal inspection andthe leak found.

CALENDAR-BASED

INTERNAL9

INVESTIGATION

18.998

I14

NOTIFICATION

28.086

N SPECTIONSReturning to the calendar-based approach, what happens when a tank programmed to be taken out of service as

required by the operator’s own maintenance plan cannot be released, for whatever reason?

SUMMARYThe integrity of storage tanks is apressing issue for all parties concerned.More so now than ever before, operatorsare being pushed to prove the integrityof their storage tanks to reduce the riskto the environment as well as gaininggreater control over their assets.

All the methods discussed in this paperserve a common purpose within ourindustry. The advantages of using precisemethodology such as a precision mass-monitoring system are clear and cananyone afford not to understand exactlywhat is happening in their tanks?

SGS UNITEDKINGDOM LTDSGS is the world’s leading inspection,verification, testing and certificationcompany. Oil, Gas and Chemicals(OGC) Services are the market leadersin independent, specialised andtechnical services to the oil, gas and chemical industries.

SGS OGC employs a multi-disciplinaryteam of the best qualified andrecognised experts. Our range ofsolutions allows you to maximise your operations while reducing limitingfactors such as risk, time and cost.

At SGS Oil, Gas and Chemicals we are committed to providing a dedicated,professional service that will make a real difference to you.

SGS United Kingdom LtdRossmore Business EstateInward Way, Ellesmere PortCH65 3ENUnited Kingdom

Tel: +44 (0)151 350 6666Fax: +44 (0)151 350 6630Email: GB.OGC.Enquiries@sgs.comWWW.UK.SGS.COM/OGC

SOURCESThis white paper acknowledges thefollowing sources:

• API 650 – Load Combinations.

• API 653 – Tank Inspection, Repair,Alteration and Reconstruction,Fourth Edition.

• API RP 651 – Cathodic Protection ofAbove Ground Storage Tanks.

• EEMUA 159 – Best Practice Inspectionof Above Ground Storage Tanks.

SWITCH ONTO YOUR OPTIMISATIONSTRATEGY WITH US, VISITWWW.UK.SGS.COM/OGC CONTACTGB.OGC.ENQUIRIES@SGS.COM OR CALL +44 (0)151 350 6666

WWW.SGS.COM

© S

GS

SA 2

011.

ALL

RIG

HTS

RES

ERVE

D. S

GS

5500

-091

1