FULLTEXT01 korozyonnn111

7/27/2019 FULLTEXT01 korozyonnn111

1/45

UPTEC F10 006

Examensarbete 30 hpFebruari 2010

Design and dimensioning of pressure

vessel for a marine substation

Lars Eriksson


2/45

Teknisk- naturvetenskaplig fakultetUTH-enheten

Besksadress:ngstrmlaboratorietLgerhyddsvgen 1Hus 4, Plan 0

Postadress:Box 536751 21 Uppsala

Telefon:018 471 30 03

Telefax:018 471 30 00

Hemsida:http://www.teknat.uu.se/student

Abstract

Design and dimensioning of pressure vessel for amarine substation

Lars Eriksson

This thesis presents the mechanical design and dimensioning of a pressure vessel,which is to be used as housing for a marine substation in a wave power park. Aconcept for generation of electricity from ocean waves is being developed at theDivision of electricity at Uppsala University. The concept is based on the use of apermanent magnet linear generator, placed on the seabed, connected via a line to abuoy at the surface. The generated electricity from a group of generators istransmitted in sea cables to a marine substation where conversion and transformationtakes place before the electricity is transmitted to shore. To reduce the risk of waterleakage, the gas pressure inside the marine substation is larger than the surroundingwater pressure. The substation can be pressurized before submersion, which requiresthe housing to be designed as a pressure vessel. The pressure vessel has beendimensioned with formula based methods according to EN 13445, the Europeanstandard for unfired pressure vessels. The construction has been based on modifying

a standard pressure tank. The housing has been designed for installation and sealing ofa large number of electrical connectors. The connectors have been placed in a waythat allows for future cable coupling with remotely operated vehicles and simplifiesmaintenance of the substation. Another design consideration has been to facilitatesubmersion by reducing the buoyancy of the substation.

ISSN: 1401-5757, UPTEC F10 006Examinator: Tomas Nybergmnesgranskare: Mats Leijon

Handledare: Magnus Rahm


3/45

Svensk sammanfattningDetta examensarbete omfattar mekanisk konstruktion av ett tryckkrl till ett marintstllverk som ska anvndas fr vgkraft. Vid avdelningen fr elektricitetslra vidUppsala Universitet utvecklas ett koncept fr generering av elektricitet frnhavsvgors rrelser. Konceptet bygger p en linjrgenerator placerad p havsbottnen

som via en lina drivs av en boj vid ytan. Linjrgeneratorn r direktdriven vilketinnebr en enkel mekanisk konstruktion men ocks att elektriciteten som genererashar varierande frekvens och amplitud. Fr att elektriciteten ska kunna anvndas ielntet behver den konverteras till vxelstrm med konstant frekvens. Fr att inte fstora resistiva frluster i transmissionskabeln som gr in till land, behverelektriciteten ocks transformeras till hgre spnning. Konvertering ochtransformering av elektriciteten frn en grupp av linjrgeneratorer sker i ett stllverkplacerat p havsbottnen i nrheten av generatorerna.

Lysekilsprojektet gr ut p att testa detta vgkraftskoncept under realistiskafrhllanden i havet utanfr Lysekil. Fr nrvarande r tre linjrgeneratorerinstallerade och inkopplade till ett stllverk p 25 meters djup. Parken ska utkas tilltotalt tio linjrgeneratorer och tv stllverk de nrmaste ren. Detta examensarbeteberr det andra stllverket i Lysekilsprojektet.

Kraftelektroniken och transformatorerna i stllverket behver skyddas frnhavsvattnet och fr att minska riskerna med eventuellt vattenlckage r lufttrycket istllverket hgre n det omgivande vattentrycket. Detta kan stadkommas genom attkonstruera stllverkets inneslutning som ett tryckkrl som innan sjsttning kantrycksttas med hgre tryck n vattentrycket vid arbetsdjupet. Inneslutningen mste dtla ett inre tryck p ca. 3 bar vilket innebr att regelverken fr konstruktion, kontroll

och certifiering av tryckkrl mste fljas.

De olika tryckkrlskomponenterna har dimensionerats med formelbaserade metoderfrn SS-EN 13445 Tryckkrl, ej eldberrda, som r en del av den europeiskatryckkrlsstandarden. Metoderna fr dimensionering har implementerats i Matlab ochkonstruktionen samt ritningsframstllningen har utfrts i SolidWorks. Vissasimuleringar med finita elementmetoden har ocks utfrts med SolidWorksSimulation, fr att underska hllfastheten vid lastfall som inte tcks in av deformelbaserade metoderna.

Utformningen av krlet har utfrts med avseende p att sjsttning och terhmtning

fr underhll ska vara s enkel som mjligt. Stllverket har ett tjugotalgenomfrningar fr kablar kopplade till undervattenskontakter. Konstruktionen harfrberetts fr att kontaktering i framtiden ska kunna utfras med enundervattensrobot. Konstruktionen har baserats p en vertikalt stende trycktank medkupade gavlar som har en volym p 5 m. Att tillverka inneslutningen genom attmodifiera en frdig trycktank r ett kostnadseffektivt alternativ till att specialbestllaett tryckkrl. Krlet har frstrkts med frstyvningsringar fr att tla strre yttre tryck.Den nedre gaveln har avlgsnats och bytts ut mot ett lock som fsts i en skruvflns.Krlet har utrustats med kabelgenomfringar i form av stutsar som r frberedda frkabelfrskruvningar, samt lyftglor dimensionerade fr att lyfta stllverket och dessfundament vid sjsttning. Mjligheten att anvnda en platt nedre gavel har ocks

underskts, vilket har frdelen att krlet blir mer kompakt som i sin tur medfr attfundamentet kan gras lttare.


4/45

Om liknande framtida stllverk skulle placeras p strre djup, och trycksttas innansjsttning, skulle inneslutningen behva tla strre inre tryck, och mjligheterna attdimensionera fr detta har underskts. Mjligheten att tryckstta krlet undersjsttningen diskuteras ocks, vilket skulle innebra att krlet behver tla betydligt

mindre inre tryck och inte behver tryckkrlcertifieras. Tre konstruktionsalternativpresenteras, som skiljer sig t med avseende p tryckkrlets infstning tillfundamentet samt den nedre gavelns konstruktion. De slutligakonstruktionsritningarna har sedan frdigstllts baserade p ett av dessakonstruktionsalternativ.


5/45

Nomenclature and abbreviationsd [mm] Diameterr [mm] Radiuse [mm] Required thicknessP [MPa] Calculation pressure

f [N/mm2] Nominal design stress [N/mm2] Normal stressReH [N/mm

2] Upper yield strengthRm [N/mm

2] Tensile strengthS - Safety factorz - Joint coefficient [%] StrainE [MPa] Modulus of elasticity - Poissons ratiom [kg] MassV [dm3] Volume [kg/dm3] Density

AC Alternating CurrentDBA Design By AnalysisDBF Design By FormulaeDC Direct CurrentFEM Finite Element MethodLVMS Low Voltage Marine SubstationMABE Maschinen und Behlterbau GmbHNDT Non-Destructive Testing

PED Pressure Equipment DirectiveROV Remotely Operated VehicleWEC Wave Energy Converter

Unit conversion for pressure:1 MPa = 1 N/mm2 = 10 bar


6/45

2

Table of contents1 Introduction............................................................................................................4

1.1 Background....................................................................................................41.2 The Lysekil project ........................................................................................6

1.2.1 ROV cable coupling...............................................................................6

1.2.2 Installation and retrieval of marine substation.......................................71.3 Aim of thesis ..................................................................................................7

2 Theory....................................................................................................................72.1 Pressure vessel regulations ............................................................................72.2 Control and inspection regulations ................................................................82.3 Design of components..................................................................................11

2.3.1 Cylindrical shells, internal pressure.....................................................122.3.2 Cylindrical shells, external pressure ....................................................122.3.3 Dished ends, internal pressure .............................................................142.3.4 Dished ends, external pressure.............................................................142.3.5 Flat ends ...............................................................................................152.3.6 Flanges .................................................................................................152.3.7 Openings ..............................................................................................162.3.8 Lifting eyes ..........................................................................................17

3 Method .................................................................................................................174 Conditions for the design.....................................................................................17

4.1 Manufacture of the vessel ............................................................................184.2 Protection against corrosion.........................................................................18

4.2.1 General corrosion.................................................................................184.2.2 Galvanic corrosion...............................................................................19

4.3 Mounting of components .............................................................................19

4.4 Connection and sealing of cables.................................................................204.4.1 ROV adaptation ...................................................................................204.4.2 Cable sealing........................................................................................204.4.3 Placement of openings .........................................................................21

4.5 Buoyancy of the substation..........................................................................214.5.1 Using a flat lower end ..........................................................................21

5 Results..................................................................................................................225.1.1 Material ................................................................................................22

5.2 DBF calculation results................................................................................225.2.1 Shell without stiffeners ........................................................................225.2.2 Shell with stiffeners .............................................................................23

5.2.3 Flat end.................................................................................................245.2.4 Flanges .................................................................................................245.2.5 Nozzles.................................................................................................255.2.6 Lifting eyes ..........................................................................................275.2.7 Dimensioning for larger pressure.........................................................28

5.3 FEM analysis results ....................................................................................285.3.1 FEM analysis of reinforcement for footings on the dished end...........285.3.2 FEM analysis of the bolted flat end .....................................................29

5.4 Buoyancy of the substation..........................................................................316 Discussion............................................................................................................32

6.1 Summary of calculation results....................................................................32

6.2 Design alternatives.......................................................................................336.2.1 Placement of nozzles............................................................................33


7/45

3

6.2.2 Lower end and foundation ...................................................................336.3 Future development .....................................................................................34

7 Conclusions..........................................................................................................358 References............................................................................................................36Acknowledgements......................................................................................................37

Assembly drawings of the vessel.......................Appendix 1Drawing of the flat end..Appendix 2


8/45

4

1 Introduction1.1 BackgroundA concept for generation of electricity from ocean waves is being developed at thedivision of electricity at Uppsala University. The concept is based on linear generatorsbeing driven by floating buoys that acts as point absorbers. A linear generator consistsof a translator mounted with permanent magnets that moves vertically inside a stator.The stator contains a three-phase coil winding and is placed in a water-tight housing,standing on the seabed. A point absorber is a buoy of small size compared to theaverage wave-length [1]. The buoy is connected to the translator of the lineargenerator with a line. The waves set the buoy and translator in motion and voltage isinduced in the stator winding. A point absorber connected to a linear generator iscalled a Wave Energy Converter, WEC, as seen in Figure 1.1.

Figure 1.1 Conceptual drawing of a wave energy converter, Division of Electricity, UU

A general challenge for wave power technology is survivability. During storms, theaverage power level in waves can reach 50 times the overall power level [2]. Themechanical design of WECs must consider extreme weather conditions. In the wavepower concept developed in Uppsala, most of the sensitive equipment is placed on theseabed and is in that way protected from the direct impact of waves.

Another advantage of this system is the mechanical simplicity. Many previous

wave power concepts use standard generators [3]. This means that the slow motion ofthe waves must be mechanically converted to a fast rotating motion. In the concept


9/45

5

from Uppsala University, the generator is designed for the wave motion. Themechanical energy conversion is minimized and instead the electric output is modifiedwith power electronics. Standard electronic components can be used, that can beexpected to need very little maintenance compared to a mechanical energy conversionsystem.

The electricity generated from a WEC will have a varying frequency, amplitudeand phase order, dependent on the ocean waves [2]. This means that the electricitymust be converted with power electronics before it can be used in the grid.Transmitting the electricity from each WEC directly to shore in individual cableswould lead to large losses because of the relatively high current and low voltageoutput from the linear generators. By transforming the electricity to higher voltagenear the WECs, losses can be reduced. Conversion of the generated electricity toconstant frequency AC and subsequent transformation to higher voltage, can takeplace in a Low Voltage Marine Substation, LVMS, placed on the seabed in the closevicinity of the WECs. A visualization of this concept can be seen in Figure 1.2.Several WECs can be connected to one LVMS where the generated electricity from

each linear generator is rectified. The combined DC is then inverted to 50 Hz ACbefore transformation to higher voltage. The output from the LVMS will be through asingle three-phase AC cable.

Figure 1.2 Visualization of a wave power plant, Seabased Industry AB

The electrical components in the marine substation must be protected from water. Aneffective method of avoiding water leakage is to have the gas pressure inside themarine substation slightly higher than the surrounding water pressure. This way anyimperfection of the seals on the housing will result in gas leaking out while water isstill prevented from leaking in. Leakage can be detected by monitoring the gaspressure or moisture content in the vessel. If the vessel is to be pressurized beforeinstallation on the seabed, the internal gas pressure will be several times higher thanthe atmospheric pressure and the vessel must be designed to withstand this high

internal pressure. In case the vessel should for some reason loose the internal


10/45

6

overpressure before or during submersion, the design should also be able to withstandthe external water pressure at the seabed.

Another alternative is to pressurize the vessel during the submersion. That waythe maximum pressure difference between the in- and outside of the vessel can belimited to a much lower value than the water-pressure at the operational depth.

1.2 The Lysekil projectA project for testing and evaluating the wave power concept is currently taking placeoutside of Lysekil on the West coast of Sweden. The first wave energy converter wasinstalled in 2006 and was connected to a measuring station on shore. Two moreWECs and a Low Voltage Marine Substation, see Figure 1.3, are now installed. Theplan is to install another seven WECs and one more LVMS. When everything isinstalled, the first LVMS and the seven new WECs will be connected to the newLVMS. The generated electricity from all ten WECs will be transmitted to shorethrough a single AC cable.

Figure 1.3 Pressure vessel for the first LVMS, the dished lower end with underwater connectors infront, Division of Electricity, UU

1.2.1 ROV cable couplingOne large expense of the research project is the diving for the installation ofequipment under water. Diving also requires good weather which limits the time whenit is possible to work. A solution to these problems is to use a remotely operatedvehicle, ROV, for the work under water. This means that the equipment of the wave

power plant, including the marine substations must be adapted to ROV operation. Onetask that ROVs could be used for is cable coupling. This requires that the connectors


11/45

7

on the substation are positioned so that they can be reached by the ROV. It might alsobe necessary to have an installation that allows the ROV to dock with the marinesubstation [4].

1.2.2 Installation and retrieval of marine substationThe marine substation and its concrete foundation will be lowered down to the seabedfrom a ship. For this it is necessary to consider how the lifting will take place. Thefirst LVMS was designed with the lifting points on the foundation. The new LVMSwill be designed with lifting eyes which will make the lifting procedure easier.

The submersion and elevation of LVMS at sea could be simplified if thesubstation with the foundation has small positive buoyancy. That means the structurefloats, but can be submerged by adding a relatively small amount of extra weight.

1.3 Aim of thesisThe aim of this thesis is to design the pressure vessel for the second low voltagemarine substation in the Lysekil project. It will be placed on a depth of 25 m and havean internal pressure of 3 bar. If the vessel is to be pressurized before submersion, theregulations for pressure vessels must be followed when designing the housing.

The marine substation will have electrical inputs from seven wave energyconverters. The size must be large enough to contain transformers and powerelectronics for processing the electricity from the connected wave energy converters.It will also have an AC input from the first LVMS, an AC output and a number ofsensor equipment outputs. The connectors must be mounted in a way that is adaptedfor coupling with remotely operated vehicles. The marine substation should be

designed to make the lifting and installation on the seabed as simple as possible.

2 Theory2.1 Pressure vessel regulationsThe European standard concerning unfired pressure vessels, EN 13445, is a part of thePressure Equipment Directive (PED) prepared by the European Committee forStandardization. In 2002, PED was given the status of national standard in Sweden,replacing the previous Tryckkrlsnormen. The Swedish version SS-EN 13445 ispublished by Swedish Standards Institute (SIS). The following parts have been used inthis thesis:

Part 1: General Part 2: Materials Part 3: Design Part 4: Fabrication Part 5: Inspection and testing

Limitations of EN 13445

The design rules in EN 13445 are primarily intended for non-cyclic loads, whichmeans that the number of pressure cycles are lower than 500. For cyclic loads,

methods for fatigue evaluation are given. EN 13445 applies to unfired pressurevessels with maximum allowable pressure greater than 0.5 bar but can also be used for


12/45

8

lower pressures. It applies for maximum allowable temperatures where creep effectsdo not need to be considered. A number of applications are excluded from EN 13445,such as transportable pressure equipment, equipment for nuclear use, riveted vessels,pipelines etc.

2.2 Control and inspection regulationsThe PED gives manufacturers the freedom to choose between many ways ofdesigning, manufacturing and inspecting pressure vessels, as long as they satisfy therequirements of the directive1. A notified body is an organization responsible forinspection, testing and certification of pressure vessels. It is allowed that themanufacturer can carry out some inspection activities, though one of the criteria forthat is a quality assurance system, approved and monitored by the notified body. Asystem of 13 modules determines which of the control activities can be carried out bywhom, see Figure 2.2. The factors that determine which combination of modules to beused are: the hazard category of the pressure vessel, whether the manufactures havean approved quality assurance system and finally whether unit or serial production iscarried out.

Hazard category of the vessel

The hazard categories I-IV are dependent on the fluid in the vessel, volume andmaximum pressure [6]. Fluids are divided in two groups. Group 1 comprisesdangerous fluids that are explosive, flammable, toxic or oxidizing. Group 2 comprisesany other fluids. A vessel designed for gases of group 2 is categorized by the diagramin Figure 2.1. The Figure shows that a vessel with volume over 1000 l and maximumpressure between 0.5 and 4 bar is of category III.

Figure 2.1 Hazard categories of vessels for gases of group 2 (non-dangerous fluids) [6]

1

Inspecta, The Directive for Pressure Equipment,(2009-04-02)


13/45

9

Relevant modules of the PED

If design and manufacture are carried out as unit production, without an approvedquality assurance system, and the pressure vessel is of category III, modules B1 (ECdesign examination) and F (product examination) are to be used [6]. However, it isalso allowed to use a module from a higher hazard category but the category remains

the same2

. The equivalent for category IV is module G (EC unit verification). SeeFigure 2.2.

Figure 2.2 The control modules of the PED1

2The pressure equipment web-site of the European Commission, (2004-04-13)


14/45

10

The relevant modules (B1, F and G) states, among other things, that:

The manufacturer must apply to a notified body for unit verification or an ECdesign examination. The application must include technical documentation toenable an assessment of the conformity of the pressure equipment with thePED.

The notified body must examine the technical documentation with respect tothe design and the manufacturing procedures, assess the materials, approve theprocedures for the permanent joining of parts and carry out the final inspectionand proof test.

The notified body must affix its identification number and the manufacturermust affix the CE marking to the pressure equipment.

EC design examination certificates or certificate of conformity with the PEDissued by the notified body must be kept by the manufacturer for ten years andmust be made available on request.

Testing groups and extent of non-destructive testing

The testing of pressure vessels is dependent on which testing group the vessel isdesigned for [5]. There are four groups and groups 1-3 are each divided into two subgroups. The testing group determines the allowed materials, maximum thickness,allowed welding process, service temperature range, joint coefficient and extent ofnon-destructive testing (NDT) of welded joints. Group 1 has the least limitations butrequires the most extensive testing, while group 4 has the most limitations butrequires the least testing. In group 4 generally no NDT is required, with very fewexceptions.

When designing the pressure vessel, the testing group must be considered. If thematerial, maximum thickness, welding process and temperature range meets the

requirements, it can be possible to place the vessel in a higher testing group, and lessNDT is required. If that is the case, the joint coefficient to use must be adjusted to thetesting group. The joint coefficient is 1 for group 1 and 2, 0.85 for group 3 and 0.7 forgroup 4. Lower joint coefficient means that the required thickness of shells will belarger for a given pressure.

Final assessment and proof test

A pressure vessel designed and manufactured according to EN 13445 has to besubjected to final assessment consisting of:

Visual and dimensional inspection Examination of the documentation Proof test Examination after proof test Inspection of safety accessories

The proof test is normally a hydrostatic pressure test, since pneumatic testing ispotentially much more dangerous. The proof test takes place after all fabrication andinspection have been performed, however operations that influence the inspectionpossibilities such as painting, lining, galvanizing etc. shall be carried out after theproof test examination. The test pressure shall be the greatest of:

t

astf

fPP 25.1= [MPa] (1)


15/45

11

and

st PP 43.1= [MPa] (2)

where Pt is the test pressure measured at the highest point of the vessel in test

position, Ps is the maximum allowable pressure,fa is the nominal design stress at testtemperature andft is the nominal design stress at maximum allowable temperature.After the final assessment, the vessel shall be marked with direct stamping or apermanently attached nameplate.

2.3 Design of componentsEN 13445-3 gives rules to be used for design of pressure bearing components, such asshells of various shapes, flat walls, flanges etc. For these components the Design byFormulae (DBF) method is generally followed. DBF means that formulas are given to

find stresses which have to be limited to safe values. General prescriptions for Designby Analysis (DBA) are also given to evaluate designs not covered by DBF methods.The following sections describe the DBF methods and basic equations for differentpressure components given in EN 13445-3.

Conditions for the calculations

For non-austenitic steels (carbon steel is non-austenitic while stainless steel often isaustenitic) the maximum allowed nominal design stress is as follows. For operationloads:

=

4.2;5.1min20//2.0 mtp RR

f [N/mm

2

]. (3)

For test loads:

05.1

/2.0 tp

test

Rf = [N/mm2]. (4)

For test conditionsfis replaced byftest in all equations in the following sections. Theminimum 0.2%-proof strengthRp0.2 above can be replaced by the upper yield strength

ReH if the former is not available in the material standard. The nominal stress shall be

multiplied by 0.9 for testing group 4.

Explanation of thickness definitions:e required thicknessen nominal thicknessemin thinnest possible manufacturing thicknessea calculation thicknessc corrosion additione absolute value of negative tolerance of nominal thicknessm margin for thinning during manufacturinge

ex extra thickness up to nominal thickness

Relationships between thickness definitions are showed in Figure 2.3.


16/45

12

Figure 2.3 Thickness definitions for shells [4]

2.3.1

Cylindrical shells, internal pressureThe required thickness of a cylindrical shell subjected to internal pressure is given by

Pzf

dPe e

+

=

2[mm], (5)

where de is the external diameter, P is the calculation pressure andz is the jointcoefficient. For a given thickness the maximum internal pressure is given by

m

a

d

ezfP

=

2max [MPa], (6)

where ea is the calculation thickness and dm is the mean diameter.

2.3.2 Cylindrical shells, external pressureEN 13445-3 gives methods for verifying that the dimensions of a cylindrical shell areadequate to withstand a given external pressure. Both unsupported shells and shellsreinforced with stiffeners are covered.

For unsupported shells the pressure where the mean ring stress reaches the yieldstrength is

m

ay

r

eP

=

[MPa], (7)

and the theoretical elastic instability pressure is

m

am

r

eEP

= [MPa], (8)

where is the normal stress, ea is the calculation thickness, rm is the mean diameter,E

is the modulus of elasticity and is the mean value of circumferential elasticstretching at collapse. The collapse pressure Pris given by the experimentally


17/45

13

determined curve [7] in Figure 2.4, which shows Pr/Pyas a function ofPm/Py. Thecondition

S

PP r< [MPa] (9)

must be satisfied for the shell to withstand the external pressure P, where Sis a safetyfactor.

Figure 2.4 Pr/Py plotted against Pm/Py. Curve 1 is for cylinders and cones, curve 2 is for spheres anddished end [4]

Shells can be reinforced by stiffeners which are rings welded to the shell on theoutside or inside. Stiffeners are categorized as light or heavy. A light stiffener is e.g. aflat ring and a heavy stiffener can be a flange. For light stiffeners calculations must bemade considering interstiffener collapse, elastic instability and maximum stress in thestiffener. The calculation of interstiffener collapse is for large diameters exactly thesame as the calculation of unsupported shells above, but using the distance betweenthe stiffeners as the unsupported length. The calculation of elastic instability is madeto ensure that the external pressure is lower than the elastic instability pressure whichis the sum of the pressure from the stiffener and a certain length of the shell.Calculations are also made to ensure that the maximum stress in the stiffener is lowerthan the nominal elasticity limit of the material.

For heavy stiffeners equivalent calculations are made but only for the stiffeneritself, independent of the support from the shell. This means that more material isneeded but the calculations are simpler.


18/45

14

2.3.3 Dished ends, internal pressureA torishperical end is a type of dished end which consists of a central part withinternal radius r1 and a torodial knuckle with internal radius r2, see Figure 2.5. Akorbbogen end is a torishperical end where r1/de=0.8 and r2/de=0.154.

Figure 2.5 Geometry of a torishperical end

For a given geometry the maximum internal pressure shall be the lowest of:

a

ads

er

ezfP

5.0

2

1 +

= [MPa], (10)

)2.075.0( 1 i

ady

dr

efP

+

=

[MPa], (11)

and

825.0

2

5.1

1 2.075.0111

+=

ii

abdb

d

r

dr

efP [MPa], (12)

where

5.1

/2.0 tp

b

Rf = [N/mm2], (13)

ea is the calculation thickness and di is the internal diameter. The factor can bedetermined by iterative calculations and is dependent on ea/r2 and r2/di.

2.3.4 Dished ends, external pressureThe dimensions of torisherical ends subjected to external pressure is verified the sameway as unsupported cylindrical shells with the differences that

e

ay

r

eP

=

2[MPa] (14)

and


19/45

15

2

221.1

e

am

r

eEP

= [MPa], (15)

where ea is the calculation thickness and re is the external diameter.

2.3.5 Flat endsThe required thickness of a circular welded flat end is calculated by

f

PdAe iwelded = [mm], (16)

where di is the internal diameter of the shell andA is a form factor dependent ofgeometry, stress limits, pressure and fatigue situation of the connection between theflat end and the shell. If the vessel is subject to non-cyclic loads the required thickness

will be considerably smaller than for cyclic loads [7].The required thickness of a bolted flat end with a full face gasket is calculated by

f

PCebolted 41.0= [mm], (17)

where Cis the bolt circle diameter. The required thickness of the flange area is givenby

boltedflange ee 8.0= [mm]. (18)

2.3.6 FlangesA number of different flange designs are covered in EN 13445-3. Narrow face flangeshave the gasket inside of the bolt circle and have no contact between the sides of theflange outside the bolts. Full face flanges have contact of the sides over the entireflange surface.

If a flange has a compressible gasket, both sides of the flange are separated fromeach other by the gasket. The compression of the gasket, which determines its sealingability, is dependent on the force from the bolts [8]. If a flange has an o-ring sealingthe sides of the flange have metal-to-metal contact and the compression of the gasket

is not dependent on the bolt force. An o-ring sealing is placed in a groove in one ofthe flange sides. The surface finish in the groove is critical for the sealing ability ofthe o-ring.

A distinction is made between flanges where the inner diameter of the flange isthe same as the inner diameter of the shell, which is called smooth bore and flangeswhere the diameters differ, called stepped bore. The connection between the flangeand the shell can be in form of a machined taper hub, otherwise fillet welds arenormally used. Different types of flanges are shown in Figure 2.6.

The required thickness of a full face flange with metal-to-metal contact is

( )hR

dnCfMe

=

6 [mm], (19)


20/45

16

whereMR is the balancing radial torque in the flange at the bolt circle, dependent ofthe dimensions of the flange and the internal pressure, Cis the bolt circle diameter, nis the number of bolts and dh is the diameter of the bolt holes

Figure 2.6 Cross-sections of three different types of flanges

2.3.7 OpeningsOpenings in shells often require that the area around the opening is reinforced tocompensate for the reduction of the pressure bearing section. The reinforcement canconsist of a nozzle, a ring in the opening, a plate on the shell or increased thickness ofthe shell around the opening. One method of reinforcement is to use nozzles. A set-onnozzle is a piece of pipe placed on the hole and welded to one side of the shell. A set-in nozzle is placed in a hole and welded to both sides of the shell. The method used inEN 13445-3 is called the pressure-area method. It is based on verifying that thereactive force from the material in the reinforcement is equal or grater than the loadfrom the pressure. The reactive force from the material is the sum of the product ofthe average membrane stress in each component and its stress loaded cross-sectionalarea. The load from the pressure is the sum of the product of the pressure and thepressure loaded cross-sectional areas.

The general equation for the reinforcement of an opening is

( )( ) ( )

( )ApApApPPfAfPfAfPfAfAf

bs

obboppsws

++

+++

5.0

5.05.05.0(20)

wherefob=min(fs;fb) andfop=min(fs;fp).

The terms in the equation above are defined as following:Afrefer to stress loaded cross-sectional areas effective as reinforcementAp refer to pressure loaded areas

The following subscripts apply to the terms above:b a nozzlep a reinforcement plates the shell

- additional pressure loaded area for an oblique nozzle connection


21/45

17

Some of these symbols are explained in Figure 2.7.

Figure 2.7 Cross-section of a set-on nozzle showing pressure- and reinforcement-areas [4]

2.3.8 Lifting eyesThe method for calculation of shells subject to local line loads can be directly appliedto lifting eyes without reinforcing plates. Lifting eyes with reinforcing plates can becalculated as a superposition of local line loads. The force acting on the lifting eye is

calculated and compared to the maximum allowed forces on the shell in longitudinaland circumferential direction.

3 MethodMost of the pressure vessel components have been dimensioned using DBF methodsgiven in SS-EN 13445-3. MATLAB have been used for the calculations. IndividualMATLAB programs have been written for each of the sections of SS-EN 13445 thathave been used. In comparison with manual calculations this gives the possibility toalter the input values in a much easier way. Iterative calculations can also be madevery easily.

For design and creation of drawings the 3D CAD software SolidWorks have beenused. 3D models have been created with dimensions from the DBF calculations andfrom drawings of the pressure tank which the design is based upon, discussed furtherin Section 4.1. FEM analysis on the 3D models has been performed usingSolidWorks Simulation.

4 Conditions for the designSome design features used on the first LVMS are very basic and also applies to thesecond LVMS. The pressure vessel must have an opening large enough to install the

components inside. To minimize the potential damage if water leakage should occur,all openings should be as close to the bottom end as possible. The most obvious way


22/45

18

to achieve this is to design the pressure vessel as a vertical cylinder where the bottomend is removable. The openings for cables can be placed either on the removable lidor on the lower part of the cylindrical shell. The ends of the cylinder can be dished orflat; tough a flat end would have to be considerably thicker and heavier to withstandthe pressure. The pressure vessel will be attached to a concrete foundation. It must be

big enough to stand stable on the sea floor considering the buoyancy of the vessel.

4.1 Manufacture of the vesselThe design of the pressure vessel will be based on a pressure tank manufactured by

Maschinen und Behlterbau GmbH(MABE) in Germany and retailed byEsska teknikin Sweden. The reason for this is economical; to buy a mass-produced vessel and thenmodify it will be cheaper than to manufacture a single customized vessel. Thestandard pressure tanks from MABE are available as both horizontal and verticalmodels in sizes from 50 to 10000 liters. The tanks have dished end of korbbogen type.The vertical 5000 liter model is available with diameter 1400 and 1600 mm. The needfor space in the marine substation has been evaluated and the 5000 liter model withdiameter 1600 mm has been selected as the most suitable. All MABE tanks areavailable as models designed for a maximum pressure of 11 or 16 bar. The lessexpensive 11 bar model will be used as both will have a wall thickness more thansufficient for the marine substation.

The MABE tank has four footings on the bottom end. These could be used toattach the tank to the foundation. Some reinforcement might be needed as the footingsare not dimensioned for the drag force they will be subjected to when the substation islifted. Another alternative is to mount the foundation with longer beams attached tothe side of the cylindrical shell. The tank also has lifting eyes but they are

dimensioned for lifting only the empty tank and will probably be too weak for liftingthe substation with its foundation.The MABE tanks are fitted with connection sockets for use with compressed air

and a manhole. These features will not be needed on the marine substation. Thebottom dished end will also have to be separated from the rest of the shell. The choicewill be to either order a tank without these features or to order a standard tank andthen modify the unneeded features while the other necessary modifications are made.The latter alternative will involve some unnecessary work but which alternative is themost inexpensive is not obvious since a specially ordered tank will result in a higherprice.

4.2 Protection against corrosionBeing submerged in seawater, the substation will be exposed to a very corrosiveenvironment. Protection from uniform corrosion and galvanic corrosion must beconsidered.

4.2.1 General corrosionGeneral corrosion means that a metal surface reacts with water and forms iron oxide,also known as rust. The reaction occurs uniformly over the entire surface and resultsover time in decreased thickness of the material. General corrosion can be avoided bycoating the surface e.g. by painting or choosing a corrosion-resistant material. Thematerial choice must however also consider the conditions for galvanic corrosion.


23/45

19

4.2.2 Galvanic corrosionGalvanic corrosion occurs when two electrochemically dissimilar metals are incontact and surrounded by an electrolyte. The difference in electric potential willresult in a current flow trough the electrolyte. The metal with the higher potential willact as the anode and corrode faster, while the metal with lower potential will act as the

cathode and corrode slower. The rate of galvanic corrosion on the anode can be muchhigher than the general corrosion rate.

Electric potentials can be compared with an anodic index; lower index means thatthe metal is nobler. Some examples of anodic index for different metals are seen intable 4.1.

Table 4.1 Anodic index for various metals3

Metal Anodic index [Volt]

Gold 0

Brass 0.40-0.45

Stainless steel 0.50-0.60Plain carbon steel 0.85

Zinc 1.25

Magnesium 1.75

The rate of corrosion is also dependent of the areas of the metals. Small anodic areaand large cathodic area means the corrosion rate on the anode will be higher. For thisreason, small details such as bolts and nuts should never be of a less noble metal thanthe rest of the structure [9].

One way of avoiding galvanic corrosion is to use metals with similar potential.For a salt water environment the difference in anodic index should be less than 0.15 V3

. The pressure vessel steel used for the substation is a plain carbon steel and cantherefore not be allowed to have contact with i.e. stainless steel or brass, whilesimultaneously having contact with water. Other ways to avoid galvanic corrosion isto either isolate the metals from each other or isolate the metal from the electrolyte.

So-called cathodic protection means that sacrificial anodes of metal with highelectric potential, e.g. zinc, are used [9]. The sacrificial anodes are in contact with theelectrolyte and attached in a conducting way to the structure. The anodes will thencorrode while the rest of the metal in the structure will be protected.

The protection against galvanic corrosion applies to the design of the substation ina number of ways. One example is the choice of material in the cable glands(discussed further in section 4.4.2), which are available in brass, stainless steel or

plastic. The two metals both have lower electric potential than the steel in the shell.This means that the corrosion rate of the shell would be increased. It would bedifficult to isolate the cable glands from the shell or from the water so the best choicewith aspect to corrosion is to use plastic cable glands. The vessel will also be paintedwith several layers of anti-corrosive paint and sacrificial anodes of zinc will beattached to the outside of the shell.

4.3 Mounting of componentsThe electrical components can be mounted either on the inside of the shell or on aframe placed inside the vessel. In the first LVMS, the components are mounted on flat

3 Engineers Edge, Galvanic Corrosion, (2009-06-05)


24/45

20

steel plates welded to the inside of the cylindrical shell. This concept will be used andfurther developed in the new LVMS. The components with high losses should be indirect contact with the shell to get sufficient cooling. The internal structure must bedesigned so that these components will have a large contact area with the inside of thecylindrical shell. Other components with less need for cooling can be mounted on an

external structure that is inserted into the vessel.

4.4 Connection and sealing of cables4.4.1 ROV adaptationThe LVMS must be adapted to the use of ROVs for cable coupling. This means thatthe type of connectors to be used as well as their placement must be considered. Onthe first LVMS the connectors are placed on the lower dished end, normal to the shell.This means that they all have an individual angle which makes it difficult for a ROVto reach them. Mounting the connectors directly on shell of the pressure vessel is then

not a good alternative. Having the connectors placed horizontally with the same angleis a less problematic set-up for cable coupling with an ROV [4].

It has been decided that the connectors will be mounted on one or several plates,placed on the foundation or on the vessel shell. The sealing and the connector for eachcable will then be separated. This solution is flexible since the ROV adaptations onlyhave to consider the connector plates and not the pressure vessel itself.

4.4.2 Cable sealingThe sealing of the WEC input cables and the sensor equipment cables will consist of

cable glands screwed into the nozzles, see Figure 4.1. Cable glands rated IP68, whichmeans waterproof up to a pressure of 15 bar, are available in different materials andfor different diameters of cables4. Available cable glands with an outer connectionthread M40 and a sealing range for cables of diameters 20-32 mm would be suitablefor the cables mentioned above. The nozzles will also be filled with a sealingcompound to give extra protection against leakage. For the AC input from the firstLVMS and the AC output, thicker cables might be necessary. How the sealing aroundthese thicker cables will be designed has not been decided, but a maximum innerdiameter of 70 mm for the nozzles has been specified.

This means that two sizes of nozzles must be designed. At least 20 nozzles withinner diameter 40 mm will be needed for the WEC and sensor cables, and two nozzles

with inner diameter 70 mm for the AC in- and outputs.

4 Pflitsch blueglobe cable gland,


25/45

21

Figure 4.1 Disassembled cable gland4

4.4.3 Placement of openingsThe cable openings can be placed either on the bottom end or on the lower part of thecylindrical shell. The main advantage of openings on the lower end is that the level ofpotential leakage is as low as possible since the openings will be below the flange.The disadvantage is that the closing and opening of the lid will be more complicated

since the cables go through it. Placing the openings on the cylindrical shell means thatall components and sealing around cables can be mounted independently of the lidand shorter cables can be used. The disadvantage is that the higher level of potentialleakage means that the space available for mounting components safely will besmaller.

4.5 Buoyancy of the substationThe pressure vessel has a volume of around 5 m3 and the weight of the vesselincluding the internal components will be much less than 5 tonnes, so the vessel will

have positive buoyancy, it will float. The weight of the concrete foundation must givenegative buoyancy enough to make the substation stand steadily on the sea floor. Theneeded mass of the concrete foundation to give zero buoyancy can be calculated by

1

=

concrete

water

vesselwatervesselconcrete

Vmm

[kg]. (21)

It would be suitable to give the foundation a weight that gives the entire substation apositive buoyancy of a few hundred kilos. Adding this relatively small weight will

submerge the substation. This makes it possible to submerge and elevate the LVMSwithout the need for a ship with a large crane.

4.5.1 Using a flat lower endIf the cable openings are placed on the cylindrical shell, the space inside the lowerdished end will be unused. An alternative design could then be to use a flat lower end.A flat end would have to be much thicker than a dished end to withstand the pressure,but for the relatively low design pressure, it could be a viable solution. The advantagewould be a smaller volume and larger weight of the vessel, which means lessbuoyancy and that the overall weight of the vessel with its foundation could be lower.

If a flat end is used there is no need for footings underneath it and the vessel could bebolted directly to the foundation outside of the flange.


26/45

22

5 ResultsThis section presents calculation results for different pressure vessel components to beused in the design of the LVMS. An overview of the components can be seen inFigure 5.1. All calculations have been made according to DBF methods in SS-EN13445. Some FEM analysis results for components not covered by DBF methods arealso presented.

Figure 5.1 Overview of pressure vessel components for two possible design alternatives

5.1.1 MaterialAll components are calculated to be constructed of pressure vessel steel P265GH withthe following specifications:Material number 1.0425 according to European standard EN 10028-2Reh = 265 N/mm

2

Rm = 410 N/mm2

E = 212 GPa = 0.3

The material will not be subjected to high temperatures that significantly affect thestrength during operation.

5.2 DBF calculation results5.2.1 Shell without stiffenersCalculations have been made to find the maximum internal and external pressure forthe given geometry. The maximum internal pressure must be higher than 0.3 MPa for

operating conditions and higher than 0.45 MPa for test conditions. The maximumexternal pressure must be higher than 0.3 MPa.


27/45

23

The data for the geometry is taken from a drawing supplied by MABE. Acorrosion allowance of 1 mm and a thinning allowance of 0.4 mm have beenconsidered for the shell calculations. The joint factor z = 0.7 have been used whichgive the most conservative results and allows the vessel to be placed in testing group 4for the inspection.

Stress limit for operating conditions is according to Eq. 3: f = 0.9Rm/2.4 = 153.8N/mm2

Stress limit for test conditions is according to Eq. 4: ftest = Reh/1.05 = 252.4 N/mm2

Geometry of the cylindrical shell

External diameter de = 1600 mm Nominal thickness en = 7.1 mm Calculation thickness ea = 5.7 mm for operating conditions and ea = 6.7 mm

for test conditions

Geometry of the torishperical ends

External diameter de = 1600 mm Internal large radius r1 = 1280 mm Internal knuckle radius r2 = 246.4 mm Nominal thickness en = 6.8 mm Calculation thickness ea = 5.4 mm for operating conditions and ea = 6.4 mm

for test conditions

The calculated maximum allowable internal and external pressures for the

unsupported shell are shown in Table 5.1. The results show that the cylindrical shellmust be reinforced to withstand the required external pressure.

Table 5.1 Calculation results, maximum allowable pressures for the unsupported shell

Internalpressure [MPa],operatingconditions

Internalpressure [MPa],test conditions

Externalpressure [MPa],operatingconditions

Externalpressure [MPa],test conditions

Cylindricalshell

0.77 2.12 0.11 0.15

Dished end 0.89 1.77 0.48 0.66

5.2.2 Shell with stiffenersSuitable stiffeners have been designed by testing different geometries. The maximumallowable pressures must be larger than 0.3 MPa and the stress in the rings must belower than the yield strength 265 N/mm2.

The chosen design uses two rings on the outside of the shell. The unsupportedlength is the distance from the flange at the bottom end, which can be considered aheavy stiffener, and the dished end at the top. The lower 2/5 of the dished end heightof 430 mm is considered unsupported.


28/45

24

Geometry of the stiffeners

Axial height 10 mm Radial width 55 mm Unsupported length between stiffeners 658.3 mm External diameter of shell de = 1600 mm Calculation thickness of shell ea = 5.7 mm

The calculation results in Table 5.2 show that this geometry reinforces the shellenough for the external pressure.

Table 5.2 Calculation results, maximum allowable external pressures for reinforcedshell, and stresses in the rings

External pressure to avoidcollapse between rings [MPa]

External pressure to avoidelastic instability [MPa]

Stress in rings[N/mm2]

Operating Test Operating Test Operating Test

0.36 0.50 1.59 2.16 235 167

5.2.3 Flat endCalculations have been made to find the required thickness of a bolted flat end to beused instead of the lower dished end of the pressure vessel.

Section 10 in SS-EN 13445-3 covers bolted flat ends with narrow- and full facegaskets but not o-ring gaskets. Required thickness of flat end bolted to a full faceflange with a soft gasket covering the entire surface can be calculated with Eq. 17. Abolt circle diameter C = 1680, gives a required thickness of e bolted = 28.9 mm. The

required thickness of the flange part will then be 0.8 e bolted = 23.1 mm and thislower thickness must be limited to a region outside of a circle with diameter 1176mm.

5.2.4 FlangesCalculations have been made for full face flanges both with o-ring sealing and withsoft gasket covering the entire surface. In both cases the flange is bolted with 32 M248.8 bolts. The flanges have stepped bore, no tapered hub and are welded to the outsideof the shell. The o-ring diameter is 10 mm.

Figure 5.2 Geometry of a flange with o-ring groove


29/45

25

Geometry of flanges

Outer flange diameter 1760 mm Inner flange diameter 1600 mm Bolt hole diameter dh = 26 mm O-ring groove mean diameter 1630 mm Depth of o-ring groove 8.4 mm Width of o-ring groove 13.5 mm

For the soft gasket flange the required thickness is to a large extent dependent of thegasket factor m and minimum seating pressurey. The lower values of the gasketfactors in Table 5.3 is for soft rubber with hardness below 75 IRH (InternationalRubber Hardness Degrees), and the higher values is for rubber above 75 IRH, valuesare taken from table H.1 in SS-EN 13445-3.

Table 5.3 Calculation results, required thickness of flanges

Chosen bolt circlediameter C [mm]

Minimum thickness [mm] Chosen thickness,e [mm]

m=0.5,y=0 m=1,y=1.4Gasket coveringentire surface

1680

19.2 28.8

30

O-ring sealing 1694 17.0 20

5.2.5 NozzlesCalculations have been made to find sufficient dimensions to give enoughreinforcement for the nozzles placed on either the dished end or the cylindrical shell.The quota Pressure area / Reinforcement area (R/P) have been calculated according to

the Pressure area method described in Section 2.3.7, both for single and multiplenozzles. For single nozzles the maximum internal pressure has also been calculated.For nozzles placed on the cylindrical shell calculations must be made for both thetransverse and longitudinal cross-section, while all directions are symmetrical for aspherical shell.

The calculations also consider how the reinforcement area is affected by thedistance between the nozzles and the distance to discontinuities in the shell. For thedished end, the distance to a discontinuity is the distance to a circle 0.8 the externaldiameter. For the cylindrical shell the distance to a discontinuity is the distance to theflange and the calculations assume the shortest allowed distance to the flange, whichis 20 mm for the given geometry. For the calculation of multiple large nozzles, the

closest surrounding nozzles are assumed to be large, which is a conservativeassumption since the small nozzles give more reinforcement.

The distances between the nozzles and the distances to the discontinuities on thedished end have been calculated assuming that the nozzles are placed in four rows asin Figure 5.3.


30/45

26

Figure 5.3 Placement of nozzles on the dished end

On the cylindrical shell the nozzles are placed between the bolt holes in the flange.This means that there is room for 32 nozzles around the perimeter of the shell and thedistance between the nozzles will be 157 mm.

All nozzles are of set-on type, welded only to the outside of the shell. The lengthof the nozzles is the maximum length that contributes to the reinforcement for thegiven diameters. Making the nozzles longer will not give more strength but longernozzles could be preferable if they are to be filled with sealing compound.

Geometry for small nozzles

Outer diameter 50 mm Inner diameter 40 mm External length 15 mm

Geometry for large nozzles

Outer diameter 80 mm Inner diameter 70 mm External length 25 mm

The calculation results in Table 5.4 and Table 5.5 shows that this geometry givesenough reinforcement.

Table 5.4 Calculation results, reinforcement from nozzles on dished end

R/P singlenozzle

Pmax[MPa] R/P multiplenozzels

Small 4.5 1.3 4.4

Large 4.2 1.3 4.0

Table 5.5 Calculation results, reinforcement from nozzles on cylindrical shell

R/P single nozzletransverse

Pmax [MPa],transverse

R/P single nozzlelongitudinal

Pmax [MPa],longitudinal

R/P multiplenozzles

Small 7.7 2.3 3.6 1.1 7.4Large 7.0 2.1 2.8 0.8 6.2


31/45

27

5.2.6 Lifting eyesLifting eyes have been designed to lift the entire marine substation vertically and tolift the vessel without the foundation horizontally. The critical aspect of thedimensioning of lifting eyes is to limit the local loads on the shell. The problem is thestrength of the shell, not the strength of the lifting eyes themselves. Reinforcement

plates placed between the shell and the lifting eyes can be used to distribute the forceand torque over a larger area of the shell.

The four lifting eyes on top of the vessel are dimensioned for lifting a total massof 10 tonnes vertically. One of the lifting eyes on the side combined with one of thelifting eyes on top are dimensioned for lifting 4 tonnes horizontally.

Figure 5.4 Angles of forces on lifting eyes, w1-w3 are angles to the direction of the weight, 1- 3 areangles to the normal of the shell, distances in mm

Figure 5.4 shows assumed angles when lifting the substation in vertical and horizontaldirection. Increasing angles will increase the actual forces but will also increase theallowed forces.

Figure 5.5 Dimensions of lifting eyes [4]


32/45

28

Geometry of lifting eyes

Thickness of reinforcement plate e2 = 10 mm Length of reinforcement plate b3 = 320 mm Width of reinforcement plate b2 = 80 mm Length of lifting eye b1 = b3/1.5 = 213.3 mm Distance from load to reinforcement plate a2 = 50 mm Eccentricity of load a1 = 40 mm

The geometry is explained in Figure 5.5 and the calculation results presented in Table5.6 show that these dimensions are adequate.

Table 5.6 Calculation results, force per lifting eye

Maximum allowed forceFR,max [kN]

Actual forceFR [kN]

Lifting eyes on top, vertical lift 29.1 26.0

Lifting eye on top, horizontal lift 46.9 22.0Lifting eye on side, horizontal lift 26.7 23.7

5.2.7 Dimensioning for larger pressureIn the future it might be necessary to place marine substations at greater depths than25 m. If this pressure vessel concept is to be used for the housing, the internalpressure would have to be larger. For this reason the possibility to dimension thisvessel for a pressure of 6 bar have been briefly investigated.

The cylindrical shell and dished ends would withstand 6 bar internal pressure

without modification. The nozzles could be made slightly thicker, which would notpresent any problems. If the vessel is to hold up an equivalent external pressure, thecylindrical shell can be designed with more and stronger stiffeners. The dimensions ofthe dished end would cause some problems since it will only withstand an externalpressure of 4.8 bar. Some kind of reinforcement would be needed, however EN 13445contains no DBF methods for reinforcement of dished ends.

The required thickness of both the flanges and the flat end is proportional to thesquare-root of the pressure and this gives a required thickness of around 22 mm forthe flange and 41 mm for the flat end. The flange would have to be bolted with 42 44 bolts if type M24 8.8 is used.

5.3 FEM analysis results5.3.1 FEM analysis of reinforcement for footings on the dished endThe footings on the lower dished of the MABE tank are intended for supporting thevessel when standing upright. If the foundation is mounted to the footings they will besubjected to a drag force when the substation is lifted and this means that the shellmight have to be reinforced. According to EN 13445, reinforcement of footingsshould be made with plates, similar to the way lifting eyes are reinforced. It is alsostated that footings should be placed on the large radius region of the dished end, butthe footings on the MABE tank are placed partly on the knuckle region. This means

that the footing would have to be moved if the DBF method for reinforcement is used.


33/45

29

FEM analysis of an alternative reinforce method have been performed. The shellis reinforced by welding supports to the in- and outside, as shown in Figure 5.7. Toreduce the number of elements, only one quarter of the lower part of the vessel hasbeen simulated. Because the vessel has four footings, symmetry constraints can beused.

Figure 5.6 Stress plot of dished end with 25 mm flange. Internal pressure 3 bar and a force of 15 kNapplied downwards on the footing. The plot shows high stress on the knuckle region above the footing.

Figure 5.7 Stress plot of the same conditions as in Figure 5.6 but with reinforcements on both sides ofthe shell. The stress above the footing is reduced to approximately the same level as in the surroundingshell.

5.3.2 FEM analysis of the bolted flat endA bolted flat end in combination with o-ring sealing is not covered by the DBFmethods in EN 13445. FEM analysis of the flat end has been performed to comparethe stress levels between the o-ring option and the full face gasket option. Based on

the calculations in section 5.2.3 and 5.2.4, a flat end and flange both of thickness 30


34/45

30

mm would be sufficient for an internal pressure of 3 bar, so this scenario has beeninvestigated with FEM analysis.

To reduce the number of elements in the simulation, only a section of 1/32 of thevessel has been modeled, since the flange has 32 bolts which make it possible to usesymmetry constraints. Also only short portion of the cylindrical shell is modeled,

since the high stress on the shell will emerge close to the flange.

Figure 5.8 Stress plots showing a section of a flat end with an o-ring groove bolted to a flange on acylindrical shell. The internal pressure is 3 bar. The flat end is in direct contact with the flange outsideof the o-ring groove which results in high stress in the shell above the flange. The lower plot hasexaggerated deformation.

Figure 5.8 shows a flat end with a o-ring groove bolted to a flange on the cylindricalshell. An internal pressure of 3 bar on the flat end and shell is applied. A bolt load

from a M20 bolt is simulated in the hole. The o-ring itself is not simulated, but thiswill not influence the results very much, since it would not contribute with anysignificant force in the rest of the structure. When the flat end is subjected to internalpressure it deforms. Since the flange is in direct contact with the flat end, torque istransferred to the flange and it bends upwards. This means that the thin shell is alsodeformed and results in unacceptable high stress in the shell.


35/45

31

Figure 5.9 Stress plots showing a section of a flat end bolted to a flange on a cylindrical shell. Theinternal pressure is 3 bar. The flat end has no simulated contact with the flange outside of the boltcircle, which results in much lower stress in the shell compared to the plot in Figure 5.8. The lower plothas exaggerated deformation.

Figure 5.9 shows a flat end bolted to the flange the same way as in Figure 5.8, andwith the same internal pressure. The difference is that the flat end in this case onlyinteracts with the flange inside of the bolt circle. Much less torque is then transferredto the flange which results in less stress on the thin shell. On the other hand the stressin the middle of the flat end will be higher since it gets less support from the flange.This is not a realistic simulation since the flat end penetrates the flange at the outerperimeter. In reality, a gasket with thickness of a few millimeters would be placedbetween the metal surfaces. However this simulation demonstrates the effect of usinga soft gasket, especially a narrow face gasket. The gasket creates a flexible gapbetween the metal surfaces which make it possible for the flat end to deform slightly

without transferring torque to the flange.A FEM simulation of a full face gasket would however require a much more

complex model to be realistic. The FEM simulation of the o-ring sealing case inFigure 5.8 requires a less complex model to be realistic, because of the metal-to-metalcontact. One can because of this assume that the required thickness of the flat endwith o-ring sealing would have to be more than 30 mm, which according to the DBFmethod is an adequate thickness when a gasket that covers the entire surface is used.

5.4 Buoyancy of the substationThese calculations have been made to find the required mass of a concrete foundation

to give the substation zero buoyancy. It would be ideal to have a mass slightly lower


36/45

32

than this to give the vessel small positive buoyancy. Additional mass will then beneeded to submerge the substation and fix it on the seabed.

The results have been calculated using Eq. 21 with volumes of the components takenfrom SolidWorks. The weight of the internal components is assumed to be 1000 kg.

One of the cases presented is using a dished lower end and filling it with ballast. Thiswill increase the weight without increasing the overall volume, so the result will bethat a lower total weight is needed to give suitable buoyancy. The lower dished endhas room for approximately 1200 kg concrete.

Table 5.7 Approximate weight of foundation and total weight to give zero buoyancy

Vessel withdished end

Vessel with dished endand 1000 kg ballast

Vessel withflat end

Mass of foundation [tonnes] 4.1 2.4 2.2

Total mass of substation [tonnes] 6.7 6.0 5.4

6 Discussion6.1 Summary of calculation resultsThe shell thickness is by far enough for the internal pressure, since the vessel isdesigned for 11 bar and 3 bar is used in this application. Using the conditions fortesting group 4 gives a maximum internal pressure of 7.7 bar. The cylindrical shellmust however be reinforced with stiffeners to cope with the external pressure.

The dimensioning of the nozzles has caused no problems due to pressure loads.

The dimensions of the nozzles will instead depend more on the sealing method usedfor the cables. Dimensioning strong enough lifting eyes for the vessel is no problem ifsufficiently large reinforcement plates are used. Reinforcing the footings on thedished end to handle the weight from the foundation will not be difficult if thatalternative is used.

The dimensioning of the flange and the choice of gasket type presents somechallenges. If an o-ring sealing is used, the sealing ability is less dependent of theforce from the bolts. This could be an advantage if the foundation is mounted to thelower end, since bolts in the flange then will carry the weight of the foundation duringlifting. The load on the flange from the weight of the foundation is however verysmall compared to the load from the pressure. The cost of an o-ring sealing will be

lower than the cost of a full face gasket, but on the other hand the machining of thegroove with high demands on surface finish will be more costly than the machiningneeded to use a full face gasket.

If the flat end alternative is used, it is advisably to use a full face gasket.According to the DBF method in section 2.3.5, a thickness of 30 mm is required.However this DBF method is based on a very simplified model based on limiting thebending stress in the centre of the end so that the yield strength is not exceeded. Theflange calculation method is also simplified and does not take in account whether adished or flat end is bolted to the flange [8]. This is a quite drastic simplification sincea flat end will cause much more bending stress on the flange. The result may be thateven if the stress limits are not exceeded, the deformation of the gasket could cause

leakage.


37/45

33

If an o-ring sealing combined with a flat end is to be used, more thorough FEManalysis and possibly some other design would be needed. One way of reducing thestress on the shell could be to use a flange with a tapered hub but that would increaseboth the material and manufacturing cost. Another option would be to reinforce theflat end, which would also mean much more costly manufacturing.

6.2 Design alternativesSome different design alternatives are presented based on the results of thecalculations above. The main choices that need to be made for the final design are todecide where the openings for cables will be placed, how the lid should be constructedand how the vessel will be attached to the foundation.

6.2.1 Placement of nozzlesThe calculations show that the suggested dimensions of the nozzles give enoughreinforcement to the shell both if they are placed on the cylindrical shell or on thedished end. The alternative that should be chosen is then decided by the utilization ofthe substation, since both alternatives give a strong enough vessel. Placing the nozzleson the cylindrical shell would be the best choice since it will make the opening andclosing of the vessel much less problematic. This alternative also makes it easier toincrease the number of nozzles, if that should be necessary. The only advantage ofplacing the nozzles on the dished end is the small increase of space available inside tomount electrical components.

If the underwater connectors are to be attached to plates outside of the vessel, andthe nozzles are placed on the cylindrical shell, the plates should also be attached to the

side of the cylindrical shell so that the lid and foundation can be removed while theconnectors are still mounted on the plates.

6.2.2 Lower end and foundationThree alternatives for attaching the vessel to the foundation and for the design of thelower end are here presented and discussed:

Alternative 1: Using the dished lower end and attaching the foundation with thefootings.

This alternative involves the least modification of the manufactured pressure tank.

The extra force from the weight foundation is no problem for the flange but thefootings need to be reinforced e.g. as shown in section 5.3.1. The space inside thedished lower end will be unused and it is possible to fill the end with ballast to reducethe buoyancy. The heavy lower end would however be quite difficult to handleespecially during maintenance of the substation. The easiest way opening and closingthe vessel would be by removing the bolts in the flange and then lift the entire upperpart of the vessel, while the lower end is still standing on its footings.

Alternative 2: Using the dished lower end, removing the short footings and attachingthe foundation with long beams mounted to the side of the vessel.

This design was used for the first substation. The lid would be as light as possibleand relatively easy to handle, assuming it is not filled with ballast. One possibleadvantage could be that the lid can be removed when the vessel is standing on the


38/45

34

beams, if they are long enough, but this would make the substation considerablyhigher than for the other alternatives.

Alternative 3 : Using a flat lower end bolted directly to the foundation.This design is more compact than the other alternatives. This results in the least

buoyancy and the lowest overall weight of the substation. The flange on the vesselwould have to be thicker than for the other alternatives. The weight of the lid will behigher than an empty dished end but lower than a ballast-filled dished end.

Figure 6.1The three suggested design alternatives drawn in the same scale with corresponding sizes ofthe required foundations

6.3 Future developmentThis thesis is based on the condition that the housing for the marine substation is to bepressurized before submersion and must therefore be designed and classified as apressure vessel. The other option, as mentioned in the introduction, is to pressurize thevessel during submersion. If the pressure difference between the in- and outside of thevessel is limited to less than 0.5 bar, the vessel is not classified as a pressure vesseland does not have to be certified. The method of pressurizing during submersion has

been used for the housings of linear generators in the Lysekil project, while the firstmarine substation has been pressurized before submersion.

One aspect of pressure vessel certification that should be considered is thepossibility to make design changes. If a pressure vessel has been certified there arevery little possibilities to change the design without going trough the certificationprocess again.

If the method of pressurizing the substation during submersion is tested andverified, it will probably affect the design conditions for future marine substations. Agood pressure-bearing design with a cylindrical shell and dished ends is not anentirely unproblematic design for easy installation of electrical components. So if it ispossible to guarantee that the pressure is limited to 0.5 bar or less, the vessel designcould be radically different. One possible approach, which has not been investigatedin this thesis, is to use a rectangular vessel.


39/45

35

Another aspect of the design process for this marine substation is that themechanical design of the housing was made before much of the electrical system wasdesigned. The alternative, to design all of the interior first and the exterior later, couldpossibly result in a mechanical design more adapted to the actual need for space andfor easy access to electrical components during installation and maintenance. These

aspects will however have more importance for possible future mass-production ofmarine substations. The construction of this marine substation is a research projectand will go through many design changes during the process, which requires that themechanical design is flexible.

7 Conclusions Constructing the pressure vessel for the LVMS by modifying a mass-produced

pressure tank is a viable solution compared to acquiring a customized pressure

vessel.

The pressure vessel can be designed with a large number of nozzles for cablesealing. A design that gives sufficient reinforcement to the shell can beaccomplished relatively easy, either if the nozzles are placed on the cylindricalshell or if they are placed the dished end. Placing the nozzles on the cylindricalshell will simplify the process of closing and opening the vessel when a largenumber of cables are connected to the substation.

Preparing the design of the substation for future cable coupling with ROVscan be accomplished by placing the cable connectors on a structure outside of

the pressure vessel, instead of directly on the shell.

By filling the dished lower end with concrete ballast, the required weight ofthe foundation to achieve suitable buoyancy of the substation can be reducedby approximately 41%, and the overall weight of the substation can be reducedby 10%. By using a flat lower end instead of a dished lower end, the requiredweight of the foundation can be reduced by approximately 46%, and theoverall weight of the substation can be reduced by 19%.

If this type of pressure vessel is to be placed at a greater ocean depth,modifying the design for the double internal pressure would not cause

significant problems. However, if it is required that the vessel also must beable to withstand the same external pressure, the strength of the dished endwould be a limiting factor.

If future marine substations are not designed for high pressures which requirepressure vessel certification and instead are pressurized during submersion, thedemands for a pressure bearing design will be lower which could allow for adesign more adapted for easy installation of electrical components.


40/45

36

8 References[1] Falnes, Johannes, Ocean waves and oscillating systems, Cambridge UniversityPress, 2002

[2] Waters, Rafael,Energy from ocean waves, Uppsala University, 2008

[3] Bernhoff, Hans, Wave power compendium, third edition,Uppsala University, 2006

[4] Lundgren, Frida and Grahn, Pia, Deploying and Coupling Underwater Cablesusing Remotely Operated Vehicles (ROVs), Uppsala Univeristy, 2009

[5] Swedish Standards Institute, SS-EN 13445:2002 Tryckkrl (ej eldberrda), SISfrlag, 2004

[6] Arbetsmiljverket,AFS 1999:04 Tryckbrande Anordningar

[7] Baylac, Guy and Koplewicz, Danielle,EN 13445 "Unfired pressure vessels"Background to the rules in Part 3 Design, issue 2,Union de Normalisation de laMcanique, 2004

[8] Flintney, Robert, Seals and sealing handbook, fifth edition, Elsevier/Butterworth-Heinemann, 2007

[9] Wrangln, Gsta,An introduction to corrosion and protection of metals, Institutetfr metallskydd, 1972

[10] Sundstrm, Bengt,Handbok och formelsamling i Hllfasthetslra, KTH 1999

[11] Bjrk, Karl, Formler och tabeller fr mekanisk konstruktion, sjtte upplagan,Karl Bjrks Frlag HB Mrsta


41/45

37

AcknowledgementsI would like to thank my supervisor Magnus Rahm for all the help, the other staffmembers and students at the division of electricity who have assisted me and finallymy family and Nika for their support.


42/45

() &

% $

&'

$%

N

lUO

$

':

*1

2

7,7/(

5(9,6,21

'21

27

6&$/('5$:,1

*

0$7(5,$/

'$7(

6,*1$785(

1$0(

'(%

85$1'

('

*(

6

'

&+.'

$1

*8/$5

4$

%5($.

6+$53

),1,6+

$339'

0)

*

81/(

6627+(5:,6(

63(

&,),('

',0(1

6,21

6$5(,10,//,0(7(5

6

685)$

&(),1,6+

72/(5$1

&(

6

/,1($5

'5$:1

$

%

6&$/(

'(7$,/%

SDFNQLQ

JV\WDQ

SODQ

VYDUY

DVVRP

VSnUHWLE

RWWHQ

SODWWDQ

'(7$,/$

6&$/(

Appendix 1

Assembly drawings of the vessel


43/45

VWXWVDUQ

DV

SOD

FHULQ

J

VQLWW

XSS

LIUnQ

,7(01

2

3$571

80

%(5

'(

6&5,37,21

47


44/45

,7(01

2

3$571

8

0%(5

'(

6&5,37,21

47


45/45

5

6(&7,21$$

%

6&$/(

'(7$,/%

VSnUI|U

S

I|U

SODQK

PD[

[

5

[

[5

$

$

'(%

85$1'

('

*(

6

),1,6+

$1

*8/$5

4$

0)*

$339'

&+.'

'() &

% $

ERWWHQ

':

*1

2

1$0(

6,*1$785(

'21

27

6&$/('5$:,1

%5($.

6+$53

7,7/(

'$7(

0$7(5,$/

81/(

6627+(5:,6(

63(

&,),('

',0(1

6,21

6$5(,10,//,0(7(5

6

685)$

&(),1,6+

72/(5$1

&(

6

/,1($5

'5$:1

Appendix 2

Drawing of the flat end

Documents

FULLTEXT01 korozyonnn111