Fcs Concept Report

A CONCEPTUAL LNG CONTAINMENT DESIGN UTILIZING EXPLOSION WELDED TRANSITION JOINTS

BY CHAD N. FUHRMANN

2

1. ABSTRACT

2. INTRODUCTION a. Short History b. What is LNG? c. Future Growth in LNG market

3. DESCRIPTION OF LNG CONTAINMENT AND TRANSPORT a. Basic Types and Designs of LNG Vessel Containment Systems i. ii. iii. iv.

Containment Design Considerations Kvaerner-Moss GTT No 96 Membrane Containment System GTT Mark III Membrane Containment System

b. Advantages/Disadvantages of Membrane vs. Moss Containment Designs

4. STAINLESS STEEL AND INVAR® VS. ALUMINUM IN LNG CONTAINMENT a. Thermal Properties of Each Metal i. Invar® 1. Coefficient of Thermal Expansion 2. Coefficient of Thermal Conductivity 3. Effect of Thermal Properties on Containment Design ii. Stainless Steel 1. Coefficient of Thermal Expansion 2. Coefficient of Thermal Conductivity 3. Effect of Thermal Properties on Containment Design iii. Aluminum 1. Coefficient of Thermal Expansion 2. Coefficient of Thermal Conductivity 3. Effect of Thermal Properties on Containment Design b. Kvaerner-Moss Design Considerations and Explosion Welded Transition Joint

3

5. INTRODUCTION TO EXPLOSION WELDING OF ALUMINUM AND STEEL a. Process and Method b. Basic Properties of Explosion Welded Aluminum and Steel Transition Pieces

6. DESCRIPTION OF PROPOSED DESIGN FOR LNG TANKER CONTAINMENT SYSTEM a. Containment Cell Design i. Aluminum Interior Similar to Moss Design ii. Advantages of Design 1. Aluminum Interior with Cryogenic Advantages 2. Reduced Material Usage 3. Reduced Weight 4. Simplicity/Relative Ease of Installation b. Explosion Welded Transition Joints/“Floating” Design i. Description 1. Basic Design 2. Materials Used ii. Advantages 1. Reliable Structural Support in Lieu of Wood InsulationFilled Boxes 2. Collision Safety c. Piping Considerations

7. FURTHER CALCULATIONS a. Coefficients of Thermal Expansion and Thermal Conductivity i. Containment ii. Transition Joints iii. Vessel Support Structure iv. Insulating Foam b. Shear/Compressive Strength i. Normal Loading ii. Collision

4 8. LIMITATIONS OF PROPOSED DESIGN a. Lack of Information Available i. Current Designs Used in Naval Applications ii. Design Details are Closely Guarded b. Possible Design Flaws i. Rough Estimates For Calculations ii. Unproven Design

9. CONCLUSIONS

10. WORKS CITED

11. WORKS CONSULTED

5 Abstract This paper introduces a new conceptual design for a liquefied natural gas (LNG) vessel containment design.

This “Floating Containment System,” or FCS, takes

advantage of recent advances in explosion welding, particularly between aluminum alloy 5083 and carbon steel for use as transition joints between the aluminum LNG containments and the vessel’s hull. The new design will be studied using the coefficients of thermal expansion for each of the metals used and the advantages and disadvantages of the design will be described as applicable to cryogenic fluids in general and liquefied natural gas in particular.

Introduction Up until the 1970’s, most natural gas was burned off at wells in the oil field. When it was recognized that this gas could be used for energy production, it was stored for future use or re-injected back into oil deposits to assist in maintaining pressure within the well. Today, natural gas is recognized as an extremely efficient, cost effective and clean-burning fuel. It is being utilized in gas and steam turbines for electrical power generation, used as an alternative fuel in transportation, heating homes and office buildings and finding uses in countless other areas of life. Natural gas is a fossil fuel found, as its name would suggest, in gaseous form and is the byproduct of the anaerobic decay of organic material. The natural gas is trapped and then processed for transport and used for energy production is normally found in oil and natural gas fields but is also produced as biogas in places such as landfills and swamps. Natural gas is made up of 70-90% methane (CH4), 5-15% ethane (C2H6),

6 propane (C3H8) and butane (C4H10), as well as small amounts of carbon dioxide, nitrogen and other hydrocarbons and sulfur-containing gases. Before it is transported any great distance, natural gas undergoes extensive processing.

This processing removes acids, moisture, mercury, nitrogen and other

contaminants producing a gas that is roughly 95% methane. Demand for the fuel is growing, primarily for residential and commercial heating, power generation and for industrial heat processes. As concerns for the environment grow, demand for cleaner burning personal and public transportation will also increase the demand for natural gas.

Description of LNG Containment and Transport There are five basic sets of conditions for storage and/or transport of liquids or gases: 1. Liquid at atmospheric pressure and temperature (atmospheric storage); 2. Liquefied gas under pressure and at atmospheric temperature (pressure

storage); 3. Liquefied gas under pressure and at low temperature (refrigerated pressure storage, semi-refrigerated storage); 4. Liquefied gas at atmospheric pressure and at low temperature (fully refrigerated storage); 5. Gas under pressure (Mannan pp.4-5).

7 For ease of pumping and transport via ship, natural gas is contained as per conditions 3 or 4, depending on the containment method. For these conditions, the natural gas must be liquefied. The liquefaction process occurs when the temperature of the gas is reduced to approximately − 160°C (− 260° F ) . This extensive cooling reduces

the gas to a liquid that is one six hundredth of its initial volume making containment and transport much more economically feasible. The extremely low temperature, however, creates a unique set of difficulties that must be addressed. Any fluid with a boiling temperature of − 89.9°C (− 130° F ) or lower at atmospheric pressure ( 14.7 psia or 101.3kPa absolute) is by definition a cryogenic fluid

(2006 International Fire Code p.279) with inherent dangers and handling/containment precautions over and above those normally associated with a flammable liquid. At a temperature of − 160°C , liquefied natural gas falls well within these parameters. In addition to the precautions that must be taken in regard to safety of personnel when handling LNG, there are also rules and guidelines that must be followed regarding its containment and transport on board LNG vessels. Dry cargo or liquid petroleum vessels are constructed of various types of steel depending on design criteria and the owner’s preference and budget. However, LNG, due to its cryogenic state must be treated differently. LNG is transported in vessels specifically designed to carry it. The fluid is stored on board these vessels in specialized containments that can only be constructed of specific materials. There are two basic designs used for LNG vessels today. The general shape and specific structure of these two containment systems is very different. The first design is the Kvaerner-Moss spherical containment design. This design was introduced in 1965

8 and is the property of the Norwegian Kvaerner Group (www.akerkvaerner.com). The design consists of a spherical aluminum tank interior covered in steel with a layer of insulation in between. This design relies on an explosively welded transition joint as a means of securing the aluminum sphere to the steel structure of the hull. It is this transition joint that allows the combined advantage of the excellent cryogenic properties of aluminum and the economic and strength advantage of the vessel’s structural steel.

The second type of LNG vessel containment is the membrane design. Much like the Kvaerner-Moss design, this design consists of the membrane inner surface that is in contact with the liquefied natural gas, surrounded by thick layers of insulation. The entire containment is encased within the protective structural steel shell of the vessel’s hull. While the spherical Kvaerner-Moss design is a one piece (relatively simple) design

9 that can be simply dropped into the hull of the vessel during construction, the membrane type design is more complicated in its design and construction.

Figure 2--General Membrane Type LNG Vessel Containment Design

Within the membrane design group are two unique designs designated as the Gaz Transport & Technigaz No. 96 and Gaz Transport & Technigaz Mark III Membrane Containment Systems. In both of these containment systems, the liquid cargo membranes are independent of the vessel’s structure, supported instead by a base of insulation and plywood that surrounds the containment. Neither of the systems is anchored directly to the steel hull of the vessel. Generally, the No. 96 and Mark III Membrane Containment Systems are very similar in their basic designs. However, each system utilizes different insulation and membrane materials. The GTT No. 96 Membrane Containment System is built upon a base of stacked plywood boxes filled with perlite insulation. On top of the stacked boxes

10 is layered an Invar membrane. Invar® is an iron and nickel alloy known for its uniquely low coefficient of thermal expansion. The Invar® membrane is anchored in the stacked layers of plywood boxes.

Invar® Membrane Primary Box Level

Secondary Box Level

Steel Tank Top

Figure 3--GTT No. 96 Membrane Containment System Construction (Curt p. 19)

Stainless Steel Membrane

Triplex® Liner Sheet

Plywood Base Layers Foam Insulation

Figure 4--GTT Mark III Membrane Containment System Construction (Curt p. 21)

11 The GTT Mark III Membrane Containment System is layered similarly to the No. 96 System. However, the Mark III System consists of layers of insulating foam divided by layers of plywood and a Triplex® plastic liner. The Triplex® plastic liner serves to strengthen the overall membrane and also acts as an additional layer of insulation and as a moisture guard. On top of the insulation, a baffled stainless steel membrane is installed. Unlike the Invar® membrane of the GTT No. 96 System, expansion and contraction of the stainless steel membrane, with its coefficient of thermal expansion of 17 mm

mm K

,

must be considered in its design and construction. The baffles of the stainless steel membrane thus serve to absorb the thermal movement of the material. The Kvaerner-Moss spherical containment design and the GTT membrane containment designs each have advantages and disadvantages in comparison to one another. The spherical containment has two distinct advantages over the membrane design. First of all, the spherical tanks are simple in design and construction. They can be completely fabricated before and during the construction of the vessel and then installed in one piece on the ship. The aluminum and steel transition joint offers a degree of structural stability that is not found in the membrane design. However, the spherical shape of the containment does not fit the generally rectangular design of the vessel. While the unused cargo space on this type of design serves as an additional safety zone in the event of collision, it is also serves as dead space within the vessel that can be more economically used in the better fitting membrane design. The economic advantage of the membrane design is the primary reason that it is the preferred design for LNG transport. Because of the better use of space in the membrane containment system, it can carry more product on a vessel of similar size in

12 comparison to the Kvaerner-Moss design.

To account for this, vessels using the

Kvaerner-Moss design must be on average 10% longer than the LNG vessels using the membrane design. Thus, the construction cost is inherently lower for a LNG vessel using a membrane containment system due to its comparatively smaller size.

The size

difference between vessels also accounts for increased shipyard capacities, less difficulty with canal passage, etc. The “collision zone” created by the additional space in the Kvaerner-Moss spherical containment is a safety advantage of this particular design. The same dead space that makes the design less efficient also serves as an impact zone, protecting the spherical containments from damage in the event of a collision or grounding. The membrane containment systems create their own collision safety zone by virtue of the independent design of the membranes. Again, the membranes are not directly supported by or connected to the structural members of the vessel. In the event of a marine casualty, the layers of insulation surrounding the containments creates a cushioned safety zone around the cargo holds and the unattached design of the membrane resists the damage from being transferred from structural members to the containment.

Stainless Steel and Invar® vs. Aluminum in LNG Containment

Members of the Workforce should verify that material and equipment that are used in cryogenic applications are constructed only of materials that do not become brittle and hazardous at low temperatures (Shrouf, p.7).

13

Material Properties

304 Titanium Mild Aluminum Aluminum Stainless 321 Cp2 Steel Alloy Alloy (6061Annealed Stainless Grade 2 Invar® 1010 (6061-T6) Untempered) ASTM Annealed ASTM B CREW A269 338

Tensile Strength

MPa (20 C)

310

124

380

590

620

50,000

345

Yield Strength

MPa (20 C)

275

55

275

240

245

275

360

Elongation

Percent

12

25

20

55

55

20

57.5

Density

Kg/m3

2700

2700

7800

8000

8000

4500

8100

Modulus of Elasticity

x 103 MPa

0.07

0.07

0.20

0.19

0.19

0.10

0.15

Coefficient of Thermal /K (or C) (x 10-6) Expansion (20 C)

23

23

11

17

17

20

1.2

Coefficient of Thermal Conductivity

225

225

45

20

20-25

15.6

125.7

W/m K (20 C)

In addition to construction and design, the major difference between the Kvaerner-Moss and GTT containment designs is the type of material used that is in contact with the liquefied natural gas.

The above table illustrates the different

characteristics of a selection of different materials. Aluminum, Invar® and stainless steel are all materials used in the different containment types. The design of the containment determines the type of the material used in its construction because each of the three materials reacts differently to temperature extremes. A lower or higher coefficient of thermal expansion may be incompatible or uneconomical with a given design. Invar®, as used in the GTT Mark III Membrane Containment system is an alloy consisting of 64% iron and 36% nickel with additional carbon and chromium. The coefficient of thermal conductivity of Invar® is approximately 125.6 W

mK

at 20°C .

This value falls between stainless steel on the low end and aluminum on the high end of the three materials in question.

14 Invar®, however, is especially known for its low coefficient of thermal expansion. While the average value at low temperatures is 1 × 10 −6 mm

mm K

, some

formulations of the material have a negative value for the coefficient of thermal expansion (Incropera, p.537). The lower the value of this coefficient, the less a material will expand or contract as its temperature varies. As the lining of the Mark III system, Invar® is therefore an excellent material in that it is largely unaffected by the temperature of the cryogenic LNG. The minimal contraction of the material at low temperatures minimizes the risk of fractures due to tension stress in the material. Likewise, when the membrane is subject to atmospheric conditions when empty, the material will not expand to the point of buckling with the increase in temperature. A disadvantage of Invar® is its tendency to creep. As the material is thermally stressed over a period of time it has a tendency to deform to relieve that stress. This could lead to weakened areas in the membrane that could eventually lead to leakage or rupture of the tanks, particularly in the event of a collision. The coefficient of thermal conductivity of the membrane material will have an effect the insulation used in the membrane construction. The perlite-filled plywood boxes used in the construction of the GTT No. 96 Membrane construction help to counteract the higher thermal conductivity of the Invar®. The perlite insulation has a lower coefficient of thermal conductivity (approximately 0.025 − 0.029 W

mK

at an average temperature

of − 126°C ) as compared to other forms of insulation. Perlite by itself, however, has little or no compressive strength due to its loose form. Thus, the insulation is used inside the plywood boxes which assist in adding compressive strength to the insulation layer along with the Invar® anchors that support the unattached membrane.

15 The GTT Mark III Containment System uses a membrane constructed of stainless steel. Described briefly earlier, this membrane design utilizes a baffle-type design to prevent stress in the membrane from leading to fractures or buckling over the wide temperature differential experienced between loading operations. Stainless steel has a coefficient of thermal expansion of approximately 17.0(× 10 −6 )mm

mm K

at 20°C . This

value is higher than that of Invar®, indicating that there will be more movement in the material over the expected temperature differential, hence the need for the baffled design. In comparison to Invar®, the coefficient of thermal conductivity of stainless steel is low, approximately 45W

mK

. In this design a foam insulation product is used. Foam

insulation can have a thermal conductivity in the area of 0.039 W

mK

at a temperature of

0°C and lower at cryogenic temperatures (Pittsburgh Corning). This is higher than that

of perlite insulation, but in combination with the lower thermal conductivity of the stainless steel, is suitable for the application. The foam has the added benefit of a relatively high compressive strength, about 620kPa , thus eliminating the need for the plywood boxes and simplifying the overall

construction of the membrane.

Plywood sheets are still incorporated as layers of

separation between the foam layers as is a Triplex® sheet which further improves the thermal conductivity and strength of the insulation layer. The Kvaerner-Moss spherical containment design uses aluminum as the construction material for its inner shell. The aluminum sphere is surrounded by foam insulation.

Perlite is not practical in this design for a number of reasons. First, the

insulation-in-box type of structure is not necessary by design as the tank is supported

16 directly by the vessel’s steel structure. Secondly, the loose nature of the perlite makes it impractical for installation around the exterior of the aluminum shell. Foam, though not necessary for support of the containment, is a solid material that can be pre-formed and will remain in place around the spherical structure once installed. Aluminum and its alloys have a thermal expansion coefficient of approximately 23(× 10 −6 )mm

mm K

. In comparison to the other materials, aluminum is therefore subject

to much more expansion and contraction over the anticipated temperature range. For this reason, the Kvaerner-Moss containment design can use aluminum as its interior material while it may not be practical for the membrane design. The spherical design of this containment system allows for the tension and compression forces experienced over the temperature range to be evenly distributed over the surface of the hold. If aluminum were to be used in a membrane-type containment system, the forces experienced in the corners (stress concentrators) of the membrane containment could be excessive and may very well lead to failure. The cryogenic nature of liquefied natural gas and the thermal characteristics of aluminum require that the Kvaerner-Moss containment be designed as it is. Meaning that the spherical shape of the containment is required to withstand the stresses involved in the containment. However, this shape requires that the Kvaerner-Moss vessel be larger than membrane type vessels to be able to carry a comparable amount of LNG. There are distinct advantages of aluminum in LNG containment and transport. Despite the higher coefficient of thermal expansion, aluminum has excellent cryogenic qualities. Unlike other materials, aluminum becomes stronger at cryogenic temperatures and is not prone to brittle fractures at these low temperatures.

17 Unlike the membrane containments described above, the aluminum sphere in the Kvaerner-Moss design is directly attached to the structural steel portion of the hull. This is accomplished by the aluminum and steel transition joint located around the circular base of the spherical tank. This is the critical area in this type of containment design. Again, because of the spherical shape of the containment, the forces applied to the structure of the tank are evenly distributed. Placing the transition joint around the base circumference of the containment allows the stress of thermal expansion and contraction to be evenly distributed around the ring’s perimeter, supporting the structure without the risk of fracture or deformation. However, having the containment system directly connected with the hull of the vessel has a disadvantage. In the even of a collision, the joint between the containment and the structure of the vessel presents a potential danger area. If the support is damaged, that damage could be transferred to the containment itself.

Introduction to Explosion Welding of Aluminum and Steel

In the Kvaerner-Moss containment design, the aluminum/steel transition joint is created by a process known as explosion welding. This is a method by which two dissimilar metals are welded together using tremendous force. The resultant bond is as strong as any normal welding process but leaves the individual characteristics of the dissimilar metals intact. In this instance, the aluminum retains its excellent cryogenic properties, while the strength and toughness of the steel is also preserved.

18

Figure 5--Basic Explosion Welding Process (DMC Clad Metal Groupe SNPE p. 1)

As seen in Fig. 5, the explosion welding process begins with preparation of the two materials to be joined (specifically aluminum and steel in this instance). After preparation is complete, the two sheets are placed a specified distance apart with the cladder material (aluminum) over the backer, or base material (steel). A precise amount of explosive powder is evenly spread across the aluminum and is detonated from a predetermined starting point. The explosion front travels across the face of the aluminum at a speed of 2000 − 4000 m forcing it down upon the steel base metal at pressures s ranging from 700 − 4,000 MPa (Murr p. 186, and High Energy Metals, Inc. p. 1). The

force of collision across the plate surfaces forces out oxides and other impurities promoting a clean metal-to-metal bond. The metals are forced together at such great

19 velocity and with such force as to briefly become viscous fluids. The metals “splash” together and bond with each other via interlocking waveforms.

Figure 6--Magnified View of Explosion Welded Joint (Murr p. 188)

Figure 7--X-Ray Photograph of Aluminum/Steel Transition Piece (Photo Courtesy of Joe Kass Photography, LLC, 2007)

20 The end result of the process is a single piece of metal consisting of steel and aluminum completely fused together across their surface areas each with their original characteristics still intact. These metal pieces are commonly produced in long strips approximately 1 − 380cm wide by 1.9 − 3.45cm thick and can be used in any number of applications, including transition pieces in LNG containment systems.

Figure 8--Explosion Welded Stock (Young p. 5)

Aluminum Plate

Aluminum/Steel Transition Joint

Steel Support Structure

Figure 9--Typical Explosion Welded Transition Joint (Merrem & laPorte p. 11)

21 The aluminum and steel transition joints manufactured by this process provide many advantages over traditional fastening methods. The bond between the two different materials provides a gap-free transition that greatly reduces galvanic corrosion.

In

interior compartments, where there is no exposure to seawater or sea air, corrosion is virtually eliminated. In areas where the transition point is subject to corrosion-inducing environments, regular painting or coating can inhibit and even prevent corrosion. Furthermore, the bond created between the metals is permanent and requires no further maintenance. This is in contrast to other fastening methods between dissimilar metals, such as nuts and bolts that can loosen or wear over time and exposure to vibration, thermal expansion, loading, etc. The explosion bonding process does not affect the original characteristics of the materials.

Contraction and expansion due to temperature changes still affect the

materials according to their thermal characteristics but the continuous joint, as mentioned earlier, evenly distributes any stress that may occur due to these thermal effects on the materials as well as any loading on the members they are a part of.

MATERIAL

Steel

Chemically Pure Aluminum

Aluminum Alloy 5083

Explosively Bonded Metal (Triclad®) 80 (min) 181 (typical)

Tensile Strength (MPa ) Yield Strength (MPa ) min % Elongation min Shear Strength (MPa )

380515

65-95

275-350

205

20

125

---

27

35

17

---

---

---

---

70 (min) 94 (typical)

Figure 10--Material Characteristics (Merrem & laPorte, pp. 6-7)

22 Proposed Design for LNG Tanker Containment System

The new LNG containment design proposed here attempts to take advantage of the technological advances made in explosion welding in a new way, offering better support of the containment cell and higher cargo capacity with less danger of containment damage in the event of collision or other catastrophic failure.

Figure 11--Proposed Containment Design Utilizing "Floating" Supports

Like the Kvaerner-Moss LNG containment design, the Floating Containment System utilizes the excellent cryogenic properties of aluminum for the main component of the LNG hold. Not only is aluminum excellent for cryogenic applications, it is also very light (approximately 2700 kg

m3

). This compared to the stainless steel and Invar®

used in the membrane designs, both of which have densities of approximately 8000 kg

m3

. The decreased weight of the construction material allows for more cargo to

be carried on a vessel of a given size, making its use more economical in comparison to other materials.

23 The FCS also utilizes a rounded shape to better withstand the contraction and expansion of the aluminum with its high coefficient of thermal expansion.

The

ellipsoidal containment cell lies within a similar shaped ellipsoidal shell (Fig. 10). While this creates more dead space within the vessel as compared to the membrane design, the modified design of the FCS ellipsoidal containment allows for more cargo capacity in comparison with the Kvaerner-Moss design and, potentially, a slightly lower profile. A vessel using the ellipsoidal containment allows for a capacity equivalent to a vessel of equal size using the Kvaerner-Moss design, however, the FCS would only require two ellipsoidal containments as opposed to the four required by the spherical containments on an identical vessel.

Figure 12--Ellipsoid

As an example, a large LNG vessel using the spherical containment design can carry approximately 155,000m 3 of liquefied natural gas with four spherical tanks each measuring approximately 42 meters in diameter. Utilizing two ellipsoid tanks with the same measurement for width and height and double the length of one sphere, the volumetric comparison between the two designs is as follows:

24 Volume = 4 3 ⋅ π ⋅ a ⋅ b ⋅ c , where a = 42m , b = 21m and, c = 21m

∴V = 4 3 ⋅ π ⋅ 42 ⋅ 21 ⋅ 21 ≈ 77,600m 3 76,600m 3 × 2 = 155,200m 3

While the FCS design will not result in a straightforward increase in cargo capacity, it will result in cost savings by reducing the amount of materials required to build two containment cells as opposed to four containments of the traditional KvaernerMoss design. To further accommodate the expansion and contraction of the aluminum containment cell and assist in its support, expandable foam insulation would be used in the void spaces surrounding the aluminum containment. Again, this is similar to the Kvaerner-Moss design in that the foam insulation can be pre-cut to conform to the curvature of the containment cell. This foam will be compressed upon installation to allow for expansion as the containment contracts. Finally, the FCS also utilizes explosion welded transition joints between the aluminum of the containment and the steel structure of the vessel. Like the explosion welded joint of the Kvaerner-Moss design, these joints are integral to the structure of the containment system. However, unlike the transition joint in the Kvaerner-Moss design, the FCS design uses “floating” joints (Figs. 12&13) to compensate for the expansion and contraction of the aluminum containment. These joints surround the containment cell in broken bands and allow the aluminum to expand and contract freely in all directions while still securing the containment cell within the vessel’s structure.

25 To Vessel

Structural Steel Hardened Steel

Explosion Welded Transition

Air Gap

Aluminum Floating Joint

To Containment

Figure 13--Floating Joint Utilizing Explosion Welded Transition Piece (Cross Sectional View)

Steel Support (To Vessel Structure)

Floating Joint

Aluminum/Steel Transition

Aluminum Support (To Containment)

Figure 14--Floating Joint (Side View)

26 The joints are constructed of three different materials—the aluminum portion which is welded to the containment cell, the structural steel member welded to the vessel’s hull and the hardened steel floating joint. The lower half of the floating joint (closest to the containment) is bonded to the aluminum via the explosion welded transition piece. The hardened steel portion of the joint is designed to allow movement within the joint without being subject to excessive wear, galling, etc. The surfaces of the joint are highly polished to have a low coefficient of friction between the surfaces and machined to be press-fit together at ambient conditions (20°C ) . Together, the foam insulation and floating joint system provide a better support structure for the containment than the Gaz Transport & Technigaz membrane designs. There is no requirement for plywood structures to assist in the support of the containment.

The foam and floating joint combination provide a lasting, reliable,

maintenance free support structure. The plywood support structures found in the GTT No. 96 and Mark III membrane containment systems, on the other hand, may fall victim to problems caused by the effects of moisture and time. With the dead space surrounding the outer shell of the containment, a safety zone is built into the design in the event of a collision or grounding. The floating joint support system adds an additional element of safety. As seen in Figs. 12&13, there are several stress concentrators present in the floating joint design. Also, the floating joints are not welded at 90 degree angles to the containment. These two characteristics are safety features of the FCS design. In normal conditions, the foam insulation and individual floating joints act as a common system supporting the containment and its contents. In

27 the event of a collision, if the vessel’s support structure is damaged to the point of endangering the containment, these supports are designed to fail at the welded seams rather than puncture the containment cell.

Further Calculations

Please note that all calculations are rough estimates only.

The alloys,

insulating materials, etc. were chosen for the purpose of analyzing this conceptual design. More accurate analyses must be made for the specific materials chosen.

Before it is filled, the containment is at an ambient temperature of 20°C . As the cell fills with cryogenic LNG the temperature of the containment and the immediate surrounding area is reduced to a temperature of approximately − 160°C . Using the dimensions of the containment cell from the previous example, it can be determined approximately how much of a decrease in volume will be experienced over the temperature differential. Using the equation for the change in volume of a geometric solid, and the containment dimensions from the previous example, the change in volume of the ellipsoid containment is as follows:

∆Vellipsoid = Vo ⋅ α ⋅ ∆T

Where, Vo ≈ 66,500m 3

α alu min um ( linear ) = 23 × 10 −6 m m °C ∆T = 20 − (− 160 ) = 180°C ∴ ∆V = (77,600 )(23 × 10 −6 )(180 ) ≅ 320m 3

28

This equates to approximately 3.3 centimeters of joint travel required in the floating joints around the surface of the ellipsoidal containment cell to compensate for its contraction over the expected temperature range. This amount will vary, however, as the contraction around the cell will not be equal in all directions.

Specifically, the

contraction will have the greatest effect at the top of the cell with minimal effect at the bottom. The materials in the joints themselves will contract as well. Assuming a void space thickness of 2 meters, the actual floating joint will be approximately 1 meter from the containment.

The explosion welded transition piece in turn will be located 50

centimeters from the containment, half way in between the aluminum containment and the floating joint. The thermal conductivity of aluminum is very high (approximately 225W

m °C

) and, since the length of the aluminum portion of the support is relatively

short, the entire segment will be assumed to experience the same temperature differential as the containment. The approximate contraction of the aluminum portion of the support is therefore, ∆L = Lo ⋅ α ⋅ ∆T = (0.5)(23 × 10 −6 )(180) = 0.002m

Likewise, the hardened steel portion of the support, with a coefficient of thermal expansion of approximately 12 × 10 −6 m

m °C

, will be assumed to experience the same

temperature differential as the aluminum portion due to its own thermal conductivity

29 (approximately 35W

m °C

), the thermal conductivity of the aluminum and its short

distance from the containment. Thus, ∆L = Lo ⋅ α ⋅ ∆T = (0.5)(12 × 10 −6 )(180) = 0.001m

Finally, the 1 meter section of hardened and carbon structural steel that makes up the stationary section of the floating joint and the structural member of the support system will have a change in length that can be assumed to be negligible in comparison to the first two sections analyzed. The total joint movement is thus,

0.033 + 0.002 + 0.001 = 0.036m

Surrounding the floating joints and the containment cell is the foam insulation. Again, this insulation is compressed prior to installation to compensate for the increasing volume seen in the void space as the containment contracts. The material chosen in for this analysis is Pittsburgh Corning FOAMGLAS® insulation. This insulation provides a thermal conductivity of only 0.10 W

m °C

permeability and having a density of 120 kg

, while providing extremely low water

m3

. It also provides a compressive force of

620kPa . FOAMGLAS® is comparable to the best characteristic of the foam and perlite

insulations used in the GTT No. 96 and Mark III membrane systems (Pittsburgh Corning).

Insulation Material  Thermal Conductivity W  m ° C  

Perlite 0.14

Foam FOAMGLAS® 0.10

0.10

30 Permeability (% absorption by volume) 2-90 Density  kg 3  80-208  m  Compressive Force (kPa ) 620

0.7

0.2

32

120

310

620

Perhaps the most important characteristic is the compressive strength of the FOAMGLAS® insulation. In the FCS containment design, the insulation serves an integral role in the support of the containment.

In fact, much like the membrane

containment designs, the insulation is the main source of support for this design. For the containment cell of the dimensions described above, the surface area is approximately 10,300m 2 , the lower half of which ( 5,150m 2 ) is supported by the foam. The foam provides a total vertical supporting force exceeding 300,000 metric tons. Considering that a full containment of LNG, at a mass of about 420 kg

m3

, weighs approximately

30,000 tons, the FOAMGLAS® will provide excellent support for the containment. Perhaps the only drawback to this choice of insulation would be the weight. If the outside shell of the FCS were 2 meters wider all the way around the containment, the total volume of space filled by the insulation would be approximately 18,500m 3 .

[

]

∴V = 4 3 ⋅ π ⋅ (42 − 2 ) ⋅ (21 − 2 ) ⋅ (18 − 2 ) − (66,500 + 275) ≈ 18,500m 3

This amount of FOAMGLAS® insulation would add about 2,200 metric tons to each individual containment system.

However, if necessary, the top portion of the

containment, which does not require the compressive force of the lower half, could be insulated using a much lighter foam (polystyrene, for example) at the cost of a small additional loss of product due to evaporation.

31

 2200  18,500   2  +  2 (32) ≅ 1400 metric tons

The foam and floating joint system together also provide added protection in the event of a collision or grounding. The floating joints have additional room for movement due to the contraction of the containment. In addition to its insulating qualities, the density of the foam provides a cushioning effect that acts against the forces experienced during a collision or grounding and serves as an additional buoyant layer should the hull be compromised. The layer of FOAMGLAS® is also another barrier of defense against puncture of the containment cell. The supports themselves are installed at angles greater than 90 degrees to the perpendicular to the containment. Particularly in the areas of the containment where there would be the greatest likelihood of damage due to collision or grounding. The weld seams provide stress concentrators in addition to those outlined in the previous description of the floating joints. Additionally, the welds are weaker than the rest of the containment cell. The floating joint support structure is purposely designed to break at the weld seems and/or other stress concentrators rather than puncture the containment cell. A number of design considerations and construction materials may be incorporated in the piping systems required with the Floating Containment System. Piping may be made of an aluminum alloy similar to the containment cell material. An explosion welded transition joint can again be used to bond this pipe to a steel or stainless steel pipe as it penetrates the outer shell of the containment. Another design may utilize an expanding/contracting bellows of suitable material attaching the containment to the

32 outer shell via a standpipe. Accommodations must also be made for the recovery of gas vapors for possible use in the vessel’s propulsion engines.

Limitations of Proposed Design

While the FCS design promises advantages over previous design, it has limitations. For example, while an LNG vessel utilizing the FCS design has greater capacity than a comparable Kvaerner-Moss vessel, a vessel utilizing either of the GTT membrane designs still has greater capacity and thus, greater short term economic benefits. The FCS design is a new, unproven concept. No design like it has been tested so its economic and operational feasibility cannot be proven at this time. Likewise, the FCS floating joint system is a new application of explosion welded transition joints that has not been used before. Although LNG containment systems are not new designs, details surrounding their construction are considered proprietary. The Kvaerner-Moss spherical containment and the GTT membrane containment systems are exclusive to the respective companies that designed them. Their construction, particularly the explosion welded transition joint of the Kvaerner-Moss design is not completely known and the details are kept fairly guarded. Likewise, the process of explosion welding is by no means a new technology. It has been in existence since soldiers first noticed the copper jackets from bullets bonding to the steel armor of tanks and other armored vehicles. However, the process is still

33 untested in many applications. Many traditional fastening methods on the other hand, are tried and true in countless industries and functions. Where explosion welding technology has been tested is in many military applications.

Naval vessels have been constructed with steel hulls and aluminum

interiors that utilize explosion welded joints as transition pieces. For reasons of security, the details of materials used and results of construction testing such as sea trials are not public knowledge and are available to very few. Protected proprietary information and industry secrets make it difficult to determine whether or not a conceptual design is feasible. As construction details and results of real world industrial and military applications become a matter of public knowledge, designs can be determined to be practical or not earlier on in the design process. Until then designs can be conceived and tested only as far as the technologies and designs are known to architects, engineers, grad students, etc. All of the calculations made throughout the design process have been rough estimates based on certain assumptions. Not all of the assumptions made may be the correct. Values were used for coefficients of thermal expansion and conductivity based on general alloy versions of the materials used. They do not take into account certain factors. For some materials, the coefficient of thermal expansion changes with the temperature. As a result, further research may show that alternate materials would fare better in this application. Additionally, the details and characteristics of materials such as insulation were based off of specifications from manufacturer’s data sheets and given on their websites. These numbers are more than likely ideal values and may not prove true in application.

34 Conclusions

The proposed design attempts to use the best aspects of already proven LNG vessel containment designs.

Taking a characteristic from the Kvaerner-Moss

containment design, aluminum is the material of choice in the Floating Containment System. It is the lightest material of those analyzed allowing for a lighter and therefore more economical vessel design. Aluminum is also the material with the best cryogenic properties in comparison to the GTT membrane materials stainless steel and Invar®. The ellipsoidal containment cell design allows more product to be carried on a vessel of similar size to one with the Kvaerner-Moss design. The compressed foam insulation in the support system mimic the support structure of the membrane designs by holding the containment cell securely, while still protecting it from damage by avoiding a direct transition joint, unlike the explosion welded transition joint found in the spherical containment design. The floating joint structure adds an element of support not seen in the membrane design that increases the strength and safety of the containment system. While aspects of the Floating Containment System may prove to be impractical or not economically feasible, it is a new approach that attempts to increase the efficiency of LNG containment and transport. However, there are many more aspects of the design and the materials used that must be analyzed. As demand for cleaner, cheaper fuel sources grows the demand on the LNG market will grow. The Floating Containment System may prove to be a legitimate direction for the future of LNG vessel design and construction.

35 Works Cited

2006 International Fire Code. “Cryogenic Fluids”. Chapter 32, pp.279-284. 2006 Aker Kvaerner. History. http://www.akerkvaerner.com/Internet/AboutUs/AkerKvaernerGroup/History/History .htm Curt, Bob. “Marine Transportation of LNG”. QatargasII presentation, Intertanko Conference. March 29, 2004 DMC Clad Metal Groupe SNPE. “Explosion Clad Plate Manufacturing”. URL http://www.dynamicmaterials.com/data/brochures/EXWprocess2.pdf Incropera, Frank P.; DeWitt, David P. Fundamentals of Heat and Mass Transfer, 5th Edition, Wiley. ISBN 0-471-38650-2. August 9 2001 Mannan, Sam. Editor. Lees' Loss Prevention in the Process Industries (3rd Edition). Chapter 22, pp.1-78. Elsevier Publications. 2005 Merrem & laPorte. “Triclad®: Welding Aluminum to Steel”. URL http://www.triclad.com/pictures/triclad.pdf Murr, Lawrence E. Shock Waves for Industrial Applications. Chapter 5, pp. 170-215. William Andrew Publishing/Noyes. 1988. URL http://www.knovel.com/knovel2/Toc.jsp?BookID=775&VerticalID=0 Pittsburgh Corning. “About FOAMGLAS® Insulation”. URL http://www.foamglasinsulation.com/mechanicalspecs.asp Shrouf, Robert. Safe Handling of Cryogenic Fluids. GN470100, Issue B. October 31, 2006. URL http://www-irn.sandia.gov/corpdata/esh-manuals/gn470100/g100.htm

36 Young, George A. and Banker, John G. “Explosion Welded, Bi-Metallic Solutions to Dissimilar Metal Joining”.

Society of Naval Architects and Marine Engineers.

Proceedings of the 13th Offshore Symposium. February 24, 2004

37 Works Consulted

2006 International Fire Code. “Cryogenic Fluids”. Chapter 32, pp.279-284. 2006 Aker Kvaerner. History. http://www.akerkvaerner.com/Internet/AboutUs/AkerKvaernerGroup/History/History .htm American Bureau of Shipping. Guide for Building and Classing Membrane Tank Hull Vessels: Hull Structural Design and Analysis Based on the ABS Safehull Approach. October 2002 Avallone, Eugene A. and Baumeister, Theodore III. Editors. Mark’s Standard Handbook for Mechanical Engineers, Tenth Edition. McGraw-Hill Companies, Inc. 1999 Castaneda, Christopher. “Manufactured and Natural Gas Industry”. EH.Net Encyclopedia, edited by Robert Whaples. September 4, 2001. URL http://eh.net/encyclopedia/article/castaneda.gas.industry.us Curt, Bob. “Marine Transportation of LNG”. QatargasII presentation, Intertanko Conference. March 29, 2004 High Energy Metals, Inc. Explosion Bonding Engineering and Design Basics. March 8, 2000 DMC Clad Metal Groupe SNPE. “Explosion Clad Plate Manufacturing”. URL http://www.dynamicmaterials.com/data/brochures/EXWprocess2.pdf Incropera, Frank P.; DeWitt, David P. Fundamentals of Heat and Mass Transfer, 5th Edition, Wiley. ISBN 0-471-38650-2. August 9 2001 Mannan, Sam. Editor. Lees' Loss Prevention in the Process Industries (3rd Edition). Chapter 22, pp.1-78. Elsevier Publications. 2005

38 Merrem & laPorte. “Triclad®: Welding Aluminum to Steel”. URL http://www.triclad.com/pictures/triclad.pdf Murr, Lawrence E. Shock Waves for Industrial Applications. Chapter 5, pp. 170-215. William Andrew Publishing/Noyes. 1988. URL http://www.knovel.com/knovel2/Toc.jsp?BookID=775&VerticalID=0 Pittsburgh Corning. “About FOAMGLAS® Insulation”. URL http://www.foamglasinsulation.com/mechanicalspecs.asp Robin Materials Inc. “Invar®”. URL http://www.rmat.com/invar.html Shrouf, Robert. Safe Handling of Cryogenic Fluids. GN470100, Issue B. October 31, 2006. URL http://www-irn.sandia.gov/corpdata/esh-manuals/gn470100/g100.htm Triplex GmbH. Basic Information: Triplex Plastic Panels: Extremely Rigid and Lightweight. 2006. URL http://www.triplex-kunststoffplatten.de/english/basicinformation/plastic-panels.htm UK P&I Club. “The Carriage of Liquefied Gases”. Issue #8. February 2005. URL http://www.ukpandi.com/ukpandi/infopool.nsf/HTML/LPCtC8 Young, George A. and Banker, John G. “Explosion Welded, Bi-Metallic Solutions to Dissimilar Metal Joining”. Society of Naval Architects and Marine Engineers. Proceedings of the 13th Offshore Symposium. February 24, 2004