WO2007064211A1

WO2007064211A1 - Double-walled fluid containment system

Info

Publication number: WO2007064211A1
Application number: PCT/NO2005/000445
Authority: WO
Inventors: Pål G. BERGAN; Kåre BAKKEN
Original assignee: Det Norske Veritas As
Priority date: 2005-12-01
Filing date: 2005-12-01
Publication date: 2007-06-07
Also published as: EP1957382A1

Abstract

The present invention regards a tank for storing of fluid, a wall for use in a tank and a method for providing a tank. The wall of the tank comprises a plurality of double barrier wall segments (1) comprising inner and outer plates (2,4), arranged around the circumference of the tank, where neighboring wall segments are joined at a connecting structure (3) at respective ends of the wall segments where the connecting structure (3) being such that, in use, it transfer loads between the inner and outer plates in the respective wall segments. In the wall segments the outer plate (4) is planar whereas the inner plate (2) is curved and the curvature radius of the inner plate (2) of the wall segments, is smaller than the overall radius of the tank inner wall.

Description

Double-walled fluid containment system

The present invention regards a mainly cylindrical tank for storing fluid, which tank is a double barrier tank which can withstand temperature changes, method for producing a tank wall, and a tank wall. Three are know several types of cylindrical shaped tanks, by cylindrical one should in this application understand a form which encloses a void, which has a near circular cross section relative to a longitudinal axis. One example of a cylindrical tank is described in the applicants own patent application No 20023288, where there is described a single barrier tank which may withstand some pressure. In general there is for safety reasons a need for double barrier tanks, which are economically viable, both with regards to material costs and also time consumption when building the tank. There is also a need for tanks for storage of fluid which may easily be monitored with regards to leakage detection and also possible repair of the tank. There is also a need for tanks wherein one may test the tanks with regards to leakage before use. There is also a need for increased wall stability both with regards to strong wind and earthquake loads. There is in also a need for tanks which are relatively easy to produce, which give double barrier safety against any leakage of the stored fluid, can withstand temperature changes and some pressure, which give double barrier safety but do not require much more material consumption than a single barrier tank and at the same time are economically viable.

These needs are fulfilled and achieved with a tank according to the following claims.

According to the invention a mainly cylindrical tank for storing of fluid comprises means for filling and emptying the tank. The tank is formed with walls, floor and roof and is in normal use arranged with a mainly vertical centre axis. The means for filling and emptying the tank and the floor and roof constructions of the tank is not a specific part of the invention and a solution to provide these elements will be within the reach of a skilled person and will therefore not be explained any further in this application. It is preferred that the roof and floor form double barriers joined with a double barrier of the walls, and the floor and roof must withstand static and eventual dynamic forces acting on them.

According to the invention the walls comprises a plurality of double barrier wall segments arranged around the circumference of the tank. By wall segments one should in this application understand units forming a circle sector of the circumferential wall of the tank. These wall segments may be equal, but may also have minor differences between them. The minimum amounts of wall segments are three, but there are preferably more than three segments and there may be up to several hundred wall segments. In one embodiment each wall segment stretches from the bottom to the roof of the tank.

In another embodiment the total height of a cylinder type tank may be divided up in two or more segments in the vertical direction as well; each such vertically aligned segment may be of the currently defined double wall type or it may even include normal one layer part-cylindrical formed elements. There may be specially fitted transition elements between such segments of different type or wall segments and the floor and roof of the tank. This division in wall segments and transition segments may be done to make the production of elements more standardized and transportation of segments from the production facility to the actual site for assembling the tank as easy as possible.

According to the invention each wall segment comprises an inner and an outer plate. Neighboring wall segments around the circumference of the tank are joined at a connecting structure at respective ends of the wall segments, which ends coincide with a plane comprising the vertical centre axis of the tank. It is the inner and outer plates of neighboring wall segments that are connected and in the transition between a wall segment to a neighboring wall segment there is formed a connecting structure. The connecting structure transfer loads between the inner and outer plates of the respective wall segments. In a preferred embodiment the connecting structure comprises at least one plate formed girder with its main plane coinciding with a radial axis of the tank. There may be one main girder joining inner and outer plates of two neighboring wall segments or two mainly parallel plate formed girders. In such an embodiment each of these two girders may be joined with the respective plates of one neighboring wall segment before the neighboring wall segments are joined by joining the two girders forming a connecting structure to form the circumferential wall of the tank. The girders extend in a radial direction between the inner and outer plates of the wall segments. Independent of the solution chosen, with one or two girders formed in the connecting structure, the effect of one or two girder/girders at the joining position of two neighboring wall segments is more or less the same.

These girders may be formed in several ways, one is by forming a fluid tight and load transferring connection between the inner and outer plate, another is connecting the plates in a load transferring manner, but not in a fluid tight manner and thereby joining the partly separated voids between inner and outer plates of the wall segments into a connected void. In one embodiment some of the girders may be fluid tight and some not, forming sections of one or several connected voids between the plates forming the wall, which sections are separated from each other by a fluid tight girder. In this manner one may when have leakage detecting means in or connected with the void between the inner and outer plate, determining which section of the wall is leaking and thereby more easily perform repairs or substitute one or several wall segments and or connecting structures.

When seen in a cross section transverse to the centre axis the outer plate is fully planar and the inner plate is curved, with its midpoint (apex) closest to the outer plate. The inner plate is with this curved outwards in a radial direction from the connection points with a neighboring wall segment, when assembled to form a tank wall. The curvature radius of the inner plates of the wall segments is less than the overall radius Ri of the tank inner walls or with other words than the radius of the cylindrical shape of the total tank. The overall radius of the tank should in this application be understood to be the radius from the centre axis to a connection point between the inner plates of neighboring wall segments. The inner plate of a wall segment has a constant curvature radius, R_s i.e. the curvature radius of the inner plate does not vary along the vertical axis of the inner plate.

A main principle for the current invention is to arrive at a double walled cylindrical- like tank where the material is optimally utilised in a fully stressed design throughout the structure. To illustrate this principle we consider a horizontal cross- section of one sector of angle 2 φ where the overall cylindrical radius is R₁ , the radius of the sector radius of the inner, cylindrical wall segment is R_s , and the corresponding angle of the inner wall segment is 1 φ_s, (see fig. 2). Note, if the entire tank periphery consists of N sectors, the sector angle is

2φ = — (1)

N and the following geometric relation applies for the segment angle

φ_s. = arcsin( — ) (2a)

^Rs or, correspondingly,

K = -^- (2b) sin φ_s where L is half the inner arc length of the segment A - A' .

Figure 2 shows the definition of geometric parameters as well as the forces F₁ and F₀ acting (per unit tank height) on the inner and the outer wall, and the pressure p acting on the inner wall. The radial equilibrium condition is simply

F₁ sinfø, - φ) = F₀ sin φ = -F_g (3 a) where F_g is the corresponding (compression) force in the girder connecting the inner and the outer wall. Note that this equilibrium condition assumes that there are no shear forces in the tank walls; this is the ideal stress condition without bending and shear in the plates as is the target condition. Note also that this force acts only for one sector; there will be an additional, corresponding force from the neighbouring sector. Equation (3 a) may alternatively be written as

F φ_s = ^ + arcsin(— -sin<p) (3b)

F,

The wall forces are a direct result of the pressure p acting on the inner wall

"•'i ⁽⁴⁾

The corresponding stresses in the inner wall of thickness t_t is σ, =£ = -£- (5)

' t, Rj, ^{K }} and, similarly, for the stress for the outer wall of thickness t₀ is

σ_o = %- (6)

Global equilibrium of the double walled tank implies

p(a + R_s ) = σ,t, + σ_ot_o (7) where a is the offset of the segment origin C in relation to the common cylinder origin C. R_s is given above.

The wall strains S₁ and ε₀ are given by the constitutive equation

and ε_o = % (9)

JtL where E is the modulus of elasticity. The above strains are also influenced by vertical stresses in the inner and outer cylinder walls; however, given that the vertical stresses (and strains) are the same for the two walls their influence on the horizontal strains will be exactly the same and do not change the compatibility condition.

In order to achieve a perfect membrane condition in the inner and outer walls we must require that there is no geometric distortion from deformation of the double wall sector element. This is achieved by requiring that the strain of the inner wall is exactly the same as for the outer wall and that the girder connection between the two is relatively stiff. The latter is automatically the case when this connector girder is narrow and thick and S₁ becomes equal to ε₀ because σ_t and σ_u axe both required to be equal to an allowable stress σ_α . The compatibility condition can also easily by validated by considering total double wall equilibrium against internal pressure, which gives the same as equation (3).

All equations for defining a double wall which acts perfectly in pure membrane condition are now defined. Given that the overall inner radius R₁ and the total number of sectors N have been decided, the procedure start with deciding what ratio of the carrying of the overall hoop (ring) force should be done by the inner wall F₁ and what force F₀ should be carried by the outer wall. Deciding this ratio implies that the segment angle φ_s can be found from (3b) and the corresponding segment radius is given by (2b). A given design pressure p and σ, and σ₀ set equal to the allowable stress σ_α implies that the thicknesses t_t and t₀ , and the offset α may be found from equations (4) through (7).

As one example of implementation of the invention let it be required that F₁ should be equal to F₀ , i.e. we want that the inner wall should carry the same amount of overall hoop (ring) force as the outer wall. Using the formulas this implies that φ_^ is twice as large as φ and that the thickness of the inner wall is equal to the outer wall and both are fully stressed. All other quantities are easily obtained from the equations above. It is thus clear that, aside from the extra connecting pieces between the inner and outer walls, the amount of material needed for this double wall tank is the same as for a single wall tank. Thus, we have achieved extra safety associated with leak detection and increased failure strength without using more materials aside for the girders.

A special, limiting case is when we want F₀ to be equal to zero. In this case the solution that comes out of the current approach is a single wall cylinder with t_u equal to zero, as should be expected.

These equations may be somewhat simplified provided that the number of segments is large and, thus, the sector and segment angles are small. In such case the sine trigonometric function may be replaced by the object angle itself. The simple relationships that come out of such an assumption are

and

R. ≡ t, -R (H)

^' t_o +t, ^•

When having a fluid within the tank the fluid will act on the walls of the tank with a decreasing pressure from the bottom to the top of the tank, even with a fully filled tank. To take account for this difference in pressure acting on the walls, it is preferable that the thickness of the outer and or inner plates decrease gradually from the bottom to the top of the tank, but in the preferred embodiment still keeping the above mentioned relationship between the thicknesses of the two plates.

Leak detection means may favorably be arranged associated with the space between the two plates in the walls forming the two barriers of the tank. These leak detection means, may be arranged to detect any leakage and or determine in which wall segments, connecting structure or sections of wall segments the leakage has occurred, or even detect leakage in a specific wall segment. When there is a connecting structure comprising two girders, one will have a configuration that will create a connection void between the two girders. There may also be leakage detection means arranged to detect leakages in these connection voids created in the connection structure. Fluid connection between the different voids of the different wall segments and connecting structures may be formed by one or several apertures or by one or several recesses in the top and or bottom of the girders between the different wall segments.

By having a tank formed by wall segments, one may produce a tank wall by providing wall segments, which has a longitudinal axis, comprising of two plates, an inner and an outer plate, which wall segments are joined with a connecting structure by at least one of two opposite ends of the wall segment, which ends are parallel with the longitudinal axis, where seen in a cross section transverse to the longitudinal axis, the outer plate is planar and the inner plate is curved or with other words having a part cylindrical shape. These wall segments may be transported as single segments or a plurality of segments joined together with connection structures to the assembly site where one may assemble the plurality of wall segments and connection structures with the longitudinal axis in a mainly vertical direction, such that they form a circumferential wall of a tank with the longitudinal axis mainly parallel to the centre axis of the tank. The tank has in an assembled situation a mainly vertical centre axis, whereas the wall segments may be transported to the assembly site with their longitudinal axis in a mainly horizontal orientation and thereafter raised to a mainly vertical orientation parallel to the centre axis of the tank.

The roof and floor of the tank and the transition between the wall and roof/floor may have several configurations as for instance truss work with a double barrier, sandwich structures forming double barrier connected to the double barrier of the tank wall etc. and a skilled person will understand this. For extra safety one may apply extra metal pieces on the top of or inserted into the vertical weld seams of the inner and outer plating the vertical wall, thereby serving as possible crack arrestors for cracks caused byu the ring (hoop) stress. The invention also regards a wall for use in a tank comprising the above explained wall segments and connecting structures:

By such a configuration of the wall of a tank one has provided a tank that largly fulfills the needs or aims mentioned in the introduction.

The invention will now be explained in further detail with illustrative examples of the invention with reference to the accompanied drawings where,

Fig. 1 shows a horizontal cross section of a tank wall according to the invention,

Fig. 2 shows a principle sketch of a wall segment with connecting structures, for use in the tank in fig.1,

Fig. 3a shows a cross section of two joined wall segments with one embodiment of connecting structures,

Fig. 3b shows a cross section of two joined wall segments with a second embodiment of connecting structures, and

Fig. 4a and 4b show different configurations of the joining of two wall segments.

In fig. 1 there is shown a cross section perpendicular to a centre axis, of a schematic sketch of a semi-cylindrical tank wall according to the invention. The tank wall comprises sixteen wall segments 1,1 ', 1 " etc, joined together to form the tank wall. Each segment comprises an inner plate 2,2 ',2" which is curved outwards forming part-cylindrical shaped plates, and a planar outer plate 4,4',4". These plates, 2 and 4 of one wall segment are connected to the plates 2,4 of a neighboring wall segment through a connecting structure in this embodiment as girders 3,3 ',3" which give a connection between the inner and outer plates 2,4 and also defines an end of a curve of an inner plate 2 going over into a curve of a neighboring inner plate 2'. The radial inner point of a girder 3' connects neighboring inner plates 2,2' and another girder 3 " inner plates 2', 2", and the radial outer point of the girder 3' connects the neighboring outer plates 4, 4' and the second girder 3" outer plates 4', 4". As indicated in the figure there may be part of filling and emptying means 7 for the tank arranged in the wall. As best seen in fig. 2 the connections point of the different inner plates 2, and the radial inner points of the connecting structure in the form of girders 3 are also forming an inner radius i?,- of the tank with a centre point at the centre axis C of the tank. The inner plate 2 has a different curvature radius R_s than the inner radius R, of the tank. The outer plate has a plate thickness t₀ and the inner plates a plate thickness t_t. The fluid within the tank act on the inner plate 2 with a pressure P. Each wall segment 1 forms a double angle φ with the walls centre axis C and a double angle φ_s with a centre point C ' in the radius centre for the curvature radius R_s of the inner plate. It is otherwise referred to the general description above for further explanation of this figure.

In fig. 3A there is shown a cross section of two wall segments 1,1 ' comprising of outer plates 4,4' and inner plates 2,2'. These inner and outer plates are joined and in this transition zone between neighboring plates forming a connecting structure between neighboring wall segments 1,1 ', where in this embodiment the connecting structure is formed by a single girder 3, 3', 3". Some of these single girders 3,3' are formed with an aperture 6 forming a connected void of the voids between the plates forming the wall segments, where one for this connected void may have a leakage detection means, indicated by a box marked 5. The leakage detection means may be a pipe inlet/outlet for circulation of inert gas, pipe inlet for creating a pressure in the void, a sensor device other arrangements for detecting possible leakages. As one may see from the figure one of the girders 3" may be formed without an aperture and thereby dividing all the voids between the inner and outer plates into sections of connected voids. By this one may determine in which section of connected voids there is a leakage and perform repairs on that section of wall segments. In fig. 3B there is shown a second embodiment of two wall segments 1,1 ', where in this embodiment the wall segments with the inner and outer plates 2,4, 2' 4', are joined at a second embodiment of a connecting structure formed by a double girder structure 3a,3b, 3'a,3'b. Where these double girder structures are forming connection voids between the two girders 3b,3'a forming a connecting structure for joining two neighboring wall segments 1,1 '. In this embodiment there are arranged leakage detection means 5 in all the different voids formed by inner plates 2,2', outer plates 4,4' and girders 3a, 3b, 3'a, 3' when joining neighboring wall segments 1,1' and connecting structures.

Fig. 4a and 4b show different embodiments of a connecting structure in more detail, in the joining of two neighboring wall segments. In fig. 4a there is shown one embodiment where there in all the different corners between the girder 3 and the outer plates 4,4' and the inner plates 2,2' are arranged a mounting device 8. The mounting device 8 is specifically used during assembling of the different wall segments to keep the inner and outer plates aligned and in correct position to weld the different segments together. One mounting device comprises in this embodiment of two plate parts 8,8' arranged in a corner between the girder 3 and one plate 2,2 ',4,4', one on top of each other, as indicated by the dotted line 8', and where one plate part 8 is attached to the girder 3 and one plate part 8' is attached to the plate 2;2';4;4'. Both plate parts 8,8' comprise an orientation device, for instance an aperture, which indicates when the plates 2,2 ',4, 4' and the girder 3 are in correct orientation relative to each other to be welded together. In the case with an aperture one may insert a rod through the aperture to lock the plates 2,2'4,4' and the girder 3 together before welding. Instead of plate parts 8,8' one may have blocks or other elements with cooperating orientation devices. In fig. 4b there is shown a different embodiment where the connecting structure comprises two girders 3B and 3 A'. These two girders 3B,3A' will be connected to the inner 2;2' and outer plate 4;4' respectively, of the wall segments before neighboring wall segments are joined. These girders 3B,3A' comprise in addition several mounting devices, comprising of two rods 8 attached to one girder 3 A' and two corresponding holes 8' in the other girder 3B. The rods 8 and holes 8' will orientate and keep the elements in correct relation while the wall segments are welded together. There may of course be more rods and holes than two or there may be other configurations for the mounting devices.

These explained embodiments of mounting device are only examples of mounting devices and a skilled person will from the explanation above understand that one may form the mounting device in several different ways. The important issue is to have at least one set of cooperative parts, where one part of the set is fastened to one element and the other part of the set, is fastened to the other element, which elements one need and wants to keep in a position relative to each other while one performs a permanent fixation between the two elements.

The invention has now been explained with different embodiments, a skilled person will understand that there may be performed alterations and modifications to these embodiments that are within the scope of the invention as defined in the following claims.

Claims

1. Tank for storing of fluid, comprising means (7) for filling and emptying the tank, which tank is formed with wall, floor and roof and in use having a mainly vertical centre axis (c), characterised in that the wall comprises a plurality of double barrier wall segments (1) comprising inner and outer plates (2,4), arranged around the circumference of the tank, where neighboring wall segments are joined at a connecting structure (3) at respective ends of the wall segments, which ends coincide with a plane comprising radial axis and the vertical centre axis of the tank, where the connecting structure (3) being such that, in use, it transfer loads between the inner and outer plates in the respective wall segments, where the outer plate (4) is planar whereas the inner plate (2) is curved and the curvature radius of the inner plate (2) of the wall segments, is smaller than the overall radius of the tank inner wall.

2. Tank according to claim 1, characterised in that the radius of curvature of the inner plate (2) and other geometric parameters are chosen in such a way that both the outer and inner plates (4,2) are in a state of pure membrane stress condition and the material is exploited in accordance with the applicable allowable stress condition.

3. Tank according to claim 1, characterised in that the inner and outer plates are made of the same material and that the relationship between the thickness of the outer and inner plates (4,2) and the tanks inner radius and the curvature radius of the inner plate (2) in a cross section transverse of the centre axis of the tank is;

where R_x is the curvature radius of the inner plate, t, is the thickness of the inner plate, t₀ is the thickness of the outer plate and R₁ is the inner radius of the tanks.

4. Tank according to one of the preceding claims, characterised in that the thickness of the outer and or inner plates (4,2) decrease gradually or stepwise from the bottom to the top of the tank in accordance with fluid pressure.

5. Tank according to one of the preceding claims, characterised in that the connecting structure comprises at least one plate girder (3) with its main plane coinciding with a plane comprising radial axis and the vertical centre axis of the tank.

6. Tank according to claim 5, characterised in that the girders (3) form fluid tight contact with the inner and outer plates (2,4) forming closed voids.

7. Tank according to one of the preceding claim, c h a r a c t e r i s e d i n that there in connection with the voids in the tank wall are leak detection means, arranged to detect any leakages and possibly determine in which wall segment the leakage occurs.

8. Method for producing a tank wall, comprising providing wall segments with a longitudinal axis, comprising of two plates, an inner and an outer plate, with possibly a connecting structure by at least one of two opposite ends of the wall segment, which ends are parallel with the longitudinal axis, where seen in a cross section the outer plate is planar and the inner plate has a curvature, transporting separate single or units of plurality of joined wall segments to the assembly site, assembling of the plurality of wall segments or units of wall segments with the longitudinal axis in a mainly vertical direction, such that they form a circumferential wall of a tank.

9. Method according to claim 8, wherein the tank wall is semi cylindrical in an assembled state with a vertical center axis, where the wall segments or units of a plurality of joined wall segments are transported to the assembly site with their longitudinal axis in a mainly horizontal orientation and thereafter raised to a mainly vertical orientation parallel to the centre axis of the tank and connected to form a liquid tight double barrier.

10. Method according to one of the claims 8 or 9, wherein at least two similar wall segments are connected in the vertical direction of the assembled tank at the assembly site.

11. Wall for use in a double barrier tank, c h a r a c t e r i s e d i n that the mainly cylindrical wall comprises a plurality of double barrier wall segments (1) comprising an inner and an outer plate (2,4), arranged around the circumference of the tank, where neighboring wall segments are joined by a connecting structure (3) at respective ends of the wall segments, which ends coincide with a plane comprising radial axis and the vertical centre axis of the tank, where the connecting structure (3) being such that, in use, it transfer loads between the plates in the respective wall segments, where the outer plate (4) is planar whereas the inner plate (2) is curved.

12. Wall according to claim 11, c h a r a c t e r i s e d i n that the radius of curvature of the inner plate (2) and other geometric parameters are chosen in such a way that both the outer and inner plates (4,2) are in a state of pure membrane stress condition and the material is fully exploited in accordance with the applicable allowable stress condition.