Multilayer Pressure vessels By: P. V.

Thomas
Multilayered pressure vessels are those having the shell and heads made up of two or more separate layers (Fig -1). The concept of multilayer design was originated in second half of 19th century for artillery, and later for riveted type pressure vessels. The methodology for multilayer Pressure vessel fabrication by welding was developed by A. O. Smith Corporation, USA, way back in the 1930s. Main Components of Multilayer pressure vessels. Cylindrical portion of Multilayer vessels consist of an Inner liner made of corrosion resistant material, over which shell Layers formed from plates or sheets are added. Dummy Layer will be provided between inner shell and Shell layers, to function as secondary containment, to protect the Carbon Steel Shell Layers. Leak detection system (described later) will help in early detection of any leak. Outer protective layer is provided over the Shell layers. Heads of the vessel will be constructed from single wall forgings with an inner lining, or, as a layered construction which is comparatively difficult to construct, and is not warranted. Dummy layer and leak detection system should be provided for the Heads also. Attractive Features of layered construction compared to Single wall construction. Major attraction is the possibility to achieve uniform shell properties and stress distribution which is difficult to achieve for thick single wall. Thin plates exhibit superior mechanical properties (strength as well as notch toughness) compared to thick plates. When thin plates are wrapped gap-less, the layered section forms a virtually solid wall with uniform properties. In addition, the compression stress induced by the tight wrapping is distributed throughout the inner two-thirds of shell courses. The tensile stress resulting from operating pressure is superimposed over this pre-stress to provide an almost uniform tensile stress over the entire thickness. Materials of different properties can be used for the layers, if required, to enhance the strength characteristics. Multiwall vessels can be easily fabricated to any desired thickness and diameter, whereas, Single wall construction is limited by the available plate thickness and fabrication facilities. By suitably selecting the materials and thickness of layers, PWHT can be avoided. Therefore, the inner liner can be made of Titanium, 25-22-2 or any other similar material, without fear of any degradation due to PWHT. Whereas, PWHT is mandatory for thick walled construction, which can degrade the inner lining.

Lighter weight is achieved for layered shell sections by using high tensile thin plate materials like SA 724 Gr. B. It may thus result in saving of overall costs . The rupture mode of layered construction is only by tearing without fragmentation; the propagation of any brittle fracture/crack will stop at layer interface. Single wall construction has the risk of brittle fracture and hence may fragment to small pieces during a rupture which can fly off in all directions. Repair of leak in single thick walled vessels will be a difficult task because any corrosion cavities formed in the wall has to be repaired only by weld build-up, which shall be time consuming and needs PWHT. This will affect the metallurgy and corrosion resistance of inner liner. Cavities formed in Multiwall vessels can be filled up by installing thin plate segments ( as described later in this paper), without any PWHT. Limitations of Multilayer construction (1) : In the Multilayered vessels, high local discontinuity stresses can occur at nozzles, supports, etc. This is due to reduced rigidity of a collection of thin plates, compared to that of solid wall. This problem can be resolved by putting additional layers at these locations or by employing solid wall construction in these local areas. For instance, Flared insert type thick walled nozzles can be used and the layers can be welded to the solid body (1). Also, Solid ring can be used for connecting layered shell to layered heads and supports. Or, use thickest plate permissible without the need for PWHT. Rapid internal cooling or heating must be avoided to prevent the inner layer from getting over-stressed which will result in weld cracks or buckling. Liner stress analysis and buckling analysis is to be carried out to establish the optimum heating and cooling rates. Vacuum rating of multiwall vessels shall be less compared to thick walled construction, because, the vacuum is to be withstood by the inner liner only. In order to alleviate this problem, the inner liner shell shall be designed for expected vacuum condition. Note that the above problems, except the first one, exist in thick wall construction also, if there is corrosion resistant inner lining made of austenitic materials. MULTILAYER CONSTRUCTION METHODS. Generally four methods are accepted for manufacturing Layered shells: a. Helical coil layer method, by winding of thin plate coils in the form of a helix. b. Spirally wrapped, using thicker plates joined by welding and wrapped as a coil. c. Shrink fitted concentric multi-wall method;

d. Concentric wrapping method, explained below (adopted for Safcos Urea Reactor). Refer ASME Sect. VIII Div. 1, Fig. ULW-2.1 and Sect. VIII Div. 2 Fig. AG-140.1 for illustrations. M/s Kobelco, Japan, who has supplied more than 1000 layered pressure vessels around the world, have reportedly made a comparative study of the above methodologies(3). Their conclusion is that, of the four methods indicated above, the concentric wrapping method is the best for achieving a close fit and completely uniform shell properties. In the Concentric wrapping method, first the inner liner is fabricated to correct dimensions. At this stage, the quality of inner liner long seam weld is confirmed by NDT Methods. Then the pre-rolled dummy layer is tightly wrapped on it. The long seam of the Dummy wrap will be butt-welded, without fusing to the inner liner. Subsequent layers are wrapped over it, one over the other, in a similar way. Attachment-3 illustrates the steps in fabrication of a typical Layered Column. DESIGN OF MULTILAYER PRESSURE VESSELS: The design of layered pressure vessels shall conform to the design requirements given in ASME Sect VIII, Div. 1, UG-16 through UG-46. (4) General design considerations. The corrosion resistant inner liner (which is considered as a sacrificial layer), Dummy layer and outer Shell are not considered as part of the required minimum shell thickness. When attachments (Nozzles, supports, etc.) are welded to vessel shell, provisions shall be made to transfer the load to all shell layers. If not, only the thickness of layer to which the attachment is welded shall be considered in calculating the stress near the attachment. Torispherical layered heads are not permitted for layered vessels. Usually, the Heads shall be CS, hemispherical. Dummy lining of Heads can be weld overlay or cladding with same material as that of inner layer. The inner liner shall be designed for vacuum condition. Strength calculations for all parts shall be checked for Testing and as-erected conditions also. Thickness Design of inner Liner. Specific Code requirements are not available for liner thickness calculation. Major factor for selecting the thickness of liner is the acceptable corrosion rate for the selected material under the operating conditions. Second factor is the possibility of liner buckling due to differential thermal expansion of the Liner compared to shell layers. While calculating the differential expansion, it should be realized that the Temperature seen by liner will be higher than the shell layers due to its proximity to the operating fluids. This temperature difference has to be calculated and accounted for, by carrying out a

detailed buckling analysis. Third factor is the maximum vacuum condition that the inner layer will have to withstand. The fabrication sequence and Fabrication stresses induced by the wrapping process of shell layers over the Liner also are to be considered. Buckling Analysis for inner liner: Buckling analysis is done to study the strength of internal lining to withstand differential thermal expansion. It is assumed in the analysis that the lining deformation due to buckling shall be only inwards since any movement outward is prevented by the shell. Finite element analysis ( FEA) programs like ABAQUS are used for Buckling analysis. Internal pressure, temperature difference between lining and shell, and residual compressive stress due to the Fabrication process are the loading conditions to be assumed in the Analysis. In the FEA model, a geometrical non-linear collapse of the lining is considered. The shell and lining are to be modeled by 3-dimensional, 8-node solid element. Both ends of the analytical model are assumed as fixed because the lining and shell are welded at these locations. Gap is considered between lining and shell whereas no gap is considered between shell layers; the shell is modeled as a solid wall. In the analysis, radial inward displacement of the Liner is plotted against Differential temperature, and buckling is considered occurring when the rate of increase of radial displacement become larger. ie., buckling is assumed to occur at the temperature T where slope-change starts for the Graph. Ref. Fig-2. If residual compressive stress between shell and ling due to fabrication process is not considered in the buckling analysis, it shall be accounted for in the final results. Actual compressive stress is to be evaluated by help of Strain gauges and the Stress is to be converted to equivalent temperature using the relation : ts= / (E*) where : = Comp. Stress due to Fabrication Process. E = Youngs Modulus. = Coeff . of Thermal Expansion. ts = Equivalent temp. due to fabrication stress. Ta = Allowable Differential Temperature. T = Temp. where slope-change starts in the graph ( The units used are to be compatible. ) The allowable max. temperature differential between shell and lining without buckling: Ta = T- ts Design of Shell Layers and other pressure / non-pressure parts

Design calculations for Shell layers, pressure parts like nozzles etc., are carried out using formulas and rules in applicable Design codes (ASME Sect. VIII Div 1 or equivalent). The design basis shall consist of Operating conditions (internal pressure, external pressure, temperature, etc.,) and other loading conditions (support stresses, wind load, self weight, etc.). It shall be assumed that the total stress induced on shell or head shall be shared by each layer used for its construction, in proportion to the thickness of plates used for the layer. The total minimum thickness calculated for the layered pressure parts shall be divided by the thickness of layer plates to be used, in order to arrive at the number of layers. Selection of layer plate material and thickness shall be based on availability and ease of fabrication. If plates with higher allowable stress is used, the wall thickness and weight of the vessel can be reduced. Minimum thickness of any layer shall not be less than 18 in. (3 mm), and, the maximum thickness should be less than 19 mm, to avoid PWHT. ASME Sect. II, Roarks Book for Stress and Strain as well as AISC can be used for Strength calculation of internal parts ( Tarys, Tray supports, etc.) . In addition to these standards, JPI standard (JPI-7R-35) for design of skirts for vertical vessels can also be used in the design. Leak detection system Leak detection holes with piping are provided for early detection of leaks from Liners. Leak detection tubes ( Ref Fig.3c), usually nominal size, 2 each at 180 degrees apart are provided above and below each circumferential shell weld as shown in Fig 3a or 3b. That means , for every shell course there shall be minimum four leak detection points. All the leak detection tubes in a shell course are to be connected to a header by means of stainless tubes. Separate header is to be provided for each shell course. This will help to identify which shell course is leaking. A carrier gas, usually dry air at about 1 psig pressure, shall be connection to one end of the header and the other end shall be connected to a bottle containing leak detection solution. The carrier gas flow is to be maintained all the time and the solution is to be topped up as and when required. In the case of Urea Reactor, the leak detection solution used is the colorless Phenolphthalein solution which becomes pink in case of any Carmbamate leak. Risks in leak detection system. Many failures were reported due to inadequate design and construction of leak detection system.

Chocking or any leak in the leak detection tube system will spoil its purpose and hence should be periodically checked for and rectified.

The practice of using Steam as the carrier gas shall be dangerous. Steam will get condensed between the layers and cause corrosion. A failure due to this reason is reported in China (2). Presence of Moisture in the Air used as carrier gas can also lead to similar failure. Any crack or looseness at joint between leak detection tube and liner can cause corrosion for the Carbon Steel Shell Layers. Instead of welding of leak detection tube to dummy plate, a Bush can be welded to the dummy and the Tube screwed to the bush after shell fabrication. Threads of bush and tube shall have close fit and Teflon tape shall be used to ensure leak tight attachment. Care shall be taken to avoid chocking of the leak path by the tape. Since the layers are in close contact, the leaking fluid may take long time to travel to the leak detection points. This problem can be minimized if grooves are provided at the inner face of the Dummy layer. All groves shall lead to the leak detection points. It is not required to force the carrier gas to enter between the layers through one set of leak detection holes and collect them at the other end. This can cause corrosion initiated by any contaminants in the carrier gas. The Carrier gas supply need be only as shown in Fig. 3b above. Using an Ejector as shown in Fig. 3a may help in quicker detection of leak.

Users Design Specification ( UDS) for ordering the new Reactor. It is responsibility of the user or his agent to provide the UDS giving the intended operating conditions in detail so that the supplier / manufacturer can select the material of construction, and carry out the design / fabrication / Inspection in compliance with applicable standards. Essential contents of UDS are given in Attachment-2. The UDS is to be certified for its compliance with code requirements as per ASME Sect VII Div 2, by a Professional Engineer, registered in one or more of the United States of America or the provinces of Canada and experienced in pressure vessel design. The need for Fatigue analysis is evaluated as per ASME Sect. VII Div. 2, Ad-160.2. Typical analysis for the need of detailed fatigue analysis is shown in Attachment-1. Design inputs required from vendor Obtain Mechanical design and strength calculations for each pressure retaining part (shell layers, top and bottom forgings, nozzles, flanges etc.), nozzle reinforcement, support skirt and lifting lugs from fabrication contractor. Strength of all parts should be checked for Design, Testing and as Erected conditions. Strength calculations for Trays and its support system, Design calculations for main Bolting for service condition, Testing conditions as well as for bolting up condition, Wind load and Earthquake load calculations for vessel and skirt support / anchor bolts system, Buckling analysis, Thickness of the inner liner to withstand full vacuum, Stress evaluation (Primary, Secondary and peak stress) in accordance with ASME Section VIII Div. 2, Part AD for magnitude and location of stress intensities, and , Buckling analysis of the Stainless steel

lining (based on elastic material property model, using finite element method) also should be verified. Conclusions. Multi wall construction is best suited for Pressure vessels with extra heavy wall thickness required in Chemical process applications. Multilayer construction has definite advantages compared to single wall construction. Their in-service maintenance and repair are easier, if handled with care and skill. The disadvantages attributed to Multiwall construction can be easily alleviated by simple techniques described in this paper. References: 1. High Pressure Vesssels , by : Donald M. Fryer and John F. Harvey. 2. The explosion reason analysis of urea reactor of Pingyin, by: Wang Weiqiang, Li Aiju, Zhu Yanyong, Yao Xiaojing, Liu Yan Chen Zhonghe. 3. Technical Presentation of Kobelco Layered Vessels, by: M. Yasukawa, Kobe Steel. 4. ASME Sect. VIII Div. 1; ASME Sect. VIII Div. 2