8/3/2019 A Guideline for Design, Analysis, And Fabrication

1/3

A GUIDELINE FOR DESIGN, ANALYSIS, AND FABRICATIONOF VACUUM VESSELS FOR CRYOGENIC ACCELERATORS*Robert C. GentzlingerGrumman Corporation, Space Systems DivisionBethpage, NY 11714

Kirk E. ChristensenUniversity of California, Los Alamos National LaboratoryP.O. Box 1663, MS H821, Los Alamos, NM 87545

This paper describes a methodology fordesigning, analyzing, and fabricating vacuum vesselsthat fulfill the unique requirements of cryogenicaccelerators. Included is a design approach that allowsthe vessel to meet the intent of the American Societyof Mechanical Engineers Boiler and Pressure VesselCode, Section VIII, Division 1, both for negativeatmospheric pressure and for positive internalpressures that could result from a cryogenic systemfailure. In addition to ensuring that vessels meet Coderequirements for safety, this approach addressestraditional high-vacuum-technology criteria. Thedesign procedure minimizes analysis time by usingstandard Code guidelines, where applicable, for shellthicknesses, joint details, and penetrations.

I. INTRODUCTIONRecent advances in accelerato r technology haveallowed the design of much smaller accelerators thanthose of the previous generation. One such advance isin the area of cryogenic operation. When anaccelerator operates at cryogenic temperatures, theamount of radio-frequency (rf) power that isdissipated on the surface of the rf cavity is greatlyreduced. In addition, the coefficient of thermalexpansion for metals of interest decreasesdramatically at these temperatures, providing asignificantly more stable structure from the point of

view of thermal distortion. When an accelerator is tobe operated cryogenically, it must be enclosed in avacuum envelope, which provides the environmentnecessary for beam operations. Typically, such vacuumvessels are relatively large (on the order of four timesthe diameter of the accelerator) to provide space forphysical access to the accelerator and for associatedequipment, such as diagnostics and cryogenicmanifolding. Vacuum vessels for state-of-the-artaccelerators usually require numerous penetrations*Work support and funded by the US Department of Defense,Army Strateg ic Defense Command, under the auspices of the US_-

and openings; these are used for assembly, testing,and attaching peripheral equipment. Attempting tomeet this diverse set of requirements can be verychallenging. We have developed an approach thatcan reduce the time required to design, analyze, andfabricate vacuum vessels for cryogenic accelerators.II. STRUCTURAL ANALYSIS

Past experience has shown that in designing avacuum vessel, the analysis required to assure thesafety and proper function of the vessel can be verytime consuming. Numerical methods, such as finiteelement analysis, are usually used because thecomplexity of the design typically precludes theapplication of closed-form analytic approaches. Theseanalyses can be very time consuming because o f thelarge physical size of the vessels and because theirmany smaller features can be accurately simulatedonly be large numerical models. In addition, theapplication of external pressure to the vessel requiresconsideration of the non-linear bucklingphenomenon, which is the predominant kind offailure caused by external pressure.

Our systematic approach is described below. Inaddition to greatly reducing the analysis time neededin the design phase, this approach has also made iteasier to obtain approvals from the LaboratorysPressure Vessel Review Committee for installing andusing our vacuum vessels.

Ill. DESIGN METHODOLOGYThe basic design approach uses the ASME Boilerand Pressure Vessel Code, Section VIII, Division 1(Code) as a guideline. Because vessels that function atatmospheric pressure do not technically fit within thescope of the Code, and because the Laboratory is notrequired by any jurisdictional authority to designaccording to the Code, we attempt to make our designmeet the intent of the Code.

Department ot tnergy.0-7803-0135.8/!91$01.00 @IEEE 2263

1991 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material

for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers

or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

PAC 1991

8/3/2019 A Guideline for Design, Analysis, And Fabrication

2/3

The first step is to choose a nominal wallthickness using the procedure outlined in ParagraphUC-29 of the Code. This is a good starting point,because it can be shown that if the thickness of thevessel wall is made to be 1.5 times the requiredthickness, then all openings in the vessel (that fallwithin the limits of the Code as to size, shape, andspacing) will be integrally reinforced without theaddition o f reinforcement rings.

The next step is to add nozzles to the mainvessel. Their wall thickness can be determined in thesame way as for the vessel walls. As a matter ofpractice, it usually turns out that the thickness of thenozzles will be governed by welding requirementsinstead of by external pressure requirements. This isbecause the nozzles are by definition smaller indiameter than the main vessel and therefore needsignificantly less wall thickness. For this reason, thewall th ickness chosen is usually the one that willprovide the best match for welding the nozzle to thevessel wall.Next, the vessel wall must be checked forpenetration reinforcement according to the areareplacement rule described in Paragraph UG-37. Asmentioned above, most of the penetrations shouldalready be adequately reinforced if the main vesselwall thickness is made to be 1.5 times the Code-required thickness. But those falling outside thebounds of the Code guidelines need to be examined

and have external reinforcement added. For somevessels, such as that used for the Ground TestAccelerator (GTA) Drift-Tube Linear Accelerator (DTL)shown in Figure 1, a very large opening is required forinstallation of the accelerator modules in the vessel.This rectangular opening runs essentially the entirelength of the vessel. This opening is far outside thelimits set by the Code and must be handled differently.In this case, we added removable compression struts(Fig. 2) at the opening, which complete the load patharound the skin of the vessel.

m F~.~~~~a,- - 7 - - -#&

Fig. 1 - GTA DTL Vacuum Vessel

Fig. 2 - Rectangular Opening With RemoveableCompression StrutsFlanges must then be added to the vessel andthe nozzles. The typical flanges used for vacuumvessels are flat-faced, with metal-to-metal contactoutside the bolt circle, (because self-energizingelastomeric O-rings are used for sealing as opposed tothe compression type used for raised-face flanges).The size specifications for these flanges can be basedon Appendix Y of the Code. Appendix Y flanges areusually much thinner than Appendix 2 (raised-face)flanges for the same pressure, because the largemoments caused by raised-face flanges are eliminated.The calculations in Appendix Y can be tedious toperform , but can be simplified by the use of aspreadsheet-type program on a personal computer.Again, it usually turns out that the final thickness willbe driven by manufacturing considerations.

Welds must now be designed for attaching theflanges to the vessel and to the nozzles, and forattaching the nozzles to the vessel wall. The Code(Appendix Y and Paragraph UW-13) provides detailedguidance on weld size, taking into account theinteraction between shells, nozzles, and flanges. Wehave found it much quicker to use these guidelinesthan try to perform a detailed stress analysis o f thesecomplex regions. But some sound engineeringjudgment must also be applied. The Code bases weldsizes on the wall thickness actually being used, whichcan produce welds much larger than typically used forvacuum vessels. We usually base the size instead onthe required wall thickness.

One weld of particular interest is shown inFigure 3. This is the weld used to attach the endflange to the vessel wall. Here we use a fullpenetration weld, welded from both inside and out. Itis often pointed out that this appears to violate thestandard high-vacuum practice of skip-welding on theoutside. We have found though, that our method,which satisfies Code requirements, produces asatisfactory weld, free from virtual leaks. But caremust be taken to assure full penetration at the rootpass of the weld to eliminate any trapped gas pockets.2264

PAC 1991

8/3/2019 A Guideline for Design, Analysis, And Fabrication

3/3

Finally, pressure-relief devices are added toprotect against accidental internal pressurization,which could be caused by the rupture of the internalmanifolding carrying the cryogenic coolant. The setpoint should be set low enough to ensure thatsatisfaction of external pressure requirements stillgoverns the vessel design. Also, there should beredundant relief devices, each with sufficient capacityto vent the largest conceivable coolant leak.

IV. CONCLUSIONThe design approach described in this paper hasrecently been used on three large vacuum vessels. Ithas proven to be quite successful in meeting theobjectives of reducing analysis time, streamlining theapproval process, and producing a design that isFig. 3 -Typical Flange-to-Nozzle Configuration Weld practical to fabricate.

226sPAC 1991