Pipe joining techniques for Phase 2 upgrades highpressure

Pipe joining techniques for Phase 2 upgrades high-pressure systems Summary report from CERN’s 18 th of May 2018 workshop JEROME DAGUIN – PAOLO PETAGNA THANKS TO: GEORG V. , FRITZ M. , ANTTI O. , KAROL R. , BART V. , PAUL K. R. , PIERRE D. , ALLEN Z. , PETER R. , HANS P. , CHI-MENG L. 25. 06. 18 JÉRÔME DAGUIN - CERN 1

Outline CERN’s 1 day workshop Context of Ph 2 upgrades Leak tightness and reliability The joining techniques ◦ Brazing techniques ◦ Welding techniques ◦ Fittings Conclusion 25. 06. 18 JÉRÔME DAGUIN - CERN 2

CERN’s 1 day workshop PIPE JOINING TECHNIQUES FOR PHASE 2 UPGRADES HIGHPRESSURE SYSTEMS 25. 06. 18 JÉRÔME DAGUIN - CERN 3

CERN Workshop – 18 th of May 2018 “The workshop is intended as an informal and open discussion on all the aspects of joining techniques for CO 2 cooling system piping, including fittings, welded joints, brazed joints and special connections involving electrical breaking. ” 1 full day 8 talks ≈ 60 participants Mailing list created (contact me if you want to be added) https: //indico. cern. ch/event/721360/overview 25. 06. 18 JÉRÔME DAGUIN - CERN 4

Context of Ph 2 upgrades 25. 06. 18 JÉRÔME DAGUIN - CERN 5

The ATLAS i. TK Ph 2 upgrade Strip endcap Strip barrel 384 CLs 392 CLs Pixel outer endcap Pixel inner system Pixel outer barrel >1000 cooling loops >10000 pipe connections of all types 108 CLs ? CLs 158 CLs 25. 06. 18 JÉRÔME DAGUIN - CERN 6

The CMS TK Ph 2 upgrade TB 2 S TEDD 124 CLs 280 CLs TBPS Flat/Tilted TBPX/TFPX TEPX 109 CLs >600 cooling loops >10000 pipe connections of all types 104 CLs 32 CLs 25. 06. 18 JÉRÔME DAGUIN - CERN 7

Ph 2 upgrades numbers CMS TK + EC is around 500 k. W ATLAS i. Tk is around 250 k. W CO 2 at -35°C or below PS= 130 bar Ptest = 186 bar 25. 06. 18 JÉRÔME DAGUIN - CERN 8

Leak tightness and reliability 25. 06. 18 JÉRÔME DAGUIN - CERN 9

Leak tightness and reliability No gas vessel (pipe, etc. ) is absolutely leak tight ◦ And actually it does not need to be The leak rate must be low enough that the required operating pressure is maintained and that the lost fluid is not doing any damage to the environment or incurs a large financial loss All estimates must be seen in the context of the system size (number of joints) For what we need r<10 -4 mbar. l. s-1 (He) is entirely adequate ◦ Which is (relatively) easily achievable… 25. 06. 18 JÉRÔME DAGUIN - CERN 10

Leak tightness and reliability Reliability is the real challenging requirement we need to achieve in our systems! ◦ This is what: ◦ During installation costs money, time and nerves ◦ During operation costs acceptance How do we quantify reliability ? ◦ By a failure rate We need to establish ◦ What failure rate do we need ? ◦ How can we establish this performance for a given design/part ? 25. 06. 18 JÉRÔME DAGUIN - CERN 11

Leak tightness and reliability A verification for system larger than 10 -100 fittings is beyond our means… ◦ Use industry standards connection techniques as much as possible ◦ They have much larger use statistics ◦ The problem is that for the tube dimensions we aim for, there is no industry standard – not even for brazing or welding ◦ In the qualification think carefully about the loads for which the joining technique needs to be qualified and perform control tests using these loads ◦ Include tests at increased stress level to find faults without the need of excessive statistics Pass/fail test sample size to demonstrate a failure rate of 1 in 10, 000 Reminder: PS=130 bar Ptest on site = 1. 44 x PS = 186 bar Ptest in lab for qualifications = 2 x PS ? = 260 bar ? 25. 06. 18 JÉRÔME DAGUIN - CERN 12

Brazing techniques 25. 06. 18 JÉRÔME DAGUIN - CERN 13

Soldering/Brazing Joining of two components with a brazing filler material (BFM), whose liquidus temperature is below the melting point/range of any joined component ◦ No melting of the component material Soldering Brazing Tl_FM < 450°C Tl_FM > 450°C Soldering of stainless steel/copper: ◦ Typically used Sn. Ag 3. 5 (ISO 9453 S-Sn 96 Ag 4; Tliq = 221°C), Rm ≈ 25 MPa Brazing at high temperature of stainless steel/copper: ◦ Typically Ag-based filler metals, i. e. AWS BAg-7 (Tliq = 650°C), Rm ≈ 400 MPa 25. 06. 18 JÉRÔME DAGUIN - CERN 14

Brazing techniques Classified by heating technology: 25. 06. 18 JÉRÔME DAGUIN - CERN 15

Brazing techniques – Manual brazing Manual Brazing at Atmosphere • Heat Sources: Flame Torch (Acetylene), Induction, (Plasma, Arc…) • Working Temperature of common filler metals: 600 -800°C • Steel/Stainless Steel, Copper Alloys: I. e. Ag. Cu. Zn. Sn (650°C) • Application of flux necessary to remove surface oxides • Brazing of tube fittings: • Lap joints (5 -10 mm overlap, rule is min. 3 x Wt) • Gap clearance of joint 0. 1 -0. 2 mm on diameter • Manual process, individual qualification of personnel necessary Brazing of CMS Pixel Ph 1 tubes in CERN EN/MME workshop 25. 06. 18 JÉRÔME DAGUIN - CERN 16

Brazing techniques – Manual brazing Manual Brazing at Atmosphere Preparation of components: • Cleaning/Etching (Surface treatment) • Application of flux on brazed surfaces • Assembly and possible inertion for tubes (Ar-flush inside) to avoid oxidation on the inner wall • Brazing material normally applied as rods/wires • Brazing with avoiding overheating (can change viscosity of filler, incrusting of flux) Post treatment: • Cleaning/removal of flux from components. Mechanically and cleaning with detergent/warm water (surface treatment) Ø Flux contains components as KF, Borates etc. -> corrosive! • Visual inspection Precautions: • Ventilation of fumes (flux) which can condense on surrounding surfaces 25. 06. 18 JÉRÔME DAGUIN - CERN 17

Brazing techniques – Vacuum brazing General features of vacuum brazing Assemblies brazed in vacuum chambers (10 -2 mbar… 10 -7 mbar) Parts have to be clean (outgassing, pollution) and principally oxide-free (wetting properties) Heating performed by radiation, induction, (laser, microwave, EB. . ) ◦ Most common technology: vacuum furnaces with resistor heaters Use of vacuum compatible filler-materials (no volatile components at corresponding brazing temperatures) ◦ Most common BFM for vacuum brazing: ◦ Silver-Copper alloys (780 -950°C) ◦ Gold-Copper alloys (950 -1050°C) ◦ Nickel-based alloys (1000 -1200°C) 25. 06. 18 JÉRÔME DAGUIN - CERN 18

Brazing techniques – Vacuum brazing Advantages of vacuum brazing Disadvantages of vacuum brazing ◦ General high costs: ◦ No flux used/necessary No residual fluxing agents have to be removed/cleaned after process, no risk of corrosion induced by remaining flux (mostly acids containing fluorides and/or chlorides) ◦ Brazed parts stay clean and no oxidation occurs during brazing process Besides flux has not to be removed, the surfaces stay clean and metallic (applications for UHV and RFcavities) Specifically for furnace brazing: ◦ Low distortion of assembled pieces due to homogeneously heated parts High precision assemblies maintain their geometry and alignment 25. 06. 18 ◦ Vacuum furnace equipment ◦ Only batch production possible ◦ Preparation of all assembly parts necessary (surface treatment) ◦ Vacuum grade filler materials more expensive ◦ Long brazing cycles (up to few days from cold to cold) Specifically for furnace brazing: ◦ Complete assembly has to be heated Due to high brazing temperatures material properties will be influenced by the heat treatment (annealing, grain growth, diffusion/precipitation) ◦ Complex preparation Fixed placement of filler material, fixed positioning of assembly parts has to be assured JÉRÔME DAGUIN - CERN 19

Brazing techniques – Vacuum brazing Assembly of capillaries for CMS PIXEL Ph 1 upgrade Assembly sequence: Inlets SS-capillaries (Øo 1. 6 mm and Øo 2 mm) 1. Brazing copper sleeves to capillary/tube 2. Brazing of VCR-connector (with nut) or welding fitting Qualification samples for US-inspection 3. Final assembly with dielectrics SS-copper transition for dielectric Return copper pipes (Øo 5 mm) ss-copper transition for subsequent orbital welding 25. 06. 18 SS-SS direct brazing of VCR-connector to capillary Usage of three different BFM with decreasing melting range Brazing to ceramic with Cu-sleeves US-imaging of SS-Cu transition JÉRÔME DAGUIN - CERN 20

Brazing techniques - Soldering Soft soldering was used for the CMS Tracker Outer Barrel piping connections ◦ Single phase C 6 F 14 cooling with max 6 bar ◦ Copper-Nickel pipes 2. 2/2. 0 and 2. 5/2. 3 ◦ Joint sleeve in brass ◦ Soft solder from CERN store: Multicore Sn. Pb. Ag 62/36/2 with core flux (ERSIN 362) Final soldering was done in situ (≈50%) Developments are on-going for soft -soldering of SS pipes Soldering iron with fume extraction tip 25. 06. 18 JÉRÔME DAGUIN - CERN Tip of the soldering iron machined to a concave shape for increased thermal contact area to the pipe and the brass piece 21

Welding techniques 25. 06. 18 JÉRÔME DAGUIN - CERN 22

NASA welding method Selecting the right welding parameters for automatic welding High limit weld Weld samples Low limit weld Nominal limit weld Not fully penetrated Full penetration (all good welds) To much penetration Welding power The Low-Limit / High limit method is a good method to select the nominal welding parameters and assure that possible fluctuations do not bring the series welds out of spec. 25. 06. 18 JÉRÔME DAGUIN - CERN 23

NASA welding method Certification procedure Each welding type is pre-qualified with weld samples ◦ 5 Low limit welds, 5 nominal welds and 5 high limit welds. ◦ From each weld group the following tests are done to qualify the quality ◦ ◦ All visual inspection (in and outside), all should look similar 1 burst test (>4 x design pressure) 1 longitudinal cut and microscopic analyses 3 samples for further research if needed Production of a series of welds ◦ ◦ 3 high, 3 low and 3 nominal pre welds samples. All process welds done at nominal settings 3 high, 3 low and 3 nominal post welds samples. The pre and post weld samples are send to NASA for inspection ◦ 1 burst, 1 microscopic and 1 spare for each type ◦ All process welds are inspected visually and compared with post and pre weld samples. ◦ Dye penetrant tests on process welds Due to the automatic welding and the similarity of all welds, no further complex inspection of the welds is needed 25. 06. 18 JÉRÔME DAGUIN - CERN 24

Welding techniques – Orbital welding on SS or Ti is standard and “easy” down to 1/8” and 200µ/300µm Wt Massively used in all systems in operation and under development Requires careful alignment, good pipe preparation and Ar protection SWAGELOK M 200 Orbital welder setup @ Sheffield 25. 06. 18 LHCb VELO cooling blocks JÉRÔME DAGUIN - CERN CMS PIXEL Ph 1 Supply manifold 25

Welding techniques – Orbital welding below 200µm W t is more tricky… ◦ CMS Ph 2: 2. 5 OD/150µm Wt Ti tubes ◦ ATLAS Ph 2: 2. 5 OD/180µm Wt Ti tubes Tube cutting, preparation and alignment extremely difficult for butt welding of thin wall tubes ◦ Sleeve welding reduce difficulty Butt weld vbc IP 50 Sleeve weld (filet) Machine stability at low current is essential for thin wall tubes welding ◦ Such machines are not available OFF-the-shelves in industry ◦ Sheffield developed their own machine with VBC company Cross section through 2. 5 mm x 0. 178 mm wall tube and sleeve fitting at welded joint area ◦ Can weld down to 125µm Wt Ti tubes ◦ Argonne is proposing to add thickness welding a sleeve together with the tube ◦ Enables better alignment as well ! 25. 06. 18 JÉRÔME DAGUIN - CERN 26

Fittings 25. 06. 18 JÉRÔME DAGUIN - CERN 27

Fittings – VCR types Stainless steel fittings with metal seal ◦ Can be re-used (almost) indefinitely – gasket need to be changed ◦ Extremely reliable ◦ Leak tight Available for pipe sizes of 1/8”, 1/4”, 1/2”, 3/4” and 1” Many suppliers on the market since couple of years ◦ Swagelok, Rotarex, Hamlet, Fitok… and even Hikelok on ALIBABA !!! Female nut silver plated to prevent galling when tightening Don’t exist for tubes smaller than 1/8” ◦ Drilling of blind glands + brazing allows for use with smaller pipes Connection is quite bulky… 25. 06. 18 JÉRÔME DAGUIN - CERN 28

Fittings – Custom design Aluminum gasket being captured inside the male nut ¼” hex male nut M 5. 5 x 0. 9 threads CMS PIXEL Ph 1 FPIX design Fixed male nut induced torque transfer to tube during tightening and self untightening during pressure/temperature cycles in the assembly phase ATLAS IBL LAPP design Identical glands ¼” hex female nut with M 5. 5 x 0. 9 threads Electron Beam welds OD 2. 2 mm, wall 0. 1 mm 304 L ss tubing CMS PIXEL Ph 1 BPIX design 25. 06. 18 Laser-welding between gland tubing Enhanced version for Ph 2 TBPX/TFPX JÉRÔME DAGUIN - CERN 29

Fittings – Custom design Material choice and welding parameters are essential when designing such fittings ◦ Hot cracking issues in the development phase of the FPIX Ph 1 fittings (See Stephanie T. talk in Bonn in 2016) https: //indico. cern. ch/event/469996/contributions/2148019/attachments/1277061/1895241/FPIX_Laser_Welding_Bonn_Forum_2016. pdf Design is not as simple as a scale down of SWAGELOK VCR design… ◦ Small size induce new constraints (tightening torque, torque transfer…) Since our sampling capacity is limited, testing and validation parameters should be much more constraining than operational parameters ◦ PS=130 bar ◦ Ptest on site = 1. 44 x PS = 186 bar ◦ Ptest in lab for qualification = 2 x PS ? = 260 bar ? 25. 06. 18 JÉRÔME DAGUIN - CERN 30

Fittings – Large sizes Phase 2 plants will need up to 2” pipes ◦ Anything above 1” cannot be done with VCR Standard flanges available in DIN, ANSI or B 10 dimensions ◦ Outside diameter of flanges 184 – 216 mm for DN 50 ◦ Generally too big for our installations… SAE flanges used for high pressure hydraulic up to 400 bar (6000 PSI) ◦ O-ring replaced by CF copper gaskets ◦ Testing foreseen at CERN in the coming weeks Compact flange option to be investigated ◦ Described in the NORSOK standard ◦ Used in Oil and Gas industry 25. 06. 18 JÉRÔME DAGUIN - CERN 31

Conclusions We are about to build (very…) large systems with (very…) large number of connections of all types Wide range of joining techniques available ◦ Make the good choice based on the detector constraints ◦ Take it into consideration in the early phase of the design This is a bit more than just plumbing… 25. 06. 18 JÉRÔME DAGUIN - CERN 32