ESS RFQ B POTTIN and RFQ team CEAIRFU
ESS RFQ B. POTTIN and RFQ team CEA-IRFU
RFQ design Strategy Hydraulic diagram 3 RF codes to validate calculations Consideration of machining and assembly possibilities Cooling system
RFQ design process • We have presently done thermo mechanical calculations • To define the number and the position of cooling circuits for the RFQ, tuners, vacuum ports, extremities plates and RF loops (if necessary) • To determine the operation temperature of the RFQ and flow velocity for each circuits • To determine the hydraulic diagram and the cooling system like IPHI and SPIRAL 2 IPHI
The ESS RFQ Global Design • The ESS RFQ: • 5 x Sections for a total length = 4. 58 m • 2 x Vacuum ports per quadrant and per section except on T 3 ( hosting the RF coupler) • 3 x 80 mm-diameter pistons per quadrant and per section acting as tuners • Endcells tuning with 4 x rods • 8 x 10 mm-diameter cooling channels per section (variable length) • RFQ in Cu. C 2 and Flanges in stainless steel T 5 T 1 T 2 T 3 40 mm T 4 Cooling channels Global Schematic view of ESS RFQ Transversal view of the RFQ across the tuners View of the end section of the RFQ with the tuning rod for a quadrant
RF calculations done with HFSS Ø Model realized with CATIA: imported in ANSYS and HFSS softwares (one quarter of RFQ) Mesh in HFSS CATIA model Ø RF simulations in HFSS (calculation of cavity voltage and power density on the RFQ) Voltage (k. V) 130 120 Power density 110 100 90 Theorical voltage law HFSS calculation 80 70 60 0 1 2 3 4 5 Length (m) Power density transferred from HFSS to ANSYS workbench
Cooling Strategy of the RFQ • A peak total RF power of 1 MW can be deposited in the modules in pulse mode (~5% duty cycle) with an increase of the power density along the beam axis (Max power on T 5) • The cooling system designed to remove a peak RF power 250 k. W • Optimization of the cooling channel position, sizes and fluid velocity from 2 D calculation (COMSOL) in order: • to minimize the frequency shift between the RF on/off states (reviewer Jim Stoval’s remark) • Allow simple mechanical fabrication (minimum Cu thickness of 5 mm close to the channels) Þ An average velocity of 3. 5 m/s is allowed in the cooling channels ( empirical estimation of the heat exchange coefficient ~ 13 000 W/m 2/K) Þ Inlet Water temperature per section (for all sections) = 25°C • Vacuum port are brazed on the RFQ sections (all in Cu. C 2): good heat evacuation towards the cooling channels • Tuners are insulated from the RFQ by SS flanges (not a good heat conductor) => cooling added using a brazed coaxial pipe • End plates can be cooled Inlet T 1 Outlet Inlet T 2 Outlet T 3 Inlet T 4 Outlet T 5 Layout of the cooling system for the ESS RFQ View of a quadrant and the cooling channels View of an end plate and its the cooling channel View of a tuner and the coaxial cooling pipe
Numerical Modelling: • Thermal and mechanical calculation are done in 3 D on the ¼ geometry of the RFQ (as in HFSS): • Calculation with ANSYS/HFSS: direct mapping of the RFQ power = 10 475 W with the same geometry • Take into account 3 D effects (vs. 2 D) : contacts, finite length of cooling channels, end cells, grids and tuners • No natural convection is added (with surrounding) • No cooling is implemented in the end plates => Not Necessary (see later) • If no cooling is implemented in tuners: Temperature reaches 260°C locally = > Necessary Zoom on the mesh of T 5: properly map the power density from HFSS to ANSYS thermal and mechanical analysis • No direct contact between sections: Heat transfer through flanges only • No direct contact between tuners and RFQ sections • Grids are brazed locally on the RFQ section T 1 T 2
Energy balance and Water Temperature Rise: • Global energy balance for the RFQ and for each section were consistent between HFSS and the ANSYS thermal analysis. Energy dissipated in the cooling channels in the # sections (duty cycle 5%) Water temperature rise in the cooling channels in the # sections Flow rate Inner/Outer channel =18 L/min Energy dissipated in the cooling elements – ANSYS vs. HFSS balance
Thermal Results: RFQ Sections T>22°C (Jim Stovall request : «The RFQ should run warm to avoid condensation and long term corrosion problems » ) T 5 (max power deposition) • Temperature almost uniform over a wide portion of the pole line • Temperature peak in regions with the highest magnetic field
Thermal Results: Vacuum Grids, Pistons, Flanges & End Cells • Maximum temperature field are found on the last section T 5 and the elements within: Grid # 10 (T 5) Tuner # 15 (T 5) Stainless steel flange # 15 (T 5) End cell + rod @ backend of RFQ
Mechanical Results: RFQ Sections • Only stresses due to thermal load are computed (no gravity or supports are included): • The copper is annealed (brazing process) => Yield stress ~ 35 MPa • Stainless steel elements in 316 LN => Yield stress ~ 170 MPa • Symmetry conditions and a central fixed point (center of the RFQ @ T 3 ) constitute the boundary conditions. Two directional deformations of interest: • The longitudinal deformation along the beam axis: limited effect on the frequency shift ion gat ropa p am Be Longitudinal deformation in the RF sections
Mechanical Results: RFQ Sections Maximum equivalent stresses Transverse deformation • Transverse deformation does not exceed 16µm in the ESS RFQ for this set conditions. • The maximum equivalent stresses in the Copper is at a vacuum grid port and is ~12 MPa which is < ~2/3 of the yield stress => no plastic deformation in the RFQ • The 3 D estimate of the deformation ratio between the pole tip and the cavity wall height varies between 3 (front end) and 2. 6 (backend) along the RFQ => Consistent with the 2 D results
Mechanical Results: Tuners Transverse deformation of the Tuner (15 th) surface in interaction with the cavity does not exceed 13µm. Localized max equivalent stress on the Pistons (Cu): 12 MPa < ~2/3 of the yield stress =23 MPa Max equivalents stress on the flange (stainless steel): 55 MPa < ~2/3 of the yield stress =115 MPa. Transverse deformation for the 15 tuners
Mechanical Results: End cell with rod Transverse deformation the end plate with its rod Relative longitudinal deformation of the end plate with rod. • With no cooling in the end plates, the rod deflection around its axis does not exceed 5µm (only 20 W deposited on both plate on average)
Prototyping plan • A cavity with 2 RF loops + 1 pumping port + 1 adjustable tuner • RF loop: validation by conditionning with a new RF system in CEA definitive ESS RF loop RFQ conditionning faster (RF loop already conditionned) • Adjustable Tuners: the same geometry during adjustment and operation (same perturbation) – Definitive position just after adjustment : No delay of machining between adjustment and final position – Adjustment possible during the operation or after the transport • To validate the industrial assembly for adjustment tuners and pumping ports • The delay is 8 months (RF loops) only for machining, not with administration delay! – • • Be careful of the delay of management’s decision between CEA and ESS The only interface between the RFQ and prototypes is the hole size : it’s possible to start the RFQ machining before the prototypes validation Schedule adjustment RF study already done and mechanic design in progress
Interfaces • We need to have interfaces and discussion between LEBT and MEBT because those parts are essentials for mechanical assembly , alignment…. • Interfaces with building is very important concerning the RFQ integration (and the warm linac in general) • It is necessary to have an infrastructure meetings 2 or 3 times per year – It is neccesary to exchange about the strategy of handling (crane, elevator…), assembly, alignment and test inside the tunnel… – The buildings group can’t know what is specific (stability and fragility) about accelerator equipments like RFQ… • It is not possible to change the building design everyday !!
Interfaces • We must define the interfaces between RFQ and infrastructure, specially about : cooling system and RF distribution : the waveguide – coaxial transition included with RFQ – Work in progress • We need to access to 3 D building pictures to simulate the RFQ integration inside the ESS tunnel (space around the RFQ : RF waveguide, vacuum…)
Diagnostics and C&C: early in kind contribution • • • « there was an overarching agreement between CEA and ESS to proceed on the early in-kind, and it was left to the technical people to finalise the technical details. » We have the « management authorization » to increase the collaborative technical work Source and LEBT – The work is in progress – Be careful of the delay of the ESS system standardization choices with the schedule impact behind • Doppler shift – Start of technical discussion with ESS – First technical definition in June (Design, C&C, Tests…) during a diagnostics meeting in CEA • EMU – We are already working with Catania (mechanic and integration) and ESS (C&C) – ESS timing specification is important (end of this year) – Mechanical design is in progress
CEA In kind contribution • Agreement between CEA and ESS is in progress for in kind contribution • A possible validation in June • First hard order possible in September • Schedule time incompressible – Impact on Source and LEBT schedule • EMU : 17 months First test in Catania without EMU ? – Possible flexibility for RFQ between prototype plan and RFQ machining : critical path
- Slides: 20