ATLAS Upgrade Thermal Management Group

Document Sample
ATLAS Upgrade Thermal Management Group Powered By Docstoc
					ATLAS Upgrade Thermal Management Group

Future ATLAS tracker Thermal Management Research Programme
BNL, CERN, Genova, Glasgow, KEK, LBNL, Liverpool, Marseille, NIKHEF, Oxford, Prague, QMUL, RAL, Sheffield

This document outlines a 4 year research programme towards the thermal management of the future ATLAS tracker at the luminosity-upgraded SLHC. It describes our present understanding of the tasks ahead, with a first attempt of identifying R&D issues and a work plan outlining a possible sharing of activities within the working group1. The aim of this programme is to deliver the technology choices needed for the thermal management of the future ATLAS tracker. At the end of this programme the chosen technologies will be verified with tests of super-module prototypes in a realistic environment (similar-to-final cooling plant etc.). At present this programme is at the stage of a letter of intent. None of the groups on this proposal has any funds for ATLAS upgrade thermal management activities yet, but most of them are applying for funding with their national funding agencies right now. Review and approval of this document through the ATLAS upgrade management will support this effort. Participation in the proposed programme for all groups is contingent to the availability of funding.

Based on the positive experience during the ATLAS construction and commissioning so far we propose to use evaporative cooling as the basic strategy for thermal management. While retaining this basic strategy, the task of thermal management of the future ATLAS tracker will have to go beyond the existing system. The overall scaling in size of the tracker will result in an increase in the total power to be removed as compared to the ATLAS SCT and Pixel cooling system. In addition, the finer segmentation, in particular in the planned inner-strip region, could considerably increase the power densities. Finally, the operating temperature of the future tracker might need to be lowered significantly to keep thermal runaway and noise under control. All this has to be achieved while the space available for services for the future tracker will stay essentially the same as for the ATLAS ID and a significant fraction of those might have to be re-used because of scheduling and access constraints. The size of the future ATLAS tracker and the limited time available for the upgrade will require segmentation into robust units (supermodules) of as yet undefined shape and size to allow for efficient, probably distributed assembly. Good integration of services and high modularity will also facilitate this task. Any design serving the thermal management of the future ATLAS tracker will need to contribute to these goals.

Definition of task
Thermal management of the future ATLAS tracker involves the transport of heat from the on-detector heat sources (sensors, ASICs, etc.) to an off-detector heat sink, together with the maintenance of a suitable environment. We have identified four groups of activities within this task: 1. Transfer of heat from the tracker super-modules to an off-detector heat sink: Due to the long distances only physical transport of a coolant can provide the necessary heat conductance. Our experience during the SCT assembly with the evaporative cooling technology

The contents of this note have been developed in a workshop of the working group in September 2006 (


chosen for the ATLAS SCT was very positive we plan to use this basic technology in the future tracker as well, albeit with a different coolant and evaporating pressure. 2. Transfer of heat from the heat sources (electronics and sensors) to the coolant: Space and material constraints will define how close the coolant can be brought to the heat sources. However, there will inevitably be a solid heat transfer path between the heat sources and the coolant, which will need to be designed carefully using recent advances in material science and avoiding the need for mechanical interfaces. One of the crucial components of this path will be the cooling pipe, which in addition to its thermal properties will also play a structural role within the super-module. In the worst case the cooling pipe might have a different coefficient of thermal expansion, which will have to be absorbed somehow by the mechanical structure. However, the cooling pipe could also add moment of inertia to the supermodule and it will be the goal of an integrated super-module design to make use of this potential. 3. Thermal environment: Taking into account the low operating temperature of the future tracker condensation is an issue within the tracker volume, where the dew point has to be kept sufficiently low, as well as on the outside, where the temperature of the tracker outer skin and any pipe surfaces have to be kept above the dew point of the experimental hall. Thermal neutrality might be less strong a constraint as in the present system due to the change in neighboring systems with other thermal requirements. 4. Control, monitoring and interlocks: The cooling system will have to operate within a large dynamic range of total power (zero to full detector power) and, potentially, operating temperature and a suitable control strategy has to be developed. Sensors and control circuits have to be found, which can operate in the high-radiation environment. Critical fault conditions have to be covered by hardware interlocks. The possible solutions are constrained by several boundaries: 1. Time: As justified elsewhere the R&D programme for the future ATLAS tracker has to be completed by 2010 to allow for a timely evaluation of the overall feasibility of the ATLAS upgrade project and decisions on the build-phase. Subtracting construction and testing time for prototype super-modules this allows for about 2-3 years of real R&D. Every delay in obtaining funding will reduce this period even further. 2. Existing services and ATLAS operation: Access to many of the existing services is very difficult and replacement might increase the interval between the shut-down of LHC and the start-up of the high luminosity upgrade. The future tracker might therefore have to reuse existing services to a large extent. 3. Radiation environment: Not only will the radiation put severe demands on the radiation-hardness of components used in the thermal management, but it will be actually the driving force behind many of the requirements for the cooling system, the most prominent being the operating temperature (through the requirement of thermal stability of sensor operation) and the cooling power (through the number and density of detection elements required to allow for a robust pattern recognition and track reconstruction).

We have started to maintain a list (Table 1) of the requirements for the various aspects of thermal management within the future ATLAS tracker. Many of these requirements cannot be quantified now and will only become clearer as the project advances. Nevertheless they have been listed already to indicate our understanding of their relevance in the future. Several of these will have to be developed in collaboration with other aspects of the overall upgrade project.
Table 1: Requirements for the thermal management of the future ATLAS tracker. Item Requirement Tsensor max = -27°C - ? (where HV power consumption becomes negligible) Safe from thermal runaway if you increase Based on/assumptions

 Strip sensor temperature


 Pixel sensor temperature  Temperature uniformity  Coolant temperature

power by factor 2 Needs to be defined ΔT < ? Δpcooling pipe < ? bar Tcoolant = Tsensor - ΔTthermal path Coldest point (sustained) T > ? Coldest point (instantaneous, incl. faults) T >?

Mechanical deformation, electronics performance derived

 Temperature dynamic    
range Power dynamic range Strip power consumption/channel Pixel power consumption/channel Strip Power density (2d) 0-100% Pchannel = 2.7 - 3.5mW Ptotal = Nchannel × Pchannel Pchannel = 70μW Ptotal = Nchannel × Pchannel p = Pchannel × Wmodule/(Wstrip × Lstrip) Current estimate

30% lower if power substation at end of supermodule Non-flammable Non-toxic Dielectric Legal regulations Overall leak rate < Reuse present SCT/TRT services from? Reuse monophase cable cooling? Heat flux to cryostat wall < ? Interface to beampipe? Convective heat transport must not adversely affect sensor temperatures Services Entire tracker runs with one cooling system Minimize number of thermal volumes TDP < Tcoolant - 15°C

 Pixel power density (2d)  Coolant properties

   

Filtering Leak rate Final stage oil-free Services

Check power neutrality

 Thermal neutrality

 Thermal enclosure     
segmentation No condensation inside tracker Outgassing rates Feedthrough leak rates No condensation outside tracker Electrical insulation

Sufficient margin for operation

Tsurface > 15°C Insulation (breaks) off supermodule Insulation (breaks) within supermodule? Insulation to modules? Non-recoverable failure rate × affected fraction < Catastrophic fault modes protected by hardware interlocks Recoverable failure rate × downtime × affected fraction < ? Maximum accumulated luminosity = ? fb-1 Flux through supermodule (incl. coolant) Maximum accumulated dose coolant Maximum accumulated dose supermodule components Maximum accumulated dose control & monitoring components (incl. readout) Cooling components within supermodule:

 Detector safety

Operational safety Operational safety Reliability

 Radiation hardness

 Material


 Cost

< ? X0 Other components: < ? X0

In the following we will outline the specific research activities we plan to address within the four main sections outlined above. In brackets we list groups which have stated an interest in particular activities or part thereof. Note that these statements of interests are at a very early stage and will need to be revised once the organization of the future collaboration and funding is better understood.

Transfer of heat from the tracker super-modules to an off-detector heat sink:
1. Services: Many of the existing services of the ID cooling and gas systems are difficult to access and their exchange would require significant efforts and time. It might therefore be desirable if not mandatory to re-use at least a fraction of the existing services. For this an inventory of the existing services together with an evaluation of their access situation has to be created. One issue with the existing pipework is the very low pressure specification: the specified design pressure is 25 bara [1] and no pressure tests have been performed exceeding this specification. Any cooling loop with a large difference between the evaporation and condensation (or environmental) temperature, as expected for the Future Tracker cooling system, will also have to cope with a large pressure difference and it is quite possible that the requirements for the send/return pipeworks of the future cooling system will exceed the present value in any scenario. 2. Coolant (Marseille, NIKHEF): In evaporative cooling systems the choice of coolant is primarily driven by the evaporation 14 temperature of the coolant. In the12future tracker this temperature will need to be lower at least in NH3 C3F8 12 10 the intermediate strip region to keep thermal runaway under control.10At present we have identified 8 8 two potential coolants for a coolant evaporation temperature below -35°C: 7.3 bara 6.6 bara 6 6 a. C2F6: So far our experience with the perfluorocarbon C3F8, which has very similar 4 4 properties, has been very positive during the commissioning of the ATLAS ID. 2 2 0.5 bara b. CO2: This coolant is being0 used in the cooling systems for the LHCb VELO and in the 0.7 bara 0 -45 15 20 -60 -45 -40 -20 0 40 -60 -20 0 40 AMS detector. IndustryT is also 15 20 showing increasing-40 interestT in this coolant. [°C] [°C] Saturation curves for both coolants are shown in Figure 1.
p [bara]

p [bara]

80 70 60

Critical point

p [bara]


p [bara]

Critical point 27.1 bara

30 25

50 40 30 Freezing 20 8.3 bara -60 -45 -40 -20
T [°C ]

50.8 bara

20 15 10

5.5 bara

10 0 0 15 20 40 1.2 bara

4.5 bara -75

5 0


-30 T [°C]




R&D on these coolants will involve the definition of viable cooling cycles for the required mass 30 25 flow, taking into account the specifications of available components (compressor etc.). Once the 20 basic feasibility of these cycles has been established, small demonstration systems will be built. 15 One aspect of the coolant search will be the study of the behaviour of the coolants under 10 irradiation. The implications of high potential fault pressures and the mass flows on the cooling 5 1 bara 0 pipe and ultimately the material budget of the super-modules will be evaluated for the two -196 -210 -190 -170 -150 -130 coolants. T [°C] 3. Coolant properties (Prague): Coolant studies do require a good understanding of a large range of thermodynamical properties for the coolant. While some data tables for common substances do exist in the public domain, these have to be verified experimentally.
p [bara]

Figure 1: Saturation curves for CO2 (left) and C2406 (right). F Critical point N2 35


4. Open evaporative cooling (Genova, RAL): Most experience of evaporative cooling has been of the closed circuit type, where a discrete number of cooling tubes is collected by a single exhaust tube. The main disadvantage of the closed type layout is that the exhaust tubes compromise the active coverage and also represent a significant mass contribution, especially when combined with exhaust vapour heaters. A possible new development will be an open system where fluid is directly evaporated by the active components and the vapour is allowed to freely circulate in the detector volume and exit at a single point. Such a system obviously would have a strong impact on the design of the thermal environment on the super-modules. As this is a new approach such a programme would have to start with a demonstration of the basic feasibility. 5. Cold/warm pipes (Marseille): The temperature of the outer surface of the pipes leading to and from the detector must stay above the dewpoint in the experimental hall. We want to evaluate two alternative scenarios for the temperature of the coolant during transfer: a. Warm pipes: On the exhaust line the coolant has to pass through heaters when leaving the thermal volume of the tracker to achieve a temperature above the outside dew point. We will investigate the space and power requirements of return heaters for the future tracker. While in principle the coolant could be supplied to the detector at warm temperatures, this would sacrifice a significant fraction of the available latent heat and would require an increase in mass flow. The usual technique is therefore to pre-cool the liquid. Obviously this should be done close to the thermal enclosure. Again two options exist:  An external heat exchanger, where the heat is removed by a separate cooling loop, possibly integrated in a bundle with the regular feed pipes, which has no other purpose than providing the heat sink for the pre-cooling (also known as “coldnosing”).  An internal heat exchanger, where the coolant before the throttling is in thermal contact with the return flow of the same loop and the heat is used to boil off the remaining liquid and superheat the return gas. b. Cold pipes: In this scenario the coolant will be brought to and from the detector at temperatures significantly below the dew point. An insulation layer around single or multiple pipes has to provide the temperature gradient. For this option insulation materials will have to be evaluated (e.g. Cryogel, Cryoloft). This activity will be closely linked to item 2 above and to item 1 in the Control section. 6. Test facility design and manufacture (CERN): Tests of the various aspects of thermal engineering will require smaller cooling plants at various sites within the thermal management working group. We believe that producing a few standardized cooling plants using approved parameters will help portability of results and reduce parallel effort to design, manufacture and learn to operate such local systems. Tests of the thermal engineering and other aspects of the supermodule design will require smaller cooling plants at various sites within the tracker design project. Ultimately these systems could be used in a distributed manufacture-cum-QA scenario for the future ATLAS tracker. This activity could involve either design and manufacture of the test plants or specification/design within our group with subsequent manufacture in industry.

Transfer of heat from the heat sources (electronics and sensors) to the coolant:
1. Highly thermally conductive materials (Glasgow, Oxford, Liverpool, QMUL, RAL, Sheffield): Survey and assessment of relevant existing highly conductive materials (specified to be those with thermal conductivity > 100W/mK), as well as further advanced materials that are becoming commercially available. This requires measuring, where necessary, mechanical moduli and thermal and electrical parameters, and assessing practical aspects of machinability, and the production techniques that would be necessary for their use in large scale industrial fabrication as module components. This also involves surface preparation techniques for materials, as these are key to producing the ultimate quality for good thermal interfacing. -5-

2. Thermal interface materials (QMUL): Survey and assessment of thermal interface materials (TIMs), required for all potential demountable intra-module components, and for ultimate interfacing with extraction heat sinks. This is a rapidly evolving product field where developments are expected over the period of this programme, particularly in applications involving carbon nanotubes and new exotic composites. 3. Interface resistance measurements (QMUL): Vital understanding and measurements on interface resistance are a separate area of engineering research but one which forms an integral part of this programme. We are expert in this work and would carry out all necessary interface measurements for intra and extra-module component interfaces. 4. Small prototypes (Glasgow, Liverpool, Oxford, QMUL): Following the progress on the previous items we will assess possible combinations of materials and TIMs and construct small sample structures for test measurements, aimed at assessing mechanical robustness (through extension measurements), thermal transfer (through video thermograms) and intra-layer adhesions and contacts (with ultrasonic microscope scanning). These will be used to benchmark the FEA analysis in 6. 5. Fittings (LBNL, RAL, Sheffield): Although strictly speaking not a part of the thermal path a critical question for the design of the cooling pipe is the connection to the feed and return system of the off-detector cooling plant. Reliable, small, low-mass and low-cost technologies will be needed, which do not introduce any strain on the various system components. Commercial solutions can be bulky and are made from stainless steel, but new commercial solutions like hybrid st/stl/Al VCR flanges have become available. Electron beam (EB) or laser welding of very thin tubes to these flanges would be researched in collaboration with the Welding Institute (TWI) and EB Engineering Ltd. Long-term testing will be performed to verify the applicability in the future tracker. 6. Thermal and mechanical simulations (BNL, CERN, Glasgow, KEK, Liverpool, Oxford, QMUL): The work on the previous work-package components will be paralleled with thermal and mechanical simulations of complete module structures, used to assess heat path designs that are worthy of prototyping. They will also allow optimisation of the thermal and mechanical properties to achieve modules with minimum mass to optimise physics performance. The simulations will be particularly important in the study of thermal runaway, which will provide significant input to the definition of the operating temperature of the cooling system for the future tracker. 7. Radiation tests (Oxford, Sheffield): Any possible solution will require understanding of its performance after irradiation. For some components (e.g. thermally conductive grease, etc.) the long-term performance after irradiation will have to be demonstrated. Where possible we will perform these irradiation studies in collaboration with radiation tests within the framework of other ATLAS upgrade work packages (e.g. opto-electronics). 8. Large-scale prototypes (Glasgow, Liverpool, Oxford, QMUL): Finally, as a preparation to the manufacture of a super-module prototype we will build full scale prototypes of the cooling pipe/thermal interface assembly, using final or close-to-final connection techniques. These will be used to verify our expectations of the mechanical and thermal performance.

Thermal environment:
1. Moisture influx: To achieve a dew point of -50°C the moisture content of the gas in the thermal tracker volume has to be 39ppmV. This is about a factor 10 smaller as in the design for the ATLAS ID (dew point target -30°). At the same time the tracker volume will be approximately four times larger. An obvious solution to this is to increase the flow of dry gas into the thermal volume by a factor 40. However, should this be prohibited by costs or feed line size restrictions we will have to reduce the moisture influx from detector material outgassing, diffusion through the thermal enclosure or leakage through service feedthroughs. -6-

2. Thermal enclosure: The present practice within the ATLAS detector is thermal neutrality for all sub-systems. However, the future tracker will be radially limited by a polyethylene moderator within the solenoid/calorimeter cryostat and strict thermal neutrality might not be required any more. At the inner radius the tracker will be in contact with the beam pipe. There the heat sink generated by the tracker cooling must still allow for bakeout of the beampipe. To achieve thermal insulation at the thermal enclosure we will investigate passive insulation, active insulation with outside heaters only and active insulation with outside heaters and an inner surface cooling. 3. Segmentation: Any segmentation of the thermal environment will obviously follow the division of the future tracker into separate units for installation. While large thermal volumes are preferable for the reduced material further sub-division might be necessary to achieve thermal neutrality between different tracker parts, even during commissioning and under fault conditions. 4. CFD simulations (CERN): Moisture and convective heat transport within the thermal volume will be simulated to optimize thermal barriers and the location and flow rate of inlet and outlet pipes for dry nitrogen.

Control, monitoring and interlocks:
1. Mass flow control (Marseille, NIKHEF): The cooling system has to operate within a large dynamic range (zero to full load). To achieve this three strategies are possible: a. Controlled mass flow: A feedback loop between the return fluid and a flow regulator in the feed line will control the mass flow to slightly more than the flow required to cool the detector at any given time (also during start-up or at partial load). Different technologies for this feedback are available and will be evaluated using the test systems described above. This strategy would allow for reduced return heater sizes (in the optimal case the heaters only need to superheat the return vapour from the evaporation temperature to the dew point limit as given by the vapour heat capacity). b. Fixed mass flow with full power heaters: The mass flow is set to slightly more than the mass flow required to remove the heat from the fully powered tracker. Any excess latent heat is then boiled off in the return heaters, which in this scenario have to be large enough to provide a power equivalent to the maximum load of the tracker. To evaluate this option we will study heater geometry and power scenarios. c. Fixed mass flow with reduced heaters: As in the previous strategy, but with smaller heaters. In the case of a sudden reduction in the load this system would sense the drop of the temperature in the return line and shut down the cooling loop (together with all the detector modules served). To understand the applicability and necessary heater power for this scenario we will investigate the system stability (including detector load variations during data-taking and calibration). 2. Throttling (NIKHEF, Prague): Various technologies exist to achieve the pressure drop to define the point of evaporation in the cooling loop. There are passive (capillaries, orifices, porous plugs, etc.) and active, where the latter allow for the regulation of the mass flow within the circuit as well (e.g. thermostatic expansion valves, etc.). We will evaluate issues like reliability, filter requirements, control possibilities, reproducibility and radiation hardness. 3. Sensors, regulators and valves (Oxford, Prague, RAL): Again, different sensor technologies exist to measure (mass) flow, humidity, dew point, pressure, temperature and other operating parameters. We will evaluate their performance and applicability for various locations within the system. We will also investigate the possibility of a direct measurement of the vapour quality, something which is not presently available commercially. In the design of the super-module cooling components we will strive to achieve good integration of the necessary sensors in the overall design (mounting and readout connections). -7-

4. Operational safety: Various risks potentially destructive to equipment do exist in connection with the thermal management (cooling power failure, condensation, etc.). Efficient hardware interlocks will have to be developed against these risks. 5. Personal safety: Personal safety needs to be taken into account from an early stage as it is likely to have an impact on the design of the cooling system. 6. Reliability (Oxford): Non-fatal faults still reduce the capability of the ATLAS experiment to take data. We will therefore attempt to maximize reliability already at the design stage by quantifying non-fatal fault rates and recovery times, followed by a quantitative reliability analysis.

Decisions and Milestones
The proposed schedule for this programme is geared towards the assembly of a supermodule prototype together with a realistic external cooling plant and an appropriate representation of the thermal environment in 2010. The steps towards this goal are a pre-stage of decisions in 2007, with first major decisions taken mid 2008 and final decisions taken in mid 2009. In particular we have so far identified the following decision milestones (incl. issues to be studied towards the decision):

Pre-stage 2007:
   Re-use of existing services (needed for following decisions). Feasibility of open evaporative cooling. Decision to pursue.

First decisions mid 2008:
Coolant choice: o Definition of cycle (Compressor, cold/warm pipes, maximum (fault) pressures), o Supermodule cooling pipe requirements and material, o Number and cross-section of external pipes, o Radiation effects, o Cost (incl. environmental legislation). Controlled or constant mass flow: o Controlled mass flow: Scheme of control, response time, required overhead in cooling power, and radiation hardness of control components. o Constant mass flow: Number, size and power requirements for heaters. Throttling: Reliability, required filtering, control possibility, reproducibility, and radiation hardness. Cold or warm pipes: o Insulation material properties: space, radiation length. o Internal hex: Operating parameters for evaporation temperature, power consumption, (“efficiency”), size, and pressure drop. o External hex: space, radiation length. o Heaters: Number, size and power requirements. Fittings: Leak rate, reliability, material (radiation length), and space. Thermal interface materials: Thermal conductivity, radiation hardness, and mechanical properties. Thermal/gas enclosure: segmentation, requirement for thermal neutrality, and active cooling (heaters only or cooling & heaters) or passive cooling. Sensor technologies (Flow, vapour quality, humidity, dew point, pressure, temperature, others): Locations, radiation hardness.


 

   

Final decisions mid 2009
 Supermodule cooling pipe: Thermal expansion, material (radiation length), cross-wall thermal conductivity, structural properties, and connection techniques. -8-

Documentation and coordination
We intend to keep regular meetings (every 2-3 months). Documentation will be centrally accessible through an upgrade thermal management website. First materials accessible from this web site, in addition to the contents of this document and minutes from the first workshop, will be a collection of experiences with the assembly and commissioning of the ATLAS ID cooling system and software tools and coolant and material data tables. Coordination of activities within the ATLAS upgrade community will be through the upgrade project office, of which this working group is a sub-group. Interface definitions will be coordinated by the project office. Interface definitions relevant for the work of the thermal management working group will be available on the website. The R&D effort towards a future ATLAS tracker within the project office will be organized so that all aspects of thermal management of the future tracker supermodules will be developed within this working group. Contact with experts on thermal management within the CMS collaboration has been established. Within both collaborations work on upgrade issues is just commencing and no details of possible joint efforts have been discussed yet. However, we do agree on the usefulness of close collaboration.

In this document we have outlined a 4 year R&D programme for the thermal management of the future ATLAS tracker within the framework of the overall ATLAS upgrade organization coordinated by the upgrade project office. The next step towards accomplishment of this programme is the acquisition of funds by the contributing groups. This will be supported by approval of this proposal through the ATLAS upgrade management.

[1] 671138v.2 DO-23054_Technical_Specification, 671138v.2 DO-23054-TS-ATLAS, ATC-TL-EP0001v.5.


Shared By:
Jun Wang Jun Wang Dr
About Some of Those documents come from internet for research purpose,if you have the copyrights of one of them,tell me by mail you!