Cooling Working Group (CWG) Minutes #8
Date of the meeting: 
17/06/1997 
Place of the meeting: 
40 S-A01 
Present: 
G.Barbier, P.Bonneau, M.Bosteels, P.Bouvier, A.Carraro, A.G.Clark, P.Fontaine, T.Gisler (STAUBLI), E.Hodin, J.Inigo-Golfin, E.Lessmann, R.MacKenzie, W.Van Sprolant, B.Vuaridel.
Agenda:

Short Summary:
A comparative and very detailed study was made on the different types of water cooled power cables. It seems that the system of cooling with water is the most interesting, and for these cables in his case, will be of a special construction.
Since 2 years, Geneva University has carried out tests which have made available a technology which will promote heat transfer and thermal stability of the SCT barrel for ATLAS; carbon structure, heat transfer plate in TPG, oval and sliding exchanger, Leakless system with a mixture of water/methanol are the main thermal characteristics with the Si modules stabilized at -5oC with the incoming fluid at -14oC.

Presented also where the main details of a cooling circuit which allows a lot of freedom in the design of piping, and which gives a good distribution of fluid travelling through the different channels. The principle characteristics are the conveying of pressurised fluids, reduction of pressure through the entrance capillaries, and a sub atmospheric pressure in the exchangers. This kind of circuit is applicable for any fluid.
Detailed Minutes:
As the minutes of the previous meeting were not available, there are no comments to be made. It is largely admitted in LHC experiments that each sub-detector should have a zero thermal balance and this applies also to their power supply cables travelling through other sub-detectors, which dissipate by the Joule effect a considerable power. Therefore, a certain number of cables will be cooled and E.Hodin is working on this problem for the CMS ECAL detector, which is situated relatively near to the center of the experiment (Figure 8-2-1) . A complete report on this was given. Each low voltage bunch powering the 36 modules for the ECAL will dissipate 550[W] and it is foreseen to cool them down at room temperature until their entry into HCAL (Figure 8-2-2) . The other characteristics are a density of 250[A/cm2] and an admissible electric loss of 0.05[V/m].A considerable number of solutions have been envisaged: All the different configurations are grouped together in  Table 8-2-1 , and  Table 8-2-2 shows the section in of cables used by the cooled bunch, following the option chosen, with the distribution between the conductor, cooling and insulation.
It seems that we do not gain very much with cooling by Liquid Nitrogen and it is evident that it would be more expensive to use this instead of cooling with water.
On paper, the most interesting solutions appears to be the bus-bar or hollow conductors, but we now must envisage the problems of manufacture and cost in industry. This is the next step for the CMS/ECAL group. P.Bouvier and B.Vuaridel presented the first results of tests carried out by Geneva on the cooling of the SCT barrel of ATLAS (note that the options are not always ATLAS baseline choices). Their presentation was in fact a summary of a technical note (INDET-No-166) available from ATLAS.
At the same time that they developed a prototype for the SCT barrel (Figure 8-3-1) , the Geneva team started, about 2 years ago, cooling tests on the Si modules  (Figure 8-3-2) . The idea was to study and to validate the solution of cooling (~-10oC) keeping in mind the very large dimensional stability necessary for the SCT barrel. The test installation consists of a Leakless circuit connected to a fridge unit with measurements of flow rate, pressures and temperatures (Figure 8-3-3) . A test trial began using prototype tubes of 1m long at room temperature continuing in a cold enclosure regulated to -10oC with a real module 1.6m long (Figure 8-3-4)
The first measurements were carried out on a round Aluminium tube of 3[mm] OD, 2.6[mm] ID and 1[m] long with a power of 40[W] followed by an oval tube 9.2X2[mm] OD, 0.25[mm] thickness, 1.6[m] length and 60[W] (Figure 8-3-5) . The fluid used was a mixture of demineralized water plus 33% glycol. The differences in behavior between the round and oval tubes are shown in  Figure 8-3-6 ; the DT corresponds to a difference in temperature of the liquid between the entrance and the exit. The round tube shows an abrupt change of efficiency corresponding to the change from laminar to turbulent flow; the oval tube stays with laminar flow and with a better efficiency in this zone because of a bigger exchange surface. It is all the more important that , to work with a Leakless system, a limit must be put on the pressure drop (we allow in general 200[mbar] for the exchanger), hence the flow rate. However, we note that the tube dimensions mentioned above proved to be too small as the viscosity of the water/glycol mixture (Table 8-3-1) was underestimated and caused a too high pressure drop. Afterwards, a mixture of water/methanol 24% (Table 8-3-2) was used giving a much higher flow.
There are the results obtained with the oval tune of 1.6[m] and with a charge of 60[W]:
 
Mixture
Tin [oC]
Tout [oC]
Tair [oC]
Flow [l/h]
DP [mbar]
DT [oC]
DTth [oC]
water/glycol
-9.4
-5.3
-10.7
13.6
280-295
4.1
4.2
water/methanol
-9.8
-7.4
-1.1
22.3
293
2.4
2.5
-16.1
-13.1
-11.5
18.4
326
3.0
3.1
It is noted that the DT measured are very near the calculated ones (DTth) and that the flow measurements are now made through the DT, as the flowmeter with Pelton wheel had many problems at low temperature.
The behavior of the pressure drop on the same tube with the mixture water/methanol at different temperature is shown in  Figure 8-3-7 .
From these tests it is evident that the oval tube should have interior dimensions of 2X8.75[mm] to meet conditions of 200[mbar] of pressure drop with the water/methanol mixture.
The team is also involved in studying carefully the heat transfer within the Si module and has tested different configurations and different materials. The best result were obtained with TPG TC1050 (see CWG Minutes#5 , point 4) as sandwich material, and with a Goretex blanket (Figure 8-3-8) . The distribution of temperatures in the module is shown in Figure 8-3-9 .
Also used in the tests was the grease used for contact purposes, which permitted the shrinking of the cooling tube, without disturbing the Si modules fixed on their carbon structure. A Silicone Dow Corning 340 grease, with a thermal conductivity of 0.84[W/m.K], was selected and put through a series of radiation tests without showing any specific changes.
These tests, therefor gave the Geneva team the opportunity to validate the structure, various materials and the Leakless cooling system for Si modules mounted on their carbon fiber supports.
At the moment the team is testing a group of 8 prototype modules in the cold room in order to optimize the manifolding which would distribute well the cooling fluid between the different tubes. Later on, the prototype will go through full metrology tests at the moment of cooling from room temperature to -15oC to control its stability. Corrosion tests are also foreseen as nothing has been done yet on this subject.
  M.Bosteels presented a design of cooling circuitry intented to be used in large detectors, which allows a very large flexibility for the pipe and tube installations. This type of circuitry is presently used for the liquid radiator for the barrel RICH of DELPHI. It consists of bringing the fluid under pressure right up to the entrance of the detector and to reduce the pressure through capillaries to a negative value inside the detector. The exchangers in the electronic are working at subatmospheric pressure, but the capillaries give a very great freedom in the pipe layout. Nevertheless, the return of the fluid remains a delicate point, as the air-pockets must be avoided and that the pressure drop is relatively limited Figure 8-4) .
The type of circuit has enomous advantages: It was discussed of a heat transfer fluid, Syltherm XLT. A certain amount of transfer fluid exists on the market and a 3M document makes comparisons between their performances as compared to other traditional fluids Table 8-5.pdf .
As this was the last meeting before the summer break, M.Bosteels expressed his wishes that the Cooling Working Group, without changing its interests and advisory role, will start to look more seriously to R&D. This will of course need more personnel and material. It is up to each individual to give thoughts to these questions.

The next meeting will take place in September; the date and agenda have not yet been aranged.
 

CERN/P.BONNEAU/21/08/97