search

nagimov / cyclotron-cryopanel

Mathematical model of the cryogenic pumping system of TRIUMF cyclotron
GNU General Public License v3.0
1 stars 1 forks source link

Make a model for the tank transfer lines #52

Open savreline opened 7 years ago

savreline commented 7 years ago

Obtained literature papers such as

Experimental Studies of MLI Systems at Very Low Boundary Temperatures by I. E. Spradley, T. C. Nast, D. J. Frank

that discuss thermal conductivity of MLI.

savreline commented 7 years ago

Went to see samples of MLI in the helium building. They consist of fiberglass and Mylar. Found the following correlation for k for this type of MLI.

k = Cs * N^1.56 * (T_h + T_c) + Cr * eps * (T_h^4.67 + T_c^4.67) / N * (T_h - T_c)

savreline commented 7 years ago

Measured (using a precision caliper) that thickness of mylar is 0.01mm and thickness of fiberglass is 0.18mm. So we get a density of MLI of about 50 layers per cm.

savreline commented 7 years ago

Came up with rough calculations based on paper by Spradley, Nast, Frank ... see attached

TransferLineHeatLosses.pdf

heatlosstl

savreline commented 7 years ago

Found original drawings of the cryo-line section in the design office. They show 15+10 layers of Al & glass insulation. Will adjust calculations.

savreline commented 7 years ago

Started making a model in COMSOL of the cryo-line heat loss, based on the drawings found in the design office.

solutuion7 solutuion4 solutuion8

savreline commented 7 years ago

Cannot figure out how to integrate flux through a boundary in a new version of COMSOL. Message sent to the COMSOL support team, should get answer within 24hr. Solved the model in COMSOL 3.4 got heat loss of 13W for the Helium line and 6W for the Nitrogen line.

savreline commented 7 years ago

COMSOL support team was great and explained how to integrate boundaries in the new version of COMSOL. Updated model:

modeldec16

savreline commented 7 years ago

However not sure if COMSOL is taking vacuum into account when calculating heat transfer. COMSOL support team is not sure either. They said to just make a custom material and enter k manually. However, CoolProp calculation do not seem to show any difference in k as a function of pressure!?!?

airprop

savreline commented 7 years ago

Seems that CoolProp & Refprop are unable to calculate thermal conductivity at low pressure. Refprop gives an error, while CoolProp doesn't.

Temp solution is to use k=1e-10 for vacuum. Updated COMSOL simulation with such values. Get heat loss of 0.28919W/m for Nitrogen, 0.36375W/m for Helium. This translated into 10W over the entire line for Nitrogen and 12.5W for Helium over the entire line.

solutiondec21

Assume emissivity of Al and polished SS was 0.05

savreline commented 7 years ago

Some information about thermal conductivity of gases at Russian wikipedia

savreline commented 7 years ago

Based on the above simulation, calculated that we lose about 20 % of quality of N2 over the 30m line!

cryolineloss