I took a deep dive into Ekin 2.6 to understand heat tranfer between various joint types, specifically pressure joints and grease joints. This was motivated by the fact that modeling of the cold plate and inner shield temperatures seemed to be missing some key physics.
Heat conductance through grease joints is dependent on the contact area of the surfaces:
, where = the heat conductance. The heat conductance of a grease contact with area 1 cm^{2} can be found in Attachment 1, indicated by the region between the blue and red boxes (our temperature range).
Heat conductance through pressure joints is dependent on the force holding the surfaces in contact:
, where the heat conductance of a pressure contact with force of 50kgf is found in Attachment 1, indicated by the orange box. (Megastat pressure contacts are between the cold plate and shields, so the Al-Al contact was referenced.)
[Added] Heat conductance through varnish joints are similar to grease joints, in that the conductance scales with contact area of the joint. (It follows the same formula as above for grease joints.) The heat conductance of a varnished contact with area 1 cm^{2} can be found in Attachment 1, and note that it is higher than that of greased contacts.
My efforts this week went into incorporating these heat links into the model. This required splitting up components into various parts (verging on a finite-element approach), since every joint is considered in addition to the elements themselves. A few notes:
1. For now, I assume the force between the inner shield and cold plate is 50 kg, for simplicity. Therefore, the heat conductance being used for this pressure joint is 4.5 W/K (from chart), and I am using this constant value across temperatures until a better estimation can be found.
2. For the grease joints, I estimate the heat conductance at 1 cm^{2} to be ~ 1 W/K, for the 50-300K range.
3. The contact between the outer shield and cold plate is not actually uniformly touching, as noted in 2706. I am not sure how to estimate the actual force between the surfaces, so I will add this as a fit parameter in the model.
The new model with this incorporation still needs some adequate debugging, but I felt these were vital steps to ensure we get realistic answers regarding cooling power of the copper bar vs. LN2. Once I feel the model can be trusted, I will follow up with analysis of the new optimized cooldown. |