30 November 2001

Presentations

Minutes

A conference call was held at 10am on Friday, November 30. The purpose of the call was to review analyses of alternate cooling schemes to determine if there was a more attractive option.

The reference cooling option features a chill plate located in the center of each double pancake.  The chill plate is cooled by gaseous helium at 80K.  The helium is introduced to the chill plate at numerous inlets located along the perimeter of each double pancake.  Two concerns with this approach have been noted, including...

The concerns noted above bring two issues into focus.  First, would liquid nitrogen cooling offer significant advantages over gaseous helium (or gaseous nitrogen)?  Second, would cooling the pancakes from the outside offer significant advantages over cooling between the pancakes?

HM Fan performed a series of ANSYS runs to address the issue of cooling from the outside.  His conclusion was basically that each pancake had to be in contact with a cold surface.  Cooling the outer pancake through the inner pancake did not appear feasible with 15 minute rep rates.  Conceptually, the interior location (our reference design) is the most efficient, but this appears to necessitate interrupting the structural continuity of the winding pack.  Alternatively, Fan showed that if we put copper strips inside and outside the winding pack, we could efficiently conduct the heat out of the winding pack.  In this approach, the slip plane (Teflon tape?) would be provided between the winding pack and the copper strips.  The winding pack could be potted as a monolith.  This approach has the important advantage of cooling the shell as well as the winding pack.  This is important for cooldown as well as normal operation.

Fred Dahlgren presented an analysis of liquid nitrogen cooling.  He assumed a liquid nitrogen pressure of 250 psia (1.73 MPa) and a temperature of 77K.  His conclusion was that adequate heat transfer could be provided with low flow rates and pressure drops.  In his base case, he assumed a 5 psi pressure drop, which resulted in a 3.8 gpm flow rate (11 ft/s flow velocity).  The temperature rise along the 15' length was barely noticeable - on the order of 1K.  The heat transfer was dominated by internal conductance.  There was a concern noted about the heat transfer coefficient used.  Following the meeting, Dahlgren reviewed his calculations and documented how the heat transfer coefficient was calculated.  His conclusion that the heat transfer was dominated by internal conductance remained unchanged.

The saturation temperature at 250 psia is about 113K.  This would permit a temperature rise of 35K before getting into the 2-phase flow regime.  A lower operating pressure would provide less temperature headroom, but the lower operating pressure might be easier to accommodate.  For instance, an operating pressure of 78 psia (5.3 atm) would still have a saturation temperature of 95K, which corresponds to an 18K window between the inlet and saturation temperatures.

With liquid nitrogen cooling, only a single cooling channel per double pancake would be required.  (Two might be desired.)  An uninterrupted cooling tube could be soldered into a channel in the shell.  Connections to this cooling tube could be made outside the shell.  This should significantly enhance the reliability of the design.  If a minor leak did occur, it probably would not matter because we would be leaking inexpensive nitrogen as opposed to helium.  Failures, if they did occur, would probably be located at cooling connections which would now be located outside the shell where they would be accessible, rather than inside the shell where they would not be accessible.

An alternative to liquid nitrogen or gaseous helium cooling would be gaseous nitrogen.  For normal operation, this appears inferior to both of the above.  Relative to liquid nitrogen, the density is down by 3 orders of magnitude.  Relative to gaseous helium, the inlet temperature must be at least 10K higher than the 77K (or lower!) that is possible with liquid nitrogen or gaseous helium.  To get the inlet temperature down to 88K requires dropping the inlet pressure to 3 atm to avoid 2-phase flow.

However, even if liquid nitrogen is the coolant of choice for normal operation, gaseous nitrogen is probably the coolant of choice for cooling the device down to 90K from room temperature.  We need to consider if we can gracefully cool the device down to 90K by circulating gaseous nitrogen through the same cooling tubes that we would use for liquid nitrogen to cool the device the rest of the way down.

Please forward any comments or corrections to reiersen@pppl.gov

(last edited on 12/04/2001 09:04 AM )