2013-08-19 Design Meeting Notes

Isolation Valves

  • Free surface on hot well makes it physically impossible to pressurize the hot leg (hence the containment). Same with free surface on stand pipes on the cold leg.
  • Need isolation valves for containment and CTAHs. For hot well, use gate valve to seal entrance from hot leg (see picture).
  • In hot well, for each primary pump, have a plug for inlet to isolate the pump. This way, one loop can be operational while the other is offline.
  • Other option: 2 hot wells, with 2 hot leg valves, and dam between the 2 wells to prevent communication between the 2 if needed (with additional requirement that elevation change in the wells across all operating conditions shouldn’t be too high). In that case, have 2 parallel hot legs coming out of the vessel all the way to each well. This way, one can perform maintenance on one well while keeping the other full for decay heat removal.
  • Also option to have one hot leg coming to a hot well with 2 compartments, 2 seats, 2 gate valves (see picture).
  • On cold leg, stand pipe might look like “cold well” with cold trap filter. Filter must be at higher elevation than drain tank. Isolation valve can be another gate valve (like hot well isolation valve). Design cold well to overflow into hot well.
  • Surface area in hot well must be large to reduce level swell. Not that critical everywhere else in the system (including cold leg stand pipe). In cold leg stand pipe, add neutrally buoyant volume displacer (graphite?) to reduce salt inventory (see picture). This will have a thin diaphragm for water hammer (if explosion in CTAH).

Decay Heat Removal

  • MSBR had a DRACS for the drain tank, using flibe (so that if there was a leak, they wouldn’t contaminate primary salt)
  • For FHR, use DRACS loop with salt, transfer heat to water, then to air stack. No need for huge air stack since some decay heat will just heat up and boil water.
  • See picture: if following CIET scaling, NDHX would end up being in the space between top of reactor cavity and control rod drive missile shield.
  • Need to isolate water from NDHX to limit parasitic heat removal.
  • Water would be on tube side of NDHX.
  • The water will be at 100°C (heat removal through boiling). This way, not too low temperature (avoid freezing) and not too high temperature (worse heat transfer). Air cooled condenser will condense water back (at 100°C). Any steam that doesn’t condense will be exhausted not to generate extra pressure in the system.

White Paper 1 – comments from Syd Ball

  1. There is a big concern in modular HTGRs about adequate mixing of the core outlet coolant that gives this “integral average” temperature referred to in this section.  As the molten salt coolant moves through the core, I’d guess the power and flow distributions would be such that a fair amount of uneven heating of the coolant would occur.  Since the overall core delta-T is small compared to HTGRs, the temperature fluctuations would tend to be much smaller; however, because of the much better heat transfer via the salt, the effects of the fluctuations on the structures it encounters might even be larger.  Might be worth looking at.
  2. Note that the Japanese (HTTR) had a huge amount of trouble with their water cooled cavity cooling system – to the point that they “would do almost anything to avoid using a water-cooled RCCS.”  As the FHR design gets further along, I’d suggest reviewing their tales of woe.
  3. Regulatory foundation:  At some point, it would be useful to review the IAEA work in this area.  There was lots done for modular HTGRs [at least in my experience – & maybe similar stuff for fast reactors, not in my experience base].  Also work done on generic advanced reactors.  See IAEA TECDOCs 1366 & 1570. 1366: Advanced nuclear Plant Options to Cope with External Events. 1570: Proposal for a Teachnology-Neutral Safety Approach for New Reactor Designs.
  4. General discussions about PRA (in 4.1): there was I think a real good point made long ago by Bob Budnitz (NRC) about how PRAs should be thought of for the passively safe reactor designs, where “failure probabilities” need to be assigned to passive systems [e.g., how do you fail a heat transfer coefficient?!].  My favorite VG on this idea is attached.

Syd Ball Figure_PRA for Passive Systems