12-19-2013 Design Meeting

Flow from the central reflector, as currently designed, is potentially a source of lots of bypass flow.
You could change the geometry of the slots to increase the flow resistance.
The main problem with this type of bypass flow will mean that some of the flow coming out of the core region will be above 700 degrees.
if the delta T is 20%, that’s a problem, because the core outlet will have to e mixed.
thermal power is generally flow rate * delta T

There are two different atws events.
one is LOHS : the pumps though, are still running. so, we have the capability to tune the amount of bypass flow going through the DHX. The flow pattern in this case is that the flow going through the core is 98% of the flow. 2% of the flow may be going through the DHX heat sink. However, we may not be removing heat efficiently in the DHX because it’s designed to avoid being a parasitic heat sink. A good question is what kindof parasitic heat load we get out with the DHX? Another question is whether we want those pools to be constantly be sitting there at 100C? A lower temperature may be better to avoid the concrete degradation. Having a liner with active cooling for the water pools is good. That feature would also be a good example for best practice because you can add a leak checker.

the other LOFC is less of a problem because of natural circulation.

Mike : For completeness, there’s another ATWS scenario that includes just one pump shutting down. This is an imbalanced situation in which you may see in the pumps or other features as well.
We may need to think about the single component failure accidents as a design basis concern. For this reason, the situation in which one pump shuts down is no longer a BDBE, but, rather, a DBA.

Per: We have the question about when to interrupt the supply of energy to something. So, since there are many diverse ways to interrupt power to something, good failures are failures that occur during power interruption and fail to a safe state. Still, there is sometimes the situation in which, mechanically, the component won’t fail successfully to a safe state. Also, there may be something that keeps power interruption from happening successfully, so we want to try to make sure that the safety systems are effectively physically secured because all you have to do is to interrupt malevolent actors before thety can break into containment to access equipment that they could destroy to disable expected power interruption and other safety situations.

Nico : Using Tommy’s decay heat curve, the pools can be 30 cubic meters large. This is based on 24 hours of use, using two tanks and full evaporation. This is without any re-condensing.
Per: We also have to think about the way the freezing problem is handled. You may be able to do this by activating some sells and allowing air flow only into some cells under cold ambient conditions. In this case, then, you just cool a subset of the condenser surface.

Mike: could we use our very nice heat sink to help the condenser behavior ? Or, use our very nice heat source to help the condenser behavior in this freezing situation? or
Per: Maybe we could just actively cool the secondary liner system between the pools and the concrete. Additionally, the cooling water provided to the motors may be provided by some other heat management system in the building. (fan coolers, HVAC system… ) We also should have redundant ways to cool the reactor cavity, as an investment protection system.
Raluca: Do we have a plan for such a redundant manner of concrete cooling.
Per: We could have cooling tubes that are fed by more than one cooling system. You could have three tubes that alternate going up each of the flat surfaces that alternate between three trains of service water supplied by different manifold pipes. The more difficult thing is how do we provide cooling to the thermal shield.
Raluca: That’s a good project for the 170 class!

Raluca: So, for the design of the DRACS, in addition to a safe failed state, should we also design for no-power instrumentation.
Per: First, what is a safe failed state for the DRACS? It’s likely to be a fully operational state. That means that in the longer term, we need to have energy to support the dracs and possibly need heat to avoid freezing in the dracs and heat exchangers. It may be possible to design the dracs in such a way that they fail elegantly when the salt starts to freeze in them. The dracs will certainly stop heat removal from the core.

Raluca: If you froze the downcomer, then you could potentially have freezing in the core.
Per : Temperature measurements will be important. The question is, can we measure flow?
Raluca:  What about pressure drop?
Per: The delta Ps might not be an easy way to determine flow.
Raluca: It might be that the difference in elevation between the surfaces, plus the pressure drop would give the flow.
Per : That’s worth checking

Tommy: Whatever happened to us using dowtherm for freezing experiments?
Per: It may be difficult to argue that dowtherm would behave like salt.

Does sinap have a reports repository in english?

We’ll want to focus in comsol on the biggest losses.
the ratio of DHX pressure drops to the core in the Annular AHTR was no more than two.
This points to having a strong basis for characterizing the DHX pressure drops, etc.
We’re likely to see the same stratification in natural circulation.
We should see reynolds drop for lower power levels as well.
Once we do this with better analysis, then there will be a reasonable basis with which to do some parametric analyses and sensitivity studies to see where we can make changes and gain efficiency.
Comparison between the DHX and TCHX will be important too.
Do we want to use twisted tubes in the TCHX?

Per: Discusses the plan for filling ciet.
Question: how do we pump this container if we want to empty it? (there is a little pump that can be used to prime it before siphoning.)
Also, how do we purge gas out of it?
Raluca: you can do it through ciet.


2013-09-20 Design Meeting Notes


  • DHX sits at bottom of metal lid
  • Arrangement on plate for modularity
  • Large radius elbows for DRACS loop to allow for flexible inspection instruments to be inserted
  • DHX to TCHX centerlines will be ~6m, which is a distortion compared to scaling from CIET (would be ~8.5m)
  • Each DHX should remove 2% of nominal power. 2/3 failure logic.

Hot well

  • Level difference between two hot leg penetrations in hot well creates a seal loop. 0.5m of head can be accommodated if isolation valve fails.
  • Look at calculation for hot well height required above penetrations for thermal expansion (600°C to maximum accident condition temperature) + level swell from pump operation.

Reactor cavity cover – refueling deck hatches

  • Need to have a center circular hatch to pull out center reflector
  • Hatch for reactor vessel
  • Hatches for 3 DRACS (see picture)
  • Seal for DRACS hatch could be integrated in frame design of DRACS
  • Because of high number of penetrations in cover, may need to have it be steel
  • Missile shield above has fewer penetrations and could be concrete

2013-09-20 Cavity Covers

Level swell

  • Design objective: 2.0 m of head from cold leg to hot leg (level swell in cold leg standpipe)
  • May need full 2.0 m available inside reactor vessel to accommodate full level swell (therefore need to increase vessel height to 13 m), although probably lower because the level swell in the control rod insertion channels doesn’t take head losses from cold leg and downcomer
  • If reactor vessel’s height increases by 1 m, control rod drive must be even longer (top must remain uncovered) (see picture)

2013-09-20 Vertical Stacking


2013-09-20 Vertical Stacking 2

Primary Coolant Routing for Defueling Chute and DHX

Meeting: Nicolas & Raluca

Top of DHX to Hot Leg

Connect top of hot leg to top (inlet) of DHX, above faulted level 1B. (because for FL2, heat removal is done through cavity wall cooling system, and salt level just needs to be sufficient to cover the fuel pebbles).

  • Faulted level 1: due to loss or removal of coolant from the hot and cold legs; at the bottom of the hot and cold leg reactor vessel penetrations
  • Faulted level 2: due to a reactor vessel break and primary coolant filling the free space in the reactor cavity, 0.6m below Faulted Level 1 (per PB-FHR-DESIGN-CALC-004-00A)
  • Faulted level “B”: due to drop in coolant temperature and volume shrinkage, Calculation: 7.2 m^3 [number needs to be updated] salt volume in the RV below Faulted Level 1. Assume salt temperature drops from 900oC to 600oC (same as the assumption in PB-FHR-DESIGN-CALC-004-00A, for dimensional change of reactor vessel in drain volume calculation). This would lead to 8% volume drop, so 0.6 m^3. The level drop of 0.6 m^3 would be 10 cm. (same cross-sectional area assumption as PB-FHR-DESIGN-CALC-004-00A)

Ensure symmetry among the three DHX to hot leg manifolds.

Note: It is important to have cooled downflow and heated upflow in the DHX.

DHX Inlet Manifold

To prevent sudden failure of the DHX when the level faults below the DHX inlet pipe, the top DHX coolant inlets can be designed such that as the level gradually drops the flow rate also gradually drops; (this might be at the expense of some loss in the effective elevation of the natural circulation loop, or might be mediated by internal detailed design of the DHX shell and tubes.) A manifold with most of the pipes at the top of the DHX and gradually fewer and fewer going down, would implement this concept.

Bottom of DHX to Downcomer

Done through penetrations in the core barrel, connecting to the bottom of the DHX shells. no need for a manifold here, just one hole sufices (although mechanical design considerations will need to be discussed here).

Top of the defueling chute to DHX

A penetration in the inner wall of the outer top lid will connect the top of the DHX to the annulus formed by the inner and outer top lids. This annulus will then be connected to the top of the two defueling chutes. The use of the annulus is important for flow redistribution because the DHXs are all located in one quadrant of the core, and the defueling chutes are located at 180oC from each other. From the annulus to the defueling chute, flow can be routed axially, or axially with a radially outwards component, depending on the axial distribution of the coolant ports connecting the defueling chutes to the annulus between the two top lids.

Flow through the defueling chute is downwards during forced circulation and upwards during natural circulation. Allowing upwards flow through the defueling chute in NC is important, to prevent formation of stagnant hot pockets at the very top of the defueling chute.

Fluidic diode functionality

The fluidic diode function needs to be implemented immediately after the dhx inlet manifold. If the check-valve option is chosen, then check valves can be put on the larger manifold lines, and some smaller manifold lines are kept without check valves, to ensure some amount of bypass flow.

We figured out that we can use the DHX bypass flow during forced circulation to provide cooling to the defueling chute. Calculations of the flow resistance in the different branches are needed to verify this, but it might be possible for all of the bypass flow to be routed through the defueling chute. This is achieved by using the two-branch check-valve concept to meet the diode functionality. The check-valve branch would connect the core to the top of the DHX. The small free bypass branch would connect the top of the defueling chute to the top of the DHX.

If it looks good, then we can ask David to sketch some of these (relatively minor) changes in SW, and from there the next steps are (1) pressure drop calculations (Alex can add these to his current set of DP calcs), and (2) COMSOL modeling of the forced and natural circulation cooling modes for the defueling chute.




Flibe Inventory

Flibe inventory will be calculated based on volumes of the FHR sub-systems and components containing the primary salt. A list of these sub-systems and components is provided here. Some volumes can be directly computed from the current FHR CAD model. When geometry/volumes are unknown, the sub-system or component has been italicized and sources for future calculations are provided where possible.


  • Active core region
  • Defueling chute

Reactor internals

  • Coolant injection channels
  • Control rod channels
  • Inner reflector bypass paths (estimate from graphite blocks geometry, thermal expansion and resulting gap size)
  • Pebble injection lines
  • Free space below pebble bed (will depend on pebble injection method as well as divider plate geometry)
  • Outer reflector bypass paths (estimate from graphite blocks geometry, thermal expansion and resulting gap size)

Cold leg

  • 2 x CTAH to drain tank (need to locate drain tank)
  • 2 x stand pipe (need to estimate level swell based on flow velocity and properly size the stand pipes)
  • 2 x drain tank to reactor cavity wall (need to locate drain tank)
  • 2 x reactor cavity wall to reactor vessel wall
  • 2 x reactor vessel to downcomer coolant distribution system
  • downcomer

Hot leg

  • Hot salt collector ring
  • Hot salt extraction (collector ring to reactor vessel wall)
  • Hot salt extraction (reactor vessel wall to reactor cavity wall)
  • Hot salt extraction (reactor cavity wall to hot salt well) (need to locate hot salt well)
  • Hot salt well (need to estimate volume; based on level swell, what else?)
  • 2 x hot salt well to CTAH (need to know relative locations of hot salt well and CTAHs)

CTAH (need to estimate CTAH tube bundle total volume)

DHX (primary side)

  • Hot leg to DHX (need to figure out pathway)
  • DHX (shell side) (need to adapt numbers from MSBR heat exchangers; look at Cristhian’s RELAP model; implement heat transfer coefficient multiplicator based on twisted tube HX performance)
  • Check valve
  • Valve to downcomer (need to figure out pathway; this will be a bypass path under normal operation)

DRACS loop

  • DHX (tube side) (need to adapt numbers from MSBR heat exchangers; look at Cristhian’s RELAP model; implement heat transfer coefficient multiplicator based on twisted tube HX performance)
  • NDHX (tube side) (need to adapt numbers from European Breeder Reactor and other metal-cooled reactors designs; look at Cristhian’s RELAP model)
  • DRACS piping (total inventory will be based on required elevation difference between the DHX and NDHX centerline and pipe size to drive natural circulation; performance will be based on a 2/3 rule and 2% nominal power extraction [or should this be 6%?], hence 1% nominal power [or 3%?] per DHX; need to determine design average and delta T between hot and cold leg of the DRACS loop, based on DHX LMTD)

Fuel handling system

  • Defueling mechanism
  • Storage space (use wet canisters? room filled with flibe? should we use some multiplication factor of the core salt inventory? Mike is working on this)

2013-08-06 Design Meeting Notes

  • Inventory of primary salt: make a list of sub-systems and estimated volumes
  • DHX: need to generate a RELAP model. For twisted tube heat transfer enhancement, use a multiplier, or perhaps other correlation? Ask Ed, read Chinese paper.

Hot and cold leg routing:

  • Hot leg comes straight through fire brick and cavity; cold leg has a 90° elbow and comes parallel to hot leg.
  • Piping is short enough that no supports are necessary inside the cavity. Pipes will be supported by whatever fitting is used when going through the cavity wall.
  • Pipe slopes for draining: cold leg drains to drain tank. Need isolation valve not to drain whole cold leg. On the hot leg, high point is the hot well with primary pumps.
  • Need to calculate friction losses to know salt elevation above downcomer from swelling.

Reactor cavity:

  • Need to design for heater rods sliding down through holes in the reactor vessel flange (2 heaters per cavity block?)
  • Cover gas: only inside reactor vessel, or also in reactor cavity?
  • Inspection of outside of reactor vessel: have a wide enough gap to be able to insert a probe. Or use heater rod channels? (remove them and insert probes in their place)
  • Rationale to use a skirt for reactor vessel support: requirement to have pipe penetrations not move a lot: their level will not move, and under thermal expansion, the rest of the vessel will change dimensions axially.
  • Need to design the insulating material in the cavity to bear some compressive load without damaging the reactor vessel.


  • Next meeting: level swell; salt volume; full pull length for control rods etc.

2013-08-02 Design Meeting Notes

Inner reflector design

  • Update: instrumentation channels should go all the way through (top to bottom).
  • In addition to tie rods in lobe corners, some of the channels will be instrumented with neutron flux measurement.
  • Bottom of buoyant control rod channels must be a blind hole, so that pressure differential on the control rods will induce proper snubbing. Need to calculate required length of blind hole below perforated region of the control rod channel (will affect rod length, therefore stacking problem).
  • Must be careful with stiffness of the channel to cope with pressure differential.
  • Consider having some kind of porous structure at the bottom of the channels to ensure that the rods stop without damaging the structure itself if the channel walls are cracked and the pressure differential is not enough to stop the rod. Need a way to monitor if structures have been damaged.
  • Have another inlet flow chanel in the thick bottom part of the reflector that merges into the control rod channels (more flow area, lower flow velocity). Refined flow areas and pressure differential calculations must be performed. Need to take into account entry losses at the bottom of the inflow channels.
  • Add small channels (1 cm diameter) for flow outlet from the control rod channels. Adjust number and spacing of these channels based on desired flow distribution (based on pebble bed dynamics). Probably 2 columns of small channels per control rod channel.
  • Tip of control rods: load with neutron poison? Amount of graphite? Need to make sure not to insert positive reactivity feedback (based on FHR core neutronics analysis)
  • Plan for the Fall: work on COMSOL coupled stress and fluid analysis (Jae? Alex?)



  • Taper entrance to the combustor (curvature radius should be 1/4 – 1/3 of line radius).
  • Secondary hot duct to vent stack should be smaller (can take higher velocity).
  • Annulus around combustor should be larger (1/4 of total area). Wall should be thin (at pressure equilibrium, takes highest temperature).
  • Diamond structures (insulation) should be thinner to be flexible.


  • Thickness of outer insulation will depend on heat losses and need to have outer wall temperature remain below regulatory limits (wall cannot be >50°C for workers protection). Based on fiberglass insulation thermal conductivity and natural convection heat transfer coefficient.

Core design

  • Block with hot leg can be larger to accommodate size of hot leg (low dose rate).
  • Hot duct liner should extend a bit down into the graphite reflector.
  • Need to analyze thermal stress transient when flow reversal through DHX.
  • Need to have small gap (1 cm?) between fixed and movable metal rings in upper core structure.
  • Flow direction in defueling chute: probably want it to be downards so that upper core structures don’t get heated up, but will make defueling more complicated (pebbles cannot be entrained in flow outside of the core).
  • Shell structure (above cold leg, below hot leg): get elevation transition closer to hot leg.
  • Need to work on detailed design of reactor vessel/fire brick type of insulation/liner/concrete.