First Wall Heat Loads - PowerPoint Presentation

Slide1

First Wall Heat Loads

Mike Ulrickson

November 15, 2014Slide2

Outline

Introduction

Plasma Scrape-off Layer Profile

Implications of First Wall Heat LoadsConclusions

2Slide3

Introduction

The ITER Blanket System successfully passed final design review in April 2013

The implications of the ITER design parameters and constraints must be taken into account for DEMO type machine designs.

The lessons learned include:

The far Scrape-off layer power profile due to ELMs is a very strong driver of the design of the first wall of a blanket system and a potential major problem for reactor design.

Electromagnetic forces are a strong design driver when coupled with the heat loads on the First Wall.

3Slide4

A Comparison of FW Requirements

During the EDA on ITER

Plasma power flows only to the divertor.

The first wall sees only radiation and charge exchange (no parallel conduction).

There is no halo current flow to the FW

No FW shaping is required.

During FDR on ITER (>2004)

ELM transport causes a long tail on power transport in the plasma edge.

From 1-5 MW/m

2

heat loads are expected on the FW due to parallel conduction.

During disruptions halo currents flow to the FW.

Strong shaping is required.

4Slide5

A Comparison of the Edge Power Profile (EDA and FDR)

5

First wall region

 Slide6

Consequences of Parallel Power Flow

Field lines are incident on the FW at a shallow angle

The FW must be precisely aligned with the toroidal field

Any location where field lines could intercept a surface that is nearly normal to the field must be protected by FW shaping (recessed below the magnetic horizon).

This reduces the effective area of the FW and increases the peak heat flux but avoids a disastrous X10 heat flux.

Halo current flow makes the fingers on FW panels run in the toroidal direction.

6Slide7

Regions Of High FW Heat FluxTop of the machine due to the second x-point (particularly difficult region to shape FW because of poloidal field null). This region has the lowest neutron wall load. Perhaps 25% of poloidal circumference.

Inner and outer mid-plane due to plasma startup and ramp down (perhaps only one or the other). These regions have the highest neutron wall load and are most critical for breeding. Size depends on operating scenarios.

7Slide8

A Schematic Section of the Final FW

8Slide9

Consequences of the FW Changes

Copper must be used in the heat transfer layer under the Be plasma facing surface because of the high heat flux.

Even with extensive slitting of the FW, large moments and forces are produced in the FW during disruptions.

Thick structural members are required to bear the loads imposed.

The overall thickness of the plasma facing portion of the FW is about 60-80 mm.

A structural beam is added on the SB side of the FW to resist radial torque (~100 mm thick).

Some regions of the SB are very thin due to shaping and structural considerations.

9Slide10

Implications of High Heat Flux FW

In order to determine the effect of 1-5 MW/m

2

on the FW for future devices, we performed some 1D steady-state thermal analysis for different material choices

We assumed:

A Be plasma facing layer

70 C coolant inlet

A heat transfer coefficient of 4.4 w/cm

2

K at the coolant solid interface

We adjusted the thickness of the heat transfer layer and Be tile to keep the surface temperature of Be equal to half the melting point.

The assessment was done for both 5 and 2 MW/m

2 surface heat flux.

10Slide11

Maximum Thickness for 5 MW/m

2

11

PFM

HS

coolantSlide12

Implications of 5 MW/m

2

The minimum thermal conductivity for the heat removal layer is 0.8 to 1.0 W/cm K.

Slightly higher conductivity is needed for He cooling because of higher pressure and more complicated heat removal structures.

For liquid metal coolants corrosion must be taken into account.

The set of acceptable materials include:

Copper alloys

Molybdenum alloys

Tungsten alloys

Aluminum alloys

12Slide13

Implication of 2 MW/m

2

The minimum acceptable thermal conductivity is about 0.5 W/cm K

The first wall can be about 3 times thicker than for 5 MW/m2

.

Detailed design must be done to determine if materials with marginally acceptable thermal conductivity (RAFS, V alloys) would be usable.

The first wall support structure will still be thick due to halo and eddy current loads.

13Slide14

Implications for a Reactor

For FW heat flux in the range of 1-5 MW/m

2

due to plasma conduction approximately 20% of the blanket volume will not be breeding material (this is in addition to volume lost to coolant manifolds, in vessel coils, and blanket supports).

This is due to plasma contact and the associated disruption induced mechanical loads.

The FW both absorbs neutrons and reduces the maximum energy of the neutrons (reduced breeding efficiency).

Some tritium breeding may be recovered by increasing the blanket thickness.

Tritium Breeding ratio is in danger of being less than one.

14Slide15

Effect of Transient Heat Loads on FW

Very thin first wall structures have very small margin for high transient heat loads

The consequences are melting, stress failure of the joint between plasma facing surface and heat removal layer.

Disruptions are often accompanied by run-away electron beams.

The predicted run-away electron beams on ITER are sufficient to cause melting of the plasma facing surface of the FW

Run-aways may even affect FW panels when there is no plasma contact because of relativistic orbit shifts.

15Slide16

Conclusions

Plasma conduction along field lines to the FW will require a major change in breeding blanket design.

Electromagnetic forces during disruptions and the related need to perform remote maintenance argue for smaller blanket modules

This means volume must be reserved for coolant manifolds (and possibly internal control coils).

The space reserved for blankets must be increased by 20-30% compared to the typical values for machine planning

.

16