the design challenges of nuclear tokamaks - lessons ... · the design challenges of nuclear...

The design challenges of nuclear tokamaks - lessons learned so far T N Todd

The design challenges of nuclear tokamaks- lessons learned so far

T N ToddJanuary 2018

Part 1 Nuclear issues


Nuclear considerations of DT fusion power–

– Breeding the tritium!

– Release of the tritium fuel (during processing, storage, or use in the reactor);

– Release of easily-levitated radioactive dust from plasma-facing components;

– Activation of the reactor structure aggravating maintenance procedures;

– Neutron-induced damage to, and elemental transmutation within, structural materials

– Nuclear heating and damage to diagnostics (engineering and plasma) used for machine control;

– Minimisation of radwaste


Nuclear tokamak enabling technologies

• Tritium Breeding Blankets

• Tritium fuel-cycle process plant

• Remote Handling technologies

• Diagnostics and Control systems

• Low activation alloy development


• Contributions of nuclear tokamaks considered so far– TFTR

• Coolant activation

• Remote handling

– JET• Tritium compatibility

• Activation

• Remote Handling successes & problems

– ITER • Neutron streaming, port design

• Activation of structures

• Activation of coolant, HX location

• Civils require shielding specifications

• Nuclear heating

• Diagnostics issues

- DEMO• Tritium breeding

• Diagnostics problems

• Fast Remote Handling

• Material damage

• Material selection

• Radwaste control


Comparison of radiological impacts on JET-ITER-DEMO First Wall (approximate!)

*About 5x the neutron fluence due to virgin 14MeV neutrons alone, due to geometric “Sec(θ)” effects and scattering in the blanket

In epoxy, polyimide, glass, carbon & alumina, there is ~(0.1-1.0)x10-15 Gy/(n/m2)

Radiation problems rise with the power and life-time of the machine

Fusion radiation in operations

Site TInventory, grams

AverageMWyr/m2

Peak n/m2 >0.1MeV*

dpa He appm

Operating Gy/sec

Shut-down Gy/hr

JET 1997

20 10-7 1x1019 10-6 10-5 n: 40Ɣ: 100

0.01

ITERall life

4000 0.3 4x1025 2 20 n: 200Ɣ: 500

500

DEMO3 FPY

6000? 6 8x1026 50 500 n: 300Ɣ: 500

10,000


The main fusion reactions yield high energy products:

http://www.kayelaby.npl.co.uk/atomic_and_nuclear_physics/4_7/4_7_4.html

D + D 50% → T (1.0MeV) + p (3.0MeV)

D + D 50% → 3He (0.8MeV) + n (2.4MeV)

D + T → 4He (3.6MeV) + n (14.0MeV)

Fusion reaction basics


Apart from 6Li, fusion blanket reactions require those fast neutrons:

http://www.kayelaby.npl.co.uk/atomic_and_nuclear_physics/4_7/4_7_4.html

3He(n,p)3

H

http://www.oecd-nea.org/janis/book/

3He from DD reactions and T decay is usually neglected but is a strong thermal neutron absorber

Fusion reaction cross-sections


Tritium Recycling at JET: the Plant

T2

T2

HTO Export

•

Twice damaged by HF attack due to

trace quantitie

s of F in

input s

tream


Tritium Recycling at JET: the History

In seven years, total injectedinto torus 105g, total lost fromstacks only 0.04g, a few % of the authorised discharge limit.

In all JET operation, the DD reactions have made ~3.1020 neutrons and ~2mg of tritium.

Since this figure was produced, a tritium half-life has gone by – and another 50g has recently been bought


● Tritium is a low-energy beta emitter, with a half-life of 12.3 years

● Irradiation by such low energy betas is not a biohazard (blocked by the dead outer layers of skin, or a few mm of air), but the tritium is readily absorbed by breathing the gaseous or aqueous forms and then it moves around the body and “irradiates you from the inside”, until lost with a ~14 day metabolic half-life.

Tritium is a significant radiological hazard


Tritium precautions at JET – some examples

● Flourine-bearing compounds decompose to hydrofluoric acid which attacks parts of the tritium recycling plant, hence certain types of insulation are banned (and leaks from the TF and divertor coil cooling systems can be troublesome)

● The torus hall and basement walls and the floor between them have large penetrations for cables, transmission lines and utilities – all must be labyrinthine or shielded to inhibit neutron streaming, and sealed with proprietary demountable blocks etc. to prevent atmospheric circulation

● The tokamak complex is held at a negative pressure and exhausted to a monitored stack.

● Long-term discharges are also checked with small vegetable plots and rain-water collectors scattered around the periphery of the site, routinely assayed for tritium content

● Tritium readily exchanges with ordinary hydrogen in polymers etc. so:

- Painted surfaces are undesirable in the torus hall (radwaste issues)

- Elastomer seals are “transparent” to tritium over timescales like hours

- Plastic bags and boxes etc. do not provide barriers to tritium



● Tritium readily exchanges with ordinary hydrogen in polymers etc. so:

- Cover-gloves worn over cotton and surgical gloves during BA suit operations in the torus need to be changed every 15-20 minutes



● All vacuum vessel boundaries must be continuously welded or double-sealed with metal gaskets – no elastomer seals

● All vacuum pumping lines “ditto”, including the roughing and exhaust lines

● Diagnostics windows must be double-window geometry with a 0.5bar neon filled, monitored interspace

● Waveguide and RF transmission line “windows” must be double, with the interspace pumped.

● Everything potentially connected to the torus must be evacuated to the Active Gas Handling System building (J25)

● Tritium is stored on uranium beds in J25 and only released in small batches to plena (in the tokamak building) used for fuelling the tokamak or neutral beams

● The Regulator tries to reduce annual discharge limits to slightly above best trends, but maintenance work needs much higher limits (as initially agreed)


DT reactors will require tritium breeding

• Readily available tritium supplies are very limited

• High power, long pulse DT facilities must breed their own tritium


Reduction in TBR with added realism

Bar chart showing the reduction of TBR from 1.8 to 1.04 upon including the internals of the DCLL blanket and its surrounding components.

Bar chart showing the reduction of TBR from 1.8 to 1.04 upon including the internals of the DCLL blanket and its surrounding components.L. A. EL-GUEBALY, et al., FUSION SCIENCE AND TECHNOLOGY, VOL. 61, 2012.

The design challenges of nuclear tokamaks - lessons learned so far T N Todd16

TBR: effects of divertor option and FW cladding

3D model (First wall materials: Eurofer, Tungsten)

FW thickness: 0.2cm, 1cm, 2cm

Zero / single / double divertor(s)

Tritium Breeding Ratio (0/ 1 / 2 divertors)

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

0 0.5 1 1.5 2 2.5

FW Thickness (cm)

TB

R

no divertor - eurofer

no divertor - tungsten

one divertor - eurofer

one divertor - tungsten

double divertor - eurofer

double divertor - tungsten

S. Zheng, personal communication, CCFE Physics & Technology talk, 5th July 2013


Blanket designs:Engineering complexity

First Wall(F82H)

Tritium BreederLi2TiO3, Li2O

Coolant water (25MPa, 280/510oC)

Neutron MultiplierBe, Be12Ti

Surface Heat Flux:1 MW/m2

Neutron Wall Load: 5 MW/m2

W coating for FW protection

LiPb inlet

LiPb outlet

He manifolds

4He-cooled Lithium-Lead breeder (CEA - France)

Water-cooled Pebble-bedbreeder (Japan)

Not re

levant to ARC all-li

quid blanket concepts!! (

You hope!)


Breeding blanket design – DD start-up?● What is the “breeding” capability of a blanket designed for 14MeV DT neutrons if

only 2.45MeV DD neutrons are available, so the Li7, Be and Pb do nothing?

Zheng, S. et al, Fusion reactor start-up without an external tritium source, FED 103 (2016) p13-20

Trit

ons

crea

ted

per

plas

ma

neut

ron

Typical factor ~0.62

● So if the T that is created is fed back into the fuelling stream, it is possible to start up a DT reactor on pure DD fuelling, i.e. with no requirement for buying any start-up T (~€300M for 6kg of T) !!

T plant recycle time (hours)

24105

3.5

(TBR=1.2, burn-up fraction = 5%)

See also Konishi, S. et al, “Myth of initial loading tritium: modelling DEMO fuel system in power ascension tests”, 29th SOFT, Prague, Sept 2016


The seven key reasons why DD start-up works so well are:

DD TPR is ~0.62 of the DT TBR (for many blanket types checked);

start-up and current drive power alone will achieve the full nominal DT plasma temperature and density if P-rad/P-plasma ~15%,not ~95% as at full DT power;

the DD rate coefficient is ~80 times lower than DT, but for each branch, and both are productive for T;

the DD reactions burn no T whereas the DT ones do, so compare (1+TPR)/2 ~0.85 (averaging the two branches) for DD, to (TBR-1) ~0.15 for DT;

the pure DD reaction rate is proportional to ne

2/2 whereas the optimal DT

reaction rate is proportional to ne

2/4;

the expected site inventory limit for tritium and the nominal DT burn-up rate are so low that the T plant has to be able to reinject the bred T within ~1 day to respect that inventory, so bred T can almost continuously be added to the DD fuelling, greatly accelerating the T inventory rise.

nuclear regulators require a progressive “Power Escalation” taking at least several months from first nuclear operation to full power anyway.

Breeding blanket design – DD start-up


Fusion reactors have many more neutrons >10MeV

Fission & fusion reactor neutron flux spectra


The fast neutrons create more transmutations and hence gas

~10x more H is produced than He, but it more readily diffuses out

M Gilbert et al, “Scoping of material damage with FISPACT-II and different nuclear data libraries: transmutation, activation, and PKAs”, TM on Nuclear Reaction Data and Uncertainties for Radiation Damage IAEA Headquarters, Vienna, Austria, 13-16 June, 2016

Fission & fusion reactor neutron flux spectra


“vital” for steels “useful” for steels Acceptable for a fusion reactor

DEMO Materials Environmental basis(III) Structural Materials(courtesy Dr Robin Forrest – CCFE, IAEA)

Divertor: plasma -facing

SiC composites ?


Environmental issues: “Reduced activation” steels

There are also “Oxide Dispersion Strengthened” (ODS) variants -

In these, nanoscale Y2O3 particles:

act as He, H sinks and reduce defect rate,

strengthen the alloy and reduce creep.

Currently only small ‘experimental’ batches have been made.

‘Reduced Activation Ferritic-Martensitic’ steels are under development: Ta replaces Nb, V replaces Ti W or V replaces Mo

Cr replaces Mn … up to a point. Avoid Ni, Cu, N

RAFM steels will be ‘cool’ enough for simple recycling and re-use after ~ 50-100 years storage (after ~ 5 years service in the reactor first wall).


DEMO Structural materials:swelling of steels

Tirr = 400°- 500ºC

Body-centred-cubic (BCC) alloys (e.g. Ferritic steels, Vanadium alloys) exhibit the lowest swelling.

BCC alloys therefore are first option for a Fusion reactor structure.

BCC materials are however subject to radiation embrittlement – difficult operational scenarios.

However ferritic steels are magnetic: Eurofer μR~50, F82H μR~15


DEMO Structural steels: Embrittlement of RAFM steels

0 10 20 30 40 50 60 70

0

50

100

150

200

250

300

KLST DBTT (FZK, NRG) ISO-V DBTT (SCK)

Tirr =300-330°C

D

BT

T (

°C)

Dose (dpa)

Fusion relevant RAFM steels have good long-term stability, but:– become brittle and radiation hardened if irradiated at temperatures ≤ 300 - 350ºC

A Ductile Brittle Transition Temp (DBTT) ~ 150° - 200ºC is unusable for a reactor – DEMO may operate at above 300°C but thermal stresses on cool-down for a shutdown would potentially crack the vessel.


DEMO Structural steels:Evolution of High Temperature FM steels outside fusion

Source – Steve Zinkle, ORNL

• But there is no “reduced activation” variant yet


DEMO Structural Materials issues Example of Fe lattice damage caused by one “cascade” induced by a 150keV Fe ion recoiling from a single neutron impact.

Size of simulation cell: 475 Å; 6.75 million atoms

Nearly all the displacements are Frenkel Pairs (vacancy + interstitial atom) which recombine rapidly (faster with higher temperature), but still hundreds remain, together with more complex crystal defects.

K. Nordlund, TEKES –

University of Helsinki: December 2012


DEMO Structural Materials issues

The nature of high-energy radiation damage in iron, E Zarkadoula et al, Journal of Physics: Condensed Matter, Volume 25, Number 12

Example of Fe lattice damage caused by a 500keV Fe ion recoiling from a single neutron impact. Initial temperature 300K.

Duration of video ~20ps.

Nearly all the displacements are Frenkel Pairs (vacancy + interstitial atom) which recombine rapidly (faster with higher temperature), but still hundreds remain, together with more complex crystal defects.


DEMO Structural materials:Could we use copper alloys?

Copper alloys have extremely high thermal conductivity, but are rapidly degraded in conductivity and ductility by neutron irradiation, as with steels dependent on dpa rather than the neutron spectrum.

Copper alloys will probably be used in the “Early DEMO” divertor but could not survive in a high neutron fluence reactor.


Reweldability of Stainless steel 316L(N)-IGK. Asano, J. van derLaan, MAR, 2001V

The challenge of transmutation helium


Structural activation is largely sensitive to thermal neutrons:


Structural activation cross-sections


Structural activation can be very strongly dependent on trace impurities

DT machine port contact dose rate

projection with normal steel cobalt

content (0.1%)

...or with fission- grade stainless steel

cobalt content (0.02%)


DEMO Passive safetyExample of reactor simulation from EU study [5]

• Accident scenario:

– complete loss of coolant (decay heat removed by passive conduction);

– no active safety system (passive safety expansion volume only);

– no intervention for many weeks;

– tritium and dust in-vessel inventories assumed mobilised and available to permeate/leak through containment barriers.

‘Model A’ max temperature

Time histories

Te

mp

era

ture

(°C

)

Time (days)• No melting of the structure

• Radioactivity release 1-18 mSv (cf 5 mSv pa EU background – no evacuation)


Fusion reactors need superconducting magnets, which require minimal neutron heating and neutron damage/transmutation.

Also, the superconductor has a copper “stabiliser” jacket which has to be repeatedly reannealed (to ~20°C) to recover its conductivity (mostly!)

Superconductorcritical current

Copperresistivity

WEBER, H. W., “RADIATION EFFECTS ON SUPERCONDUCTING FUSION MAGNET COMPONENTS,” Journal of Modern Physics E Vol. 20, No. 6 (2011) 1325–1378

Neutron damage effects


Yamada, H., “NEUTRON-INDUCED HELIUM IMPLANTATION IN HELIUM COOLANT PIPES OF FUSION REACTORS”, Journal of Nuclear Materials 103 &104 (1981), p 615-618

Neutron spectrum cut-off:

Fusion neutron wall loading 1MW/m2, He gas at 100bar, 550°C, SS316 pipe wall:

Blistering was predicted at ~1018/cm2, after ~3FPY

Slo

win

g by

gas

co

llisi

ons

Neutron damage effects – helium recoil in gas coolant


Another source of neutrons is 17N (from 17O(n,p)17N in the reactor coolant,

17O being ~0.038% of natural O).

The 4.17s half-life can allow the 17N to emerge from the reactor hall

17N neutron spectrum

Neutrons

17O(n,p)17N cross-section


Even very low fluxes of neutrons impact electronics in detectors, processing electronics and digital logic in personnel access and machine control systems

E.g. ~6 “Single Event Upsets”/hr from only 0.4n/cm2-sec above 1MeV in a UAV control system [NASA, memory area ~6cm2] http://www.nasa.gov/sites/default/files/files/SMIII_Problem6.pdf

They create ionisation by recoil nucleus motion and by activation decay products

The recoil motion also damages the semiconductor crystal structures (with <14nm scales now appearing in chip design)

http

s://e

n.w

ikip

edia

.org

/wik

i/13 0

_nan

omet

er

Mes

seng

er, G

. & A

sh, M

., T

he e

ffect

s of

rad

iatio

n on

ele

ctro

nic

syst

ems,

Van

Nos

tran

d R

einh

old

Com

pany

Inc,

198

6

10 Si lattice constants

Neutron damage effects


Another source of gammas is 16N (from 16O(n,p)16N in the reactor coolant).

The 7.13s half-life can allow the 16N to emerge from the reactor hall, potentially affecting electronics and operations staff

16N gamma spectrum

The 16N Calibration Source for the Sudbury Neutrino ObservatoryM. R. Dragowsky et al; http://arxiv.org/pdf/nucl-ex/0109011.pdf

6.1MeV

7.1MeV


Gamma rays

16O(n,p)16N cross-section


Another source of gammas is 19O (from 19F(n,p)19O in the FLiBe).

The 26.5s half-life can allow the 19O to emerge from the reactor hall, potentially affecting electronics and operations staff.

19O decay scheme


Gamma rays


Gamma radiation lifts electrons from closed shells into conduction band, making insulators become conductive (a flux effect, not fluence)

Gamma rays


Fusion reactor plasmas require sophisticated diagnostics

…but there are many challenges... Non-nuclear (as in existing machines) Nuclear


Fusion reactor plasmas require sophisticated diagnostics

…but there are many challenges...


The fusion products create “radiation capture” reactions with other ions in the plasma, producing characteristic gammas useful for core plasma diagnosis.

E.g. D/T ratio from ratio of t(p,γ)4He to d(t,γ)5He (by gamma spectroscopy)

Ste

ady

Sta

te -

flatVarious n,T profiles

Stead

y Sta

te -

peak

ed

Eγ~17MeV

Eγ~20MeV

Kiptily, V. G., “On the core deuterium–tritium fuel ratio andtemperature measurements in DEMO”, Nucl. Fusion 55 (2015) 023008 (7pp)

Diagnostics – some useful effects


“Alpha power dominates power input to plasma”

Burn Control – a simplified picture

Alphas drive H(r)

H(r) drives p(r)

p(r) drives jBS(r)

jBS(r) drives q(r)

q(r) + p(r) drive MHD

MHD alters transport

Then add divertor heat flux control (e.g. impurity seeding) and its effects on the main plasma!

Transport and H(r) set p(r)

q(r) controls banana orbits

Banana orbits set H(r) shape

p(r) drives alpha birth rate and its profile

All types!!


Neutrons and gamma rays are very bad for close-in plasma diagnostics!

Both types of radiation are ionising and therefore generate conductivity flux (Radiation Induced Conductivity)...

...and radioscintillation in optical media (lenses, fibre-optics, detector windows)They create heating sufficient to require active cooling provisions, even in ITER

The neutrons create displacements per atom (dpa), which:– Create defects in optical components (lenses, fibre-optics), hence darkening– Derange nano-structures in detectors and electronics– Precipitate metals in ceramics (Radiation Induced Electrical Degradation)– Create trapping sites for helium (hence swelling)– Embrittle ductile conductors

The neutrons activate and transmute atoms in the sensors and cables etc., which:– Creates thermocouples in same-alloy connectors (Radiation Induced EMF)– Alters doping in sensors and electronics

Neutrons and gamma rays


Fusion requires very energetic fuel species nuclei (~50keV, ~500M°C) and generates much faster neutrons than fission (~3MeV - 14MeV)

The harder neutron spectrum permits a greater range of transmutation reactions and activation, of interest in minimising radwaste, and it also creates ~30x more He appm/dpa in the surrounding structure, bad for swelling, rewelding, embrittlement, conductor resistivity...

...and it is more aggressive in activating any oxygen isotopes in the reactor coolant, producing short-life radioactive 16N and 17N which leave the reactor hall with the coolant and require flow delay features and shielding to protect staff and electronics

But the fast neutrons are necessary to achieve neutron multiplication and hence tritium breeding ratio >1, vital for the reactor function

Fusion has no long-lived radioactive daughter products but the activation of the structure makes low activation materials attractive, to reduce radwaste and decay heat

The neutrons have to be well shielded from the nearby superconducting magnets, to prevent degradation of the superconductor and excessive resistance rise in the associated copper stabiliser jacket

Other radiation effects


Fusion also creates very energetic charged products, important for heating the plasma but not significant as a radiation hazard

These fast reaction products (p, d, t, 3He, 4He) can be lost from the plasma if born in unfavourable orbits, creating some damage to the first wall, but the loss flux is measurable and informative for the machine operation

The fast products can undergo secondary reactions with characteristic gamma ray emission, also useful for plasma diagnosis

Beta emission is of most significance as radiation from the tritium fuel, which drives most of the reactor safety case analyses.

Beta emission in the activated structure contributes to decay heat but gamma radiation dominates Health Physics for maintenance and radwaste issues

The energetic neutron and gamma radiation is extremely bad for most plasma diagnostic systems, but can be characterised in spectrum and emission profile to assist machine control

Other radiation effects


o TFTR pioneered high power DT operation in controlled fusion research and lessons regarding activation and RH compatibility can be noted.

o JET is not licensed by the NII but has radiological limits agreed with the EA and a very comprehensive RH experience in the MCF fusion context

o ITER will demonstrate many aspects of a high-gain, nuclear-licensed tokamak, but has a very much lower gain, neutron fluence and tritium requirement than DEMO.

o DEMO will have all the same non-nuclear requirements on its load assembly and diagnostics as any presently operating tokamak of considerable size and complexity.

o It will also have a number of nuclear requirements and constraints, currently little recognised by the wider fusion community, but becoming standard thinking for the ITER team.

o Low activation materials, resistance to neutron damage and erosion of the first wall and divertor represent key materials problems.

o A traditional belief that many kg of T had to be made available for each new DT reactor to start up seems to be wrong – DD start-up appears to be entirely feasible..

o Remote handling is key to achieving adequate availability and together with overall plasma heating system efficiency is vital for economic viability.

Conclusions


END


Effect of Li6 enrichment on TBR

Variation of TBR with Li6 content for 10mm first wall cladding in W or Eurofer

D-T (HCPB - one divertor)

1.1

1.2

1.3

1.4

1.5

0 10 20 30 40 50 60 70 80 90 100

Li6 enrichment (%)

TB

R

FW-W(1cm)

FW-Eurofer(2cm)

S. Zheng, personal communication, CCFE Physics & Technology talk, 5th July 2013


(For EPR)

???

Fission reactor start-up follows IAEA guidelines – no sudden jump to full power


Fission reactor nuclear commissioning timescales(data from World Nuclear News, http://www.world-nuclear-news.org/sectionhub.aspx?fid=798)

0 5 10 15 20 25 30 35 40 450

5

10

15

20

25

30

21st century fission reactor escalation durations

Time (criticality to full power), months

Run

ning

to

tal o

f re

act

ors

First Of A Kind, or if things go wrong...

Typically four or five months


Neutron shielding by concrete


Fusion reaction products are not radioactive, but the reactor structure (steel, tungsten...) becomes activated, as in fission plant

Fission and fusion radwaste decay

the design challenges of nuclear tokamaks - lessons ... · the design challenges of nuclear...

Documents