Title: Aspect Ratio Optimization of
1Aspect Ratio Optimization of Burning Plasma
Tokamaks
Pietro Barabaschi ITER International Team
Garching Prepared for IEA Workshop on
Optimization of High-b Steady-State
Tokamaks February 14-15, 2005 General Atomics
2Introduction
- The ITER EDA.. developed
- -needed design solutions,
- -enabling technologies,
- -and knowledge base
- BUT, the Tokamak is a complex system and for its
optimisation it requires a detailed understanding
of all interplays of design drivers (and the
devil is in the details!) - We do not have the basis and criteria to design
(or even more to optimise!) a fusion reactor
today, most notably we are missing - Plasma burn demonstration
- Adequate understanding of Plasma
- Practical viability of fusion
- Beta (device optimisation)
- Reliability, Availability, Maintainability
(R.A.M.) - Materials
- Divertor power exhaust
- Higher performance structures/SC
- Until we understand and develop all these points
the cost optimisation of a reactor may not be
realistic
3Tokamak Design Machine parameters
- For a SC Tokamak, given
- -Desired Plasma performance Q, burn time,
of shots - -Plasma Boundary conditions q, ngwmax, bNmax,
k, d, HH - -Physics Criteria t, ngw, PLH, Beta
- -Engineering Criteria Stress, loads, SC
criteria, times and solutions for maintenance,
Access to Plasma (diags, HCD), Nuclear criteria,
Design solutions - Only Aspect ratio (or Peak Field in magnet) is
left free - However, allowable k and d are function of R/a
for divertor space, plasma shape and position
control. Access to plasma is function of ripple
requirements and R/a - NBIn the case of Steady State tokamaks also the
safety factor may be an optimisation parameters. - A System code is normally used to study options
combining physics rules with engineering design
knowledge. It can only be used if a more
detailed design of a particular type of machine
has been done and the experience / knowledge has
been implemented
4Example Nb3Sn SC Design Criteria
- Temperature Margin from the max predicted T at
any point to the local current sharing T. - Tmarggt1K with FP plasma
- Stability (Heat transfer to Helium)
- Hot Spot Temperaturelt150K
- 1 and 2 determine the amount of SC strand, 3
determines the amount of additional copper and
overall conductor size - Main significant system interactions with Magnet
(local cost, System size and cost, VV (stress due
to Fast Discharge)
? Strand 1 lt CunonCu lt 1.5 RRR 100 Tcom
18K Bc28T Jc (12T,4.2K,?-0.25) 650 A/mm2
5Plasma Performance Criteria
q95
- BUT.
- All 3 main scaling relationships have little real
physics basis!! Optimisation have big limitations
and uncertainties! - It is likely that theres an interplay between
H, shaping, q, nGW,
6Typical results from the System Code Study
- The main machine parameters and the cost change
with increasing aspect ratio in the following
way - The toroidal Field, the Magnetic Energy increase
with Aspect Ratio - The minor radius and the Plasma Current decrease
with Aspect Ratio - However, the cost of the machine stays constant
over most of the Aspect Ratio range investigated
7Machines with different R/a Elevation View
8Machines with different R/a TF Magnet section
9System Analysis Design Drivers
- Radial Build D ? 10cm R ? 18cm C ? 60kIUA
- Shielding (heating, damage, reweldability)
- CS Magnet
- TF Magnet
- Assembly and tolerances
- Elongation k95 ? 0.1 R ? 17cm C ? 80kIUA
- BUT (Stability, VDEs, SN-DN control, Divertor
space, flexibility) - Triangularity d95 ? 0.1 R ? 10cm C ? 100kIUA
- BUT (SN-DN control, Divertor space, Sawtooth R,
Magnet loads) - Safety Factor q95 ? 0.1 R ? 5cm C ? 50kIUA
- BUT (HH degradation, Disruptions loads, Magnet
loads) - Confinement H ? 0.1 R ? 12cm C ? 130kIUA
- (Note DC/DR not constant)
10A Power Reactor
- Where is a Power Plant quite different from an
experiment? - Additional problems
- In physics
- Beta(density, peaking),
- Steady State
- In engineering
- Remote Maintenance
- Current drive efficiency
- Reliability
- Availability
- Power exhaust
- Materials
- With some simplifications
- In physics
- Experimental Flexibility
- In engineering
- Disruptions/VDEs?
- Diagnostics
- Heating methods
- Fatigue
11Aspect Ratio related issues
- Relation between shape and density limit is far
too simplified. Effects of shaping must be
included. - Beta limit is NOT an invariant of aspect ratio
- limit f(A)
- At low A the relative distance between plasma and
wall is less (RWM) - Achievable shaping is NOT an invariant of aspect
ratio - Natural elongation increases at low A
- Available space for divertor at given d increases
at low A - Relative (and absolute) distance between plasma
and PF magnet reduces at low A. This impact shape
control (in particular in SN) as well as plasma
vertical controllability. - In steady state tokamak q95 is also a free
optimisation parameter (HHf(q95) ?) - All these effects are crucial for the
optimisation of steady state tokamak and when
included point to reduction of value of optimal A
12Important (but often forgotten) Engineering Issues
- Access to plasma for HCD (and maintenance of
internals) . In particular tangential for NBI
(ans shinethrough issues). Check carefully at
High A!! - Space available for divertor
- Thermal and EM loads on Blanket a function of A
(PF almost constant but TF not!) - Cost and replacement reqs of internals is a
function of complexity (thermal and mechanical
loads) - TF discharge parameters is important for VV
stresses (and cost) at high A. - Space for water cooling and pipe extraction
- Again all the above issues, when included, tend
to benefit low A - (note for Cu devices with limited pulse duration
e.g. FIRE High A benefits largely from
reduction of required pulse length with reduced a)
13Conclusions
- We cannot optimise a steady state tokamak today.
- First we must understand the underlying PHYSICS
and gain experience in construction, operation,
and MAINTENANCE of a large super-conductive
tokamak. - Warning Simple scaling optimisation of steady
state tokamak typically points to relatively high
A (e.g. 3.5 - 4) due to low current requirements
and high IBS fraction. This may be well be an
illusion. Real engineering and more accurate
physics basis may well reverse this conclusion.