Presentation is loading. Please wait.

Presentation is loading. Please wait.

Path to a High-Availability MFE Power Plant John Sheffield ISSE—U.Tennessee-Knoxville.

Similar presentations

Presentation on theme: "Path to a High-Availability MFE Power Plant John Sheffield ISSE—U.Tennessee-Knoxville."— Presentation transcript:

1 Path to a High-Availability MFE Power Plant John Sheffield ISSE—U.Tennessee-Knoxville

2 Questions to be answered on the path to the DEMO-1 What is the product, electricity, hydrogen, support of fission…? What is the physics basis? What is known and what is yet to be confirmed, and how will we confirm it? What is the expected plant availability (capacity factor). The basis for this estimate: scheduled operational time for each component, and its expected replacement time? How is unavailability assigned, the mean time to failure, the meantime to repair or replace for each component? What is presently unknown and how will we obtain the data required to confirm the assumptions?

3 Questions to be answered on the path to the DEMO-2 What is the basis for component cost estimates and how will we confirm these estimates? What materials will we use and from where do we expect to obtain the data base for performance at the required heat, plasma, and neutron fluxes and fluences—steady state or repetitively pulsed as appropriate? How long is the expected development path? What characteristics make this approach the best (one of the best) to achieving commercial fusion energy?

4 This Talk Looks At AVAILABILITY: What is required for each component of a DEMO and where will the data base come from? STAGING: Can an FNF or Pilot Plant be staged? - Pulsed and steady state operation? - Varying impurity level?

5 Cost of Electricity - COE COE = (C cap F CR + C fuel + C OM )/(P enet x 8760 (hrs) x f cap ) + Decom. C cap = Construction costs including interest charges during construction. F CR = Fixed charge rate. C fuel = Fuel costs –includes lithium blanket replacement. C OM = Operations and Maintenance – likely higher than fission because of additional complexity. P enet = Net electric power. f cap = Capacity factor. Decom. = Annual decommissioning charge. Capacity factor is the least certain parameter.

6 Availability -1 The availability of a system depends upon the failure rate of its components and the time to repair or replace them. When the availability (Fav ) is high e.g., > 0.80, a simple formula may be used to describe it. Fav ≈ (1 - Sch)(1 -  i N i MTTR / MTFB) Where Sch is the fraction of time scheduled for maintenance. N i is the number of components of subsystem ( i ). MTTR is the mean time to repair or replace subsystem (i). MTFB is the mean time to failure of subsystem (i). It is assumed that the probability of component failure is negligible during a downtime. I am not distinguishing here between availability and capacity factor—the ratio of the integral of power over time divided by the full power times time.

7 Availability - 2 For this example I will assume that Fav = 0.8 and Sch = 0.05 –> 0.10. Then  i N i MTTR / MTFB = 0.158 → 0.111. For 10 systems each with 10 components and if the mean time to repair or replace were 1 month, then we would require a mean time to failure of each component 53 → 75 years. If we argue we deal with only the 10 systems then it would be 5 → 8 years per system. Again, the question is simple, Whatever availability is required, where’s the data going to come from? Be realistic. Note what happened with Fission!




11 Essential DEMO Design Elements Design for Remote Handling. Make it as Simple as Possible. Use Redundancy. If you have to make it Bigger to Reduce Stresses and/or Provide Better Access, Accept it.

12 Issue of Replacing a TF Coil I am not convinced that it will be possible to replace an SC toroidal coil on a reactor in any reasonable timescale i.e.,  1 year. Maybe it would be a good idea to view the TF coil set like a reactor pressure vessel. It can be replaced, but only every 20 years or so.

13 Considerations What part of an SC TF coil set would fail? What is the data base? Unless there is a fundamental design problem, as at CERN, do we expect a coil to fail after the shakedown phase? An SC coil does not suffer corrosion problems like a water-cooled copper coil. Pulsed field and radiation damage could be a problem. This argues for steady state operation and hefty shielding. Current feeds could be main area for concern, but could be made accessible.

14 Testing Conditions High Power Density 14 MeV neutrons Hot Walls High Duty Factor or Steady State DEMO-level of Fluence

15 Priorities, Gaps and Opportunities: Towards A Long-Range Strategic Plan For Magnetic Fusion Energy, FESAC October 2007. Top Level DEMO Goals Integration and Scalability to a Commercial Power Plant: 1. Use the physics and technology anticipated for the first generation of commercial power plants as an integrated system 2. Be of sufficient size for confident scalability (>50%-75% of commercial). Reliability 3. Demonstrate robotic or remote maintenance of fusion core. 4. Demonstrate routine operation with minimum number of unscheduled shutdowns per year. 5. Ultimately achieve an availability > 50% and extrapolate to commercially desired levels. Safety and Environmental Impact: 6. Not require an evacuation plan. 7. Generate only low-level waste. 8. Not disturb the public’s day-to-day activities. 9. Not expose workers to a higher risk than other power plants. 10. Demonstrate a closed tritium fuel cycle. Economics: 11. Demonstrate that the cost of electricity from a commercial fusion power plant will likely be competitive.


17 ITER Contributions I find it surprising that ITER was given low ratings for Fuel Cycle, Safety and Maintainability. But it may be because of low duty cycle and limited power flux and fluence. I believe that the key issue for ITER will be whether it runs reliably, once activated. Having effective remote handling will be crucial to achieving its goals. For this reason, I would give it a 2 rating on maintainability.

18 Other Sources of Data Today: Small experiments  Alc-CMod, Asdex-U, DIII-D etc. D-T operations  JET Steady State  East, K-Star, Tore Supra Stellarators  LHD, WVII-X Future: Not D-T but ITER-like  JT-60SA D-T, 400s – Steady State  ITER Missing: Simultaneous high power density, steady state, hot walls (only EAST?), with and without D-T. Ultimate configuration?

19 Some Candidate Facilities A high power, hot wall facility capable of testing walls and divertors with various levels of impurities e.g., NHTX, Vulcan. A D-T burning, high power density, high duty factor tokamak (ST) or stellarator? e.g., VNS or Pilot Plant. An interesting question is whether the main testing could be undertaken in stages in one (or two) facilities. Either way, I believe that all of the above facilities—and more?--are necessary to complement ITER. Just look at how many facilities were used for ITER R&D.

20 Should we allow for Pulsed Operation? How much should one rely on non-inductive start-up? Non-inductive start-up in ARIES-AT is estimated to take 1.5 hours. On the other hand, ITER takes about 3-5 minutes to raise the current and fusion power. The likelihood is that a Pilot Plant would have an intermediate start- up time e.g. 10 to 20 minutes, and need to provide some or a lot of the initial volt-seconds (~ 1.5 LI) inductively. Is fully non-inductive operation possible with acceptable current drive power? Issue of operation of a Pilot Plant at lower  N where the bootstrap current and non-inductive current drive do not provide all the current. This suggests to me that there need to be volt-seconds available for a partially inductively-driven flat-top—of a few hours?

21 What pulse length and down-time? What pulse lengths should be available when bootstrap current and non- inductive current drive cannot provide all the current, e.g. at intermediate  N levels? A study at Argonne, Ehst, D.A., Jardin, S., and Kessel, C. “Starlite,” ANL/FPP/TM-284, 1995 looked at the consequences of repetitively pulsed operation for a tokamak reactor. They concluded that pulsed operation would be acceptable if there were a massive heat storage system available to maintain acceptable temperature gradients in the blanket during the down time (I recollect that their study assumed that it would be desirable to restrict gradients to ~ 50 O C across the blanket).Kessel, C. How many cycles the plant should experience in its lifetime. For example,15- 30 minutes down might have 5-10 minutes to shut down and reload the transformer and 10-20 minutes to start-up. Incidentally, if the system has a double null divertor it might be simpler to reverse the current on successive shots. Pulse length > 2 hours might be a good idea ~ 10 pulses/day, ~ 1000/year.

22 Some Key Points The following areas should be explored: - Maximum permissible current density/field in the solenoid. - Extent one can rely on non-inductive start-up. - Issue of whether non-inductive flat top is possible or too restrictive on operating space. - Operation with less than full non-inductive current drive. - Necessary down time and limit on number of cycles to keep stresses at acceptable level.

23 Staged Pilot Plant Simple analytic formulae make it easy to understand how a Pilot Plant might have staged operation. For example, initial operation could be under ITER-like conditions. Ultimate operation could be under ARIES-AT-like conditions, if the physics holds up.

24 Numerous studies made of adding impurities at the plasma edge to increase radiation losses (plasma mantle) and alleviate plasma material interaction problems [e.g., Gibson 1978, Watkins 1981]. The ARIES-AT tokamak [Najmabadi et al] is a notable example of a power plant in which impurities were added to the core to reduce the conduction losses to the separatrix and spread the radiated power more benignly over the wall. A design study, such as ARIES-AT, may be based on assumptions that aggressive physics parameters will be obtained. But experiments designed to test PMI issues will have to be more conservative to give assurance that testing at interesting parameters can be done in the presence of a large fraction of power radiated by impurities. This means conservative values for H H and balance between inductive and non-inductive current drive.

25 Where and how does one gain by adding core impurities? The answer lies in the example given by ARIES-AT, where it was desirable to shed the fusion power as radiation and spread it over the wall, rather than having it conducted across the separatrix. The approach was possible in ARIES-AT but not in ITER because ARIES assumed that H H (1.4x) and  N (2.4x) could be larger.

26 EXAMPLE In a given tokamak at a given Q, H H   N 0.38 [1 + 5/Q - f R ] -0.31, where f R is the fraction of core alpha power radiated. Consequently, everything else being equal, H H has to increase when the fraction of radiated power is increased. A comparison of ITER and ARIES-AT is in the table with rough figures to illustrate the point.

27 ParameterITERARIES-AT NN 2.25.4 Q1040 fRfR 0.20.5 H 1.01.4

28 Increase tokamak size a small amount One can always make a tokamak big enough that there will be more than enough power to satisfy conduction losses, because the condition for the core power to exceed conduction losses is given by (ignoring direct alpha losses, and with fixed plasma shape). (P  + P a –P R )/ P  = [1 + 5/Q - f R ] ≥ Const.  N 1.23 /(R 4 B 2.16 H H 3.23 ). For example, in a power plant with Q = 40, a change from f R = 0.2 to 0.5, may be accommodated with an increase in R of a factor of about 1.1. (Note that, with a fixed field on the toroidal coil and fixed coil-plasma separation, B will also increase with an increase in R.) Everything else being equal, the power flux to the wall will increase by x1.1. But the conducted heat flux will decrease by 70%. This kind of approach is considered in the European 9 m tokamak reactor study. Hopefully, it will not be necessary to go so far.

29 Tokamaks that will test PMI issues will need to be able to add impurities “The key tests will involve trying various combinations of core and mantle radiation using a variety of impurities; in order to establish the optimum conditions for a DEMO. The tokamaks in question include non D-T, PMI-testing machines such as NHTX (Goldston et al) and Vulcan, and the D-T VNS and Pilot Plant. Generally, in designing these, efforts have been made to make them as small as possible to test variously, the nuclear technologies, PMI effects, and Q eng ~ 1 at the minimum useful wall loading. For the PMI program one should make test facilities large enough to handle uncertainties in being able to achieve certain necessary plasma conditions. This includes having a lot of auxiliary power to be able to handle high core radiation loss scenarios and vary P/P L-H in the range of interest to the DEMO. Both ARIES-AT and NHTX have P/P L-H ≥ 6.

30 CONCLUSIONS Preparing for the DEMO will require many facilities to establish the data base on component reliability and maintenance. Flexibility will be crucial to cover all the interesting regimes of operation. These facilities will need to have large margins in volt-seconds (tokamaks), confinement, heating, and the ability to handle high impurity radiation levels. Much more detailed analyses are needed of how they will provide the component availability data.

31 Acknowledgements I appreciate receiving useful advice and information from Rob Goldston, Wayne Houlberg, Stan Kaye, and Dale Meade.

32 Back-up Slides

33 Important Parameters Fusion power P F (MW) Plasma profiles n = n 0 (1-r 2 /a 2 )  n and T = T 0 (1-r 2 /a 2 )  T Fusion fuel fraction n D /n e, where n D = n T (10 20 m-3 ) and the impurities ∑n Z Z. f G the fraction of the Greenwald limit used. Greenwald limit n G = I/(  a 2 )(10 20 m -3 ) T = T e = T i (10keV)— these are volume averaged  = a/R, , d, q 95,  the gap between the plasma and inner leg of the TF coil, and f a = a w /a the scrape-off layer thickness (assumed to be constant around the plasma). B m (T) the maximum field on the coil and R m (m) is TF coil inner leg outer radius I = 2.5(BR)  2 f 1 /q 95 (MA) BR = B m R m where f 1 = {1 +  2 (1 + 2  2 – 1.2  3 )}{1.17 – 0.65  }/{1 -  2 } 2

34 Fusion Power P F = 0.247 f(  n,  T ) f 1 2.0 (q 95 )-2 f 2 2 (f 3 f 4 ) -2  N 2  4  (BR) 4 /R (MW) f(  n,  T ) = {(1 +  n ) 2.(1 + 2  n + 3  T ) 2 / (1 + 2  n + 2  T ) 3 } x{(1 +  n +  T ). (1 + 2  n + 2  T )} 2 /{(1 +  n ).(1 + 2  n + 3  T )} 2 The beta limit is given by =  N I/aB (%) + 2 + ∑ = n e T.f 3.f 4 where f 3 = (1 +  n )(1+  T )/ {(1 +  n +  T ) and f 4 = 1 + 2 f 2 +  / R = R m /[1 – (1 +f a )  -  /R] (m)

35 Confinement and Power to Sustain Plasma ITER H-mode scaling is  E = 0.1444 H H I 0.93 B 0.15 n 0.41 M 0.19 R 1.97  0.58  X 0.78 P -0.69 = 4.737Ra 2  /P (s) Note that in the ITER-FEAT studies  X /  ≈ 1.1, and with M = 2.5 P = [1.023 H H -1 f G -0.41  N I -0.34 (BR) 0.85  1.24  0.22 ] 3.23 (MW) For Q =20, P a = P/5 and 0.2P F /P = 0.8

36 Balancing Alpha Power Against Power Requirement With the formulae for P F, P, I, and f G < 1, and T = 1 ( 10 keV ) 0.2P F /P = 4.43 x 10-3 (Plasma)x(Shaping)x(Impurities)x(Size) Plasma Parameters H H 3.23  N 0.094 f(  n,  T )/(q 95 3.1 f 3 3.32 f 4 3.32 ) Shaping f 1 3.1  3.5  0.29 Impurity Depletion (n D /n e ) 2 Size (BR) 3.68 /R where f 1 = {1 +  2 (1 + 2  2 – 1.2  3 )}{1.17 – 0.65  }/{1 -  2 } 2

37 Staging a Pilot Plant Example Case: R/a =3.5,  = 2.2,  = 0.8, q 95 = 3.7, B m =10 T,  = 0.8, R = 4.23 m, a = 1.21 m, H H = 1.1.  n = 0.4,  T = 1.0, I = 12.2 MA, V.s = 144. Start with ITER-like limits and design to stage to ARIES-AT-like limits if accessible.

38 NN fGfG I BS /II CD /I  F (hrs) P eg (MW e ) Q eng 2.50.590.490.060.41401.0 3.00.700.570.070.52001.2 3.50.820.650.070.72701.3 4.00.900.720.081.03501.45 4.50.900.790.112.04401.6 5.00.900.860.13S.S.5401.7

39 V.s required Required V.s is  req = L p I + 0  tr R p Idt + R p I t F t F = the flat-top time, L p and R p are the plasma inductance and resistance respectively. I is in Amps. 0  tr R p Idt ≈ 0.5 L p I (V.s). L p ≈  0 R[ln(a c /a) + 0.25} (H), where (a c /a) is the ratio of the average poloidal coil radius divided by the average plasma radius. R p ≈ {2.6 x 10 -4 Rg r Zeff ln(  e )}/a 2  T e0 1.5 (Ohms).

40 NN P F (MW) P  (MW) P (MW) P aux (MW) P/R (MW/m) P/A w (MW.m -2 ) P n /A w (MW.m -2 ) 2.5 400 80 98 18 23 0.331.1 3.0 580115138 23 33 0.461.6 3.5 790157186 29 44 0.622.1 4.01030205240 35 57 0.802.8 4.51300260301 41 71 1.003.5 5.01600321367 46 87 1.224.3

Download ppt "Path to a High-Availability MFE Power Plant John Sheffield ISSE—U.Tennessee-Knoxville."

Similar presentations

Ads by Google