3 ms·

Also interesting is the Wendelstein 7-X experiment, where they build a stellarator with a unique geometry, compared to the tokamak reactors. [1] https://en.wik

Also interesting is the Wendelstein 7-X experiment, where they build a stellarator with a unique geometry, compared to the tokamak reactors. [1] https://en.wikipedia.org/wiki/Wendelstein_7-X https://en.wikipedia.org/wiki/Wendelstein_7-X

willis936 7y ago
Ah quasi-omnigeneity. If anyone could explain it I would appreciate it. I know that optimizing for it "minimizes radial drift", as does quasi-helical-symmetry and quasi-axis-symmetry depending on your model for particle drift. I just don't have a grasp of omnigeneity and haven't found an easily consumed explanation for it yet.
The_rationalist 7y ago
Can scientists quantify how much of energy efficiency gains can be made from current hypothetically suboptimals tokamak geometry?
- willis936 7y ago
  Tokamaks are different beasts than stellarators. They have a toroidal current induced by a central solenoid (literally a transformer). There are pros and cons to this. For one, you get a lot of ohmic heating and some focusing benefits from the current. However, there are plasma instabilities and there needs to be electrical feedback systems operating with microseconds of latency. This isn't trivial when dealing with huge currents and voltages at high inductance. Additionally the transformer can only induce current in one direction for a finite time (you cannot ramp current to infinity). No one has figured out how to make tokamaks steady state. The best hope is for runs of a few days and 80% uptime. It's undetermined if this is "good enough". Stellarators have traditionally lagged tokamaks in terms of Lawson criterion performance, but they've also been less funded because tokamaks have performance advantages. Stellarators are steady state and don't risk the same number of plasma instabilities. Germany is so confident that it's the right approach that they built the largest magnetic confinement device ever to test optimizing for quasi-omnigineity. Stellarators in general are becoming closer to the mainstream approach and given another 5 or 10 years you'll likely hear about them as much as tokamaks (at least in terms of non ITER research).
  - hedvig 7y ago
    >operating with microseconds of latency. What are the numbers if you know, of the speed or rate at which plasma can destabilize?
    - willis936 7y ago
      There is a lot of reading out there on the various types of instabilities, but generally the timescale for these effects are in the microsecond to millisecond range. I think particles move toroidally in the transsonic range, which depends on the temperature (which is a gradient from the core). This would be in the range of 300-1000 m/s in a device with a circumference of a meter or less (for most science machines made so far). I think it only takes a handful or orbits for an instability to grow. It depends on how much abnormal magnetic shear there is. I’m not comfortable saying more. https://www.energy.gov/science/fes/articles/zero-tolerance-tokamaks-eliminating-small-instabilities-they-become-disruptions https://www.energy.gov/science/fes/articles/zero-tolerance-t... http://www.ccfe.ac.uk/assets/Documents/POPVOL17p082509.pdf http://www.ccfe.ac.uk/assets/Documents/POPVOL17p082509.pdf
      - willis936 7y ago
        I can’t edit this post for some reason so I’ll just put my correction here. I attended a presentation today that, at one point, showed toroidal flow rate over radius in a 1m major radius device. It ranged from 30 to 100 km/s. There are many effects operating at different timescales. Turbulence is among the shorter timescales, making it difficult to simulate and why it is only now the current focus of many researchers.

willis936 7y ago

Ah quasi-omnigeneity. If anyone could explain it I would appreciate it. I know that optimizing for it "minimizes radial drift", as does quasi-helical-symmetry and quasi-axis-symmetry depending on your model for particle drift. I just don't have a grasp of omnigeneity and haven't found an easily consumed explanation for it yet.

The_rationalist 7y ago

Can scientists quantify how much of energy efficiency gains can be made from current hypothetically suboptimals tokamak geometry?

willis936 7y ago
Tokamaks are different beasts than stellarators. They have a toroidal current induced by a central solenoid (literally a transformer). There are pros and cons to this. For one, you get a lot of ohmic heating and some focusing benefits from the current. However, there are plasma instabilities and there needs to be electrical feedback systems operating with microseconds of latency. This isn't trivial when dealing with huge currents and voltages at high inductance. Additionally the transformer can only induce current in one direction for a finite time (you cannot ramp current to infinity). No one has figured out how to make tokamaks steady state. The best hope is for runs of a few days and 80% uptime. It's undetermined if this is "good enough". Stellarators have traditionally lagged tokamaks in terms of Lawson criterion performance, but they've also been less funded because tokamaks have performance advantages. Stellarators are steady state and don't risk the same number of plasma instabilities. Germany is so confident that it's the right approach that they built the largest magnetic confinement device ever to test optimizing for quasi-omnigineity. Stellarators in general are becoming closer to the mainstream approach and given another 5 or 10 years you'll likely hear about them as much as tokamaks (at least in terms of non ITER research).
- hedvig 7y ago
  >operating with microseconds of latency. What are the numbers if you know, of the speed or rate at which plasma can destabilize?
  - willis936 7y ago
    There is a lot of reading out there on the various types of instabilities, but generally the timescale for these effects are in the microsecond to millisecond range. I think particles move toroidally in the transsonic range, which depends on the temperature (which is a gradient from the core). This would be in the range of 300-1000 m/s in a device with a circumference of a meter or less (for most science machines made so far). I think it only takes a handful or orbits for an instability to grow. It depends on how much abnormal magnetic shear there is. I’m not comfortable saying more. https://www.energy.gov/science/fes/articles/zero-tolerance-tokamaks-eliminating-small-instabilities-they-become-disruptions https://www.energy.gov/science/fes/articles/zero-tolerance-t... http://www.ccfe.ac.uk/assets/Documents/POPVOL17p082509.pdf http://www.ccfe.ac.uk/assets/Documents/POPVOL17p082509.pdf
    - willis936 7y ago
      I can’t edit this post for some reason so I’ll just put my correction here. I attended a presentation today that, at one point, showed toroidal flow rate over radius in a 1m major radius device. It ranged from 30 to 100 km/s. There are many effects operating at different timescales. Turbulence is among the shorter timescales, making it difficult to simulate and why it is only now the current focus of many researchers.