1
00:00:00,510 --> 00:00:11,070
So when you hear standard sort of public talks about fusion, the picture that they show you at the beginning is this one.

2
00:00:12,360 --> 00:00:15,810
So and they tell you either this is what we're going to build.

3
00:00:16,530 --> 00:00:29,849
Right. So this is the machine. And you might add this in in at least in terms of sort of funding and and story, as it were, as it's unfolding,

4
00:00:29,850 --> 00:00:36,900
this this image and this the thing behind this image kind of dominates the diffusion world.

5
00:00:37,170 --> 00:00:43,090
Well, you might ask. The question, where's the theoretical physics here?

6
00:00:44,080 --> 00:00:51,640
Right. So all this stuff, this is clearly going to be a great engineering feat of 21st century civilisation,

7
00:00:51,730 --> 00:00:59,110
possibly, actually the way it's going 22nd century civilisation. But.

8
00:01:00,040 --> 00:01:06,279
But. But. Well, despite appearances, this is.

9
00:01:06,280 --> 00:01:11,230
No, not all quite so nice and control and well organised as you would think.

10
00:01:12,610 --> 00:01:24,640
Somewhere deep inside that the you know, you might ask where is it that the that this sort of, you know, of wild, primal, basic things are.

11
00:01:26,380 --> 00:01:33,850
And that, of course, is what theoretical physics does while primal things the basic instincts of nature, if you like.

12
00:01:33,850 --> 00:01:40,690
Right. So this is in polite society. They call them fundamental laws of physics, universal physics, that sort of thing.

13
00:01:41,020 --> 00:01:46,180
So where's all this? Well, here it's in pink. It's it's it's this.

14
00:01:47,310 --> 00:01:52,330
It looks nice and and and thing, but it is the beast.

15
00:01:53,890 --> 00:01:59,730
So, so, so, so, so I think I think the bottom line is that is that, you know,

16
00:02:00,190 --> 00:02:07,660
theoretical physicists are only useful in the in this program if they can engage well, engage tame the beast.

17
00:02:08,230 --> 00:02:20,100
So let's discuss the beast. Let's let's strip out all the engineering and let's realise some very, very basic thing about, about what's inside there.

18
00:02:20,470 --> 00:02:28,330
Right. So you have a toroidal plasma. You have a plasma. And the basic feature of that plasma is that it's hot in the middle and cold at the edges,

19
00:02:29,350 --> 00:02:32,890
and it has to be like that because it has to be hot in order to fuse.

20
00:02:33,160 --> 00:02:42,549
And it has to be cool because, you know, it's it's if we're talking about it being at 100 million degrees in the middle,

21
00:02:42,550 --> 00:02:47,080
it shouldn't be it certainly shouldn't be a hundred million degrees at the edges because we don't want it to melt the device.

22
00:02:48,100 --> 00:02:51,340
Right. So so we have to separate themselves from the beast somehow.

23
00:02:52,300 --> 00:03:01,870
And so we're maintaining this plasma in this state, which is if you're a theoretical physicist, you don't actually think in terms of pictures.

24
00:03:01,870 --> 00:03:05,350
It will be lots of pictures in movies in my talk. But in fact.

25
00:03:05,860 --> 00:03:13,659
Right, you have to realise that in, in, in the theoretical physics department everybody thinks in terms of plotting Y versus X fundamentally.

26
00:03:13,660 --> 00:03:21,250
Right. So, so, so, so the really interesting picture here is not this one but that one, and that's temperature versus radius.

27
00:03:21,400 --> 00:03:26,020
Right. So so radius goes from zero to out and the temperature goes from hot to cool.

28
00:03:26,260 --> 00:03:30,129
And so there is a temperature gradient in the system.

29
00:03:30,130 --> 00:03:33,310
The system is out of equilibrium. Remember your statistical physics.

30
00:03:33,310 --> 00:03:37,660
If you still remember your undergraduate days, the system is out of equilibrium.

31
00:03:38,290 --> 00:03:46,600
And so it's potentially a dangerous situation because because nature wants to meet, always wants to be in equilibrium.

32
00:03:46,930 --> 00:03:51,850
So the question for a physicist is, well, how hot does it get?

33
00:03:51,870 --> 00:03:55,360
Or more practically, how hot can we make it?

34
00:03:55,570 --> 00:03:59,260
How much of a temperature contrast can you create? Well, let's.

35
00:03:59,530 --> 00:04:06,100
And so this is where theoretical physics starts. And theoretical physics starts on the level of second year undergraduate physics,

36
00:04:07,120 --> 00:04:11,500
which which I hope some of you have at least some sort of vague recollection of.

37
00:04:12,340 --> 00:04:17,350
I assume this, but I'll remind you of all the all the relevant things.

38
00:04:17,800 --> 00:04:22,570
So, so let's start with what we teach our students in the second year undergraduate physics course.

39
00:04:22,690 --> 00:04:26,560
Right? We teach them the heat equation. Right? So so we tell them that, you know,

40
00:04:26,590 --> 00:04:34,840
if you want to know what the profile of the temperature somewhere is and remember the reason it gets hot is because you heat it in some ways,

41
00:04:34,840 --> 00:04:38,559
which I won't go into Steve might in in in his talk.

42
00:04:38,560 --> 00:04:42,010
So so you you have to solve the heat equation, right? Well,

43
00:04:42,010 --> 00:04:51,850
the heat equation is the temperature by the T equals two heat diffusivity times the plus in of T plus whatever sources and sinks of heat you have.

44
00:04:51,850 --> 00:04:55,150
Right. So this is s stands for heating minus cooling.

45
00:04:55,160 --> 00:05:00,280
Should there be cooling as well? Right. And you might say, all right.

46
00:05:00,280 --> 00:05:05,740
Well, the problem is the problem is simple. We look at it in steady state.

47
00:05:06,040 --> 00:05:12,609
So we eliminate the time derivative. We solve this in the Taurus. We know all our inputs, we know the boundary conditions.

48
00:05:12,610 --> 00:05:18,009
So we get the temperature profile. We hand over the solution to the engineers, and then we move on to thinking of dark matter,

49
00:05:18,010 --> 00:05:22,899
quantum entanglement, brief history of time, the meaning of life, and, and, and the rest of it,

50
00:05:22,900 --> 00:05:26,860
which is what we're supposed to be thinking about as sort of, you know,

51
00:05:27,070 --> 00:05:31,990
coastal in ivory tower theoreticians in the Rudolf or centre for threats, theoretical physics.

52
00:05:32,860 --> 00:05:40,000
But before we do that and of course, once we have built the fusion machine and have the unlimited source of.

53
00:05:40,070 --> 00:05:47,510
Energy. It's, you know, the whole world will be a centre for theoretical physics and it's, you know, everybody will be thinking of nothing of that.

54
00:05:47,510 --> 00:05:51,260
But but that these all these mundane concerns would disappear.

55
00:05:51,380 --> 00:05:56,510
But before that happens. Let's let's ask a quick still an undergraduate question.

56
00:05:57,530 --> 00:06:04,310
And that question is, what's this? Right. Because clearly so if I knew how I heat the plasma and we all know what the blossom is.

57
00:06:04,580 --> 00:06:10,010
But the interesting question here is this deep, right? Because depending on what that is it.

58
00:06:10,130 --> 00:06:16,640
Tell me in the end. Right. What, what that those temperature gradients are.

59
00:06:16,790 --> 00:06:23,479
Right. You see the D multiplies the gradients of T and so D effectively normalises links here.

60
00:06:23,480 --> 00:06:30,260
So, so in the end. Right. You know, how big are, you know, if you tell me that, that, that the temperature contrast has to be 100 million degrees.

61
00:06:31,460 --> 00:06:37,550
How, how big a device I have to build and so on will all sort of depend on what that is.

62
00:06:37,790 --> 00:06:42,080
Well, how do we calculate this? Well, again, let me remind you how to calculate heat diffusivity.

63
00:06:42,740 --> 00:06:48,440
It's the reason he diffuses. And that's what we tell undergraduates.

64
00:06:48,950 --> 00:06:53,959
There is a it's true, but at the same time it's a lie. And then you'll still see in the moment.

65
00:06:53,960 --> 00:06:57,230
Y Well, it's, it's, it's not a lie. It's not the whole truth.

66
00:06:57,770 --> 00:07:03,440
So the not the whole truth is the it right? Are you have you have hot particles in the middle.

67
00:07:03,440 --> 00:07:11,849
They random walk their way through the tokamak. They're, they're in thermal motion so they're moving randomly when they well each of them is in

68
00:07:11,850 --> 00:07:14,930
the first instance moving the straight lines and they collide with another particle,

69
00:07:15,170 --> 00:07:25,340
change their direction. And so it's like, you know, drunk in the famous story of a drunken man progressing from a pub out into the world at some rate.

70
00:07:26,300 --> 00:07:34,700
Right. So so so the the particle random walks its way eventually out from the core to the to the edge and brings its energies with it.

71
00:07:35,240 --> 00:07:40,790
And so and so there is a diffusion of energy now that is just the diffusion coefficient.

72
00:07:41,660 --> 00:07:45,350
And the diffusion coefficient is simply the ratio of the mean square displacement

73
00:07:45,350 --> 00:07:49,130
of the particle by the time over which that means square displacement happens.

74
00:07:49,400 --> 00:07:57,950
In other words, you could write it as the mean square particle velocity times, the time between collisions and the when particles move.

75
00:07:57,950 --> 00:08:03,919
Similarly, the velocity is basically the speed of sound in that in that medium.

76
00:08:03,920 --> 00:08:07,520
And so this is like the speed of sound times, the mean free path of the particle.

77
00:08:07,520 --> 00:08:10,910
This is the standard undergraduate estimate of the diffusion coefficient.

78
00:08:11,150 --> 00:08:18,229
It's actually in the in applause in the magnetised plasma and talking about this is actually on true even on that level.

79
00:08:18,230 --> 00:08:27,020
In fact, it turns out that because what Felix has explained to you, the particles actually gyrate around the magnetic field.

80
00:08:27,020 --> 00:08:32,240
So they can't actually move very far across the magnetic field. So all of that is only really true along the magnetic fields.

81
00:08:32,720 --> 00:08:39,500
So if you orient your magnetic fields in such a way that there is no magnetic field taking anybody out radially,

82
00:08:39,500 --> 00:08:46,040
which is what you care about the radial transport, then actually your step size for the particle is the radius of the lawnmower generation.

83
00:08:46,040 --> 00:08:54,949
This is that Felix talked about so in fact it's built to x is is the lawnmower radius and and this is still the time between collisions.

84
00:08:54,950 --> 00:08:59,689
So so you get this is this is the basic estimate that you might get for for the plasma.

85
00:08:59,690 --> 00:09:05,300
And you know since since since we're theoretical this is here we're certainly not going to substitute the numbers.

86
00:09:07,790 --> 00:09:14,870
But I'll tell you that this turns out to be way, way, way, way, way, way, way too small to explain the transport that we see.

87
00:09:15,620 --> 00:09:23,659
Okay. So in fact, when when fusion devices were first designed, Steve might say something about this in this talk by Lyman Spitzer, they, you know,

88
00:09:23,660 --> 00:09:28,520
when they thought when they basically estimated things based on collisional transport, you know,

89
00:09:28,550 --> 00:09:33,710
they came up with a device that would that would have been almost a tabletop arrangement.

90
00:09:34,820 --> 00:09:37,820
So. Well, it's you know, the reason it's no longer a tabletop arrangement,

91
00:09:37,820 --> 00:09:43,010
but that huge monster of engineering that you saw is is precisely because this does not work.

92
00:09:43,400 --> 00:09:47,060
Now, why doesn't it work? So let's look closer.

93
00:09:47,330 --> 00:09:53,989
So let's go back to this to this toroidal plasma, which is hot in the middle and cold on the outside.

94
00:09:53,990 --> 00:09:58,610
This all looks nice and and and and smooth and and kind of pain, doesn't it?

95
00:09:59,780 --> 00:10:03,380
Well, in fact, it's nothing of the kind. Let me show you a movie over.

96
00:10:03,530 --> 00:10:09,170
This is admittedly a computer simulation of of of what happens to a plasma like this,

97
00:10:10,400 --> 00:10:16,160
which was done for to model the operations of the 3D Tokamak in San Diego.

98
00:10:17,180 --> 00:10:23,959
And this is it. I'm not showing it because that tokamak is is any way any better or worse than any other tokamak.

99
00:10:23,960 --> 00:10:27,740
But they do have the absolute best movie making machine there in Southern California.

100
00:10:28,520 --> 00:10:33,889
And and so their movies are unrivalled. And so here's the movie.

101
00:10:33,890 --> 00:10:39,950
Let's let's see if this actually works. So so they start this simulation.

102
00:10:40,310 --> 00:10:44,360
And this is this is actually this is actually density of the plasma in the Tokamak.

103
00:10:44,630 --> 00:10:55,490
And you see that it all breaks up in this in this sort of seething, boiling, unholy, chaotic mess, which is called turbulence.

104
00:10:56,060 --> 00:11:06,709
So so what this is meant to demonstrate is that when you when you put something out of equilibrium, what tends to happen is,

105
00:11:06,710 --> 00:11:14,180
you know, nature dislikes temperature gradients and because they are they constitute a lack of equilibrium.

106
00:11:14,180 --> 00:11:21,370
And so it can try to drive the system unstable and to to try to rectify this situation for itself.

107
00:11:21,380 --> 00:11:26,750
I'll show you in great detail how it will in some detail, how it exactly does this.

108
00:11:26,960 --> 00:11:35,180
The point that I want you to carry out of this picture for now is that the whole thing is moving rather violently and chaotically.

109
00:11:35,930 --> 00:11:39,950
Right. So actually what we should do is not second year undergraduate physics,

110
00:11:39,950 --> 00:11:50,210
but third year the third year undergraduate physics fails you what happens to temperature when it's not just being diffused by collision,

111
00:11:50,360 --> 00:11:53,690
but when it's actually in a moving medium. Right.

112
00:11:53,840 --> 00:12:03,350
And what happens is, is, on the face of it, a fairly fairly harmless what happens is that the the heat equation develops another term.

113
00:12:03,860 --> 00:12:13,310
And that's the term. It's you don't guarantee what this describes to you is the is the velocity of the of the plasma motions,

114
00:12:13,400 --> 00:12:16,970
not the velocity of the particles, but the mean velocity of the motions at any given point.

115
00:12:17,390 --> 00:12:21,860
And what this term describes is the advection of the velocity by these motions.

116
00:12:21,980 --> 00:12:26,360
Okay. These are chaotic motions. No, no.

117
00:12:26,360 --> 00:12:32,989
Let's do the following thing. Right. We're going to see that temperature has a mean profile.

118
00:12:32,990 --> 00:12:37,510
I mean, obviously, now that everything is is is fluctuating, the actual profile,

119
00:12:37,550 --> 00:12:43,190
the temperature will be all that will be old jacket and have the little fluctuations and so on.

120
00:12:43,400 --> 00:12:49,340
But on the average, we still want to have a situation when in the middle things are hot and on the outside the things are cool.

121
00:12:49,370 --> 00:12:55,490
Things are cool, right? So, so there will be some mean profile of tension which, which I'll call, which I'll call T bar.

122
00:12:56,060 --> 00:13:01,160
And so that and that will be the average of the temperature in the plasma over.

123
00:13:01,310 --> 00:13:07,600
Let's see these fluctuations and these fluctuations will assume have are fast in time.

124
00:13:07,610 --> 00:13:12,709
So what you saw there is that you have you have all these fluctuations things things are changing very,

125
00:13:12,710 --> 00:13:18,680
very fast and will assume that they're changing faster than the mean quantities change in the class.

126
00:13:19,010 --> 00:13:23,400
Okay. So, so if you like, I've just been working on constructing.

127
00:13:23,810 --> 00:13:28,370
Those of you who have done fourth year undergraduates will remember what, what, what mean field theory is.

128
00:13:28,820 --> 00:13:35,320
Right. So? So I've embarked on constructing some kind of an effective, meaningful theory of of of what's going on.

129
00:13:35,330 --> 00:13:39,379
So I will think of my temperature as being composed of the mean profile and the

130
00:13:39,380 --> 00:13:43,040
fluctuations and the fluctuations will be small compared to the mean temperature.

131
00:13:43,050 --> 00:13:48,640
So so it doesn't it doesn't change very much, but they're fast and the mean profile is slow.

132
00:13:49,230 --> 00:13:57,139
Now, that means that I can if I average that equation and also assume that the difference of velocity zero,

133
00:13:57,140 --> 00:14:01,910
which I can do because these motions tend to be subsonic and therefore incompressible.

134
00:14:02,960 --> 00:14:10,280
So. If I average that equation above, what I find is an equation for the mean profile.

135
00:14:11,060 --> 00:14:14,629
And so this is just the average of that term.

136
00:14:14,630 --> 00:14:18,350
This is the average of that term, the source, the heating of the plasma.

137
00:14:18,350 --> 00:14:21,590
I assume that I control that and that and that that's all happening slowly.

138
00:14:22,910 --> 00:14:30,290
And the only interesting new place here is this divergence of the average of the velocity times temperature.

139
00:14:30,770 --> 00:14:41,390
Okay, so this thing here now to calculate this, we need to know about the fluctuations because if you think of T as being t border plus delta t,

140
00:14:41,630 --> 00:14:47,810
then then the average of u t is the average of u times t bar plus the average of you delta t.

141
00:14:48,080 --> 00:14:53,270
And if I assume for now that there is no mean motion of the plasma, Steve, in this talk,

142
00:14:53,270 --> 00:14:56,479
we'll actually talk about situations when there is mean motion of plasma.

143
00:14:56,480 --> 00:15:01,340
But for now, we can see that there's no mean motion of the plasma. And if there is no mean motion, this is zero.

144
00:15:01,550 --> 00:15:07,430
And all of that average is the is the average between the velocity fluctuations and the temperature fluctuations.

145
00:15:09,140 --> 00:15:13,030
Well, so let's calculate that average. How do we calculate this?

146
00:15:13,100 --> 00:15:17,000
So I'm going to do a sort of a split of theoretical physics and then switch back to free quality.

147
00:15:18,260 --> 00:15:21,860
So now how do I do this?

148
00:15:21,860 --> 00:15:26,900
Well, here's a simple way of doing this. Not entirely rigorous, but I think this is good enough.

149
00:15:27,620 --> 00:15:30,319
So so this is the it's the average of you times,

150
00:15:30,320 --> 00:15:35,510
t and I'm going to take this T and get it out of this equation, which I will just integrate over time.

151
00:15:36,530 --> 00:15:38,150
Okay. So there's the U.

152
00:15:38,300 --> 00:15:45,770
And this integral is simply the integral of all those terms of time from zero to T and it's that's the reaction term, that's diffusion.

153
00:15:45,770 --> 00:15:53,930
So that's sort of in now in under this integral everywhere, I will simply approximate temperature with its mean value.

154
00:15:54,920 --> 00:15:59,540
Okay. I will also use the fact that it's changing very slowly.

155
00:15:59,540 --> 00:16:03,919
So, so, so so what happens at previous times is not that different from what happens,

156
00:16:03,920 --> 00:16:10,700
but at the current time you might even think of it as being constant if we ever manage to construct a fusion device that operates in steady state.

157
00:16:11,630 --> 00:16:16,730
So if we do this, then all these terms become zero.

158
00:16:16,730 --> 00:16:20,660
Because basically those are just in terms and it's the average of you and that's zero.

159
00:16:20,660 --> 00:16:29,060
So the only known non-zero term here will be when we average these fluctuations that time T with the fluctuations at some previous time.

160
00:16:30,580 --> 00:16:34,790
And so what I end up with is the following expression for you don't see this.

161
00:16:35,330 --> 00:16:45,920
So, so what I find is that it's proportional to the gradient of the mean temperature times this time integral of the correlation function.

162
00:16:45,980 --> 00:16:51,110
This is again this is I think fourth year undergraduate physics, the correlations or maybe even third.

163
00:16:51,980 --> 00:16:57,889
So it's correlation function of the velocity attempted with the velocity time t prime this quantity if you

164
00:16:57,890 --> 00:17:03,740
look at where it is in the equation is a heat flux and that that's what's called the turbulent heat flux.

165
00:17:05,240 --> 00:17:11,420
Now I can stick this quantity into the air and get this equation.

166
00:17:12,740 --> 00:17:20,510
And finally, I will remember that all my mean things only vary along the radial direction.

167
00:17:21,080 --> 00:17:26,510
So it's all actually one dimensional. So all these divergences become derivative suspect radius.

168
00:17:26,810 --> 00:17:33,920
And what I end up with basically is an additional bit in the diffusivity.

169
00:17:33,920 --> 00:17:38,030
So I end up with an equation that looks just like the equation I had before,

170
00:17:38,330 --> 00:17:42,170
which was the equation for, for, for the, for the diffusion of temperature.

171
00:17:42,740 --> 00:17:50,059
But now, besides the collisional diffusion, I also have this thing called the turbulent diffusion in this D turbulent is equal

172
00:17:50,060 --> 00:17:54,170
to this correlation function of the turbulent fluctuations integrate over time.

173
00:17:55,220 --> 00:17:58,940
So this is what the effective mean field theory for our system looks like.

174
00:18:00,740 --> 00:18:05,629
It's it's it's basically just the there are you know, I should say at this point,

175
00:18:05,630 --> 00:18:09,710
there are all kinds of bells and whistles and so on, which I'm sweeping under an enormous rock.

176
00:18:11,120 --> 00:18:14,930
So but fundamentally, that's the idea of what of what one does.

177
00:18:15,590 --> 00:18:19,489
So so it looks like this. There is a there is an equation for the temperature profile.

178
00:18:19,490 --> 00:18:23,390
You'll set this this to zero to get to get steady state.

179
00:18:23,660 --> 00:18:27,470
And in order to solve this equation, you need to know what this is.

180
00:18:27,680 --> 00:18:35,810
Now, this general idea, you know, we gave our morning of theoretical physics a rather grand title.

181
00:18:38,820 --> 00:18:42,150
But that title has too important had two important components.

182
00:18:42,450 --> 00:18:48,090
One was that plasma was an unlimited source of energy. The other was that it was a normal state of all matter in the universe.

183
00:18:48,450 --> 00:18:48,750
Right.

184
00:18:48,810 --> 00:18:56,190
So this actually, you know, because we're theoretical physicists, because we, you know, we want to have a kind of an ivory tower sight to our lives.

185
00:18:57,060 --> 00:18:59,850
You know, you might say, well, is this is this just for the talk?

186
00:19:00,060 --> 00:19:06,210
This is actually quite the universal idea to treat plasma systems, which is generally physical systems this way.

187
00:19:06,570 --> 00:19:09,870
I mean, the world is full of systems where there's something happening in the mean.

188
00:19:09,870 --> 00:19:14,370
And also and also there are some fluctuations on top in the evolution of the mean things

189
00:19:14,760 --> 00:19:19,410
depends on how the only statistical properties of the fluctuations that they're embedded into.

190
00:19:19,830 --> 00:19:23,430
For example, my favourite example, right, if you didn't want to be a fusion physicist.

191
00:19:23,670 --> 00:19:28,680
But if you're instead you're a plasma astrophysicist or just an astrophysicist, even not a plasma physicist,

192
00:19:28,950 --> 00:19:34,680
then you know that the largest plasma object you can find in the universe are called clusters of galaxies.

193
00:19:35,250 --> 00:19:40,350
And clusters of galaxies, despite the name, contain mostly dark matter and hot diffuse plasma like galaxies.

194
00:19:40,920 --> 00:19:44,190
Galaxies and clusters of galaxies are, you know, rather an afterthought.

195
00:19:45,630 --> 00:19:51,480
So they look like this in optical, they look like that in X-ray.

196
00:19:52,800 --> 00:19:59,129
They are megaparsec sized objects. But the important thing, though, is not that they're so grand and big and so on.

197
00:19:59,130 --> 00:20:03,780
This is actually most of the visible. I mean, apart from the dark matter, this is most of the matter in the universe.

198
00:20:03,780 --> 00:20:07,139
And this is actually just a blob of plasma. Right?

199
00:20:07,140 --> 00:20:12,150
So this is the normal form of matter. It's rather exclusive that we are here, not plasma.

200
00:20:13,590 --> 00:20:21,210
So and it's also rather presumptuous of us to to then feel we can claim it, but we are presumptive question so we're going to do it.

201
00:20:22,170 --> 00:20:26,999
So anyway. Well, the again for a theoretical this this is a nice pretty picture, right?

202
00:20:27,000 --> 00:20:28,320
So astronomers like that kind of thing.

203
00:20:28,950 --> 00:20:35,189
But but actually for a theoretical physicist, the interesting fact about this is that this this is roughly radially symmetric.

204
00:20:35,190 --> 00:20:38,700
And this has a temperature profile. This is what its temperature profile looks like.

205
00:20:38,820 --> 00:20:42,900
It's hot in the middle. It's actually not as hot as a tokamak. So, you know,

206
00:20:43,920 --> 00:20:53,280
but with only just three tens of millions of degrees still fully ionise and it's colder in the middle and colder on the outside and hot in there.

207
00:20:53,280 --> 00:20:59,130
And why that is, is also a mystery. And the way you solve that mystery is basically by doing that kind of theory.

208
00:21:00,390 --> 00:21:03,750
So and of course, you know, you think it's laminar. No, it's not.

209
00:21:03,750 --> 00:21:10,110
Remember, it's all very turbulent. If you look at the temperature fluctuations in that in that environment, they look like this.

210
00:21:10,110 --> 00:21:14,459
It's again, in what we missed. So so it's you know, these are rather universal methods.

211
00:21:14,460 --> 00:21:20,520
Let's go back to, uh, to, to focus on the future a bit more.

212
00:21:20,820 --> 00:21:27,000
And well, first of all, the turbulent transport that I've derived through the turbulent diffusion coefficient turns out to be much,

213
00:21:27,000 --> 00:21:31,590
much larger than the collisional diffusion coefficient, and it's much faster.

214
00:21:31,860 --> 00:21:35,700
And, you know, fundamentally on a sort of a wishy washy level, it's faster.

215
00:21:36,150 --> 00:21:41,639
I mean, of course collisions will relax you to equilibrium. You know how people teach you in statistical physics.

216
00:21:41,640 --> 00:21:45,390
People say, oh, you know, if things are out of equilibrium, things will eventually go into equilibrium.

217
00:21:45,390 --> 00:21:49,800
And, you know, the universe will eventually go into a boring, homogeneous state and etc., etc.

218
00:21:50,100 --> 00:21:57,270
Right. The reason this is not as depressing as it sounds is because, you know, you can still hope that it might happen slowly.

219
00:21:58,470 --> 00:21:58,800
Right.

220
00:21:59,550 --> 00:22:08,850
So and and collisional relaxation does happen quite slowly, but nature actually tends to be rather impatient with with slow equilibrium and processes.

221
00:22:09,630 --> 00:22:13,230
And so you won't typically wait for slow collisions to relax the system.

222
00:22:13,230 --> 00:22:16,770
And so and so the transport becomes, becomes fast.

223
00:22:16,950 --> 00:22:21,809
It's the same kind of situation as, you know, if I switch on the radiator in the back of this room.

224
00:22:21,810 --> 00:22:31,440
Right, and try to and try to ask how fast the equilibrium of temperature will happen between the hot radiator and the I don't know, cold outside.

225
00:22:31,860 --> 00:22:36,420
Right. The answer to this is not based on Collisional transport, because there will be convection, there will be motions.

226
00:22:37,020 --> 00:22:41,099
Right. So, you know, like, again, nature will be impatient and and, you know,

227
00:22:41,100 --> 00:22:47,610
because I'm standing here waving my hands, I'm generating turbulence, which will help to transport as well.

228
00:22:48,180 --> 00:22:54,060
So so what we need to know, so this is the primary thing we need to know, right?

229
00:22:54,090 --> 00:23:02,610
This, this quantity. And what we basically wanted to be able to do is to predict what it is as a function of everything.

230
00:23:03,210 --> 00:23:07,680
Right? Everything being local equilibrium quantities, for example, the gradient of temperature.

231
00:23:07,680 --> 00:23:10,770
You know, this might depend on how much temperature gradient there is.

232
00:23:10,890 --> 00:23:15,870
In fact, it will depend on how much temperature gradient there is because if there is no temperature gradient,

233
00:23:15,870 --> 00:23:20,670
things will not be turbulent and single things will just be boring and stay put and so on.

234
00:23:20,910 --> 00:23:23,850
It is the temperature gradient that makes flows more unstable.

235
00:23:24,030 --> 00:23:29,040
I won't describe exactly how that happens, but that that turbulence happens because that the energies,

236
00:23:29,790 --> 00:23:34,020
the free energies injected into plasma via this temperature green.

237
00:23:34,380 --> 00:23:40,860
So what tends to happen experimentally. And on some level, theoretically and also in numerical simulations,

238
00:23:40,860 --> 00:23:47,970
is that this token diffusivity is a function of the temperature gradient basically behaves like this, it starts off well.

239
00:23:48,030 --> 00:23:53,610
Sorry, this I should have said to bring Diffusivity plus collisional is of the total transport.

240
00:23:54,450 --> 00:24:02,160
So it starts off as collisional transport which is which is very tame and slow and harmless and well understood and will give us a tabletop device.

241
00:24:02,730 --> 00:24:07,950
And then there is some critical temperature gradient. And beyond that critical point, it just shoots up like crazy.

242
00:24:07,950 --> 00:24:11,940
This is cool. This is turbulent transport. It's fast through in this poorly understood.

243
00:24:12,210 --> 00:24:21,240
Not understood. I'm sorry. It's so and it's this is this is a property that that in the language of the game is called

244
00:24:21,240 --> 00:24:25,140
stiff transport in the sense that the moment you go beyond a certain critical gradient,

245
00:24:25,320 --> 00:24:28,830
it just shoots up like crazy. You can't really go very far beyond that.

246
00:24:29,430 --> 00:24:34,530
So the name of the game practically is to move that critical gradient around anyway.

247
00:24:35,190 --> 00:24:40,590
So Timmins is the enemy. In order to kill your enemy, you have to understand it first.

248
00:24:42,060 --> 00:24:49,320
And we also have to understand it because you have to keep your self-respect as a Homo sapiens.

249
00:24:50,400 --> 00:24:56,280
So anyway, so since we have to understand it, let's, let's take a look at it again.

250
00:24:56,700 --> 00:25:03,179
Right? So what I showed you before was, was a movie of simulated turbulence, right?

251
00:25:03,180 --> 00:25:10,620
And there is always there is always a degree of how should I put it, wariness that you should have.

252
00:25:10,620 --> 00:25:14,630
When you look at simulations of anything, who knows what the computer is showing you?

253
00:25:14,640 --> 00:25:18,959
I mean, you know, but in the end, physics is an experimental site.

254
00:25:18,960 --> 00:25:21,630
So here's what turbulence looks like experimentally.

255
00:25:21,630 --> 00:25:28,650
This is a measurement of turbulence using something called the Blue Emission Spectroscopy System on the Moscow Air column.

256
00:25:29,490 --> 00:25:34,310
And this is done by Anthony Field, who is one of my collaborators of column.

257
00:25:34,350 --> 00:25:39,870
And the movie actually was done by my student, Yong Kim, who was in Oxford until recently,

258
00:25:39,870 --> 00:25:46,500
and he's now a professor in South Korea who are also serious about turbulence and fusion.

259
00:25:47,430 --> 00:25:52,860
In fact, they're very serious. So. All right. So this is so this is what experimental evidence looks like.

260
00:25:52,860 --> 00:25:56,760
This is density condensed. This is like a little window inside. You can't do the whole tokamak, right?

261
00:25:56,940 --> 00:26:03,719
So this is a tiny sliver inside the Tokamak and well, let's let's see what they look like.

262
00:26:03,720 --> 00:26:09,060
Well, these basically look like blobs that move around right randomly.

263
00:26:09,840 --> 00:26:18,870
So. Well, what does this tell us about this turbulent diffusion coefficient, the fact that they look like blobs that move around?

264
00:26:19,770 --> 00:26:26,310
Well, first of all, let's let's look at this expression and let's try estimate on a very hand-waving level what this is actually equal to.

265
00:26:26,910 --> 00:26:30,240
Well, you know, it has two powers of velocity and it has a time here.

266
00:26:30,510 --> 00:26:33,569
Right. And you see, this is a correlation function of resources.

267
00:26:33,570 --> 00:26:41,110
So so if you if you basically take if I if I take some point here and if I take the velocity at this point,

268
00:26:41,110 --> 00:26:46,410
the time T and multiplied by the velocity at this point that the of time multiply one by another.

269
00:26:46,680 --> 00:26:52,650
If the difference between the times is long enough, that would be zero, right. Because they'll be uncorrelated from each other.

270
00:26:53,130 --> 00:27:01,230
Right. So, so this function peters out to zero over some timescale which we will call the correlation time of the turbines.

271
00:27:01,230 --> 00:27:09,300
And so the two event diffusion coefficient is the roughly speaking, the size, the square of the velocity times, its correlation time.

272
00:27:09,780 --> 00:27:12,990
Well, what is its correlation time when again, if you looked at this movie,

273
00:27:13,290 --> 00:27:18,390
you might have convinced yourself that, you know, you have these blobs, or we call them ages sometimes.

274
00:27:19,230 --> 00:27:27,930
And the time that each of these blue globes moves is sort of comparable to its size, divided by the velocity at which it moves around.

275
00:27:28,170 --> 00:27:32,610
So in the way it breaks up, you know, it's, you know, you didn't see any coherent things.

276
00:27:32,610 --> 00:27:37,770
It just went like that, right? You know, what you saw is right.

277
00:27:38,130 --> 00:27:43,620
Something like that. Right. So so it's a basically the correlation time is something like the turnover time of the globe.

278
00:27:44,130 --> 00:27:47,580
So if the correlation time is that so l will be a disease agent.

279
00:27:47,580 --> 00:27:53,520
I call them edges and globes and whirls and twirls and swirls and and so these are all the same things.

280
00:27:54,150 --> 00:28:01,650
So so this is the research. This is this is the velocity. If I plug that in, I can come up with a couple more expressions for this for this to happen.

281
00:28:01,650 --> 00:28:06,660
Transport, for example, that can be the square of the size of the 80 divided by the correlation time.

282
00:28:06,870 --> 00:28:11,909
And it is actually evocative of what I explain to you about the Collisional transport, right?

283
00:28:11,910 --> 00:28:18,210
Remember how the collision transport, the particle carrying energy goes a distance, collides with another particle,

284
00:28:18,360 --> 00:28:22,950
gets deflected, goes another distance in the random direction and so on, and so diffuses well.

285
00:28:22,950 --> 00:28:31,589
What happens here? This is again, the random walk argument, but now a particle carrying energy is carried by this motion until that motion

286
00:28:31,590 --> 00:28:36,120
breaks up and then it's picked up by another motion that's random and and moves.

287
00:28:36,120 --> 00:28:38,620
Moves off. And then it's being. Another motion and so on.

288
00:28:38,860 --> 00:28:44,350
And so the step so is L and the correlation and, and the time it takes is the correlation part.

289
00:28:44,800 --> 00:28:50,290
So again, it's, it's a, you know, I could have actually admitted that derivation just showed you that derivation to show you.

290
00:28:50,290 --> 00:28:54,820
I'm a theoretical physicist, but fundamentally, I put it just written this and said, well, it's make sense, doesn't it?

291
00:28:55,330 --> 00:29:00,250
So in other words, you could write that as velocity times this random walk.

292
00:29:00,250 --> 00:29:03,220
And so and so again is there is a diffusive process here.

293
00:29:05,950 --> 00:29:12,910
Well so I have was sort of surreptitiously surreptitiously started introducing all sorts of notions.

294
00:29:15,180 --> 00:29:18,210
To do with the nature of turbulence. I talked about Globes and AIDS and all that.

295
00:29:19,230 --> 00:29:24,600
So what I'm going to do in the rest of this lecture is I will attempt a very basic and non-readers introduction to what turbulence is.

296
00:29:25,290 --> 00:29:35,670
Right, because because even if you weren't interested in fusion or indeed even plasma physics, turbulence by itself is quite a fascinating phenomenon.

297
00:29:36,240 --> 00:29:42,450
It's a routinely it's almost become banal to refer to it like this.

298
00:29:42,450 --> 00:29:45,510
But it's true that it's the last great unsolved problem of classical physics.

299
00:29:48,000 --> 00:29:55,500
I think it's was it Heisenberg who reportedly told that, you know, answer the question as to what he will ask God when he meets him.

300
00:29:55,710 --> 00:30:03,450
If you can only ask one question, and I think the question was about the structure of turbulence and was more precise in that he

301
00:30:03,450 --> 00:30:09,629
wanted to know the precise scaling of and what the correlation functions of turbulence with,

302
00:30:09,630 --> 00:30:16,620
etc., etc. But you know, I don't know if God would have time for that, but but there are insights to be gained from God and nature,

303
00:30:17,400 --> 00:30:23,730
or rather nature pending our meeting with God, but about, about the structure afterwards.

304
00:30:24,030 --> 00:30:27,660
So let me let me let me talk to you about turbulence generally.

305
00:30:27,840 --> 00:30:32,430
Not even necessarily implausible. I'll come back to plasma later. How did it all start?

306
00:30:32,700 --> 00:30:37,260
Well, turbulence, like everything else. It was first discovered by Leonardo da Vinci.

307
00:30:38,010 --> 00:30:42,899
Right. You know, if you've heard enough of these of these talks, you know that that everything goes back to.

308
00:30:42,900 --> 00:30:46,469
Then after that, every every engineering thing goes to Leonardo da Vinci.

309
00:30:46,470 --> 00:30:51,000
Everything goes. In fact, Leonardo da Vinci invented the Tokamak, right?

310
00:30:51,030 --> 00:30:57,320
I mean, you can fly, you know, if you leave through his notes, you will see a picture of a tourist with with, with.

311
00:30:57,360 --> 00:31:02,819
In fact, yes, he did so and he tried to build it.

312
00:31:02,820 --> 00:31:06,809
But the Duke of Milan instead wanted an enormous equestrian statue of himself.

313
00:31:06,810 --> 00:31:11,590
And so and so they couldn't afford to take him up. And so there we are right now.

314
00:31:11,670 --> 00:31:14,819
You know, governments at all times fundamentally want the same thing.

315
00:31:14,820 --> 00:31:17,520
And that's an enormous question, equestrian statue of themselves.

316
00:31:18,240 --> 00:31:23,280
So our goal is to persuade them that the Tokamak would be an enormous equivalent statue of themselves.

317
00:31:24,130 --> 00:31:29,190
So anyways, this is this is from the notes of Leonardo da Vinci.

318
00:31:29,190 --> 00:31:33,630
Right. So he he's been Leonardo was very good at observing reality.

319
00:31:34,500 --> 00:31:43,170
And one of the things he noticed this, I suspect, is a is a picture of a sewage pipe discharging its contents into into,

320
00:31:43,170 --> 00:31:48,420
I don't know, River Island or something like that. So so but what he noticed was the following.

321
00:31:48,420 --> 00:31:52,950
He said, observe the motion of the surface of the water, which resembles that of here.

322
00:31:52,950 --> 00:31:55,530
It does resemble that of here. Remember, he had a lot of hair.

323
00:31:56,460 --> 00:32:01,290
It's it's it's it has two motions of which one is caught is caused by the weight of the hair.

324
00:32:01,590 --> 00:32:04,920
The other by the direction of the curls or eddies.

325
00:32:05,970 --> 00:32:09,960
Thus the water has eddy motions, one part of which is due to the principal current.

326
00:32:10,470 --> 00:32:16,570
The other to the random and reverse motion. So he noticed for the first time that there was this separation of scales between there was

327
00:32:16,590 --> 00:32:22,440
a mean feel theory and fluctuations in both in the in the in the in the motion of water.

328
00:32:22,680 --> 00:32:27,060
He had met several of these sketches. There are all these he was pretty good at this sort of thing.

329
00:32:27,900 --> 00:32:31,229
So so anyway, he noticed all these all these eddies.

330
00:32:31,230 --> 00:32:42,300
And this, of course, is. So the basic idea that he came up with was that there is a mean laminar flow, which then breaks up into disordered motions.

331
00:32:42,360 --> 00:32:51,570
Somehow it doesn't stay stable. Right. So this actually is a universal phenomenon, these these ageing motions being created in in various fluids.

332
00:32:51,800 --> 00:33:00,210
Right. Here's a spectacular picture of an eddy. This is turbines in the wake of a of an aeroplane descending to Heathrow.

333
00:33:01,080 --> 00:33:05,230
So that's what they look like. These turtles. This is this is an air.

334
00:33:05,730 --> 00:33:07,610
This is this is the same thing.

335
00:33:07,620 --> 00:33:18,390
This is a close up of the of the great red spot of Jupiter, observed by by not by the other man, but gathered in this spacecraft.

336
00:33:18,600 --> 00:33:25,229
Galileo could not see in near-infrared. So and you see, there are these there are these overall currents.

337
00:33:25,230 --> 00:33:29,370
And then there are all these there are all these eddy motions on top of them.

338
00:33:29,940 --> 00:33:39,750
So this is very close by. These are called the giant radio loaves of flux, a million light years across like that.

339
00:33:40,110 --> 00:33:45,060
Right. So so very little example, a very large piece of turbines.

340
00:33:45,180 --> 00:33:51,299
Here's another spectacular astro picture. This is this is a so in the middle is something like a supernova,

341
00:33:51,300 --> 00:33:55,440
not quite the supernova that something like a supernova went off and there was a dust cloud around.

342
00:33:55,440 --> 00:34:01,050
And and there's a shockwave that hit this dust cloud. And they thought it all went spectacularly chaotic.

343
00:34:01,830 --> 00:34:07,290
And so there you can see the there is an overall emotion there, also all these emotions.

344
00:34:08,970 --> 00:34:13,770
So this is the universal. It's been seen in millions of places.

345
00:34:13,770 --> 00:34:17,400
This is a familiar. A picture? Probably. Right. So. So. Cosmic turbulence.

346
00:34:18,180 --> 00:34:23,459
Van Gogh could see it with a naked eye. Most of us need, need, radio, X-ray, infrared, that kind of thing.

347
00:34:23,460 --> 00:34:26,580
But anyway, this is how it happens in the Tokamak.

348
00:34:26,580 --> 00:34:30,810
Great. So this is a movie which was made by Edmund Haycock,

349
00:34:30,930 --> 00:34:38,060
who is who did his Ph.D. here in Oxford and still here as a as a junior research fellow at the modern college.

350
00:34:38,370 --> 00:34:43,379
So this is this is how turbulence happens in a in a plasma.

351
00:34:43,380 --> 00:34:46,260
This is a simulation of a small piece of plasma. You see, it's first.

352
00:34:46,500 --> 00:34:52,380
First, there are these irregular flows that are created and they all break up into into a mess.

353
00:34:53,430 --> 00:35:04,620
So, Miss, let's talk about this mess and let's let's let's try to understand this almost sounds like a strange thing to say, the structure of chaos.

354
00:35:05,370 --> 00:35:08,400
Right? But there is structure to chaos.

355
00:35:08,550 --> 00:35:11,970
So let's talk about the structure of these fluctuations.

356
00:35:15,360 --> 00:35:21,480
So this is a very famous picture. This is the biggest numerical simulation of terms ever made.

357
00:35:22,800 --> 00:35:27,930
A while ago already it's it was done in Japan on a machine called Earth Simulator,

358
00:35:27,930 --> 00:35:32,040
which at the time was the biggest machine in the world by Yukio Haneda.

359
00:35:32,040 --> 00:35:36,929
This is just fluid turbulence being simulated. So and this is sort of a field of turbulence.

360
00:35:36,930 --> 00:35:42,030
This is this is know fluid being turbulent in a box. And so you see all these messy fluctuations.

361
00:35:42,240 --> 00:35:50,220
And one of the interesting things about these fluctuations that you notice is that if you take a piece of that box and zero in on that piece,

362
00:35:50,490 --> 00:35:55,260
what will happen? I'll show you a succession of frames as you zero in further and further and further.

363
00:35:55,260 --> 00:36:05,790
Right. So so if you just look at this box, a box, this sub box looks like this, and then we'll take another box inside it and it'll look like that,

364
00:36:05,790 --> 00:36:10,379
and then we'll take another so box inside it and it'll take and it'll look like this.

365
00:36:10,380 --> 00:36:15,090
And what you basically see in all of that, then it then runs out of resolution at that point.

366
00:36:15,870 --> 00:36:25,920
But what you see in this succession of pictures is that it's not just a mess with a correlation length, it's a mess on multiple scales.

367
00:36:26,580 --> 00:36:32,430
Right. So what happens is, isn't just chaos. Chaos is actually a relatively tame and easy thing to do.

368
00:36:33,970 --> 00:36:37,590
Not to have is because turbulence is disorder on multiple scales.

369
00:36:38,130 --> 00:36:42,060
Right? That's that if you like. If you, you know, if you want the definition of two, that's it.

370
00:36:42,690 --> 00:36:49,739
Disorder of multiple scales. So, so, so what you see is this multi scale disorder and now we have two.

371
00:36:49,740 --> 00:36:56,040
Oh, well, since I'm overdosing on movies these days, here's, here's, here's what it looks like in plasma.

372
00:36:56,970 --> 00:37:00,630
So this is a very spectacular this is also a movie from Southern California.

373
00:37:02,070 --> 00:37:08,340
So this is a this is a picture of a sort of a simulation of plasma turbulence on multiple scales.

374
00:37:08,340 --> 00:37:16,950
Look look at the it's quite interesting to stare at look at the variety of structures that create you see you see, you have you have large structures.

375
00:37:16,950 --> 00:37:21,320
Then you have small structures and top of them. Then you have small structures on top of small structures and so on.

376
00:37:21,330 --> 00:37:24,780
So it is a, it is, it is a truly multi scale beast.

377
00:37:26,970 --> 00:37:29,190
So how do you characterise such a thing.

378
00:37:29,370 --> 00:37:35,370
I mean if you just think, think of it for a moment, not as a theoretician but even as an experimentalist, right?

379
00:37:35,460 --> 00:37:42,210
So imagine you're handed a system like that. And you have to do something about it.

380
00:37:42,220 --> 00:37:48,160
You have to you know, you know, before you start understanding something, you have to decide what is it that you're going to understand about it.

381
00:37:48,160 --> 00:37:51,819
You have to measure something about it, right? Because in the end, what are we doing?

382
00:37:51,820 --> 00:37:56,320
We're predicting some kind of Y versus X dependence of something, right?

383
00:37:56,320 --> 00:38:01,899
So what should we measure? Well, one of the standard things about turbulence is,

384
00:38:01,900 --> 00:38:10,150
is that if you measure the energy of these motions and if you then look at this energy scale by scale, what you tend to find is a power low.

385
00:38:10,660 --> 00:38:14,799
Right? So there is a so it's power laws galore and in intervals.

386
00:38:14,800 --> 00:38:20,020
This is this is this is the energy spectrum in the in an air jet.

387
00:38:20,020 --> 00:38:24,309
So this is this is disturbance in air has this power low spectrum.

388
00:38:24,310 --> 00:38:31,510
This is low look. So that's what power laws look like. Straight lines, the slope is minus four thirds.

389
00:38:32,470 --> 00:38:35,510
And then this is density.

390
00:38:35,530 --> 00:38:39,280
This is called this picture is famous. It's called the Great Power Low in the Sky.

391
00:38:39,550 --> 00:38:47,080
This is a density fluctuations in the interstellar medium on the range of scales that goes that goes.

392
00:38:47,080 --> 00:38:51,490
This is about ten orders of magnitude of scale separation.

393
00:38:51,640 --> 00:38:54,340
And over that scale, scale. So there's an overall parallel.

394
00:38:55,180 --> 00:39:03,550
There are many more in the solar wind spacecraft measure a turbine fluctuations find a point actually tends to be the same power load that you find.

395
00:39:04,060 --> 00:39:09,070
You always find this 5/3. This is a measurement from a galaxy cluster, right?

396
00:39:09,070 --> 00:39:18,100
So so this is just outside the Earth's orbit and this is this is a megaparsec away and the in the megaparsec across as well.

397
00:39:18,940 --> 00:39:22,900
So and here's here's here's a picture from a tokamak.

398
00:39:23,290 --> 00:39:31,720
So this is a simulation by Michael Michael Barnes and and Felix Parra, which was done here in Oxford in 2011.

399
00:39:32,890 --> 00:39:37,209
So this is this is turbulence in the plasma. It's slightly different quantities plotted.

400
00:39:37,210 --> 00:39:42,070
That's point seven, four and 5/3. And this is actually a measurement of turbulence.

401
00:39:42,070 --> 00:39:49,720
These these blue lines, the red line, the simulation, the the blue lines is a measurement on the French document go towards the graph.

402
00:39:50,770 --> 00:39:56,110
So again, power loss. So so what let's let's let's try to make sense of that.

403
00:39:56,980 --> 00:40:01,330
Why is it that turbulence is multi scale and why does it have these power laws?

404
00:40:01,870 --> 00:40:09,429
So fundamentally, this has to do with the way in which a nonlinear system will process energy that's injected into it.

405
00:40:09,430 --> 00:40:13,570
Right. So that's a that's a sort of a key executive summary, if you like, of what I'm about to say.

406
00:40:14,020 --> 00:40:20,880
So let me let me give you a simple example of how that works. Let's start again with third year undergraduate physics, right?

407
00:40:20,920 --> 00:40:27,670
So if we're just talking about turbulence in water or gas, it satisfies something called the Nova Equation,

408
00:40:27,880 --> 00:40:31,450
which is which is actually just a momentum equation for moving fluid. Right.

409
00:40:31,450 --> 00:40:35,050
This is velocity, the rate of change of velocity acceleration.

410
00:40:35,320 --> 00:40:44,200
This term says that the velocity of the moving fluid elements that direct each other, then move each other around, that's a pressure gradient.

411
00:40:44,800 --> 00:40:55,150
There is viscosity, which is dissipation of energy. And I will put in some unspecified F, which is which is going to stand for injection of energy.

412
00:40:55,150 --> 00:40:58,959
You know, I put a spoon in it. Let's start doing something like this.

413
00:40:58,960 --> 00:41:04,450
That's f so if I do this well, what's the energy of this system?

414
00:41:04,630 --> 00:41:11,170
The energy of the system is just the kinetic energy density times the times the square of velocity integrated over space.

415
00:41:12,130 --> 00:41:18,550
Right? So that's, so that's the energy of the system. And I can derive from that equation, I can derive an evolution equation for this energy.

416
00:41:18,550 --> 00:41:22,660
So I can, I can ask what is devoid of that.

417
00:41:22,930 --> 00:41:26,950
It turns out that if you tried to. So you do that by multiplying that equation by those two.

418
00:41:26,950 --> 00:41:30,370
And it so it turns out that if you do that,

419
00:41:30,850 --> 00:41:36,910
all the contributions from the nonlinear terms and from the pressure disappear and what you find is a simple balance.

420
00:41:37,450 --> 00:41:44,080
What you will find is that the change, the rate of change of energy is equal to the power that that you inject,

421
00:41:44,620 --> 00:41:51,340
which is just you don't have time to integrate that over the volume minus the power you dissipate,

422
00:41:51,820 --> 00:41:56,979
which is viscosity times integral over the square of the gradient of the velocity.

423
00:41:56,980 --> 00:42:01,270
So that that's this term multiplied by velocity. Okay.

424
00:42:02,020 --> 00:42:05,200
So, so that's that's that's the and it's unsurprising, isn't it?

425
00:42:05,200 --> 00:42:08,559
I mean, you know, you put it in, then it dissipates, right?

426
00:42:08,560 --> 00:42:13,930
That's that's the balance of energy in steady state.

427
00:42:13,930 --> 00:42:18,400
These things have to have the balance right. Whatever whatever you put in has to be dissipated.

428
00:42:18,820 --> 00:42:25,330
All right. Well, let's estimate these terms. If we estimate this terms, we can do this purely dimensionally.

429
00:42:25,510 --> 00:42:29,919
We can ask, well, how much power are you injecting into turbulence? Well, just dimensionally.

430
00:42:29,920 --> 00:42:34,870
Right. So if things move at a certain rate, which I call you are a mass, the mean square velocity,

431
00:42:35,050 --> 00:42:39,380
if they have a certain density, if, if my books has a certain size, just the.

432
00:42:39,440 --> 00:42:44,120
Mention that you could convince yourself that if you wanted to concoct a quantity that has the dimensions of power,

433
00:42:44,420 --> 00:42:50,270
it'll look like really you cube over l this l is sort of the size of the system, or rather the size of the spoon.

434
00:42:51,110 --> 00:43:01,659
That's, that's during the turbulence. If you calculate the dissipation just from here, it's density times, viscosity times,

435
00:43:01,660 --> 00:43:05,650
the square of velocity divided by the square of the size because there's a greater power.

436
00:43:06,910 --> 00:43:16,930
And if you compare these two terms, divide one by another, the ratio will be velocity times scale divided by viscosity.

437
00:43:17,200 --> 00:43:20,710
This is this is a dimensionless number known as the Reynolds number.

438
00:43:20,860 --> 00:43:23,770
And it turns out that in most of these systems, it's very, very large.

439
00:43:24,820 --> 00:43:32,500
So there is an imbalance seemingly between the injection of this equation, and nature has to find a way to correct that imbalance.

440
00:43:33,130 --> 00:43:39,160
How does it do that? Well, it does this the following way, the way to make this.

441
00:43:39,160 --> 00:43:43,750
So it's the imbalance. The imbalance is is in the direction of making dissipation smaller.

442
00:43:44,740 --> 00:43:49,090
So how do you make dissipation large? Well, dissipation has gradients in it.

443
00:43:49,090 --> 00:43:52,150
See its viscosity times of the question. Right.

444
00:43:52,270 --> 00:43:59,139
So I can make dissipation large by making scale smaller if I have smaller scales that have larger gradients.

445
00:43:59,140 --> 00:44:04,330
And so even the small viscosity coefficient will go a long way to dissipate all my motions.

446
00:44:04,480 --> 00:44:12,610
And indeed I can even dimensionally estimate how small that scale has to be in order for this to be equal to that.

447
00:44:12,880 --> 00:44:19,570
And that again is by a split of dimensional analysis, its density times, cube of viscosity divided by the injection rate.

448
00:44:19,750 --> 00:44:24,250
All of the one quarter power in that sense turns out to be the size of the system.

449
00:44:24,520 --> 00:44:27,850
Times the Reynolds number to the minus three quarters. This is called the cumulative scale.

450
00:44:28,720 --> 00:44:36,580
In that scale is much smaller than the scale of the system because the Reynolds number is much greater than the scale than one.

451
00:44:37,030 --> 00:44:40,900
And so we have a scale of injection at which we're putting the energy in.

452
00:44:41,170 --> 00:44:47,650
We have a scale of dissipation in which this energy is dissipate, and then somehow that energy has to get from large scales to small scales.

453
00:44:48,190 --> 00:44:54,009
These intermediate scales is something called the initial range. Now, how does it do that is the next question, right?

454
00:44:54,010 --> 00:45:00,010
How do you you set up a large motion. How do you convert that large motion into tiny little things?

455
00:45:00,700 --> 00:45:03,580
Well, it turns out that in most cases, in nature,

456
00:45:03,820 --> 00:45:12,910
the way you do this isn't by somehow directly transferring energy from an enormous swirl into tiny little blips of fluctuations.

457
00:45:13,210 --> 00:45:18,940
What you do is you you make a large motion. Then that motion breaks up into motions that are half the size,

458
00:45:18,940 --> 00:45:24,070
and then it breaks up into motions that are half the size of them, and then half again and again and again.

459
00:45:24,280 --> 00:45:34,390
So this is called the turbulent cascade. And this is an idea that was probably first due to Louis Fry Richardson.

460
00:45:36,010 --> 00:45:41,329
This is this is an illustration of what I've just said. You start with large with large motions, you break into small motions of breaking.

461
00:45:41,330 --> 00:45:43,240
It's more of the motions and small motions and so on.

462
00:45:44,350 --> 00:45:52,149
In terms of, you know, these spectral power laws, you could think of energy going from large scale to smaller scale, smaller scale, smaller, smaller.

463
00:45:52,150 --> 00:45:55,810
So this is key. The wave numbers are large, came in small scale.

464
00:45:56,440 --> 00:45:59,950
He actually didn't put it that way. He put it in inverse.

465
00:46:01,030 --> 00:46:04,929
He said big whirls have little whirls that feed on their velocity and little will also have lesser rules.

466
00:46:04,930 --> 00:46:11,880
And so on. The viscosity easy to remember he this is this this is actually a piece of plagiarism.

467
00:46:11,920 --> 00:46:16,780
Well, maybe not quite. It's a piece of being inspired, cross-disciplinary.

468
00:46:18,430 --> 00:46:25,389
So he was he was inspired. Cross-disciplinary by Jonathan Swift, who was inspired.

469
00:46:25,390 --> 00:46:33,760
Cross-disciplinary by by Hook and Levin Hook, who in those days won one in Britain, one in Holland invented the microscope.

470
00:46:34,000 --> 00:46:38,170
And we're looking at all kinds of thing. And once the intensity microscope, they put all kinds of things under it.

471
00:46:38,500 --> 00:46:42,070
And what they found is this amazing, microscopic world that you couldn't see.

472
00:46:42,220 --> 00:46:48,879
And in particular, you could see, you know, two things. They looked at the fly and then the fly had little flies on top of it and so on.

473
00:46:48,880 --> 00:46:50,230
And so Jonathan Swift said,

474
00:46:50,230 --> 00:46:56,410
So naturalist observe a flea had smaller fleas that only prey and these have smaller yet to point them and so proceed on the ad infinitum.

475
00:46:56,410 --> 00:47:00,850
It's kind of a cascade that's every point that he's quite this bit by him that comes behind.

476
00:47:01,030 --> 00:47:05,889
That's not true about the theoretical physics, so it's only the points that have that property.

477
00:47:05,890 --> 00:47:10,270
Theoretical physicists are collegiate, but no backstabbing whatsoever.

478
00:47:11,440 --> 00:47:20,019
No. The last bit that that that I want to talk to you about is how do you actually make this all more more quantitative right.

479
00:47:20,020 --> 00:47:26,220
So we'll go from verse to mathematics and nobody was better it going from poetry to mathematics.

480
00:47:26,230 --> 00:47:32,080
And then then Komarov, a great Russian mathematician who died in 1987.

481
00:47:33,160 --> 00:47:39,250
And that what he said is this is actually one of the greatest contributions to physics by a mathematician.

482
00:47:39,250 --> 00:47:44,830
I think this this this bit and the and and it all and it all fits on one slope.

483
00:47:45,460 --> 00:47:49,000
So so what did he do?

484
00:47:49,180 --> 00:47:52,690
He basically said, well, let's let's think of the system in the following way.

485
00:47:52,930 --> 00:47:55,160
First of all, let's let's. Let's think.

486
00:47:55,610 --> 00:48:01,790
I mean, if you notice if you noticed what I showed you about this picture, they all had power loss and they all had the same power.

487
00:48:02,360 --> 00:48:09,200
So there was a suggestion there that there's some universal physics going on in in the process.

488
00:48:09,200 --> 00:48:13,130
Right. So. So McConnell said, well, let's assume that there are no special systems.

489
00:48:13,490 --> 00:48:17,030
That all systems are sort of roughly the same. That everything is homogeneous.

490
00:48:17,030 --> 00:48:21,439
There are no special locations in the system. So so it's the turbines in one place.

491
00:48:21,440 --> 00:48:28,729
So the system is no different from turbulence in another place statistically that everything is isotropic, meaning there are no special directions.

492
00:48:28,730 --> 00:48:33,320
We'll consider a system where that way is the same as this way, in the same exact way, in the same exact way.

493
00:48:33,620 --> 00:48:40,340
And everything is local in scale space in the sense that as you break up these motions,

494
00:48:40,550 --> 00:48:47,660
there is nothing special about any scales in between the injection scale and the dissipation scale that that is sort of a natural assumption to make.

495
00:48:47,840 --> 00:48:53,840
And then what we wish to predict is, you know, if you if you thought about how you would measure turbulence, well,

496
00:48:53,840 --> 00:49:00,740
what you would probably do is you would stick two probes into it, registers the velocity at one probe rich to the velocity and another probe.

497
00:49:01,490 --> 00:49:07,340
Right. And then ask how the difference between these velocities scales with the distance between the probes.

498
00:49:07,910 --> 00:49:11,900
Right. That's a that's a very natural thing to ask about this.

499
00:49:12,170 --> 00:49:18,350
And and so and so what we want to predict is this build that you at scale out as a velocity difference.

500
00:49:19,220 --> 00:49:26,690
And so then at each scale you could say because there is look, it's each scale can only depend on things that happen at the same scale.

501
00:49:26,960 --> 00:49:30,110
And so you could see the energy associated with that scale,

502
00:49:30,110 --> 00:49:35,569
which is zero times the square of that quantity divided by some time that it takes for

503
00:49:35,570 --> 00:49:42,410
that structure to break up must be equal to the power that's flowing through your system.

504
00:49:42,680 --> 00:49:51,620
Because you've injected the certain amount of power there, you're going to dissipate it at a much smaller scale, at all the intermediate scales.

505
00:49:51,620 --> 00:49:56,690
You just have to transfer it from one structure to another to it to have a small scale so dimensionally.

506
00:49:57,050 --> 00:50:03,890
This should be the case. And then again, dimensionally the only thing that this time can be.

507
00:50:03,890 --> 00:50:09,680
And I've already introduced that notion sort of surreptitiously, you know, when I talked about typical diffusion dimensionally,

508
00:50:09,680 --> 00:50:14,990
what that time has to be is the size of the structure divided by the velocity at which it which it turns over.

509
00:50:15,530 --> 00:50:25,580
And so if you plug that into here, what you find is rho delta, you cube overall equals to the injected power, which is a constant.

510
00:50:25,790 --> 00:50:33,380
And so they'll tell you the velocity difference across a scale is proportional to the one third power of the scale.

511
00:50:33,650 --> 00:50:37,730
So it's almost a miracle, right? Something out of nothing. We started with a complete mess.

512
00:50:37,940 --> 00:50:44,329
We've just predicted a scaling law for, you know, based on on all these assumptions.

513
00:50:44,330 --> 00:50:47,600
This this corresponds to the incidentally, to the 5/3 spectrum.

514
00:50:47,810 --> 00:50:51,190
I can if anybody wants to know how exactly that works out, explain it.

515
00:50:51,500 --> 00:51:00,170
But the point is that we now know that the typical amplitudes depend on scales in the following way.

516
00:51:00,860 --> 00:51:05,360
Well, recall, you know, going back to the transport recall.

517
00:51:05,360 --> 00:51:10,700
Well, I told you two of them diffusivity is u times l, right.

518
00:51:10,880 --> 00:51:12,530
Very basic on a very basic level.

519
00:51:12,530 --> 00:51:22,670
I mean this estimate that means that so and we now know that you is proportional to L and so the whole thing is proportional to L to the fourth.

520
00:51:22,690 --> 00:51:26,360
Third. So so it's bigger when the scales are bigger.

521
00:51:26,630 --> 00:51:32,209
What this tells us is that the largest scale eddies will make the largest contribution to the twin.

522
00:51:32,210 --> 00:51:34,940
Transport is perhaps intuitively obvious, but.

523
00:51:35,090 --> 00:51:42,170
But with the nice conclusion, the interesting practical question about the answering of which I think you'll hear from Steve is,

524
00:51:42,620 --> 00:51:48,200
is, is what that scale is and how fast these edges are and what you can do to reduce both of those things.

525
00:51:49,520 --> 00:51:53,600
Well, before I conclude, let me tell you that things are not quite as simple as I have described.

526
00:51:54,530 --> 00:51:59,450
Uh, there are complications. First of all, to consider the tokamak is not homogeneous.

527
00:52:00,260 --> 00:52:02,310
Conditions vary with radius, right?

528
00:52:02,330 --> 00:52:09,890
So if this is a tokamak, this is a picture of a typical tokamak took two months would actually be different depending on where you are in this area.

529
00:52:09,970 --> 00:52:15,500
So what we do is we theorise of simulate locally on the magnetic surfaces we have,

530
00:52:15,620 --> 00:52:20,389
we take these magnetic fields that wrap around tourists and that in fact,

531
00:52:20,390 --> 00:52:26,870
what we do is, is, is we take something called the flux tube that moves, that wraps around the door.

532
00:52:27,240 --> 00:52:31,640
And then we look at the turbulence in that curvilinear flux to improve linear coordinates.

533
00:52:31,910 --> 00:52:37,489
There is an absolutely spectacular movie that I have to show you, which was also made by Edmund Haycock.

534
00:52:37,490 --> 00:52:42,030
It's what theoretical physics does to this to this flux tube.

535
00:52:42,050 --> 00:52:46,610
It's so. So how do you reduce it to something we can understand?

536
00:52:46,610 --> 00:52:51,020
Well, it gets transformed into a box and there we go.

537
00:52:51,950 --> 00:52:55,470
Right. And so that when we. All these boxes.

538
00:52:55,710 --> 00:53:02,460
They were not actually boxes in the context of the document. They were those flexed and done in accordance.

539
00:53:03,390 --> 00:53:10,049
Secondly, it's not isotropic. Things are actually very stretched out along the magnetic field.

540
00:53:10,050 --> 00:53:14,550
So all these structures I talked to you about, they're not just little girls and squirrels.

541
00:53:14,550 --> 00:53:17,610
They are actually sort of worms. Right.

542
00:53:17,610 --> 00:53:20,120
So they're whirls and swirls in this direction. So.

543
00:53:21,180 --> 00:53:26,250
But but the law, you see, they have they have a very long coherence length along the magnetic field.

544
00:53:26,790 --> 00:53:31,980
So that brings in lots of interesting new physics. And finally, this is not even happening in three dimensional space,

545
00:53:32,310 --> 00:53:37,440
because if you remember the end of Felix's talk, plasma is actually a six dimensional object.

546
00:53:37,440 --> 00:53:42,480
It's it's described by the probability distribution function of particles which depend on positions and velocities.

547
00:53:43,080 --> 00:53:45,390
This lives in the six dimensional phase space.

548
00:53:45,810 --> 00:53:55,110
And so the small scales that I talk to you about are created not as I described, for water or air in space, but in three space.

549
00:53:55,140 --> 00:54:03,390
Right. These are six dimensional small structures. And and you go through large gradients in both velocities and in space.

550
00:54:04,170 --> 00:54:09,660
You don't want to know what the complication of before is. Anyway, story so far.

551
00:54:09,690 --> 00:54:15,840
Let me summarise. So we want to build a machine to tap the energy that fuels the stones.

552
00:54:16,860 --> 00:54:22,560
And inside that machine, we have a plasma that's locked in a cage and kept out of equilibrium,

553
00:54:23,250 --> 00:54:31,860
but inside called on the outside it rattles its cage, it breaks into whirls and swirls, and it's trying to regain equilibrium this way.

554
00:54:32,130 --> 00:54:38,010
It's rather good at doing that. And so to keep it and to keep it hot, we must stay in the nonlinear beast.

555
00:54:38,400 --> 00:54:44,460
And so that, you know, it's a beast, let's see if this is going to work. So this is a this is what I'm going to show you.

556
00:54:44,520 --> 00:54:49,320
I hope it's loud enough. So so this is this is what the transmission sounds like.

557
00:54:50,610 --> 00:54:57,030
This is Anthony Field, who is a experimentalist, have called them put a piece of plasma to different observations of sound.

558
00:54:57,690 --> 00:55:28,320
And this is this is what it is. Let's see. Um.

559
00:55:30,470 --> 00:55:31,320
So it is a beast.