I read yet another article about this issue in the paper today which, like all the others, does not explain why a solar storm should damage the grid. After some searching, I found a rather technical slideshow that seems to explain at least some of it. From what I could understand of this, the problem is that the neutral line in a 3-phase system is grounded, and solar storms induce something called "geomagnetically induced currents" along the surface of the Earth. These cause what is effectively a DC current flow, up one grounding point and down another one, hundreds of km away - clearly not something the transformers and grid wires are designed for. What I did not understand is there was some further talk of excess induced magnetic fields, saturating transformer cores and thereby causing magnetisation of the bodyshells of transformers, creating heating from eddy currents and....damage. But if these currents are either DC or changing very slowly I don't understand where the eddy currents come from. Can anyone confirm or correct the above and explain about the eddy current issue?
My limited understanding is that large coronal mass ejections deform the magnetic field and the deformation varies quickly over time due to varying density of the charged particles hitting the magnetic field. The deformation means the power lines are in a moving magnetic field so there is an induced current in the lines. The part about the transformers is news to me. It seems far fetched that there would be that much of a difference in the magnetic field over the size of a transformer. But I've been wrong before!
Ah, that is a different mechanism entirely, then. The slide presentation I found talked about induced currents along the surface of the ground. The eddy currrents in transformer shells were due to the iron cores of the transformer becoming saturated, as a result of the extra magnetism induced by this current flow (through the neutral grounding connections), forcing magnetic flux outside the cores, into other components. Here is a link to the presentation: https://www.swpc.noaa.gov/sites/default/files/images/u33/finalBoulderPresentation042611 (1).pdf Where's billvon when you need him? Please Register or Log in to view the hidden image!
Hopefully someone will come along that can clear this up, maybe theorist will come back and straighten this all out for us.Please Register or Log in to view the hidden image!
It's guaranteed he'll be back. Just a question of how long he leaves it. Please Register or Log in to view the hidden image! But I'm getting intrigued by this solar wind issue. I'm now wondering how it is that an emf is induced along the surface of the Earth that is not offset by an equivalent emf being generated in the transmission wires, that would prevent a current flowing. All a bit mysterious. If you come across anything that sheds light on this, do share it.
Yes, although it's not the neutral - it's the actual current in the transmission lines. The frequencies range from 1Hz to much lower than that. The AC portion tends to saturate the transformer since transformers care about volt-seconds, and if you make the "seconds" part of that formula very long, it doesn't take much potential to saturate the transformer's core. The DC portion flows from the lines through the transformer to ground. That also saturates the transformer by moving the operating point of the core very far in one direction. Then, once the core is saturated, the windings no longer present much impedance to the incoming power, and currents much greater than the transformer was designed for can flow. This causes a lot of heating. On the plus side this effect can be protected against by overcurrent devices. The really insidious damage happens when the core saturates but the windings don't see excessive amounts of AC current - because then the overcurrent devices don't trip but the transformer still overheats. AFAIK 99.99% of the damage comes from the DC and VLF AC current induced in transmission lines and not any currents induced in the transformer itself.
Thanks for this. So the "core" issue (as it were Please Register or Log in to view the hidden image!) to take away is saturation of the magnetic cores of transformers by this very low frequency AC or quasi-DC induced current. I can certainly see that if the core is saturated the thing ceases to absorb power from the input side and then the current will run out of control, rather like a short-circuit. What now intrigues me, though, is the mechanism by which which solar storms induce these currents. I found this article on Wiki, which may bridge the apparent difference between what you are saying and the presentation I read: https://en.wikipedia.org/wiki/Geomagnetically_induced_current According to this, the solar storm causes the Earth's magnetic field to vary, which induces an electric field at the surface of the Earth. This will cause a Geomagnetically Induced Current (GIC) to flow "in any conducting structure grounded in the Earth". It seems from the article that it is indeed induction in the Earth that is responsible for the GIC in electricity systems and that it is the grounding of these systems that is the cause of the trouble. From then on, it is all about transformer saturation and the damage this can create, along the lines you describe. The volt-second relationship you refer to is not something I knew about. It looks from this article as though there can be ways to mitigate the effects through introducing reactive components. It also suggest that grids at high latitudes are more susceptible, due to the greater degree of magnetic field fluctuation closer to the magnetic poles, I suppose. Cool.
An obsolete and often misleading notion best avoided: http://bobweigel.net/wiki/images/1/18/Falthammar_MovingFieldLines_2007.pdf
?? Moving magnetic fields are how motors and generators work. Without a constantly moving magnetic field, motors would not be able to generate torque.
A rotating field pattern is not the same as saying the field itself is rotating. Apparently you never read the article linked to in #8. Maybe this one makes it easier: https://physics.stackexchange.com/questions/217332/moving-magnetic-fields
This is a subtlety I had not appreciated. I'll need to read the articles more carefully to understand the point being made. I can see that a rotating field pattern can be formed like a wave, i.e. by the field vector values at each point in space varying with time. As with a wave, the "medium", i.e. the field, is not moving, just varying et each point. However where I'm struggling is with the how this helps in the case of a conductor moving through a uniform magnetic field and having a current induced in it. Classically, one speaks of it "cutting flux", though this is obviously just a graphical representation. But here the field values are not changing, i.e. there is no "wave", but a current is still induced. I'll read further.......
No question there are situations where thinking of the B (or E) field of a moving magnet (or capacitor) as moving with the source makes it easier and can give correct results. But not consistently correct physics in general. One compelling example illustrating when it fails is that of a parallel-plate capacitor in the form of two circular discs with a small gap between them. Charged to give an almost perfectly uniform E field normal to the disc faces, and spun up to a constant angular velocity about the axis of symmetry. If the E field lines really moved with the rotating charges, one would have a radial B field varying in strength in direct proportion to the radius r from spin axis, at that location. This would blatantly violates the Maxwell-Gauss law ∇.B = 0. http://www.maxwells-equations.com/ The correct analysis is that B is generated by circumferential convection currents in the two charged discs. Yielding then the correct result satisfying ∇.B = 0. There are various other instances where 'moving fields' give erroneous results, but one is enough here. The moral is to always rely on Maxwell's equations together with the Lorentz force expression. One caveat that some will take exception to - it's possible to show that under certain conditions the fully relativistic field equations do predict subtle deviations from two of the Maxwell equations. I won't be elaborating here. [Missed answering your query re 'flux cutting' when a conductor moves through a uniform magnetic field. Formally the magnetic Lorentz force law handles that, F = qv x B. In the rest frame of moving conductor, it 'sees' an electric field E' = v x B' that results from differential Lorentz contractions of conduction charge linear densities, in opposite portions of the current source(s) generating the uniform B, having current components parallel to v. A square current loop as source of B, with two opposite sides parallel to v helps to see that.] Given the concentration on transformer action earlier, it's worth noting that for an ideally wound toroidal transformer with secondary the outer winding, there is exactly zero 'flux' cutting' across secondary windings owing to primary + core magnetic field. Instead one simply has E = -dA/dt, where A is the magnetic vector potential at the secondary windings.
Sure, the relativistic treatment of it makes it fairly clear why induction occurs in the uniform field scenario I was talking about. I have to say that the twin, but apparently separate, ideas that both changes in magnetic flux density, and the "cutting" of flux, both give rise to induction of a current, always rather bothered me at school in the 6th form. It did not seem obvious how to harmonise the two. But I just got on with it. Anyway this is all a digression from the thread topic, really. But still interesting.
Well there is a fundamental distinction in that curl E = -dB/dt generates an emf around a circuit, whereas 'flux cutting' of a circuit in uniform translational motion in a uniform applied B does not. [a rotating circuit will have such an emf induced but again, recourse to the magnetic Lorentz force is the recommended approach.] I have seen silly explanations from qualified physicists as to how a Faraday homopolar generator is supposed to be derivable from curl E = -dB/dt, but it cannot be. The analysis has to be done in parts to get a consistent picture. OK back to Solar Storms.
The field itself is rotating. Take the simplest example - a permanent magnet. It has a magnetic field associated with it. You can draw it as field lines. You can graph it via magnetic flux density. (In both cases it is important to remember it is a vector, not a scalar, field, so such representations are necessarily simplified.) When you rotate the magnet, you rotate the field. You could say "no, it's not the field that's rotating, it's the field PATTERN that's rotating" - but that's akin to saying "you are not rotating the magnet - you are rotating the PATTERN of atoms that make up the magnet." That's not really valid. Now look at an electric motor. The field generated by a three phase motor (synchronous, induction, BLDC, any three or more phase motor actually) and you will see a field identical to the rotating magnet case. The field itself, in whatever visualization you want to use, will rotate.
I'm well aware of that EE view of rotating fields, and I once saw it that way too. It's not the view of competent physicists, and I gave a specific example in #12 where the 'moving field' pov leads to gross error. That first linked to pdf article gives more examples. That second linked to webpage provides a simplified definition of field that doesn't allow it to 'move' or 'rotate'. Of course the magnet rotates in your example, but the field does not. If you want consistent physics not ad hoc physics, that's the pov to be adopted.
To this chemist, this looks like one of these numerous cases in physical science where one has two models, one of which is simpler and gives correct results in most cases encountered in practice, and a more sophisticated one that has to be invoked from time to time where the simpler one breaks down. But I see, from reading further, that the mechanisms involved in induction via relative motion and that by rate of flux change do really seem to be different things, whatever model one uses.
A fair summary. Depends. In the case of straightforward transformer action between magnetically linked primary and secondary windings, a change of reference frame won't reveal any neat relativistic explanation and in fact just adds complication. Instead work from the basic field definitions in terms of φ (scalar potential) and A (magnetic vector potential): E = -∇φ -∂A/∂t B = ∇ × A Of which only E = -∂A/∂t is relevant for transformer action. Which situation remains true for the different case of time varying flux linkage via rectilinear relative motion of a coil wrt say one end of a magnet. Provided coil relative motion is strictly along the magnet magnetization axis. An interesting hybrid situation is that of the familiar 'primitive alternator' where a typically rectangular coil uniformly rotates within a uniform applied B field such that the flux linkage is periodic, i.e. rotation axis normal to applied B field. Back in #14 recourse to the formal magnetic Lorentz force expression was invoked. To try and explain it in the rest frame of either coil leg running parallel to the rotation axis, one finds the contributions to induced emf are generally split between both rhs terms in E = -∇φ -∂A/∂t. Relative contributions depending on the geometry of applied B source. I won't elaborate - leave it as an exercise to ponder.Please Register or Log in to view the hidden image!
