I recently had a paper about electrodynamic braking accepted by Wind Engineering. Electrodynamic braking is a fancy term for stopping a wind turbine by short-circuiting the generator, usually with some simple resistance across the terminals. The work was inspired by a system failure back in 2010. I realised that although some people in the industry knew something about why this often doesn't work as expected, many manufacturers had a poor understanding of the limitations of the technique.
The first rule of thumb is, don't do it for power control - although it's an attractive possibility, most radial small wind turbine generators have a peak torque capability well below the instantaneous torque exerted by a small wind turbine rotor in high wind speeds, i.e. at around 15 m/s and higher. If you have complete confidence in your understanding, then it can be possible to use electrodynamic braking as part of a control strategy but you should think about an axial flux design.
The aim of the paper is show how straightforward analysis can help designers understand the limitations of electrodynamic braking.
Several well-known wind turbine manufacturers have deployed electrodynamic braking as part of wind turbine control strategies. Available systems include the Xzeres Skystream 3.7 (3.72 m rotor diameter, 2.1 kW rated at 11 m/s) series (The Air and Skystream systems were developed by the famous Arizona-based Southwest Windpower. The company closed down in 2013.) and the larger Evance R9000 (5.5 m rotor diameter, 4.7 kW rated at 11 m/s) . The Evance R9000 uses electrodynamic braking of an axial flux generator in conjunction with a passive pitch control system.
At this point, I'd encourage you to look at the paper!
The paper went through a few revisions and there are many details omitted.
One of the central points is that it's enough to know that the position of the peak short-circuit torque relative to RPM can be changed by adding or subtracting a resistance across the generator terminals; a straight line traced by the peak torque as the resistance changes represents the limit of the generator's braking ability.
Many simplifications were made - for example, it's right to think that adding in a resistive component across the terminals results in energy dissipation as heat in the resistive component rather than the stator windings. However, as a worst case estimate, which is what is under consideration, it is sensible to assume that most of the heat will be dissipated in the generator. This is the case when then short circuit happens at the generator, the most common way of implementing electrodynamic braking.
It's also an assumption to say that there's an even split in heating between the magnets and the windings. It's the windings that dissipate the energy in reality but for the purposes of this paper, i.e. considering the edge cases, an even split is not a bad way to think about it. It amounts to considering the stator windings and the magnets as a single unit. It is the case, after all, that the rotor magnets do heat as a direct result of braking due their proximity to the windings.
In reality, a short-circuit is a pretty transient event, with very unsteady currents. The aim was to show how a sustained braking event at an almost constant RPM, which almost happens in real situations, can result in a failure event and to show how a more traditional applied mathematical approach can provide useful insight. The speed does not drop.
The mid-point of the air-gap was used as an imaginary point of action for the resulting forces for two reasons: (i) it is nicer theoretically to imagine the forces acting here; (ii) all that is important for scaling is a fixed reference radius. The mid-point is a nice reference point.
There is also the assumption that the number of poles remains constant with increasing radius.
Finally, the higher winding voltages associated with higher velocities and higher winding resistances or inductances - there are more wires - both mean reduced short-circuit currents. The result can be a deviation, in reality, from the estimates presented here.
Nevertheless, the analysis is useful as a rough guide to short-circuit braking performance. Nothing replaces the need for rigorous testing but a little calculation up-front can save time and a lot of money.