Thermal expansion of winding coils in a rotor 3

[Shark96]

I wanted to know if the winding coils in a generator rotor expand radially due to thermal effect? Axial expansion is known but wanted to know about radial expansion due to temperature at high currents.

 

[electricpete]

I'm not a generator specialist. Some thoughts fwiw

The differential temperature and differential coefficients determine the fractional expansion and the copper conductors would be predicted to expand more than the iron core on both counts.

The predicted expansion in the radial direction is a lot smaller than axial because the fraction applies to a smaller dimension.

For the large 4-pole turbine-driven generator rotors that I'm vaguely familiar with (smooth rotor, hydrogen cooled) the rotor turns are actually hollow copper conductors cooled by hydrogen on the inside. So that means even if there is constraint from outward expansion, the resulting stress is somewhat limited by the hollow construction.

The stress generated by radial expansion can be relieved to some extent through expansion in the axial direction if axial expansion is not restricted.

 

[waross]

Anecdote Alert!
An interesting failure of the exciter on a small gen set.
The exciter armature was good when cold.
The exciter armature partially shorted when hot from running.
The symptom was excessive sparking at the brushes.
Normally sparking may be minimized by shifting the brushes, but that did not help in this instance.
This was a remote site in the Yukon Territory.
The armature was taken to a shop in Whitehorse.
In the shop, the armature was tested by heating in an oven.
It did not fail, and the shop declined to rewind an armature when they could not locate a fault.
Back on site with the armature re-installed.
Performed a brush neutral set-up.
That is,energize the field with 120 Volts AC and shift the brushes for null voltage.
Ran the machine. After a short while, the sparking started.
Re-did the brush neutral setup.
The brushes still arced.
Re- checked the brush neutral point.
The test indicated that the neutral was set wrong.
Eventually, we left the exciter field energized with 120 Volts AC and slowly turned the engine over while monitoring the voltage across the brushes.
As the engine (and the armature) rotated, the voltage across the brushes varied. It should have remained constant.
We took the armature back to the shop and described the test.
The shop accepted our results as conclusive and rewound the armature.
Conclusion.
In service, heat was generated in the windings which then transferred the heat to the iron.
As the windings heated first, there was differential expansion and the slight movement of the windings relative to the iron caused the short.
In the oven, the iron and the copper heated up together and there was no differential expansion and so the winding did not fail under the oven test.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[edison123]

Radial expansion of copper in smaller HP motors is not a cause for concern. For large MV motors and generators of MW capacity, it's a standard practice to provide ripple springs below wedges to accommodate radial expansion both in stator and rotor.

Muthu

http://www.edison.co.in

 

[electricpete]

I didn't realize there are ripple springs in the rotor (it may be the case, as I said I'm only vaguely familiar with our generator).

 

[wcaseyharman]

Not a designer, but have been the customer engineer on several stator and rotor rewinds of 50MW to 100MW generators in the last few years, so not an expert at all, but perhaps have enough experience to maybe be able to answer the question.
As was mentioned before, in the stator the armature coils (or bars depending on design) are held in place with epoxy wedges, spacers and commonly epoxy “wavy” pieces called a “ripple spring.” Any expansion will slightly compress the spring. Specs call for the spring be compressed somewhere between 80 and 90 percent (if I recall correctly).
On a high speed 2 pole 3600 RPM round rotor the metal wedges that hold the rotor coils are commonly loose to account for expansion(depending on the rotor design). They are held right by the centrifugal force at running speed. This has caused us issues when the machine is on “turning gear” (a slow roll to keep the turbine shaft from having a bow) for long periods as the winding moves as the rotor turns. This can abrade the winding, creating conductive dust and possible shorted turns. I think some other designs, particularly when the generator is used as a starting means, have springs under the wedges to keep them tight at slow speeds - I have been told this is done to prevent arcing between the wedges to the end (retaining) rings when the machine is started as an induction motor and the surface of the rotor carries significant current.
On the hydroelectric generator we rebuild last year the manufacturer accounted for pole copper expansion by providing some space and allowing the inner pole collar to move. To keep the pole tight 12 springs were placed around the pole collar to squeeze the pole but allow for expansion. The new design that we were provided when we sent the poles in for rewinds uses RTV between the pole collar and pole body to perform a similar function.
I apologize for all the terminology, but it’s hard to describe the internals of generators without using it. Hopefully this helps.

 

[Gr8blu]

Shark96: When anything is heated, it "grows" in all directions - X, Y, and Z. How much it grows in each is different; in general, more heat can radiate out through a larger cross-section surface, which means thermal expansion in that particular direction will be a bit less than in a direction with less exposed surface. It also happens that the less restriction there is to thermal growth, the more growth will occur in that direction (think of it as "reduced friction").

CASE 1 - A conductor in a slot. The circumferential direction is the largest area (coil height x slot length) and it has the most restriction: the steel of the core is not going to move and the coil is usually fairly tight to the slot dimension. Radially, less surface is available (coil width x slot length), but the radial build in the slot tends to use materials that are less restrictive to growth (i.e. insulation materials and/or wedges). Axially, the coil can grow pretty much unrestrictedly and the available heat radiation surface is the smallest (coil width x coil height). The net result is that the coils grows furthest in a axial direction - due to the coefficient's definition (unit growth per unit length).

CASE 2 - A coil wrapped around a magnetic medium. The largest exposed surface is the outside coil parallel to the shaft (coil height x pole length). The next largest is the end of the magnetic surface (coil height x pole width). And finally, there is the growth in the third direction (e.g. "through" the multiple coil layers, from the pole body to the outside surface). The net result is that the coil will "grow" in all three directions, since there isn't really any restriction on growth.

A good stator (and rotor) design will account for expected thermal growth in all three directions. Note that although the coil grows first (less mass, direct heat applied through current, etc.) the steel will as well until the whole mass reaches thermal equilibrium.

Converting energy to motion for more than half a century