Coil to coil clearances in high voltage generators 2

[edison123]

Does anyone know of any standards / references about the coil to coil clearances required in the end winding zone for 11 KV Generators with multi-turn diamond coils having resin rich insulation? Thx in advance.

 

[aolalde]

Edison:

There are no standards for such coil to coil clearance. Each manufacturer has researched to find proper figures of the space between adjacent end turns or diamond area.

In my experience for B-stage resin rich hot pressed coils between 10KV and 12.5 KV the space is 0.400” and for 12.5 to 15 kV the space coil to coil is 0.450”.

The larger the space the longer the coil extension, which will be limited by the end bracket or baffle proper clearance.

 

[electricpete]

I have nothing definitive to contribute. However just an observation. The breadkdown strength of air is approximately 30 kv per inch. That equates to approximately 0.5" for 15 kv (phase-to-phase).

Since the end-turn comes up to line voltage, it would seem logical that we have to provide some margin below this voltage stress to avoid corona.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[justinn4]

Epete:

In my opinion the insulation layer around each coil side modifies the electric field strength in the air portion. It is not the same field distribution as compared to two bare conductors in free air.

Besides the insulation thickness on both coil sides adds spacing from copper to copper conductors.

Certainly the field gradient is higher in the air portion, inversely proportional to the dielectric constant. Consider k=1 for air and around 6.3 for resin impregnated mica paper.

 

[electricpete]

justin

It is a good point that the insulation has a bearing on the problem.

A first pass calculation would be a parallel plate capacitive voltage divider.

Let subscript i be solid insulation and a be air.
e = permability
X = capacitive impedance
d = depth
A = area

Voltage divider
Vair = Vtotal * Xair / Xtotal
Vair = Vtotal * Xair / [Xair + 2*Xinsulation]

Xair = 2*pi*f * (da / [ea*A]
Xinsulation = 2*pi*f * ei A / di
factors of 2*pi and A will not show up in the ratio Xair/Xtotal

Vair = Vtotal * da/ea / [da/ea + 2*di/ei]

Typical values:
da = 0.5", di = 0.25"
typical insulation ei ~ 4* ea

Vair = Vtotal * da/ea / [da/ea + 2*di/ei]
Vair = Vtotal * 0.5"/ea / [0.5"/ea + 2*0.25"/(4*ea)]
= Vtotal * 0.5 / [0.5 + (1/8)]
= Vtotal * 0.5 / [0.5 + 0.125]
= Vtotal * 0.5 / [0.625]
~ 80% Vtotal
For your value ei~6*ea, the air voltage would be a higher percentage of the total.

My approximation was that Vair~Vtotal - all of the voltage drop appears accross the air. For that particular approximation we consider the distance from coil surface to coil surface and distance to copper is not relevant. For fuller solution you are right it is relevant.

There are I'm sure also very important geometry factors to consider (effect of localized stress concentration). Also if we design for an average gap, there will be some above and some below and ideally would like to prevent corona on them all.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[edison123]

Thx for your respsonses.

In this case, the coils have an non-uniform end winding clearances varying from 1 mm to 4 mm max. Also, the knuckle drop varies from 22 to 44 mm whereas as per design it should be 36 mm. I just can't accept these coils with these wild variations in shapes. Any suggestions that the coils are ok and can be used without any problems?