Totally enclosed panels

[Besoeker]

I'm bidding for a small project. It's just one 5.5 kW (about 7.5 HP) VSD running at a maximim of 4.6kW plus a few controls in a box.
The box has to be totally enclosed non-ventilated.
IP65 for UK and Europe. Probably NEMA 3X or higher.
And it has to be stainless steel.
Total losses will be around 150W all of which will have to be dissipated through the walls of the enclosure.
I'd like to limit the temperature rise inside the box to no more that about 10degC (18degF)
Box dimensions will be dictated by required surface area to dissipate the heat rather than by the components it needs to house.
If the box was mild steel with two part epoxy coating, typical for most standard enclosures, I've been there and done that. But stainless?
My question is what surface area does my stainless steel box need to have?
Any information, Imperial or metric would be appreciated.

 

[weressl]

I'm bidding for a small project. It's just one 5.5 kW (about 7.5 HP) VSD running at a maximim of 4.6kW plus a few controls in a box.
The box has to be totally enclosed non-ventilated.
IP65 for UK and Europe. Probably NEMA 3X or higher.
And it has to be stainless steel.
Total losses will be around 150W all of which will have to be dissipated through the walls of the enclosure.
I'd like to limit the temperature rise inside the box to no more that about 10degC (18degF)
Box dimensions will be dictated by required surface area to dissipate the heat rather than by the components it needs to house.
If the box was mild steel with two part epoxy coating, typical for most standard enclosures, I've been there and done that. But stainless?
My question is what surface area does my stainless steel box need to have?
Any information, Imperial or metric would be appreciated.

Have you tried Rittal?

http://www.rittal-corp.com/software/

 

[petersonra]

http://www.hoffmanonline.com/stream_document.aspx?rRID=162538&pRID=162533

looks like you need about 300 square feet of surface area.

thats one big cabinet.

btw, many vfds have the capability to have the heat sinks mounted through the wall of the enclosure. HINT.

also, a fan inside to stir up the air helps speed up the heat transfer, but I am not aware of any charts or graphs that show any correlation for temperature rise and fan size for such an install.

Why not get a 150 watt heater and put it in a box and see what the difference in rise is between a box with a fan and one without. will be a cheap experiment.

I have used liquid cooling or sealed heat exchangers for such an application.

I'm bidding for a small project. It's just one 5.5 kW (about 7.5 HP) VSD running at a maximim of 4.6kW plus a few controls in a box.
The box has to be totally enclosed non-ventilated.
IP65 for UK and Europe. Probably NEMA 3X or higher.
And it has to be stainless steel.
Total losses will be around 150W all of which will have to be dissipated through the walls of the enclosure.
I'd like to limit the temperature rise inside the box to no more that about 10degC (18degF)
Box dimensions will be dictated by required surface area to dissipate the heat rather than by the components it needs to house.
If the box was mild steel with two part epoxy coating, typical for most standard enclosures, I've been there and done that. But stainless?
My question is what surface area does my stainless steel box need to have?
Any information, Imperial or metric would be appreciated.

 

[Jraef]

Quick and dirty rule of thumb that has always worked for me is to use 4 times the volume of a NEMA 1 (IP20) ventilated enclosure. No science behind it, it's just empirical; something I learned from an engineering mgr of a VFD manufacturer 20+ years ago (Parametrics if anyone is interested) and it has never failed me, as long as the enclosure is not in direct sun.

That said though, what I have done to prove it is to use the Hoffman or Rittal software that is intended to sell you on an AC unit or a heat exchanger, and keep increasing the enclosure size until it no longer says you need one! Then take the total volume and re-adjust by what dimensions you can live with, i.e. you may not be able to accommodate a 90 inch high box, so it will have to be wider or deeper.

 

[Besoeker]

Hi guys.
Thanks all for your help and suggestions.
I did look at the Rittal software (thanks, weressl).
It uses a k factor of 5.5 (W per m^2 per degC).
I have a spread sheet which uses this figure. In truth, I probably got it from Rittal when I was making the spreadsheet - I had the Rittal software on my previous laptop but that got stolen about six months ago.
Anyway, the 5.5 is for their powder coated mild steel enclosures.

But the new unit has to be stainless steel and I don't know if the same k factor is applicable. My gut feeling is that the stainless steel unit should be better at dissipating the heat - after all, it doesn't have a layer of paint inside and out to act as insulation.

Going on that basis, I sized the panel the same as I would have done for the regular mild steel units. I had a bit of a concern that it would be larger thus more expensive than required.
However, I have submitted my bid and had feedback that it is competitive.

 

[hillbilly]

Hi guys.
Thanks all for your help and suggestions.
I did look at the Rittal software (thanks, weressl).
It uses a k factor of 5.5 (W per m^2 per degC).
I have a spread sheet which uses this figure. In truth, I probably got it from Rittal when I was making the spreadsheet - I had the Rittal software on my previous laptop but that got stolen about six months ago.
Anyway, the 5.5 is for their powder coated mild steel enclosures.

But the new unit has to be stainless steel and I don't know if the same k factor is applicable. My gut feeling is that the stainless steel unit should be better at dissipating the heat - after all, it doesn't have a layer of paint inside and out to act as insulation.

Going on that basis, I sized the panel the same as I would have done for the regular mild steel units. I had a bit of a concern that it would be larger thus more expensive than required.
However, I have submitted my bid and had feedback that it is competitive.

Under identical conditions... bare, mild steel will conduct approximately 3 times as much heat as stainless steel.

steve

 

[Jraef]

Under identical conditions... bare, mild steel will conduct approximately 3 times as much heat as stainless steel.

steve

Didn't know that. Thanks, I learned something new today.

 

[Besoeker]

Under identical conditions... bare, mild steel will conduct approximately 3 times as much heat as stainless steel.
steve

That's a good point and it made me ponder some. Then I did a few calculations.
I'm looking for a temperature difference from the inside to the outside to be around 10C (18F) or less.
The temperature difference between the inside surface and the outside surface of the panel walls is negligible for either meterial, so I think the difference in thermal conductivity is in't a key factor in this case.
The external finish may be. Heatsinks are often matt black.....

 

[weressl]

Didn't know that. Thanks, I learned something new today.

There goes your image, Karnak.......:grin:

When you look at the Hoffman data it shows aluminum and SS as being identical from the heat loss standpoint of view. Knowing the above one would assume that aluminum is even better conductor - both heat and electricity - one has to wonder if the majority of heat loss in still air has to doi with radiated heat rather than conducted heat.

Some of the ASD manufacturers install their dirve chassies in NEMA4 boxes, bolt the heat sink on the outside of the enclosure and apply heat transfer paste to conducting the surfaces. Of course the forced air fan on the heat sink has to be discarded and since the heat sink is usually on the back of the unit the mounting of the box becomes trickier. Still something to consider if this were to be a repeat design.

 

[Lcdrwalker]

I have seen advertisements for an "Electronic Cabinet Cooler" in either EC&M or Electrical Contractor mags.

 

[zbang]

The external finish may be. Heatsinks are often matt black.....

Exactly. The rules for thermal radiation and absorption are the same- a good absorber is a good radiator. Or- a black box will capture external heat much better than a shiny box, but it will also radiate that heat back out much better than the shiny. (Which is why painting a heating radiator silver cuts down on it's output.)

Since brushed SS is generally more reflective than epoxy or powder-coated metal, it's not going to be able to radiate out internal heat as well. (Ignoring any thermal insulation properties of the coatings.)

 

[Besoeker]

There goes your image, Karnak.......:grin:

When you look at the Hoffman data it shows aluminum and SS as being identical from the heat loss standpoint of view. Knowing the above one would assume that aluminum is even better conductor - both heat and electricity - one has to wonder if the majority of heat loss in still air has to doi with radiated heat rather than conducted heat.

Having done the calculations for the sheet metal, mild and stainless, I'm sure you are right. Plus an element of convected heat as the air in contact with the outside surface warms and rises. The temperature difference across what is a relative thin sheet of metal is so small as to be negligible.

Some of the ASD manufacturers install their dirve chassies in NEMA4 boxes, bolt the heat sink on the outside of the enclosure and apply heat transfer paste to conducting the surfaces. Of course the forced air fan on the heat sink has to be discarded and since the heat sink is usually on the back of the unit the mounting of the box becomes trickier. Still something to consider if this were to be a repeat design.

External heatsinks is something we have done on larger inverters but for something this small, it just isn't worth the extra labour.
We also use water cooling, particularly for large rectifiers in anodising plants.
Atmospheres with sulphuric acid and aluminium heatsinks are not good bedfellows.