VSDS drive - what is the temperature rating of heatsinks ?

[rotw]

Hello,

I would like to know please is typical design temperature or max working temperature (say sufficient to trigger an alarm) of the heatsinks in VSDS system (frequency converter).

I refer to large drive vsds system of 20 MW power and above. I heard that typical design temperature is 80 deg. C and trip level is at 70 deg C, is it something in line with what you have experienced ?

Certainly this should be dependant upon munafacturer and particular design, however I am looking after typical/ most encountered values to have a rough idea.

Thanks


"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[Skogsgurra]

Your numbers are often seen in the specifications from different manufacturers. Sometimes somewhat higher and sometimes lower. But very often in that range.

A water cooled inverter usually has lower limits.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[jraef]

The heat sink temperature is however NOT the same as the ambient air temperature for operation of the drive. On large drives, that is typically 40°C, sometimes 50°C, but never more than 60°C with de-rating. At least, that's what I have ever experienced. The heat sink temperature and trip point, usually 70°C, is for the hottest point right nearest to where the power devices attach, and there is usually a temperature sensor calibrated for that value right in that hot spot.

Was there a specific reason why you asked the question by using HEAT SINK temperature?


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[rotw]

Thanks for your input.

quoted
Was there a specific reason why you asked the question by using HEAT SINK temperature?
unquoted

Reason for asking is that we have VSDS driven motor of 20MW power which is envisaged to be installed in a desertic region where outdoor temperature can reach up to 56.5 deg C.

Currently there is limited raw water in the facility so an air cooled type exchanger is considered as the only option available to cool down the VSDS primary coolant (water) circuit; however we are wondering if this solution will not cause issues from the view point of the overheating and tripping of the converter or affecting its component life cycle/durability. Thus my question.

Thank you for any orientation

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[Skogsgurra]

You really have to discuss this with the VFD supplier.

f tehe is spare capacity in the inverter/motor or if the plant can tolerate an occasional lower output, then it is not uncommon to reduce the load automatically when heatsink temperature gets critical. That possibiliy is often standard.

20 MW, Robicon?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[rotw]

Skogsgurra, thanks.

This is an LCI type VFD.

With regard to your response, no doubt this will be discussed.
The point is that such high temperature is not uncommon, means is expected to occur frequently.
Per your experience, do you think there are good reasons to worry here ? is it usual that a converter would trip due to cooling system / site conditions ?

I am not to happy with the idea that system could trip as this is one single machine in critical service.



"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[itsmoked]

On things like this the power rating is completely based on two things. The actual current capacity of the switching elements and/or the eventual temperature-rise of the heatsinking. And often the capacity of switching elements is directly related to their temperature. That's likely a 20MW VFD because, based on the expected temp rise with an expected maximum water temp the switching elements reach their derating temperature.

A typical solution would be to use, say, a 25MW VFD because it's operating rise would be less than the 20MW version based on the same load.

Agreed, you should talk to the manufacturer immediately.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[rotw]

itsmoked

Very clear explanation. In case there is a need to derate, say from 20 to 25MW, I see a huge impact on equipment cost which are way higher and way off the cost of simply buying a sort of water chiller ?

another question, when you say that:
quote
The actual current capacity of the switching elements and/or the eventual temperature-rise of the heatsinking
unquote

I thought the power rating of the converter is based on the size of the driver or does the driver power translate into current capacity in which case these are both ways to say the same thing ?

excuse my ignorance.

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[itsmoked]

Hi rotw; Most semiconductors are current rated based solely on a current that will keep them under the lifetime curve's sharply aging point.

That temperature is kept at bay by the particular package the semiconductor is mounted in and an expected heatsink performance in hauling away the waste heat.

The driver (as you state) can also limit things if it can't drive hard enough. That results in the driven semiconductor not being as fully ON as possible which causes yet more heat loss to occur. However, in a 20MW drive you can pretty much assume the designers did not skimp on the driver power and the extremely expensive semiconductor power elements are being fully driven. The only limit is going to be the heatflux rate from the semiconductor to its own case(ultimate limit) and from the case to whatever heatsink system (realistic limit), water in this case.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[rotw]

Hello Itsmoked, all,

So if I understand correctly the story can be summarized as follows:

The driver (motor) is sized according the load power requirement, which is 20 MW.
The drive system (converter) is then sized to fit the driver.

When the motor operates at full load, the current in the semiconductors is maximum and temperature of the semiconductors increases, whereby the temperature of the case is also higher.

The heat sink shall cool down the component/case below the trip or set point temperature, which is calculated based upon a certain design/lifetime criteria. The hot spot is not where the water is entering the heat sink; but when it leaves the heatsink which is where the temperature of the case is the highest and where the trip / monitoring sensor has of course to be located as this is the worse case temperature to verify against semiconductor design.

In case due to demanding ambiant conditions that do not allow a proper cool down of the heat sink to the required safe temperature level there are three alternatives as I now understand:
1/ Reduce the load, this is equivalent to loss of production and operate at partial load or eventually trip (safety protection)
2/ Use a different cooling system approach in hot location (chiller), here it is not possible to have raw water so probably use of closed circuit is an alternative
3/ Last resort is the de-rating of the converter, say to 25MW this will allow to maintain the full load operation and lifetime of the converter within the equipment specification.

Am I correct ?

Refering to semiconductors, what is most critical part are we talking about : inverter, rectifier or DC link ? i.e., Which part will be the bottleneck in view of the heatsink / high site temperature problem ?


I would like to sort out these things and have clear ideas of whats going on when talking to vendor
thanks again

"If you want to acquire a knowledge or skill, read a book and practice the skill".

 

[electricuwe]

As stated previously the typical ambient temperature rating of large drive Systems is in the range of 40..50°C. Operation at higher temperature is usually possible. The derating percentage will be highest for the components with the lowest aximum temperature at their hotspot.

Therefore, as nearly all large drive Systems today use film capacitors their limit of 105 degree (compared to power semiconductors with 125°C or 150°C) will determine the total derating of the drive, not the semiconductors.

To make the approach simpler, try to find a system that already has a hight ambient temperature without derating, e.g. 45°C or 50°C.

 

[jraef]

To reiterate, you are getting too deep into the nitty-gritty details of the HEAT SINK temperature, that is for the drive DESIGN team to consider. The finished PRODUCT, the DRIVE ITSELF will have an overall ambient air temperature design specification, with de-rating factors if necessary. But de-rating can only work up to a point because at some point, you will exceed the safe limits of the components inside. And it's not just the power devices, because if ANY component fails in the drive SYSTEM, the drive fails. So even if the power devices survive because of de-rating, what good are they if the firing circuit is fried?

As a potential user / specifier, your role should be to accurately determine the ambient operating conditions, estimate the extreme swings in those conditions in which you will still expect full operation or reduced operation etc., and clearly STATE those as requirements to the vendors. Let THEM determine how they will meet your requirements and explain to you how they plan on it with their tender offer. If you don't feel their approach fits your needs, you can always reserve the right to reject.

By the way, I would not dismiss liquid cooling out of hand. A closed loop liquid cooling system with a liquid-to-air heat exchanger would not put any undue burden on a local water resource. Again, don't micro-manage it, just state your requirements clearly and unambiguously.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[rotw]

jrae,

I agree with your statement and I thank you for the orientation.
Only clarification I would like to bring is that the cooler type is actually specified, so I am re-questionning if this has to be really specified.

If it is a given there are basically two outcomes /

- either manufacturer will deviate and comment and ultimately propose or recommend system which is fine;
- or the manufacturer will try to match the requirements and that can be misleading and resulting in a complex mitigation and is a bit ineffective for all parties if things have to be realigned all over again.

So communication at each stage is key, and proper understanding of the background may also help to tweak a bit the requirement upfront. Thus my exhaustive/detailled questions...

Thank you again for the valuable insight.

"If you want to acquire a knowledge or skill, read a book and practice the skill".