I'm developing a 1kW hardware prototype of a DC Nanogrid. The standalone DC nanogrid is supplied by a 1kW PV array and a 60V, 150 A battery bank. I'm not using any soft switching techniques.
In total there will be three converters.
1) Boost converter- For MPPT. Pmax= 1kW
2) Buck converter (for the load) - Pmax= 1kW. IL,max= 41A.
3) Bi-directional converter- Pmax= 720W, Ibat,max= 12A.
How do I go about sizing the heat-sink ? I'm attaching a block diagram of my system as well.
Dear Nishant,
there are three ways of dimensioning a heat sink:
1) Get a liquid cooling system installed, knowing to be massively oversized and be sure it will save the power electronic from overheating. Also be sure, this will be the most expensive solution.
2) Get a big air-cooled heat sink, add a few powerfull fans and consider it done. Cheep, but usually a bad and most likely noisy compromise.
3) Math, physics and lifetime of power semiconductor correltaion.
If you know the maximum output power and the accordiung efficiency of the design, you can estimate the losses.
If you have intimate knowledge of current waveforms, supply voltages, control strategy, switching frequency and semiconductor data, you can estimate the losses more accurately.
If you have an idea abou the semiconductor to be used and know about the control strategy and the schematic, you can try a simulation tool to calculate the losse.
Companies like Infineon offer powerfull online simulations, see the link.
Use is free of charge.
Once you know the losses Pv, the thermal resistance needed to dissipate the heat can be calculated by using the tolerable chip junction temperature Tvj, Based on an ambient temperature Tamb, you can calculate the thermal resitance from heat sink to ambient Rthha from Tvj=Tamb+(Rthjh+Rthha)*Pv.
Rthjh is a value from the semiconductor in use, charakterizing the thermal resistance from the chip's junction to the heat sink. It usually is a sum ot the resistance from junction to case Rthjc and the transfer of heat from the package's case to the heat sink Rthch. Datasheets of various manufacturers will get you any of these information.
Don't be to conservative on Tamb. Think about what can happen if the grid is installed in regions with 40°C in summer.
Once you get the Rthha-value, check the catalogs of heat sink manufacturers for suitable parts.
To prevent rapid aging of the power semiconductors, the termperature swing of Tvj should be kept below certain limits. What these limits are, mostly depends on the semiconductor in use, the load profile and the lifetime you expect from your design.
If you keep the maximum chip temperature below 120°C for a device rated Tvj,max=150°C and the temperature swing at maximum load at
Look at this to do the calcs yourself http://www.ti.com/lit/an/slva462/slva462.pdf
Or use a calculator https://students.cs.byu.edu/~jgoodell/heatsink3.html
As Martin mentioned above, don't be conservative with the ambient temp. If the device is indoors where you have temperature control of on a subgrade floor you may be able to get by with low ambient temperatures in the calc. If your device is outdoors, you need to also consider it's exposure to the sun. Boxes sitting in direct sun can easily reach 60oC during the hottest part of the day.
One more little detail that you'll need to consider is whether or not the devices mounted on the heat sink need to be isolated. If you need insulators between the devices and the heat sink, you will need to keep the heat sink at a lower temperature to achieve effective cooling of the devices. This will require a look at the datasheets.
Also, you should put a temp sensor on the heat sink to shut down your system if there is a thermal problem. If one device fails, you don't want to destroy all the others attached to the heat sink.
I suggest using Psim. I like a lot.
http://powersimtech.com/products/psim-modules/thermal/
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning