Fuel Cell-Powered Data Centers Could Cut Costs and Carbon

angry tapir writes "A group of Microsoft researchers believe that using fuel cells to power data centers could potentially result in an 'over 20% reduction in costs using conservative projections', cutting infrastructure and power input costs. In addition, using fuel cells would likely result in a smaller carbon footprint for data centers. The researchers looked at the potential of using fuel cells at the rack level to power servers in data centers — although they note there is a long way to go before this could become a reality (not least of the small worldwide production level of fuel cells)."

Wake me up...

[Mr D from 63]

Give me a call when those fuel cells are ready for deployment, then we can talk about all these wonderful uses. No talk about the carbon footprint of operating fuel cells?

The article mixes the use of fuel cells as a power source with efficiency improvements. The only place that makes sense is with the minor savings that may be seen by eliminating DC converters, but you will still need DC regulators which will have some losses.

A major oversight of this article is the fact that fuel cells are major heat generators, not something you want in a data center. They would need to be installed in a separated structure, therefore idea that "Rack-level fuel cells would do away with data-centre-wide electricity distribution for servers" is hard to imagine.

Installing FCs in servers/racks won't work

[orzetto]

The article does not mention it clearly, but those fuel cells are likely natural-gas powered. They are either very high-temperature cells [wikipedia.org] (800 degrees C) or low-temperature cells (70-120 degrees C) with a reformer somewhere that converts natural gas to hydrogen. In the former case you would need to handle fuel at insanely high temperatures close to a bunch of electronics (you can guess what happens at the first leak), in the second you have to handle a hydrogen distribution network, and hydrogen is a nasty gas to work with (see for example hydrogen embrittlement [wikipedia.org]); nothing that cannot be handled, but providing it to single servers or even racks? Hydrogen-proof piping is expensive, and even worse are the valves.

In any case, gas piping is never going to be as practical as power cords. You cannot bend it, coil it, join it easily, and you will need also piping to collect exhaust gases: since this hydrogen comes from natural gas, it travels with CO2, and you don't want it to accumulate in the data centre. You may also need another line to provide oxygen if the data centre ventilation is insufficient.

The argument that one would do away with power supplies is foolish: simply provide a network of DC power instead for all required voltages. FCs produce DC power, but their output voltage is unsteady and needs to be converted to the right voltage; and there are several voltages that a server requires anyway.

So, if FCs have to be, they need to be placed outside the data centre, and function as their power stations. At this point, one wonders, why should we ever consider to install FCs in power stations? Simply build a FC power station and export to the grid.

The main driver for FCs in power generation in the US is the low price of natural gas due to high shale gas production.

Re:What's a fuel cell?

[Nethemas the Great]

I have a furnace that is hooked up to a natural gas main. I have a faucet that's hooked up to a water main. I have a water heater that's hooked up to both. They all work fine with occasional maintenance. Commercial/industrial buildings just do it on a larger scale. A fuel cell would just be another appliance hooked up to a supply main. If the technology scales well--which it does--there's no reason to add all the overhead (cost/reduced efficiency) required to centrally produce and distribute. The only problem with local/hyper-local production is the business model of power companies.