The way in which data centers are powered is diversifying. Today, power availability and time to deployment are among the greatest threats to continued industry growth. Securing control over when and where power is generated is becoming increasingly mission critical.

The solution lies in selecting a reliable onsite power system. Facilities that can produce their own megawatts are able to shift energy availability from a core constraint into a strategic advantage that optimizes how the whole data center is driven.

Wärtsilä manufactures ten to 20 megawatt reciprocating engines that come in two main forms: natural gas-only and dual fuel solutions that can operate on either natural gas or diesel. “Our engines are available in a range of setups and they’re particularly large compared to the rest of the market,” adds Sean Hughes, business development manager, North America, at Wärtsilä.

In a recent episode of DCD>Talks, Hughes explores the key differences between two onsite power generation solutions: engines and turbines, and makes his case for why engine-based systems should be the go-to for managing the complex demands of modern data center power requirements.

GettyImages-999834586
– Getty Images

Fueling up

Hughes begins by outlining the fundamental differences between engine-based onsite power solutions and traditional turbines: “Turbines are essentially a mass flow technology, and that means the more air that they're able to secure, the more output they can produce, and the more efficiently they'll run.

"Engines work differently, and that offers a multitude of advantages over turbines. As such, in my opinion, they're a better technology for data center applications.”

Unlike turbines, which require steady conditions to maintain efficiency, engine-based systems can respond to minute-by-minute changes in demand without being impacted by external conditions in order to maintain consistent performance while providing low-emissions.

Effectively, superior reliability, fuel efficiency, heat rate and altitude tolerance, plus lower emissions, minimal water use, modular design, shorter lead times, and lower gas pressure requirements all give engines a competitive advantage over alternative onsite solutions such as turbines.

Efficiency and tolerance

“When we’re talking about fuel efficiency, the difference is pretty stark,” continues Hughes. “There’s a particular design application that the turbines can use, called combined cycle, which would actually provide a much better fuel efficiency, but it's not appropriate for the varying demand loads that characterize modern AI data centers. As such, we have to compare our engine solution to a simple cycle design in turbines, and in that case, it’s between ten and 15 percent more fuel efficient.”

Taking into account the huge level of investment in fuel required for data center power generation, the numbers compound over time into impressive savings.

“Secondly, probably my favorite reason why data centers should be looking at reciprocating engines rather than turbines is their better heat tolerance,” says Hughes.

As explained above, turbines require a significant amount of air in order to run efficiently and provide the output they're designed for, factors like hot climate – or simply dramatic spikes in temperature – and high altitude lead to lower air pressure and consequently negatively affect efficiency and overall output.

Reciprocating engines don’t suffer the same affliction in hot spells because they don’t work based on a mass flow rate that can quite literally change with the wind. “And as we know, data centers require just as much power, in fact, even more power on hot days than they do on a relatively mild day. So this benefit is particularly important, and the same applies to altitude,” adds Hughes.

He points to Utah, Arizona, and New Mexico as key examples of high-altitude areas that are also data center hubs where the name of the game is more power. Here, engine-based solutions are well equipped to operate reliably and efficiently in more challenging climates and elevations.

Eco mode: On

Data centers are under increasing pressure to scale up both quickly and sustainably. Facilities are under significant scrutiny when it comes to the environmental impact of operations, while local and national sustainability targets grow more intricate – adopting the driving seat when it comes to key decisions.

Wärtsilä is unique in its ability to guarantee its emissions output. The company designs and manufactures its own emission reduction equipment – a task typically outsourced to third parties – in order to reduce environmental impact without compromising on performance.

“Minimal water use is another particularly important factor,” says Hughes. “There’s a huge amount of pressure on data centers because of the enormous water usage found in certain circumstances. Luckily, engines, compared to turbines, use far less water. In fact, to provide a rough idea, consumption sits at around one gallon of water per engine per week – that’s probably less than I drink in the same week!”

Navigating bumps in the road

One of the greatest challenges for powering modern facilities comes down to the complex demand profiles of AI workloads. Demand for power oscillates dramatically, requiring a generation solution that can manage both full-throttle surges and slow-paced periods. Hughes expands:

“Demand may go from 20 percent of the plant’s capacity all the way up to 100 percent within a matter of seconds, if not milliseconds. Imagine your car trying to go from 20 to 100 miles an hour in the blink of an eye.”

To keep pace with this reality, power solutions must team up. Hughes believes the answer lies in microgrids: a combination of different technologies and interconnected energy resources that act as a single unit.

Hughes outlines how Wärtsilä is helping to fuel this synced up approach: “There’s a side to the business that packages battery energy storage systems (BESS), battery systems themselves, plus a combination of batteries and engines, along with the various pieces of equipment typical to data centers – such as UPS and perhaps super capacitors. All of these elements come together to create a microgrid that can handle extreme fluctuations in demand.

“We’re continuing to do a lot of work internally and with third parties to optimize these solutions for our data center customers.”

Additionally, when it comes to lead times, Wärtsilä’s solution is available out of the factory within around two years. “I believe that turbines take around five to seven years, and therefore we're particularly competitive once the engines are out of the factory,” adds Hughes. “You're likely looking at a construction time of around one year, and so we're very well placed to provide our customers with a competitive solution that meets their schedule.”

Full speed ahead for onsite power

Amidst a complex landscape of competing challenges and priorities, driving consistent innovation is central to securing long-term progress. So, what does the future hold for the next generation of onsite power? Hughes shares his perspective:

“I think one thing that is likely to happen is that heat generated from both engines and turbines will be better utilized, mainly because cooling is such an important aspect of data center operations.”

Today, onsite power solutions don’t typically provide a direct source of cooling. But by converting heat generated by engine-based systems, for instance, and running it through what's known as an absorption chiller to create chilled water, cooling capacity can be optimized.

“This process doesn’t require any additional fuel. It’s an enormous benefit that significantly reduces energy consumption,” says Hughes. “Essentially, you go from a fuel efficiency of around 50 percent up to around 75 percent depending on the project.”

The second innovation Hughes sees on the horizon is the transition from onsite power generation supplying just data centers with the energy they need, to these systems actively assisting the national grid when and where help is required.

Capacity held on site for data center operations isn’t necessarily being used 24/7. To maximize efficiency and push facilities to become responsible power partners, exporting spare megawatts from site to grid is the next step forward. Hughes doesn’t view this as a daunting or risky road ahead, but as a smart move that could generate an additional source of revenue for those willing to shift things up a gear.

To hear more about the advantages of engine-based power generation, watch the full DCD>Talks episode with Sean Hughes, here.