RECAP | ROUNDTABLE | Approaches to Site Power

Using traditional utility power with backup diesel engine-generators has been the standard and the norm for many years. However, many data center operators are starting to consider alternative approaches to site power delivery. Jeremiah Beaver, Consultant for Uptime Institute, and Jay Dietrich, Research Director, Sustainability for Uptime Institute, joined the roundtable where attendees discussed potential alternate approaches to site power.

As an introduction, the concept of “typical” and “non-typical” approaches to site power was reviewed.
Typical:
• Electric utility with diesel engine-generators as backup
Non-typical:
• Microgrids – Renewables, Fuel cells, Energy storage, CHP (combined heat and power) systems
• No diesel engine-generators

An attendee commented the 2 most common non-typical configurations would be:
• Fuel cells as primary along with UPS + utility as backup
• No engine-generators (EGs) and deploy large scale battery systems to displace EG functions + utility as primary
He also commented the reason for investigating these alternatives is to potentially reduce Scope 1 emissions.

Another attendee commented they are looking at solar panels as the second source + diesel rotary UPS (DRUPS). He also indicated sustainability is the driver and to bring down the utility cost. Utility in their region is already considered sustainable because their use of renewables.

A third attendee commented their 2 northern California facilities have solar installed as supplemental power, along with fuel cells as supplemental + local utility as primary. He commented they do not consider fuel cells safe and reliable enough to rely on 100%. Solar is a turnkey funded solution from an outside provider with small benefit on cost, but really hands off from a service perspective. Fuel cell is similar. However, sizing will be increased by about 25% because of module failures. Therefore, they do not consider it reliable enough to be the primary source. Service is managed by fuel cell vendor. The fuel cell vendor is contractually required to meet a certain percentage output, so this gives them some skin in the game.

Jay chimed in that when he was working for IBM, their fuel cell tests displayed more reliability problems than they expected. IBM also determined they are still not economical. Another attendee indicated they have considered fuel cells but heard of vendor maintenance issues.

It was highlighted the Key Factors in considering alternate non-typical site power (i.e., fuel cells, renewables, energy storage, etc.) are:
• Reliability
• Cost, ROI
• Sustainability
This was reiterated by one attendee. He has experience with evaluation of fuel cells as primary + grid as backup. Their findings indicated fuel cells are very expensive from a 1st cost and TCO, for both purchased power and outright owned options. The economics are highly dependent on incentives from the government. Another factor to consider, since fuel cells are in the early stages, what’s the useful life? There is no real operating history.

Jay commented that cogeneration was an option in areas where IBM didn’t have to buy electricity, thus there were no taxes on utility. Cogeneration tends to take a lot of management and operations effort, and cogeneration plants do not like to run on variable outputs. Jeremiah commented cogeneration tends to be used in areas where you have unreliable power girds.

Jeremiah then commented on the relationship between alternate non-typical site power and Tiers. As most attendees have confirmed, microgrids are viewed and typically categorized as stand-alone (primary) or supplemental. When considering for Tiers certification, the requirement is for at least 12 hours of onsite energy storage, which is difficult to meet for an onsite stand-alone microgrid. He also commented that sustainability must be looked at upstream and downstream. As an example, EGs are only polluting when they run. For batteries, consider replacement and disposal and these impacts on sustainability upstream and downstream. Same would apply for onsite microgrids (i.e., fuel cells)

A question was asked by an attendee: How many have considered multiple grid supplies, from diverse parts of the grid network, not from a single substation? He commented the concept is if you get 2 diverse feeds, you can rely on the 1 if the other goes offline, and this could mean you potentially would not need to size your EG plant to handle the full building capacity. One attendee commented they use 2 diverse feeds for all their sites, but that doesn’t preclude them from still sizing EGs for full capacity. The consensus of attendees indicated 2 diverse feeds doesn’t mean you can eliminate EGs. Also, the Tier Standard does not allow you to eliminate this onsite backup source since having 2 utility feeds doesn’t provide you 12 hours of onsite energy storage.

One attendee indicated they are looking at gas engine systems with a fast start within 45 seconds instead of diesel EGs (https://www.innio.com/en). These systems can be used as primary power in some regions to meet government laws. Also, he commented for sustainability it is better to have a gas system verses getting coal fired utility power.

Lastly, an attendee commented if you are continuing to increase dependence on onsite generation and less on utility, the utility can come back and take power away or charge for power you resource.

Overall, the roundtable was very well attended and indicates this is an important topic for data center operators. Also, based on attendee dialogue during the session it is apparent a lot of doubt remains when it comes to using microgrids, typically fuel cells.