“I believe ESIF is going to revolutionize the way renewable energy is used. It’s going to modernize the entire grid system.”
—Matt Graham, PMP
FACILITIES PHOTOS BY DENNIS SCHROEDER, COURTESY OF NREL
Sunrise at the Energy Systems Integration Facility (ESIF) at the National Renewable Energy Laboratory (NREL) in Golden, Colorado, USA
If Alexander Graham Bell encountered a smartphone in 2014, he wouldn’t know what to make of it. Yet the electric grid hasn’t undergone nearly as much transformation.
“If Thomas Edison were alive today, he’d recognize our electric grid,” says Drew Detamore, PhD, PMP, deputy director of sustainability, infrastructure transformation, engineering and operations, National Renewable Energy Laboratory (NREL), Golden, Colorado, USA. The innovation lag stems in part from how difficult it is to experiment on the existing grid. “You can’t shut the grid down, because the country needs it at all times.”
To make way for innovation, the U.S. Department of Energy (DoE), a PMI Global Executive Council member, found a clever workaround: It would build the Energy Systems Integration Facility (ESIF), a hyper-efficient research center on the NREL campus for energy scientists to develop and test renewable energy technologies and learn how to integrate them into the existing grid. At the heart of the US$135 million, 182,500-square-foot (17,000-square-meter) facility are four small-scale power grids for research. Each grid connects to multiple laboratories in the facility, allowing various electrical power generation, storage and consumption devices to interact for testing.
From left, Matt Graham, PMP, federal project director, DoE Golden Field Office; Drew Detamore, PhD, PMP, deputy director of sustainability, infrastructure transformation, engineering and operations, NREL; Ken Powers, chief operating officer and deputy lab director, NREL; and Brian Larsen, PMP, senior project manager, NREL
PHOTO BY JESSYEL TY GONZALEZ
“I believe ESIF is going to revolutionize the way renewable energy is used,” says Matt Graham, PMP, federal project director with the DoE Golden Field Office. “It’s going to modernize the entire grid system.”
PLAN WITH A PURPOSE
While sustainability is an afterthought on many construction projects, it was a key requirement in the ESIF project plans, says Dr. Detamore. In addition to cost, schedule and scope, the project team talked about energy efficiency as “the fourth constraint.”
“We set very high energy standards—75 percent less than what a commercial building of its type would normally use,” says ESIF project manager Brian Larsen, PMP. The project team also set an ambitious goal of securing LEED platinum certification—making it the only high-performing computing data center in the world with that sustainability rating. “Those goals combined made the project pretty complex.”
“We set very high energy standards—75 percent less than what a commercial building of its type would normally use.”
—Brian Larsen, PMP
NREL senior scientist Kenny Gruchalla examines the velocity field from a wind turbine simulation at the Insight Collaboration Laboratory at ESIF in Golden, Colorado.
Peregrine, the supercomputer at ESIF in Golden, Colorado
Everything from skylights in the laboratories to white ceilings and low cubicle walls helps the facility—part of the DoE’s NREL—use natural light to minimize energy consumption. While the average commercial building uses about 90,000 BTU per square foot per year, ESIF clocks in at less than 25,000 BTU.
The project team’s eye toward energy extended to the facility’s supercomputer, Peregrine, which is capable of crunching mind-boggling amounts of energy data into actionable insights and analysis.
“The researchers are using the supercomputer as a modeling tool,” says Mr. Graham. “They can simulate the grid on it, and they can actually do modeling that saves them years of R&D.”
NREL scientists studying how the wake from one wind turbine affects another wind turbine downstream can now use the supercomputer to generate images in a visualization laboratory. Rather than raw data, the supercomputer shows the wake as colorful streams coming off the wind turbine propellers. It looks like the tail of a comet or a shooting star. Then the scientists can even don 3-D glasses, enabling them to walk through the image, immersing themselves in the rendered data.
To ensure Peregrine was as powerful as possible, the project team delayed procurement until the last possible moment, hoping to capture the latest processor technology available.
“What you can plan on up front in the industry isn’t the same as what you can get when the time comes to install equipment, so we wanted to account for that,” Dr. Detamore says.
The project proposal initially required 200 teraFLOPS of computing capability. By using a phased procurement plan, the team managed to quintuple the processing power to 1 petaFLOPS, which means the computer can perform more than 1 quadrillion calculations per second. And it uses less than one megawatt of power continuously.
The central interaction space in the ESIF
Wire mesh keeps birds from crashing into the windows of the ESIF and also reduces the amount of sun hitting the east windows.
CHANGES AHEAD
The project team recognized early on that building this first-of-its-kind facility would require careful risk management and an iterative approach.
Rather than mandating all of the project specs and waiting for contractors to submit bids, the project team selected a two-phase design-build approach. In the request for proposals, the team presented a list of performance specifications that needed to be accomplished within a certain budget.
“The approach is performance-based,” says Dr. Detamore. “Instead of saying, ‘We want a 60-watt light bulb in the center of the room,’ we said, ‘Here’s what the researchers say this project should do. How you do it, that’s up to you.’”
An air-intake structure on the NREL campus pulls in air to cool the ESIF data center. Such structures help lower energy costs in the building.
“Instead of saying, ‘We want a 60-watt light bulb in the center of the room,’ we said, ‘Here’s what the researchers say this project should do. How you do it, that’s up to you.’”
—Drew Detamore, PhD, PMP
In phase one, bidding contractors worked with a firm fixed price to develop the scope through preliminary design. In phase two, they offered a firm fixed price for the final design and construction.
The researchers who met with the project team to develop the performance requirements were included in meetings at least once a week as the team moved through the design and construction of the project.
“By doing a two-phase contract, we let the contractor reduce the amount of risk by completing the preliminary design before giving a firm fixed price to finish the design and construct the project,” says Mr. Larsen. “The contractor reduces his contingency, and that money is converted into additional scope.”
That was but one example of the team’s sophisticated approach to risk reduction. When the integrated team, including members from the DoE’s Washington, D.C., headquarters and NREL, identified each of the project risks, they hovered around 40. The list included possible scheduling setbacks if government approvals took too long along with one rather unusual risk: construction work shutdowns due to possible explosives. The project site was a decommissioned U.S. National Guard training ground, which meant the possibility that construction crews might unearth unexploded mortars or cannon shells.
“We set aside US$9 million to manage risk for this project. That wasn’t an arbitrary number,” says Mr. Larsen. “We used Monte Carlo simulation on the 35 or so risks that we had identified to come up with an 80 percent level of confidence on the amount of reserves we’d have to have.”
The team also developed contingency plans. To mitigate the risk of explosives, for instance, the project plan included a rigorous site survey before construction began and hiring a subcontractor with a demolitions expert on call.
“By managing the risks, we were able to realize very few,” says Mr. Larsen.
CONVERTING CONTINGENCY FUNDS
The project team didn’t approach risk management merely as a means to keep costs and delays at bay. By carefully monitoring existing risks and periodically releasing contingency funds, the team was able to expand the initial scope.
The team used just 0.5 percent of the original project funds on realized risks, about US$500,000. When federal approval of a critical task was delayed 49 days, the project incurred a US$300,000 setback in lost work time, for instance. The remaining funds—roughly US$8.5 million—went to scope expansion.
“We reviewed and looked at the risk on a quarterly basis, if not monthly,” Dr. Detamore says. “And as risk was reduced, we took the money that was set aside to deal with that and put it back into the project.”
HEPA filtration on all of the fume hoods, for instance, started on the prioritized list of unfunded scope but only got the green light after the team was well into the construction phase.
To accommodate those late additions, the design and construction teams worked closely with ESIF’s project manager to manage both the approved project plans and the possible additions on the horizon. “Each of the desired scope additions was carefully considered,” says Dr. Detamore. “The design or construction teams would sometimes say to us, ‘You need to be able to give us a go or no-go on this particular item by this date, or it can’t be done without severe impacts.’”
NREL technicians Josh Martin and Scott Walters work in the Power Systems Integration Lab (PSIL) at the ESIF.
PHOTO BY JESSYEL TY GONZALEZ
That iterative approach is particularly impressive considering the complexity of the government project. Mr. Larsen credits buy-in from the federal project director, Mr. Graham, who had a seat at the table throughout the project. Present at all meetings, he was fully involved with changes as they happened.
“When we did have to change the plan, he already knew why and what we were going to do, so he was almost always fully supportive of the change,” says Mr. Larsen.
Of course, funding a scope addition required careful change management to make sure it delivered real value to the project. Once each change request or scope-add was initiated, the project’s change control board considered the request, carefully weighing it against the project’s baseline and earned value.
“Complexity presents a lot of risk, but if you pay attention to that early and continuously, and form teams to help you manage risk at a lower peer level, then the project can be done successfully.”
—Brian Larsen, PMP
TURNING THE LIGHTS ON
In the end, meticulous risk management and change control allowed the completed facility to outshine the original project plans. Completed in September 2013, the project beat its original schedule by a month and came in within budget and beyond scope.
“If there’s one message to convey to others about this project, it is that complexity presents a lot of risk,” Mr. Larsen says. “But if you pay attention to that early and continuously, and form teams to help you manage risk at a lower peer level, then the project can be done successfully.”
For the DoE and the larger energy sector, ESIF promises to be the laboratory pointing the way to a next-generation power grid.
“I think this project greatly contributes to the community,” says Ken Powers, COO and deputy lab director at NREL. “It is a one-of-a-kind facility that brings researchers and equipment manufacturers together toward a common goal of enhancing energy efficiency and renewable energy.” PM
LIGHTS, CAMERA, ACTION!
Check out behind-the-scenes videos of this year’s PMI Project of the Year finalists on PMI’s YouTube channel.