If you think research scientists sit around in white lab coats tossing theories around without any practical engineering experience, think again. The National Renewable Energy Laboratory (NREL) recently launched its Advanced Research on Integrated Energy Systems (ARIES) platform, a computing environment designed to model today's complex electric grid and support the development of new energy generation technologies and resilient microgrids. About two months into ARIES' tenure, NREL's Flatirons Campus—a facility that tests renewable energy systems—experienced a power outage caused by a utility transformer explosion. Because the damage occurred near the Flatiron connection to the main grid, the utility told NREL to expect several weeks without power.
Faced with the prospect of delaying their research, NREL administrators wondered whether the facility could go into self-sustaining mode until the transformer could be replaced. The lab-coat crowd replied, "Challenge accepted."
Pulling Itself Up by Its Own Bootstraps
The Flatirons Campus microgrid is designed to be grid-tied with the ability to go into "islanding" mode (disconnecting from the grid and running independently) when necessary. So, NREL engineers isolated the microgrid and used its 1 MW / 1 MWh battery to power the control center. Once that was up and running, they used ARIES’ digital twin to see how the microgrid would respond to each power source. After simulating and validating a black start procedure, they connected the 430 kW solar array and the 1.5 MW wind turbine to keep the battery charged. Other systems were brought online until the facility—which draws about 200 kW—was fully capable of resuming its work and remained so until the transformer was finally replaced, and the campus reconnected to the grid. ARIES helped engineers turn a crisis into a case study in energy resiliency.
Flatirons Campus
At the center of NREL's Flatirons microgrid sits the controllable grid interface (CGI), which tests the mechanical and electrical characteristics of renewable energy technologies both on and off the grid. It is capable of simulating various fault conditions and evaluating the microgrid's response. The CGI controls grid support systems, such as voltage and frequency regulation, reactive power (VAR) compensation, and load balancing. In addition to battery storage, two solar arrays and various wind turbines, the campus is home to a water power instrumentation lab, a three-megawatt load bank and three dynamometers that help wind turbine manufacturers test their generators under controlled conditions.
The facility's Composites Manufacturing Education and Technology (CoMET) center lets researchers test different materials and processes that may bring about the next generation of wind turbines. Engineers using the facility can design, manufacture and test turbine blade prototypes. One very intriguing aspect of CoMET is the concept of fabricating turbine blades—which are already difficult to transport due to their size—at the location where they will be installed. As turbines get bigger and more powerful, on-site manufacturing will be crucial to reducing cost, adding high-tech jobs and decreasing the carbon footprint of transporting turbine components. In keeping with NREL's mission to support green technology, the campus has a slew of electric vehicle charging stations.
ARIES
Just a few miles down the road from Flatirons is the Energy Systems Integration Facility (ESIF), which houses the computer network that enables much of NREL's research and development. ARIES, NREL's energy research flagship, was created to help engineers and scientists develop and evaluate new energy technologies of various sizes and configurations. With ARIES, engineers can optimize the integration, control and security of diverse technologies consisting of millions of devices, including renewable energy generators, conventional generators, energy storage systems, transmission and distribution equipment, electric vehicles, and smart appliances.
Private sector companies, academic researchers and government agencies can use ARIES' supercomputers and software libraries. For example, Toyota wanted to evaluate how the proliferation of electric vehicles would impact grid power quality. Through the ARIES virtual emulation environment, the company developed ways to manage EV charging to ensure grid stability. The Southern California Gas Company looked into handling renewable energy overproduction during off-peak hours and optimized methods to use electrolysis and bioreactors to convert water and waste CO2 into biofuels that can be stored and used during high-demand and low-production periods. Battery company AES was developing a hybrid peaker plant using a 500 kWh battery energy storage system (BESS) and a 240 kW photovoltaic array. ARIES was able to validate its functionality prior to the installation. In addition to ensuring reliable operation, the simulations helped reduce the company's financial risk. The Department of Defense (DOD) has a long history of using renewable generation and energy storage to build resilient microgrids that keep DOD installations running during a grid failure. The DOD and private sector partners used ARIES to evaluate renewable energy systems under different climate and market conditions and find the most economical and reliable solutions for each scenario.
Eagle Supercomputer
The ARIES platform is centered around NREL's Eagle, a supercomputer with a combined processing power of 20,000 quad-core desktop PCs. With over 2,000 servers arranged in an eight-dimensional hypercube topology, Eagle can perform 8 x 1015 floating-point calculations every second. (In geek-speak, that's 8 petaFLOPS.)
That processing power allows Eagle to perform multiphysics simulations, which take many physical factors into account. For example, engineers use multiphysics computations to analyze the relationship between heat transfer, electrical flow, magnetic fields and mechanical stresses in a device.
Other software applications that Eagle supports include:
- Fluid dynamic modeling and simulation
- Equation solvers
- Molecular modeling
- Quantum chemistry
- Materials modeling
- Statistical analysis
- Atomic-scale analysis, including quantum effects
Scientists create molecular models to simulate the chemistry of batteries, allowing them to run millions of virtual experiments in seconds. Even using NREL's older supercomputer, Peregrine, researchers were able to improve battery chemistries by simulating prospective combinations and only running lab experiments on the most promising candidates. Eagle takes that to the next level.
Microgrid and Grid Modeling
Conventional power plants burn fuel, or split atoms, to boil water and use the resulting steam to spin the generator's turbine, which produces alternating current (AC) electricity. The grid frequency—how fast the current alternates—is a function of the turbine's rotational speed, which is carefully monitored and controlled to ensure a constant frequency. As more devices are placed on the grid, the power draw puts a drag on the generator, causing it to slow down. A control system compensates by pumping more fuel into the generator, thus speeding it up.
Solar arrays and grid-level batteries produce direct current (DC) electricity, which isn't compatible with the AC grid. These generators use inverters to convert DC into AC before the power goes to the utility. (Although wind turbines generate AC, it's difficult to match their output frequencies with the grid, so it's common to convert the AC into DC and then use an inverter to generate a grid-matching AC output.) The inverters are designed to synchronize their outputs with grid frequency and to compensate for frequency dips caused by excessive loading. The problem is that when the grid goes down, there's nothing for these "grid-following" inverters to synchronize with.
Grid-tied microgrids that are capable of islanding, like the one at NREL's Flatirons Campus, need a "grid-forming" inverter—one that can generate its own synchronizing signal. The remaining inverters in the system can follow its lead while in islanding mode. When the grid comes back up, the grid-forming inverter reverts to its grid-following role.
When the utility power at Flatirons went down, the microgrid disconnected from the grid (automatically, per code) and engineers began a black start procedure. The generators were then able to power the campus and provide all grid services. Every generator on the system was inverter-based, and the smart inverters were developed, in part, thanks to the research done through the ARIES project.
The power grid was designed to carry electricity from a few large generators to numerous consumers in a one-to-many topology. Given the generation, transmission and distribution technology at the time, that was the most efficient and reliable configuration. Today, as residential rooftops become power plants and generating capacity fluctuates with the clouds and wind, a new paradigm is needed. NREL and its international counterparts are ensuring that the modern grid will be able to balance supply and demand while maintaining safety, improving reliability and reducing greenhouse gas emissions. It's a lot to ask, but scientists and engineers are up to the task.