Atom Computing and the National Renewable Energy Laboratory (NREL) recently announced a collaborative project to explore the first use of quantum computing to assist in the management and optimization of the U.S. power grid.
The U.S. electrical grid is a complex system of power generation and distribution. It consists of three regional sections—eastern, western and Texas. According to the U.S. Energy Information Administration (EIA), the grid has more than 7,300 electrical power plants, more than 160,000 miles of high-voltage power lines for long-distance transmission and millions of miles of low-voltage power lines and distribution transformers for local connections to over 145 million customers nationwide.
Challenges of computer modeling
Models must monitor each of the three regional grids 24 hours a day for efficient management and to optimize its operation. Models are critical during emergencies, when loads must be quickly shifted and rebalanced between regions, or if a failure occurs and new paths must be established to restore failed services.
Existing grid models run on NREL’s 8-petaflop classical supercomputer, which performs 8 trillion floating-point operations per second. Although supercomputers can handle today’s modeling requirements, the growth of millions of grid inputs and outputs over the next few years raises concerns for NREL scientists. In particular, there is the possibility that growth will cause the grid to become so computationally complex that the current supercomputer will be unable to handle it.
Looking to the future, the NREL research team believes that quantum computers have the potential to eventually handle the grid’s significant optimization needs, even if they do outgrow the capabilities of classical supercomputers.
Forbes Daily: Get our best stories, exclusive reporting and essential analysis of the day’s news in your inbox every weekday.
I had the opportunity to discuss this project with Rob Hays, CEO of Atom Computing. He sees quantum computing as a good match for optimizing the grid. “Optimization of the energy grid is a high-value use case,” he said. “Exploring how quantum computing can enhance grid modeling is an important opportunity. We believe this experiment establishes a foundation for an ongoing relationship between Atom Computing and NREL. The first project looks at resilience of the grid, but there are other projects applicable to quantum computing. As a national lab, NREL has a broad mission, and there are many instances where quantum could add value in support of that mission. We believe this is the start of a valuable long-term collaboration.”
Connecting quantum to ARIES
Advanced Research on Integrated Energy Systems (ARIES) is a research platform that simulates the operational complexity of the national energy system. It facilitates research and supports the development of new energy technologies. ARIES makes it possible to understand how millions of devices such as electric vehicles, renewable generation energy storage equipment and grid-responsive building technologies affect the grid.
Atom’s quantum technologies will connect to ARIES using using an API to create a so-called “quantum-in-the-loop” capability. Hardware-in-the-loop was used in past research; however, this is the first use of quantum-in-the-loop. The interface is key to facilitating communications between the quantum computer and ARIES’ power system simulations so that optimization problems can be translated into quantum circuits.
Although there are plenty of potential quantum applications to address, NREL and Atom Computing will start by focusing on simulating the failure of a switch or power line. For this scenario, Atom Computing’s task will be to provide an optimized solution using quantum approximate optimization algorithm (QAOA) or variational quantum eigensolver algorithms to re-route and restore simulated feeder lines that carry electricity from a substation to a local or regional service area. NREL will then begin to use the results to evaluate the effectiveness of quantum optimizations.
Atom Computing’s quantum platform
Atom Computing’s Phoenix platform uses nuclear-spin qubits created from an isotope of naturally occurring strontium atoms. Strontium is a neutral atom because it has equal protons and electrons at the atomic level. However, isotopes of strontium have varying numbers of neutrons. Differences in the number of neutrons produce different energy levels in the atom. Atom Computing uses unique energy levels of the isotope strontium-87 to create spin qubits.
Qubits must remain in a quantum state long enough for the machine to complete computations. Since Atom Computing’s neutral-atom qubits are natural rather than manufactured, no adjustments are needed to compensate for differences between qubits. That contributes to the machine’s stability and relatively long coherence time of about 40 seconds. Moreover, a neutral atom has little affinity for other atoms, making the qubits less susceptible to noise.
Natural neutral-atom qubits have many advantages that make them suitable for quantum computing. Here are a few of them:
- Identical qubits for consistency
- Less susceptible to noise
- Scalable to large numbers
- Flexible geometry of arrays
- High coherence times
- Wirelessly controllable by lasers
- Scientifically proven (This type of atom powers the most accurate atomic clock.)
Atom Computing is no stranger to advanced government research programs. In early 2023, DARPA selected Atom Computing to participate in its quantum computing project. As a participant in that project, Atom will receive funding and technical assistance from DARPA to move from today’s quantum computing architecture to an advanced fault-tolerant architecture sometime in the future.
Atom Computing scientists are currently working on the company’s next-generation machine. It will have more qubits, higher fidelity and other quantum improvements that will allow it to run more complex problems. I expect Atom Computing will break new ground by announcing its new system in Q4 of 2023 or Q1 of 2024 with 500 to 1,000 qubits. Based on Atom Computing’s continued emphasis on scaling, I am leaning toward the high end.
DARPA is a wild card. Although it has much to offer from a technical and funding standpoint, I think it is too early to see Atom Computing get a technical lift from working with DARPA.
Atom Computing’s next-generation commercial quantum computer will be able to model larger and more complex problem sets, adding more value to its relationship with NREL. A larger processor will also help Atom Computing attract additional customers and use cases.
Rob Hays closed our discussion with an insightful perspective on the NREL project. “There is real value in partnerships like this one between the U.S. Government and an industry partner like Atom Computing,” he said. “We bring specific technical knowledge about quantum to the project, and the government team brings its unique expertise on the energy grid. Our collaboration with NREL is an ideal public-private partnership because it adds value to both sides. It not only helps advance our quantum technology, it also helps the entire quantum ecosystem. In exchange, NREL gets a new and powerful way to support its mission of providing resilient energy grids.”
Paul Smith-Goodson is the Vice President and Principal Analyst for Quantum Computing and Artificial Intelligence at Moor Insights & Strategy. You can follow him on Twitter for current information and insights about Quantum, AI, Electromagnetics, and Space.