Behind every self-driving laboratory at the Department of Energy's Oak Ridge National Laboratory is a team most people never see. Facilities and Operations (F&O) workers are building and maintaining the infrastructure that makes autonomous science possible.
Autonomous labs run with little human intervention. Instead, they rely on robotics, sensors and automation to perform everything from pouring liquids to placing samples into ovens. They learn and adapt, with artificial intelligence making at least one decision in the process. More than a dozen self-driving labs operate at ORNL, placing the Tennessee national lab among the first research institutions in the world to create this autonomous laboratory model at scale. Embedding AI into scientific workflows is changing not only how experiments are conducted but also how research facilities are designed, operated and maintained.
"Embedding AI into our workflows and systems is completely changing how we do science," said Rob Moore, a distinguished staff scientist who leads efforts to advance autonomous science at ORNL. "That transformation impacts not only the scientific process but also the entire operations enterprise that enables it." The F&O workforce - consisting of craft workers, skilled trade professionals, facility maintenance technicians, engineers and other support personnel - provides hands-on work in labs while helping design, operate, maintain and evolve the ecosystem required for continuous autonomous operations.
Moore said F&O personnel are equal partners with research shaping ORNL's autonomous science strategy and helping the laboratory keep pace with rapidly evolving research needs. "Operations has an equal role in figuring out how to do science in this new era," Moore said. "We're collaboratively defining what the future will look like."
It takes a village to build a self-driving lab
Several skilled trades professionals provided hands-on support in setting up ORNL's autonomous laboratories.
"Hoisting and rigging professionals transported heavy robotic equipment and furnaces, while pipefitters connected up to eight different flowing gas streams available at different locations in the lab," said Craig Bridges, a chemist in ORNL's Autonomous Chemistry Laboratory, detailing the labor needed to bring the lab online. "Instrument technicians diagnosed and installed complicated wiring to enable specialized communication between equipment and control computers. Sheet metal workers connected a large, custom ventilation enclosure, expanding the scope of chemistry that can be safely performed. Carpenters constructed custom equipment fixtures, and electricians provided power, ran physical networking lines for over 10 pieces of equipment and set up a charging station for the lab's mobile robot."
Bridges added that bringing the lab online wouldn't have been possible without talented operations staff. "This work is complex in and of itself, but they had so many safety and performance aspects to consider, since experiments run while most people are sleeping," he said. "Thanks to the staff, I can now focus on delivering the science."
Infrastructure of an autonomous lab relies on diverse expertise
Autonomous laboratories depend on a tightly integrated network of equipment, controls and data systems. Experiments can run continuously. Instruments communicate in real time. Data flows directly into computational models that guide the next step.
"Supporting the kind of environment required for an autonomous lab doesn't happen overnight," Moore said. "It takes long-term planning and people who understand how every piece fits together."
F&O oversees electrical distribution, utilities, high-capacity networking and control systems that keep experiments stable and safe. Its nearly 1,300 staff members maintain the complex energy and infrastructure demands of ORNL's vast campus. Utility, security and fire protection employees work around the clock to oversee and safeguard the infrastructure. Autonomous labs add new operational demands, requiring steady power, higher data rates, uninterrupted connectivity and systems that can respond instantly to change.
Keeping pace with autonomous science requires operations teams to anticipate future demands long before new instruments or AI systems arrive in a laboratory. F&O engineers, architects and project managers ensure ORNL can grow ahead of its scientific goals. They also oversee demolition of older labs and construction of entirely new ones.
The result in the case of ORNL's autonomous labs is a new model of discovery - faster, more efficient and driven by data - to advance energy discoveries and innovations. Moore said autonomous science succeeds because researchers and operations teams develop solutions together. "They don't just oversee these systems; they help build them," he said. "We rely on that partnership to keep pace with the science."
From support to innovation, F&O staff excel
In many cases, operations staff are doing more than maintaining systems and supporting science - they are innovating.
Darren Loposser, an instrument technician who supports research facilities and fabricates control systems at ORNL, is one example. After a customer asked his team to support a project, Loposser saw the opportunity to transform an old lab space into one that would help achieve that project's goals. Plus, transforming the old lab and surrounding spaces would set the stage for future autonomous science.
His upgrades, which are in various stages of completion, include a system to connect multiple laboratories into a single control network that can operate and share information continuously. The system will allow researchers to monitor experiments, track performance and respond in real time. It supports ORNL's "Labs of the Future" vision, pioneered through the Interconnected Science Ecosystem (INTERSECT) Initiative directed by Moore. This connected system will help integrate labs into a communicative ecosystem for 24/7 achievement.
Loposser's work also creates structured data streams (dense collections of data points about an experiment or piece of equipment) that AI tools can use to analyze results, detect problems and guide future experiments.
"We're bringing the power of industrial automation to research," said Loposser, who proposed a similar approach a couple of years ago when he was still unaware of INTERSECT. "We have a lot of new staff. Their predecessors had institutional knowledge gained by years of experience, which allowed them to safely perform experiments using complex systems. My goal is to capture and automate that knowledge to provide systems that are safer and intuitive to operate."
Chris Brewer, a strategic program manager in ORNL's Computing and Computational Sciences Directorate, said he wishes more people knew about hidden heroes like Loposser. "Darren and the diverse cast of F&O staffers won't win awards or receive patents for their work. They see it as simply doing their jobs and don't seek praise, but I think all of the lab, and the public, need to know how essential their work is to the mission of pushing discovery and innovation forward at ORNL."
UT-Battelle manages ORNL for DOE's Office of Science, the single largest supporter of basic research in the physical sciences in the United States. The Office of Science is working to address some of the most pressing challenges of our time. For more information, please visit energy.gov/science . - Clint Keeton
