Skip to main content
Research Highlight

Moving molten salt: Engineers research instrumentation and controls for 700 degree Celsius environment

When it comes to a challenging application for embedded instrumentation and control, none quite beats an environment of molten salt at 700 degrees Celsius.

But that is just the application chosen by scientists at the US Department of Energy’s Oak Ridge National Laboratory to demonstrate the advantage of embedding controls—or systems that drive moment-by-moment adjustments to a machine or process to improve efficiency and reliability.

“We wanted to do something that would be very difficult, if not impossible, to do using conventional machine design techniques,” said researcher Alexander Melin of ORNL’s Electrical and Electronics Systems Research division. “But we also wanted a testbed that could benefit a great deal from embedded instrumentation and control.”

The scientists chose to investigate a pump that could tolerate placement in a fluoride salt reactor—circulating the molten salt used as a coolant. ORNL is leading DOE’s research and development effort on these advanced reactors amid renewed interest.  

The legacy approach to moving molten salt in a test loop is a sump pump hanging over the coolant tank. But although that design is sufficient for research purposes, it doesn’t scale well to a commercial application because of the challenges and expense of shaft machining and balancing, as well as bearing life, seal degradation, and low pump efficiencies. 

‘Cherry red’ and fully functional

Pumps traditionally employed in nuclear reactors are very large machines with non-variable speeds and a great deal of mass; they contain various seals, bearings, and windings—all components that would quickly degrade in molten salt.


The challenge is to build a pump that can tolerate immersion in liquid salt under harsh operating conditions, function “cherry red” as a normal operating condition at 700 degrees Celsius, and at the same time provide greater performance with less size, mass, and weight, said ORNL’s Roger Kisner, who leads the project.

The solution being explored by Kisner and Melin is to use a canned rotor pump that eliminates the need for rotating seals and mechanical bearings. Other researchers contributing to the project include Tim Burress and David Fugate.

A canned rotor pump, commonly used in the food service industry, is a seal-less design in which a thin metal sheath separates the fluid being pumped from the electrical components. The design would allow the pump to operate submerged in salt by using a material to shield the electrical motor (stator) components.

The researchers opted for a magnetic bearing suspension system, rather than mechanical bearings, to control the rotor shaft. In this design, electromagnets apply force to the shaft to control its position, since conventional sensors could not be used in the extreme environment.

Physics, math replace sensor function

The sensorless design “uses the physics of the system and some clever mathematics to detect shaft position without having a physical sensor,” Melin explained.

“Instead of doing all the design in the mechanical realm, you can virtualize these functions through electronics and embedded instrumentation and control,” Melin said. “It provides us new opportunities for increasing efficiencies and improving reliability. And it opens up new operating regimes that weren’t possible because of restrictions on materials and other types of mechanical restrictions.”

After conducting feasibility studies and crafting a conceptual design, Kisner and Melin developed a test bed to simulate and model their ideas regarding magnetic bearings. The next step is building a larger loop-size, canned rotor magnetic suspension pump that will operate in a room-temperature water environment. This will allow the researchers to look for performance issues and resiliency to disturbances and faults, Melin said.

Future research could involve building a magnetic suspension to test the pump in a liquid salt environment. 

The project is funded by DOE as part of its Nuclear Energy Enabling Technologies Program, which is developing crosscutting solutions in support of the DOE Office of Nuclear Energy’s programs.

The goal of the project is to provide scientists with a framework to develop high-temperature electromagnetic components with embedded instrumentation and control. These component designs are necessary to the development and commercialization of modern advanced reactors and could also be deployed in solar energy and future spacecraft.

Collaboration, system-level design required

Making such sophisticated embedded systems possible are technical advances such as computational system improvements, new knowledge of control theory, and sensor improvements, Kisner said. 

He noted the importance of a ground-up, collective process to designing an embedded system. “When you say you’re going to do an embedded design you’re not just embedding some electronics on something that’s already created. It’s not business as usual,” Kisner said.

“You have to conduct system-level engineering and design of the final product. You will need various people involved in the mechanical design, magnetic and thermal design, materials specification, sensors and instrumentation, control systems, electronic drive systems, and so on,” Kisner explained. “It’s not compartmentalized. So what could have been separate topical design areas now have to be integrated, and that also requires better management skills and control of a project.”

“The idea of a fully integrated embedded system fits in with our capabilities here at the lab,” Kisner said. “At ORNL, we have people who do materials, who do mechanical, who do instrumentation and electronics. We pride ourselves on our ability to pull groups of people together to do things that you may not be able to do at a university or at a single company.”

Melin concurred: “We all have our own areas of expertise, but for a project that is so multidisciplinary, the lab environment works well. If you have a materials science question, you can go to a world expert down the hall. If you have a problem with power electronics design, you can go to a world expert just down the hall. The depth and breadth of knowledge at the lab help us with these large, complex projects.”

More details on the project are available at:

UT-Battelle manages ORNL for the DOE’s Office of Science. The Office of Science is the single largest supporter of basic research in the physical sciences in the U.S., and is working to address some of the most pressing challenges of our time. For more information, please visit