HTML, HFIR team with industry to find hidden stresses in materials
Whether it's that klunky new Hummer you bought for a song, that honey of a BMW sports coupe, farm equipment from John Deere, or a pressurized water nuclear reactor, the materials from which they are made contain hidden stresses that affect safety, longevity, even energy efficiency and the environment.
Camden Hubbard and his Residual Stress User Group at the High Temperature Materials Laboratory (HTML) User Program at ORNL use neutron diffraction at HFIR (HB-2B beam line) to help industrial partners such as Metalsa and John Deere to find out where these stresses are. Their input corrects or validates the companies' computational models so that they ultimately can make better, safer, less costly, and more efficient structural materials.
In another project of importance to the country's 100+ pressurized water reactors, the team works with the Electrical Power Research Institute (EPRI) and the Nuclear Regulatory Commission (NRC) to detect stresses in the welds on pipes that conduct coolant water from reactor vessels.
The user program at the Neutron Residual Stress Facility (NRSF) at HFIR is funded by DOE's Energy Efficiency and Renewable Energy Vehicles Technologies Program. The current HB-2B beam line-which the team calls NRSF-2-is the second-generation instrument for the work they do and dates from 2006.
Residual stresses are innate to materials. When industry makes a vehicle, the materials are subjected to heat, shot blasting, or laser blasting, as manufacturers modify them to get improved performance life and corrosion resistance. "They're always playing with the process to change the residual stresses to minimize the detrimental and to maximize the beneficial," Hubbard explained.
When the metal components in a vehicle are welded together, and the vehicle hits the road for several hundred thousand miles, further stresses occur. Such stresses can be either detrimental-opening up the material and contributing to crack formation-or beneficial, compressive stresses that push the materials closer together, making parts last longer.
Today, manufacturers use computer modeling to try to predict the processes within materials. Then they arrive at HTML and work with the scientists there, using the instrument at HFIR to validate their models. "They want to see if their models are making the right predictions, so that they can then use the models with more confidence." Hubbard said.
The material scientists are the interface between the computational modelers that industry uses, and the neutron science researchers at HFIR. "They're the people who truly understand the material. The modelers work on the algorithms to make the models right. But it is the material scientists who are getting all the properties of the materials together, all the input that the model needs. It's a big team effort."
The actual science entails bringing a beam of thermal neutrons to diffract off the sample material. At HFIR, a monochromater selects one energy of neutrons to feed through the beam. "When you know the energy and you can see at what angle the neutrons are scattered, you can measure the innerplanar spacing, or despacing, of the atoms in the materials," Dr Hubbard explained.
Pressure brought to bear compresses such spaces between atoms, whereas pulling the material apart increases them. The researchers are trying to measure that change, whether there are compressive or tensile stresses on the atoms in the materials.
The changes in the material may be 100 parts per million, Hubbard said, so this is one of the more precise measurements done with neutron instruments at ORNL. Fortunately, the NRSF-2 is on the order of 50 to 100 times as efficient as its predecessor. "Now instead of a single sample, we can study 10 different samples and compare them. We can run the process faster, or run it slower, run it at higher and lower temperatures, make all kinds of variations in the process, and then see how such changes affect the residual stresses."
Three recent projects at HFIR involve industrial partnerships with Metalsa S. de R.L. (car and truck chassis frames, suspension modules, structural stampings, and truck side rails), EPRI (pressurized water reactors), and John Deere (farm and forestry equipment).
Metalsa makes 50% of the steel frames for trucks and SUVs in North America. The frame is the unifying base of the vehicle to which everything-the engine, the cab, the joints that hold the axles-is attached. As a vehicle travels, there are flexing stresses, and the manufacturer wants to know what effect boring holes in the frame to attach components will have on residual stresses in the material.
"They were studying four different processes for making the bolt holes. If you attach parts to the side rail, and it's going down the road for 500,000 miles or so, you don't want that hole to crack open and have your wheels detach. They really need to know what the crack propagation loads are that might lead to a catastrophic failure."
The team studied the four techniques and found that it is feasible to cut some holes into the frame just to reduce weight. They also found places where thinner materials may be used to reduce overall vehicle weight, resulting in saving expensive raw materials and improving fuel efficiency.
Metalsa researchers now can correlate this research at HFIR with their economic models. "They said this information-not just ours but the whole project-could lead to savings of as much as 200 pounds per truck." Hubbard said.
The collaboration with EPRI concerns ongoing questions about joints that are welded to the country's 100+ pressurized water reactors (PWRs). In PWRs, the coolant water exits the reactor vessel, typically a carbonized steel, via a large stainless steel pipe that is welded to the vessel wall with a weld called a "dissimilar metal weld."
In their inspections of PWR plants, EPRI and NRC have discovered the weld is one of the places that problems occur. The residual stresses at the joint can either close cracks and keep them from breaking, or open cracks faster. The researchers used the neutrons to look for the stress distributions so they can project crack growth and the probability of failure.
Reactors and their attachments are huge. To date, the work at HFIR has been on small models so that the neutrons can actually do the work, Hubbard explained. "Then we try to validate the models and then see if we make a bigger component, do we get the right answers, again."
Progress has been good. "I have the last of the experiments on a couple of the mockups scheduled for the next beam cycle, and then we'll process all the data and turn it over and let them compare with the models."
The third project, with industrial partner John Deere involves stresses in cast materials. When metal is melted and poured into a die, it cools differently at different locations, which sets up residual stresses in the casting.
"They need to be able to predict these stresses to ultimately predict proper performance of the part. So they bring us several parts with different casting operations, to see if their computer models for die casting are giving the right answers.
"We did neutron residual stress measurements on several of these and it turns out the data predictions from the models and the measurements that we had were pretty good-as good as I ever expected they could be."
Confident that they have it right, John Deere can now look at how to optimize the casting, at what temperature, at what speed to do the cooling, now that the computer simulations have been validated by neutron experiment at HFIR.
"This is a big-impact study," Hubbard said. Castings are heavily used throughout all transportation industries, even in jet planes. Ultimately, John Deere hopes to be able to computer model the building of an entire tractor or a road grader, with validation coming from neutron science at places such as HFIR.
Users have access to 50% of the time on the instruments, and DOE-funded science research projects have the remaining 50%. "So we have both industrial users coming, and then DOE-funded projects or NRC-funded projects, to balance the usage," Hubbard said.
"The good aspect of the user program is that they get to understand what the method can do and the accuracy and the precision of the measurements, and they get to help direct the measurements.
"User projects are good in the sense that we try to be really productive. Neutrons are expensive and we try to use them as productively as we can."
December 28, 2010