Data collected by scientists at the National Institute of Standards and Technology (NIST) over the past three years is now being used to improve the simulation of laser welding processes.
Such simulations could lead to a better understanding of laser welding, making it easier for industries to consider investing in the technology.
According to the researchers, even with the advantages that laser welding presents, the process currently makes up only a small fraction of overall welding efforts in America.
Simulations could help change this by enabling manufacturers to obtain a better understanding of the process – helping them predict what kinds of welds they can expect when using different materials and parameters.
To make such computer models however, data is needed from past experiments. Currently, that data is spread across hundreds of studies, representing decades of work from dozens of laboratories. Piecing this information together requires introducing a lot of what NIST physicist Brian Simonds calls ‘fudge factors.’
‘Modellers look through all these resources from different labs for different materials, and they kludge them together in a way that they think is most applicable to their experiment,’ he said. ‘And they say, “It's close enough.” But they don't really know.’
The NIST team is therefore attempting to build a much firmer foundation for a model using data that they say is more accurate and comprehensive than any data previously collected on the subject.
The data, collected over the past three years, encompasses everything that a welding simulator would need – the amount of power that is hitting the metal, the amount of energy the metal is absorbing, the amount of material that is evaporating from the metal as it is heated, all in real time.
According to the researchers this information is now starting to be used by computer modelers to improve simulations of laser welding processes, a necessary step to prepare the work for industry.
‘Our results are now mature enough to where academic researchers are starting to use our data to thoroughly test their computer models in a way that they just haven't been able to do before, because this kind of data hasn't been available,’ confirmed Simonds.
The researchers believe that the ultimate goal for industry is that one day, if a manufacturer has an idea about something they want to make, they can input that information into a computer which can then tell them exactly how to make it. Though this ideal scenario is a decade or more away, according to Simonds, manufacturers may start seeing a benefit much sooner as the progress made by the NIST team helps refine computer models.
Expanding horizons
As they continue to gather information, the NIST scientists are collaborating with institutes around the world to expand the dataset. This summer, for example, they will collaborate with the US Department of Energy's Argonne National Laboratory to take advantage of the lab's unique ability to do high-speed x-ray imaging of the molten pool of metal during welding. Other collaborators include Graz University of Technology in Austria, Queen's University in Ontario, Canada, and the University of Utah in Salt Lake City.
Simonds and his colleagues are also broadening the scope of their work as they direct their high-power laser beams onto metal powders instead of solids, which could directly support the community of additive manufacturing.
Where no one has gone before
Many of the techniques the researchers are using to collect the data were either designed or developed at NIST to measure novel aspects of welding. For example, until recently researchers could not gauge laser power during a weld. NIST physicists John Lehman and Paul Williams and their colleagues designed and built a device that can accomplish this using the pressure of the light itself.
They also had to get creative to sense the amount of light absorbed by the heated material, since it changes constantly. ‘You go from a rough metal to a shiny pool to a deep pocket that is essentially a blackbody,’ meaning it absorbs almost all of the light that hits it, Lehman said. The physics, he said, is ‘super complex.’
To solve this problem, they surrounded the metal sample with a device called an integrating sphere, designed to capture all the light bouncing off the metal. Using this technique, they discovered that the traditional method for making this measurement ‘severely underestimates’ the energy absorbed by the metal during a laser weld. The integrating sphere also allows the data to be measured in real time.
They also found a way to better measure the weld plume, a cloud of vaporized materials that includes tiny amounts of elements that evaporate out of the sample during welding. Detecting the exact amounts of these elements as they leave the weld would give scientists valuable information about the strength of the material that remains. However, traditional techniques fail to accurately sense the concentrations of certain elements, such as carbon and nitrogen, that exist in extremely low concentrations.
To sense these minuscule signals, NIST researchers are adapting a technique called laser-induced fluorescence (LIF) spectroscopy. The method involves hitting the plume with a second laser that targets just one kind of element at a time. The targeted element absorbs the second laser's energy and then releases it at a slightly shifted energy, producing a strong signal that is also a unique marker of that element. So far, researchers have demonstrated that LIF can sense trace elements in the weld plume with 40,000 times more sensitivity than traditional methods.
Another important aspect of the work is that researchers are conducting all of their experiments with a type of stainless steel that is a NIST standard reference material (SRM), meaning its composition is extremely well known. Using the stainless-steel SRM ensures that experiments conducted anywhere in the world can have access to metal samples with an identical composition, so that everyone is effectively contributing to one big project.
‘In 20 years from now, if somebody says, 'Oh man, I wish they had measured this,' or some new technique is invented that gives much better data than we can take today, they can go buy the SRM and tie it in to all the research we've already done,’ said Simonds. ‘So it kind of future-proofs what we're doing.’