Argonne National Laboratory (ANL) scientists have developed a new optimization tool called ‘ActivO’ to slash product design costs.
ActivO is a hybrid algorithm that leverages the strengths of two different machine learning surrogate models to help users focus on how to efficiently target computational resources.
According to postdoctoral appointee Opeoluwa Owoyele and research scientist Pinaki Pal, this new AI process could help fast track the best design and simulation for engines and other manufacturing essentials.
“ActivO runs the simulations in a very smart way and quickly identifies the parts of the design space we should focus on,” explains Pal.
“A process that used to take two to three months to give you the optimum design can now be completed within about a week.”
The ActivO approach was successfully demonstrated for use in optimizing combustion engines.
According to Owoyele, the machine learning models are designed to work cooperatively:
“Rather than run simulations that are sampled randomly, one of the models allows us to explore the design space adaptively, which essentially guides us to the regions most likely to contain the global optimum. And the other model looks in those promising regions and performs a local search to identify the exact location of the global optimum.”
“ActivO is well suited to explore the design space that is typical of combustion engines in the automotive and aerospace industries,” notes Owoyele.
“But it could also be used for design optimization for a general product. It can be readily adopted by industry to improve their design workflows, which would slash product design costs. There are significant commercialization opportunities in this regard.”
There is an even more important long-term benefit in helping lower the design costs for industry, says Pal.
“The big-picture goal is to enable energy efficiency and lower the environmental impact of these engines. So ultimately that’s why we want to design better, more innovative engines.”
Image and content: Roman Zaiets-Shutterstock/ANL
