Machine Learning Applications in Development of Metal Matrix Composites
Mentor 1
Pradeep Rohatgi
Mentor 2
Amir Kordijazi
Mentor 3
Swaroop Behera
Mentor 4
Kaustubh
Start Date
16-4-2021 12:00 AM
Description
This project reviews literature on the applications of Machine Learning (ML) in the development of Metal Matrix Composites (MMCs). MMCs are a type of composite material that consists of fibers or particles surrounded by a metal matrix. Because the use of MMCs in a product can increase its stability and lightness without compromising its performance, they have many applications in industries such as automotive and aerospace. ML is a subfield of Artificial Intelligence (AI) in which algorithms are designed to adapt and improve themselves as they acquire more data. Due to contributions from the Material Genome Initiative (MGI) and similar programs, there is currently an abundance of data in the field of Materials Science. With the use of this data and data produced from new lab experiments, researchers have used ML algorithms to optimize certain properties of materials, such as hardness, tensile strength, and surface roughness. We reviewed recent academic publications that applied ML specifically to MMC research. ML models were compared. The models were analyzed for accuracy, as well as what types of datasets they can be used on. The review suggests some models will work better for certain manufacturing methods of MMC. The results can be used to further Materials Science research at UWM.
Streaming Media
Machine Learning Applications in Development of Metal Matrix Composites
This project reviews literature on the applications of Machine Learning (ML) in the development of Metal Matrix Composites (MMCs). MMCs are a type of composite material that consists of fibers or particles surrounded by a metal matrix. Because the use of MMCs in a product can increase its stability and lightness without compromising its performance, they have many applications in industries such as automotive and aerospace. ML is a subfield of Artificial Intelligence (AI) in which algorithms are designed to adapt and improve themselves as they acquire more data. Due to contributions from the Material Genome Initiative (MGI) and similar programs, there is currently an abundance of data in the field of Materials Science. With the use of this data and data produced from new lab experiments, researchers have used ML algorithms to optimize certain properties of materials, such as hardness, tensile strength, and surface roughness. We reviewed recent academic publications that applied ML specifically to MMC research. ML models were compared. The models were analyzed for accuracy, as well as what types of datasets they can be used on. The review suggests some models will work better for certain manufacturing methods of MMC. The results can be used to further Materials Science research at UWM.
