Principal investigator: Elizabeth Holm
University: Carnegie Mellon University
Industry partners: Perryman Company
The potential for artificial intelligence (AI) to revolutionize manufacturing has received considerable attention. However, traditional industries have been slow to adopt AI, in large part due to legacy equipment and established processes. In this project, we leverage unique, materials science focused AI expertise at Carnegie Mellon University to support the metallurgical analysis processes at Perryman Company in order to reduce costs, increase productivity, and train current and future workforce members in AI applications. The properties of metal components depend on their microscopic substructure (termed the microstructure). During the fabrication of metal products, both process and quality control involve identifying of desirable and undesirable microstructural features. This process is typically performed by a human engineer who scans the sample using a microscope and points out any irregularities and flaws. This is a time-consuming, but essential, step in the metals fabrication workflow. Computer vision (CV) uses computational algorithms to capture the visual content of an image for tasks such as facial recognition and self-driving cars. By combining CV with machine learning (ML), computers can learn to make judgments about images; for example, they can distinguish a flaw from normal material. The goal of this project is to capitalize on these concepts to create an autonomous CV/ML system to evaluate microstructural images with an emphasis on identifying and flagging features that are important for process or quality control. The result will be an Autonomous Microstructural Analysis system that scans a metallurgical sample, analyzes the resulting images to identify ‘normal’ and ‘anomalous’ microstructures, and records and flags anomalous structures for further evaluation and analysis. Both due to direct participation and through on-campus interactions, this project will introduce undergraduate and graduate students to opportunities in the PA manufacturing sector, as well as prepare them to join the manufacturing workforce.