Material Systems Design with Reduced Order Methods and Mechanistic Machine Learning

Seminar | October 24 | 4-5 p.m. | 3110 Etcheverry Hall

 Wing Kam Liu, PhD, PE, Director of Global Center on Advanced Material Systems and Simulation; Walter P. Murphy Professor of Mechanical and Civil Engineering, Northwestern University

 Department of Mechanical Engineering (ME)

Abstract: As in all everyday applications, in engineering problems, the volume of data has increased substantially compared to even a decade ago but analyzing big data is expensive and time-consuming. Data-driven methods, which have been enabled in the past decade by the availability of sensors, data storage, and computational resources, are taking center stage across many disciplines of science, e.g. physics, material science and information science. Mechanistic machine learning (MML), by integrating prior mechanical knowledge and machine/deep learning algorithms, provides a new solution to longstanding scientific and engineering challenges, and will accelerate material systems design and discovery by leveraging advances in mechanics of materials science and machine learning algorithms. A general procedure of mechanistic machine learning includes high-dimensional high-fidelity data generation and collection, feature engineering and selection, dimension reduction, regression, classification, and data-driven prediction. New mathematical and computational paradigms are being developed, which assimilates multi-fidelity data (e.g., heterogeneous materials, complex microstructures, processing conditions) into large-scale simulation to speed up the computation while capturing the “hidden physics” that was not possible through conventional methods and experimentations.

Two challenging applications will be presented: composite material systems and additive manufactured alloys. Modeling of composite material systems requires a considerable database for complex microstructures (e.g., Unidirectional, woven composite). Based on geometries (features), microstructure resolution (fidelity), and other information hidden in the database, machine learning techniques can be used to construct a reduced-order database for the prediction of composite property and structure performance through a predictive mechanistic model together with machine learning. The second application is additive manufacturing. We will demonstrate a data-driven methodology for discovering low-dimensional scaling laws with the property of dimensional homogeneity by combing dimensional analysis and genetic programming for the synthesis and design of additive manufactured alloys. Two strikingly simple but universal scaling laws for keyhole depth and porosity are identified. Finally, the potential application of mechanistic machine learning for linking process-structure-properties in additive manufacturing of metals will also be discussed.

Biography: Professor W.K. Liu is the Walter P. Murphy Professor of Northwestern University, Director of Global Center on Advanced Material Systems and Simulation, President and Past President of the International Association for Computational Mechanics (IACM) (President (2014-2018) Past President (2018-2024)), Past Chair (2017-2018) (Chair 2015-2016) of the US National Committee on TAM and Member of Board of International Scientific Organizations, both within the US National Academies. Liu’s selected honors include Japan Society of Computational Engineering Sciences Grand Prize; IACM Gauss-Newton Medal (highest honor) and Computational Mechanics Award; ASME Dedicated Service Award, ASME Robert Henry Thurston Lecture Award, ASME Gustus L. Larson Memorial Award, ASME Pi Tau Sigma Gold Medal and ASME Melville Medal; John von Neumann Medal (highest honor) and Computational Structural Mechanics Award from USACM. Fellow of ASME, ASCE, USACM, AAM, and IACM.

 fma@berkeley.edu, 510-643-6527