Search for: All records

Developing an accurate interatomic potential model is a prerequisite for achieving reliable results from classical molecular dynamics (CMD) simulations; however, most of the potentials are biased as specific simulation purposes or conditions are considered in the parameterization. For developing an unbiased potential, a finite‐temperature dynamics machine learning (FTD‐ML) approach is proposed, and its processes and feasibility are demonstrated using the Buckingham potential model and aluminum (Al) as an example. Compared with conventional machine learning approaches, FTD‐ML exhibits three distinguished features: 1) FTD‐ML intrinsically incorporates more extensive configurational and conditional space for enhancing the transferability of developed potentials; 2) FTD‐ML employs various properties calculated directly from CMD, for ML model training and prediction validation against experimental data instead of first‐principles data; 3) FTD‐ML is much more computationally cost effective than first‐principles simulations, especially when the system size increases over 10³atoms as employed in this research for ensuring reliable training data. The Al Buckingham potential developed by the FTD‐ML approach exhibits good performance for general simulation purposes. Thus, the FTD‐ML approach is expected to contribute to a fast development of interatomic potential model suitable for various simulation purposes and conditions, without limitation of model type, while maintaining experimental‐level accuracy.

A large number of microstructural parameters and a wide range of transport physics impose challenges on thermal transport analysis of alloy. Herein, modern data science techniques are employed to overcome the challenges, pursuing effective calculation of thermal transport properties. This emerging approach is tested for precipitate‐hardened aluminum (Al) alloy with consideration of precipitate morphology. The finite element method (FEM) is employed to create a database of effective thermal conductivity of hypothetical Al alloys with varying precipitate morphological and thermal transport features. Using the FEM‐generated data sets, the correlation analysis is conducted to qualitatively evaluate the importance of various precipitate features. The correlation analysis identifies the surface area, average diameter, and volume fraction of precipitates as the most descriptive features for determining the thermal conductivity of alloys. Afterward machine learning (ML) models are trained to accurately predict the effective thermal conductivity. Comparing the ML predictions with effective thermal conductivity and microstructural information from experiments, precipitate thermal transport properties can be calculated, such as interfacial conductance between Al matrix and precipitate, without atomistic simulations. This research demonstrates the feasibility of data‐driven approaches for effective thermal transport calculation and the promise of the FEM‐generated data analysis for more comprehensive evaluation of metallic alloys.