About this Abstract |
Meeting |
MS&T23: Materials Science & Technology
|
Symposium
|
Ceramics and Glasses Modeling by Simulations and Machine Learning
|
Presentation Title |
Machine Learning Prediction of Heat Capacity for Solid Mixtures of Pseudo-binary Oxides |
Author(s) |
Julian Barra, Simone Audesse, Rajni Chahal, Stephen Lam |
On-Site Speaker (Planned) |
Julian Barra |
Abstract Scope |
Direct resistance-heated thermal energy storage has been proposed as a low-cost energy storage alternative. The bricks used in these energy storage systems must be composed of oxide mixtures with certain material properties, amongst them a high heat capacity. Methods to predict heat capacity in oxide mixtures face different problems, with the Neumann-Kopp rule underperforming at higher temperatures and computational methods such as CALPHAD and Density Functional Theory being computationally expensive. Machine Learning could be used as a tool for prediction, with the main impediment to its use being the unavailability of a database with the heat capacity data for oxide mixtures. For this work, we use the CALPHAD method to generate a heat capacity dataset for oxide mixtures, and use it to train and test Machine Learning algorithms to predict molar heat capacity. Some resulting algorithms obtain predictions with an r2 over 0.98 and a MAPE under 0.7%. |