Title:
Using the Bulk Oxide Content to Predict Performance of Fly Ash in Concrete with Machine Learning
Author(s):
Ley
Publication:
Web Session
Volume:
ws_S23_Ley.pdf
Issue:
Appears on pages(s):
Keywords:
DOI:
Date:
4/2/2023
Abstract:
This presentation will show the results of two different machine learning approaches to predict the strength gain, heat produced, diffusion coefficient, and resistivity between 3d and 180d. The first approach uses a kernal principal component analysis (kPCA) method to predict the performance into three categories for more than 40 different fly ashes. The method is implemented with a simple web interface where the user enters the bulk oxide content, and a prediction is made. The second method determines which individual oxides are important to make different predictions at different ages. This method can be implemented with a series of tables and could lend itself to implementation in a traditional specification.