An artificial neural network approach to predict particle shape characteristics of clay brick powder under various milling conditions

Abstract

The prediction and monitoring of shape descriptors of pozzolans generated from ball milling processes remain today unattended by researchers. Since the particle shapes of pozzolans influence the performance of pozzolanic materials in cement-based composites, an accurate characterisation of particle shapes of pozzolans is required. In this research, we use artificial neural networks (ANNs) to predict the 2-d shape parameters of clay brick powder (CBP) particles considering various milling treatments. An integrated prediction of particle shapes is developed by combining the ANN models with image analysis. Through R values of greater than 0.96 and reduced mean square errors (MSEs) for the ANN models, it is shown that our ANN models can be able to predict the shape parameters of CBP particles under various milling treatments. Moreover, the best validations of our models are achieved at the cost of less than 5 epochs. Collectively, these results increase our understanding in the prediction of the particle shapes of CBP under various milling treatments, setting the stage for additional studies, especially in other pozzolanic materials.