What to expect from heavy clay?

The need of testing the quality of brickclay arises in all brick factories, with the opening of new deposits. The analyses are both time and economically consuming, so the aim of this study was to shorten the procedure using the already known data. This study was focused on determining the usability of heavy clays, when only the raw material major elements chemical composition is determined. The effects of chemical composition, firing temperature, and several shape formats of laboratory samples on the final properties were investigated. Chemical composition of major elements was determined on the basis of classical silicate analysis. Firing was conducted in an oxidizing atmosphere, while maintaining all other experimental conditions constant, except the final temperature. Principal component analysis (PCA) was used to determinate groups of samples according to similarity of chemical composition. Prediction of compressive strength (CS) and water absorption (WA) was done by developing five artificial neural networks (ANN). The average regression coefficients r2 were used to explore the confidence level of the models. Developed models were able to predict CS and WA in a wide range of chemical composition and temperature treatment data, and the highest average r2 of 0.923 for CS was obtained, while r2 for WA was 0.958. The wide range of processing variables was considered in the model formulation, and its easy implementation in a spreadsheet using a set of equations makes it very useful and practical for CS and WA prediction. As it is known from literature, all the parameters entered this analysis are dependent on each other, but their mutual relationship was not quantified yet. Most importantly—the developed neural networks can be used on a global scale.