Machine learning optimization of bio-sandcrete brick modelling using response surface methodology

Abstract

In this study, raw grinded groundnut shell (RGGNS) was used as a fine aggregate in the brick industry to reuse agricultural waste in building materials. In this study, an experimental approach was used to examine a new cement brick with raw groundnut shells integrated with compressive strength, water absorption and dry density optimization utilizing response surface methodology (RSM). The raw ground-nut shell content improved the fine aggregate performance of the 40%, 50%, and 60% samples. The 28-day high compressive strength with the raw ground-nut shell was 6.1 N/mm2 maximum, as needed by the technical standard. Samples made from 40%, 50%, and 60% raw groundnut shells yielded densities of 1.7, 2.2, and 1.9 kg/cm3 for groundnut shell (GNS) brick, respectively. A product's mechanical properties meet the IS code standard’s minimum requirements. RSM was then utilized to develop a model for the addition of raw groundnut shell to concrete. R-square and Adeq precision values indicated that the results are highly significant, and equations for predicting compressive strength, water absorption, and dry density have been developed. In addition, optimization was performed on the RSM findings to determine the efficiency optimization of the model. Following the optimization results, experiments were conducted to determine the applicability of the optimized model.