Compatibility of sustainable geopolymer based on artificial neural network

The present investigation on concrete and cement-based substances seeks to address pressing issues such as rising carbon emissions or corroding impacts on reinforced concrete structures through sustainable means. The use of geopolymer concrete is attributecd to its exceptional sustainability and carbon reduction features. The Global Cement and Concrete Association states that cement is the third most significant emitter of greenhouse gases into the atmosphere. Geopolymer concrete is a new type of concrete that has been developed through extensive research to reduce the harmful effects of cement on the surrounding atmosphere. Geopolymer concrete's engineering characteristics (compressive strength) are often described by testing methods that require large quantities of raw material, significant amount of sample preparation time and expensive equipment. Concrete’s compressive strength is a crucial factor in ensuring its quality. The evaluated methods include artificial neural networks (ANN) and deep neural networks (DNN) based on experimentally collected data. By constructing three multilayer network models and using the Levenberg–Marquardt (LM), Bayesian regularization (BR), and scaled conjugate gradient (SCG) algorithms, the prediction accuracy of the compression was evaluated of bagasse based geopolymer paste at different ages (7, 28 and 56 days) respectively. To evaluate the effectiveness of the proposed methods, various statistical tests are used, such as R squared (R2), Root mean square error (RMSE), and Mean square errors (MSE).