Soft computing prediction of mortar strength with copper slag in aggressive environments

To develop sustainability, a major slag-based industrial waste – copper slag – is adopted in the present work to develop mortar. An experimental study is carried out under two different exposures of mortar freezing–thawing and acid attack. The parameters that vary in relation to the mortar are: the percentage of copper slag used as a replacement for fine aggregate (0 to 25%); curing period (14, 28 and 56 days); and specimens with and without epoxy resin and water repellent agent coatings. The compressive strength of mortar with copper slag exposed to an aggressive environment is analysed using an ultrasonic pulse velocity test. In addition, the authors develop two deep learning models: an artificial neural network (ANN) and an adaptive neuro-fuzzy inference system (ANFIS) to predict the compressive strength of the mortar with copper slag. The experimental results obtained are split into training, testing and validation datasets used in designing ANN and ANFIS models. The developed ANN and ANFIS models have a high correlation coefficient of 0.9917 and 0.9644 and a low mean squared error value. Furthermore, compared to the ANN model, the ANFIS model shows minimal error in predictions with a high degree of accuracy when forecasting the compressive strength of the mortar with copper slag.