Mechanical properties of sustainable self-healing concrete and its performance evaluation using ANN and ANFIS models

K. Shanthi Sri,R. Ramesh Nayaka,M. V. N. Siva Kumar
DOI: https://doi.org/10.1007/s41024-023-00345-8
2023-09-29
Journal of Building Pathology and Rehabilitation
Abstract:In order to address the challenges to repair uncontrollable cracks and to enhance the life span of civil engineering structures in a sustainable manner the technique of self-healing of cracks by introducing bacterial concrete has been developed, researched and implemented by many researchers. The present paper focuses on the overview of self-healing concrete (SHC) and conducting laboratory investigation of the prep bacteria, feeding material and water. The study experimentally investigates its compressive strength and explores an alternate mathematical approach to predict its mechanical characteristics. An attempt is made to predict the compressive strength of self-healing concrete with varying proportions of calcium lactate using an artificial neural network (ANN) and adaptive neuro-fuzzy interface system (ANFIS). It has been discovered that using bacteria improves the compressive strength of normal concrete and aids in the self-healing property of concrete. The best prediction models that can learn, compute, and solve problems with non-linear data are the artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS). The study finally compares the compressive strength values obtained from experimental work, ANN and ANFIS models and determines the best suitable model out of both ANN and ANFIS for the considered dataset in the study. The study concludes that the ANN design model produces the most accurate results with a regression value of 0.9865 and a mean square error of 3.07 in comparison to the regression value of 0.9725 and a mean square error of 3.16 obtained from the ANFIS model.
What problem does this paper attempt to address?