Predictive Model Development of Bio-Based Self-Healing Concrete using Gene Expression Programming (GEP) Technique

Abstract

A new predictive model of the crack width healing for bio-based self-healing concrete using gene expression programming is presented in this study. To model the healing of cracks, an extensive and reliable database of crack healing results of bio-based self-healing concrete is developed through a comprehensive review of existing experimental studies. The established database contains 843 crack healing results, collected from numerous published research work. After having a deep investigation of these research studies, the ten most influencing parameters including cement, fine aggregate, coarse aggregate, water-to-cement ratio, bacterial concentration, bacterial strain, immobilizer, curing type, days of healing and initial crack width are considered as the input parameters or predictors for the modeling of crack healing properties of self-healing concrete. The performance of model is assessed statistically as well as experimentally. The effective contribution of each parameter in healing the cracks is also evaluated. An empirical expression for the prediction of healing of cracks is developed. This research study is useful for the preliminary design of bacteria- based self-healing concrete.