Development of A model to Predict Maximum Crack Width and Spacing in Reinforced Concrete Beams Under Flexural Stress

Abstract

Cracking in reinforced concrete structures is unavoidable due to the low tensile strength of concrete. Even though minimizing cracks on exposed concrete surfaces is significantly important from the aesthetics point of view, it is also very important for the durability and overall performance of the structure. flexural members must be designed to minimize cracking that may occur during the structure's service life. Even though there has been a lot of experimental work on crack width over the last several decades, a literature review reveals that researchers still can't agree on which variables are most important. Part of the reason for this discrepancy is that different researchers have used different factors in their experiments. There are too many variables, and some of them are too interdependent, for a single experimental program to adequately account for all the factors. Several researchers have offered methods for predicting crack width, and an examination of these formulas reveals that they all use different sets of variables. However, experimental test results might be misleading due to concrete heterogeneity, internal and external cracks, and bond behavior.

The primary factor that contributes to the bond is the mechanical interlock resulting from the interaction between the bar ribs and the concrete. The effectiveness of such reinforcement is related to the bond properties of the bars since they must ensure uniformity of deformations of the reinforcement and concrete. Many experimental, analytical, and numerical studies were conducted on the bonding mechanism between concrete and deformed steel bars. Including as many parameters in an analytical study as possible is a key emphasis of this dissertation. An extensive data collection of maximum crack width and spacing from the literature review has been done. A synthesis is provided on the state of practice of computational methods and design criteria for predicting crack widths in reinforced concrete due to flexural stress to select the top models. A parametric study of the important factors that affect crack width and spacing on the selected computational methods from synthesis. An advanced multiple linear regression (MLR) model using R language, and artificial neural networks-based models (ANNs), are used for the prediction of maximum crack width and spacing in flexural RC members. Both models were validated with the real data from the literature review and results show a very good R2 and MSE.