Predictive models of behavior and capacity of FRP reinforced concrete columns

Abstract

This paper proposes a new model for predicting the axial capacity and behavior of Fiber Reinforced Polymer-reinforced concrete (FRP-RC) columns using a promising variant of Genetic Expression Programming (GEP). Current design codes, such as the ACI 440.1R-15 and the Canadian Code CSA S806, disregard the compressive contribution of FRP bars when used in compression members. The behavior of concentrically short FRP-RC columns has been widely investigated in the past few years; however, limited research has been dedicated to investigating the effect of load eccentricity and the slenderness ratio of FRP-RC columns. In addition, the methodologies adopted for including the effect of column slenderness remain a subject of debate, as no solid conclusions are withdrawn in this regard. In this paper, the experimental results of FRP-RC columns are gathered from the literature and used to formulate two GEP models to predict the axial capacity based on load eccentricity. The experimental data includes columns reinforced with different FRP types and subjected to concentric and eccentric axial compressive loads. In addition, the database comprises short and slender columns. The proposed GEP models are functions of concrete compressive strength, longitudinal reinforcing bars ratio, FRP bars elastic modulus, eccentricity level, and column dimensions. For the aim of comparison, a preliminary evaluation of previously suggested empirical equations/models for estimating the axial capacity of FRP-RC columns was carried out over the collected database. The proposed models showed superior accuracy in axial capacity prediction with coefficients of determination R2 equals to 0.978 and R2 equal to 0.992 for eccentric and concentric axial load, respectively. The proposed models were found to give reliable estimates of the axial capacity of columns reinforced with FRP longitudinal bars. Finally, a parametric study to evaluate the effect of each variable on the proposed models was conducted.

Neural network‐based models versus empirical models for the prediction of axial load‐carrying capacities of FRP‐reinforced circular concrete columns

“…However, there was a slight variation between the R 2 values of the developed DNN and CNN models with the other established ML models. Further, a similar approach in terms of selection of critical parameters (see Table 2) for the prediction of axial load‐carrying capacities of FRP‐RC columns was adopted by Almomani et al 76 using genetic expression programming (GEP) algorithm. The accuracy of the established DNN and CNN models ( R 2 = 0.931 and 0.955) were validated and shown a reasonable agreement with the previous GEP ML models ( R 2 = 0.978 and 0.992 for eccentric and concentric axial‐load, respectively).…”

Punching Shear of FRP-RC Slab–Column Connections: A Comprehensive Database

“…This also implies that the models are applicable to the whole range of variables. Additional development in this area can be found in recent publications [51][52][53][54].…”

Bond strength prediction of UHPC-NSC interface

“…Generally, ANN is considered a multiple layer framework with three fundamental layers: a fee (input) layer, one or multiple hidden layers, and a target (output) layer. Each layer is made up of several neurons (processing stations) that are fully tethered to the neurons in the following layer in order to replicate the human nervous system, as shown in Figure 2 [19]. Each signal travels through neurons from one layer to the next, passing through processing units.…”