Soft computing applications through artificial intelligence (AI) are becoming increasingly popular in civil engineering. From concrete technology to structural engineering, AIs have provided successful solutions to various problems and greatly reduced the computational costs while achieving excellent prediction accuracy. In this study, a review of the main artificial neural network (ANN) types used in civil engineering is presented. Each ANN type is described, and example applications are provided. As a new research contribution, a deep feedforward neural network (FFNN) is developed to predict the load capacities of post-installed adhesive anchors installed in cracked concrete, which is challenging and computationally expensive to achieve with conventional methods. The development of this FFNN is discussed, the influence of several parameters on its performance is demonstrated, and optimum parameter values are selected. In addition, a hybrid methodology that combines 2D nonlinear finite element (NLFE) techniques with the developed FFNN is briefly presented to account for real-life adverse effects in anchor analysis, including concrete cracking, wind-induced beam bending, and elevated temperatures. The results show that the developed network and methodology can rapidly and efficiently predict the load capacities of adhesive anchors installed into cracked concrete, accounting for the damage caused by the cracks, with high accuracies.

TUBEDECK, a one-way spanning voided composite slab, has been utilized in the construction field over the years to enhance the efficiency, constructability, and environmental performance of structures. TUBEDECK incorporates both cast-in-situ reinforced concrete slabs and profiled steel decks. However, there is a need to clarify the shear resistance capacity in this slab because the shear strength of the member reduces as concrete volume is eliminated to optimize flexural strength. Therefore, this study applied the artificial neural network (ANN) technique to determine the shear strength of TUBEDECK. By varying factors in the ANN features, several ANN models were developed. Out of many models developed, an optimal model was selected, having a maximum/mean relative errors of 5.1% in a dataset.

A computational framework employing machine learning (ML) is applied to predict failure mode of ultra-high-performance concrete (UHPC) beams. For this purpose, results from a number of tests on UHPC beams with different geometric and loading configurations and material characteristics are collected and utilized as an input to the ML framework. Results from numerical studies are not included in the data set due to the fact that they are highly dependent upon the adopted material models, meshing practices, as well as other assumptions used in modeling. Artificial neural network is used to predict the failure mode of the UHPC beams. Results indicate that the proposed ML framework is capable of predicting failure mod of UHPC beams with varying reinforcement and configurations, and can be considered for use in design applications. This paper aims to promote the applicability of ML for a practical engineering problem, detecting structural response of UHPC beams.

The probabilistic study of fire exposed structures is laborious and computationally challenging, especially when using advanced numerical models. Moreover, fragility curves developed through traditional approaches apply only to a particular design (structural detailing, fire scenario). Any alteration in design necessitates the computationally expensive re-evaluation of the fragility curves. Considering the above challenges, the use of surrogate models has been proposed for the probabilistic study of fire exposed structures. Previous contributions have confirmed the potential of surrogate models for developing fragility curves for single structural members including reinforced concrete slabs and columns. Herein, the potential of regression-based surrogate models is investigated further with consideration of structural systems. Specifically, an advanced finite element model for evaluating the fire performance of a composite slab panel acting in tensile membrane action is considered. A surrogate model is developed and used to establish fire fragility curves. The results illustrate the potential of surrogate modeling for probabilistic structural fire design of composite structures.

Artificial Intelligence has one of the most efficient methods for solving engineering and materials problems because of its impressive performance and can reach higher accuracy. The Deep Learning theory is an approach based on Deep Neural Networks for establishing numerical analysis and value predictions. This paper proposes a fresh approach, using a Deep Learning model for predicting the compressive strength in a particular concrete just based on non-destructive test measurements (NDTs). The model proposed is an attractive alternative to estimate the resistance of compressive strength in any structure, just taking data like ultrasonic pulse velocity, electrical resistivity, and resonance frequencies. The present work employs data science techniques to find the correlation values between the NDTs and the compressive strength effort and realized broad numerical exploration about concrete performance. An amount of 285 specimens of concrete were monitored during this research. The model proposed contains 600 neurons and uses a Rectified Linear Unit and Sigmoid as activation functions where the NDTs were established as the input data. The dataset was segmented into two groups: train and test. In order to evaluate the model, the authors tested it in a validation set with different concrete features, achieving an accuracy of 94%.