The production of fiber-reinforced concrete presents a complex reaction system, posing significant challenges in determining appropriate material proportions to achieve targeted mechanical properties. To address this issue, this study proposes novel Artificial Neural Network (ANN) models for predicting the compressive and flexural strengths of fiber-reinforced concrete using a backpropagation feed-forward algorithm. A wide range of concrete mix designs was prepared and tested using cylindrical samples for compressive strength and beam samples for flexural strength. Polypropylene fibers were incorporated into the mixes, and all specimens were cured for 28 days in a water-saturated lime solution. The results demonstrated that the ANN models produced strength predictions that closely aligned with experimental data, yielding high correlation values of 99.46% and 98.57% for compressive and flexural strengths, respectively. The best-fit models exhibited mean squared errors of 0.0024 (compressive) and 0.44 (flexural). Furthermore, parametric analysis indicated that the proposed models effectively captured the constitutive relationships among the concrete components and successfully represented the dominant mechanical behavior of the tested specimens.

o monitor the initiation of concrete cracking beyond the service life of the structure, a novel prediction model of time to cracking of concrete cover using artificial neural network (ANN) was developed in this study. Crack mitigation prevents corrosion and crack development to occur in a more rapid phase that is an essential component in performance-based durability design of reinforced concrete structures. Data available in various literatures were used in the development of the ANN model which is a function of compressive strength, tensile strength, concrete cover, rebar diameter, and current density. The neural network model was able to provide reasonable results in time predictions of cracking of concrete protective cover due to formations of corrosion products. The performance of ANN model was also compared to various analytical and empirical models and was found to provide better prediction results. Even with limitations in the available training data, the ANN model performed well in simulating cracking of concrete due to reinforcement corrosion.

Predicting the compressive strength of self-compacting concrete (SCC) is one of the complicated tasks because of its complex behavior due to the reaction of chemical and mineral content and the hydration process of cement. The distinct difference of SCC to normal concrete is its improved workability or also known as rheology that is divided into four categories namely viscosity, flow ability, passing ability, and resistance to segregation. It was proposed in this study to include the rheological behavior of SCC to the prediction of its compressive strength. Neural network was utilized for predicting the 28th day compressive strength of SCC. A 97.78% prediction rate was achieved using a feed-forward back-propagation ANN with 1 hidden layer and 8 hidden nodes. Tansig transfer function was used as activation function. The model has a Pearson R value of 0.991 and mean square error (MSE) of 3.42.

Many construction companies and individuals (construction designers) are still using spreadsheets and laboratory tests just to obtain a certain data. In the field of technologies, advancement will contribute to the improvement of designing structures in terms of usefulness and effectiveness. By using the principle of artificial neural network, this study developed a sorptivity model which gives immediate quantities with high accuracy and precision which are needed to attain appropriate sorptivity values of concrete design mix. In this study, 40 concrete samples with varying percent replacement of copper slag to sand were tested for sorptivity by following the ASTM C1585 which is the Standard Test Method for Measurement of Rate of Absorption of Water by Hydraulic-Cement Concretes. These values in turn were used in the development of the sorptivity model using Artificial Neural Network. This study used the software called Matrix Laboratory (MATLAB) to train several neural networks. Several numbers of neurons in the hidden layer were considered because there is no actual study that suggests that a certain number of nodes in the hidden layer produce the best model. A parametric testing was conducted to determine which of the parameters considered have the greatest significance to the target output. The predicted results of the best model were compared to the experimental values of sorptivity and produced a 2.36 percentage error. The study results suggest that ANN models could be used to predict the sorptivity value of a concrete sample. The model produced a good prediction result.