Prediction of Concrete Compressive Strength Using a Back-Propagation Neural Network Optimized by a Genetic Algorithm and Response Surface Analysis Considering the Appearance of Aggregates and Curing Conditions

Influence of the ANN Hyperparameters on the Forecast Accuracy of RAC's Compressive Strength

The artificial neural networks (ANNs)-based model has been used to predict the compressive strength of concrete, assisting in creating recycled aggregate concrete mixtures and reducing the environmental impact of the construction industry. Thus, the present study examines the effects of the training algorithm, topology, and activation function on the predictive accuracy of ANN when determining the compressive strength of recycled aggregate concrete. An experimental database of compressive strength with 721 samples was defined considering the literature. The database was used to train, validate, and test the ANN-based models. Altogether, 240 ANNs were trained, defined by combining three training algorithms, two activation functions, and topologies with a hidden layer containing 1-40 neurons. The ANN with a single hidden layer including 28 neurons, trained with the Levenberg-Marquardt algorithm and the hyperbolic tangent function, achieved the best level of accuracy, with a coefficient of determination equal to 0.909 and a mean absolute percentage error equal to 6.81%. Furthermore, the results show that it is crucial to avoid the use of overly complex models. Excessive neurons can lead to exceptional performance during training but poor predictive ability during testing.

Effect of Corrosive Environment on the High-Cycle Fatigue Behavior of Reinforced Concrete by Epoxy Resin: Experimental Study

Large engineering structures made of various materials, including concrete (e.g., bridges, dams, buildings, and multilevel car parks), steel (e.g., power towers, ships, and wind turbines), or others, are often subjected to severe vibration, dynamic, and cyclic loads, which lead to crack initiation, crack growth, and finally structural failure. One of the effective techniques to increase the fatigue life of such structures is the use of reinforced materials. In the meantime, environmental factors, such as corrosion caused by corrosive environments, also affect the fatigue behavior of materials. Therefore, the main purpose of this paper is to study the influence of corrosive environment on the high-cycle fatigue (HCF) behavior of concrete reinforced by epoxy resin. For this purpose, five corrosive environments with different intensities, including fresh air, water: W, sea water: SW, acidic: AC, and alkaline: AL, were considered and the laboratory samples of conventional concrete (CC) and polymer concrete (PC) were immersed in them for one month. Next, axial fatigue tests were performed under compressive-compressive loading with a frequency of 3 Hz on cylindrical specimens. Moreover, to achieve reliable results, for each stress amplitude, the fatigue test was repeated three times, and the average number of cycles to failure was reported as the fatigue lifetime. Finally, the stress-life cycle (S-N) curves of different states were compared. The results showed that polymer concrete can resist well in corrosive environments and under cyclic loads compared to the conventional concrete, and in other words, the epoxy resin has performed its task well as a reinforcer. The results of fatigue tests show that the load bearing range of 10 tons by CC has reached about 18 tons for PC, which indicates an 80% increase in fatigue strength. Meanwhile, the static strength of samples in the vicinity of fresh air has only improved by 12%.

A Novel Approach for Analyzing the Effects of Almen Intensity on the Residual Stress and Hardness of Shot-Peened (TiB + TiC)/Ti-6Al-4V Composite: Deep Learning

In the present study, the experimental data of a shot-peened (TiB + TiC)/Ti-6Al-4V composite with two volume fractions of 5 and 8% for TiB + TiC reinforcements were used to develop a neural network based on the deep learning technique. In this regard, the distributions of hardness and residual stresses through the depth of the materials as the properties affected by shot peening (SP) treatment were modeled via the deep neural network. The values of the TiB + TiC content, Almen intensity, and depth from the surface were considered as the inputs, and the corresponding measured values of the residual stresses and hardness were regarded as the outputs. In addition, the surface coverage parameter was assumed to be constant in all samples, and only changes in the Almen intensity were considered as the SP process parameter. Using the presented deep neural network (DNN) model, the distributions of hardness and residual stress from the top surface to the core material were continuously evaluated for different combinations of input parameters, including the Almen intensity of the SP process and the volume fractions of the composite reinforcements.

Prediction of the Mechanical Properties of Basalt Fiber Reinforced High-Performance Concrete Using Machine Learning Techniques

In this research, we present an efficient implementation of machine learning (ML) models that forecast the mechanical properties of basalt fiber-reinforced high-performance concrete (BFHPC). The objective of the present study was to predict compressive, flexural, and tensile strengths of BFHPC through ML techniques and propose some correlations between these properties. Moreover, the modulus of elasticity (ME) values and compressive stress-strain curves were simulated using ML techniques. In this regard, three predictive algorithms, including linear regression (LR), support vector regression (SVR), and polynomial regression (PR), were considered. LR, SVR, and PR were utilized to forecast the compressive, flexural, and tensile strengths of BFHPC, and the PR technique was employed to simulate the compressive stress-strain curves. The performance of the models was also determined by the coefficient of determination (R²), mean absolute errors (MAE), and root mean square errors (RMSE). According to the obtained values of R², MAE, and RMSE, the performance of PR was better than other types of algorithms in estimating the compressive, tensile, and flexural strengths. For example, R² values were 0.99, 0.94, and 0.98 in predicting the compressive, flexural, and tensile strengths using PR, respectively. This shows the higher accuracy and reliability of the PR technique compared with other predictive algorithms. Finally, we concluded that ML techniques can be appropriately applied to assess the mechanical characteristics of BFHPC.

The Prediction of Compressive Strength and Compressive Stress-Strain of Basalt Fiber Reinforced High-Performance Concrete Using Classical Programming and Logistic Map Algorithm

In this research, the authors have developed an algorithm for predicting the compressive strength and compressive stress-strain curve of Basalt Fiber High-Performance Concrete (BFHPC), which is enhanced by a classical programming algorithm and Logistic Map. For this purpose, different percentages of basalt fiber from 0.6 to 1.8 are mixed with High-Performance Concrete with high-volume contact of cement, fine and coarse aggregate. Compressive strengths and compressive stress-strain curves are applied after 7-, 14-, and 28-day curing periods. To find the compressive strength and predict the compressive stress-strain curve, the Logistic Map algorithm was prepared through classical programming. The results of this study prove that the logistic map is able to predict the compressive strength and compressive stress-strain of BFHPC with high accuracy. In addition, various types of methods, such as Coefficient of Determination (R2), are applied to ensure the accuracy of the algorithm. For this purpose, the value of R2 was equal to 0.96, which showed that the algorithm is reliable for predicting compressive strength. Finally, it was concluded that The Logistic Map algorithm developed through classical programming could be used as an easy and reliable method to predict the compressive strength and compressive stress-strain of BFHPC.

Influence of Meso-Scale Pore Structure on Mechanical Behavior of Concrete under Uniaxial Compression Based on Parametric Modeling

Existing concrete random aggregate modeling methods (CRAMMs) have deficiencies in in the parameterization of the mesoscale pore structure. A novel CRAMM is proposed, whose pore structure is determined by the pore gradation, total porosity, sub-porosity, and pore size of each pore gradation segment. To study the influence of pore structure on the mechanical properties of concrete, 25 mesoscopic concrete specimens with the same aggregate structure but different meso-scale pore structures are constructed and subjected to uniaxial compression tests. For the first time, the influence of sub-porosity of each pore gradation segment, average pore radius (APR), pore specific surface area (PSSA), and total porosity on concrete failure process, compressive strength, peak strain, and elastic modulus were quantitatively and qualitatively analyzed. Results indicate that the pore structure makes the germination and propagation of the damage in cement mortar show obvious locality and affects the formation and expansion of macroscopic cracks. However, it does not accelerate the propagation of the damage in cement mortar from the periphery to the center of the specimen, nor does it change the phenomenon that the ITZ is more damaged than other meso-components of concrete before peak stress. Macroscopic cracks occur in the descending section of the stress−strain curve, and the sudden drops in the descending section of the stress−strain curve are often accompanied by the generation and expansion of macroscopic cracks. The quadratic polynomial, exponential, and power functions can well fit the relationship between total porosity and compressive strength and the relationship between PSSA and compressive strength. The linear, exponential, and power functions can well reflect the relationship between total porosity and compressive modulus and the relationship between compressive modulus and PSSA. For concrete specimens with the same total porosity, the elastic modulus and strength show randomness with the increase in the sub-porosity of macropores and are basically not affected by the APR. Based on the grey relational analysis, the effects of pore structure parameters on the elastic modulus and compressive strength are in the same order: total porosity > T [k1,k2] > T [k2,k3] > T [k3,k4] > T [k4,k5] > AVR > PSSA. The order of influence of the pore structure parameters on the peak strain is: T [k2,k3] > T [k1,k2] > T [k3,k4] > T [k4,k5] > APR > PSSA > total porosity.