Next Article in Journal
Pseudo-Dynamic Test on Composite Frame with Steel-Reinforced Recycled Concrete Columns and Steel Beams
Next Article in Special Issue
Study on the Performance and Mechanisms of High-Performance Foamed Concrete
Previous Article in Journal
Magneto-Thermo-Elastic Theoretical Solution for Functionally Graded Thick-Walled Tube under Magnetic, Thermal and Mechanical Loads Based on Voigt Method
Previous Article in Special Issue
A Comparison of Machine Learning Tools That Model the Splitting Tensile Strength of Self-Compacting Recycled Aggregate Concrete
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Fatigue Performance Prediction of RC Beams Based on Optimized Machine Learning Technology

1
School of Civil Engineering, Central South University, Changsha 410082, China
2
National Engineering Research Center of High-Speed Railway Construction Technology, Changsha 410075, China
*
Author to whom correspondence should be addressed.
Materials 2022, 15(18), 6349; https://doi.org/10.3390/ma15186349
Submission received: 9 August 2022 / Revised: 5 September 2022 / Accepted: 7 September 2022 / Published: 13 September 2022
(This article belongs to the Special Issue High Performance Concrete and Concrete Structure)

Abstract

:
The development of fatigue damage in reinforced concrete (RC) beams is affected by various factors such as repetitive loads and material properties, and there exists a complex nonlinear mapping relationship between their fatigue performance and each factor. To this end, a fatigue performance prediction model for RC beams was proposed based on the deep belief network (DBN) optimized by particle swarm optimization (PSO). The original database of fatigue loading tests was established by conducting fatigue loading tests on RC beams. The mid-span deflection, reinforcement strain, and concrete strain during fatigue loading of RC beams were predicted and evaluated. The fatigue performance prediction results of the RC beam based on the PSO-DBN model were compared with those of the single DBN model and the BP model. The models were evaluated using the R2 coefficient, mean absolute percentage error, mean absolute error, and root mean square error. The results showed that the fatigue performance prediction model of RC beams based on PSO-DBN is more accurate and efficient.

1. Introduction

As important components of the global land transportation network, bridges, the most common of which are reinforced concrete (RC), play a pivotal role in improving people’s livelihoods and promoting regional economic development [1]. In the context of increasing traffic volume, RC beams are subjected to high-frequency vehicle loads for a long time. The service performance and durability performance of many bridges have deteriorated at an accelerated rate, and the fatigue life has been further reduced, which seriously affects the operational safety of RC bridge structures [2,3]. From a macroscopic perspective, the fatigue damage of RC beams is caused by the gradual accumulation of structural material damage and the gradual deterioration of properties, and the fatigue performance of structural materials determines the fatigue performance of the structure itself [4]. The adoption of scientific and reliable methods to effectively evaluate the fatigue performance of existing RC beams has become an urgent engineering problem to be solved.
At present, the fatigue performance of RC beams has been studied mainly through indoor fatigue loading tests [5,6,7]. Wang and Li [8] investigated the effect of material randomness on structural fatigue performance by conducting fatigue loading tests on RC beams. Yang et al. [9] studied the fatigue performance of RC beams after water freeze-thaw and salt freeze-thaw cycles by conducting four-point bending fatigue loading tests. There are problems such as high intensity, long cycle time, and difficulty in obtaining fast and accurate test results in a short period of conducting a large number of tests. With the continuous development of numerical simulation technology, more and more researchers have started to use numerical simulation methods to study the fatigue performance of RC beams [10,11,12]. Jin et al. [13] simulated the bond-slip behavior of longitudinal reinforcement by establishing a three-dimensional mesoscale numerical model and investigated the mechanical properties of carbon fiber reinforced polymer (CFRP)-reinforced RC beams with impact damage. He et al. [14] studied the fatigue performance of ordinary RC beams and reinforced beams by simulating impact tests, and the dynamic properties of RC beams before and after CFRP strengthening under different impact conditions were investigated. However, the fatigue life of RC beams can be predicted well by using a single fatigue analysis software, but the mechanical properties of RC beams after a certain number of cyclic loading cannot be accurately evaluated, and it is difficult to realize the whole process of fatigue analysis. Dobromil P. et al. [15] used the finite element method to simulate the fatigue damage process of concrete structures, but usually, the accurate mechanical performance analysis needs to be realized by writing special programs or secondary development of software, and the versatility is not strong. The number of fatigue loading is usually in the millions, and it is a huge workload to measure the mechanical properties of RC beams under different loading times in a turn. Therefore, it is important to combine the appropriate amount of indoor fatigue loading tests with efficient scientific calculation algorithms to effectively judge the mechanical properties of RC beams under different damage conditions after cyclic loading for engineering structural evaluation and fatigue durability research.
In recent years, with the continuous development of artificial intelligence technology, many scholars have applied artificial neural networks in engineering structure prediction or structural material research [16,17,18,19,20,21,22] with rich results. Liu et al. [23] established a BP neural network prediction model for blast safe vibration velocity of newly cast concrete structures based on BP neural network theory and selected key influencing factors such as Poisson’s ratio, and predicted the blast safe vibration velocity of concrete at different ages under two different conditions. Asteris et al. [24] used the artificial neural network method to predict the ultimate shear strength of RC beams and compared the predicted values with the experimental values as well as the values calculated by existing formulas in the code provisions, which proved the reliability and validity of the predictive performance of artificial neural networks. Cong et al. [25] used neural networks to estimate the effluent water quality at frequent changes in conditions to improve the accuracy of water quality evaluation in the wastewater treatment process. Sahoo and Mahapatra [26] used an artificial neural network (ANN) model to predict the compressive strength of concrete at different water curing days, sulfate exposure time, and fly ash substitution levels. Onyari and Ikotun [27] used an artificial neural network to predict the compressive and flexural strength of modified zeolite additive mortar, and the results showed that the compressive and flexural strength of modified zeolite mortar could be very short using a neural network model and the prediction results are more accurate.
The deep learning method, as the frontier in the field of machine learning, is one of the commonly used methods in the field of artificial intelligence. It improves prediction accuracy by building models with multiple layers and mining the features implicit in the learned data. Zhang et al. [28] proposed a general method for component life prediction under creep, fatigue, and creep-fatigue conditions based on deep learning. It has better prediction accuracy and generalization ability than traditional machine learning models. Yang et al. [29] proposed a multi-axis fatigue life prediction model based on deep learning and analyzed six sets of existing fatigue data from different materials, respectively. Hinton et al. [30] proposed a deep learning method called deep belief network (DBN), which has attracted a lot of attention from the academic community. DBN is a stack of multiple restricted Boltzmann machines (RBMs), which can map complex nonlinear relationships and have a better-unsupervised feature learning capability, and can maintain strong stability when dealing with complex data prediction [31]. Currently, DBN has been successfully applied to problems such as identification or evaluation in engineering. Wang et al. [32] based a regional landslide sensitivity zoning model on a deep belief network model with example analysis and finally compared the evaluation results with logistic regression (LR) and artificial neural network (BPNN) model evaluation results by RO curve features. Chen et al. [33] evaluated the landslide susceptibility of a region using the DBN model, generated a landslide susceptibility zoning map of the region, and compared it with shallow neural networks and traditional logistic regression methods. Xu et al. [34] identified RC beam damage based on acoustic emission and DBN.
Although the deep belief network is widely used, it has problems in that the number of hidden layer nodes is not easily determined, and the parameters, such as learning rate and the number of iterations, are mainly determined by manual experience during the pre-training process, and improperly set parameter values can have an impact on the model prediction performance. Therefore, it is necessary to optimize the DBN model further to further improve the accuracy of prediction. In this paper, a particle swarm optimization algorithm is used to search for the optimal number of hidden layer nodes. At present, regarding the intelligent search algorithms, the common ones are the simulated annealing algorithm [35], genetic algorithm [36,37], and particle swarm optimization algorithm. Among them, the effect of global optimization and computational efficiency of simulated annealing algorithm is greatly affected by parameters, genetic algorithm is difficult to converge effectively in a limited time, and particle swarm optimization algorithm is a population-based evolutionary algorithm with simple and easy-to-understand principle, fast convergence speed and global optimal solution.
In this paper, the fatigue performance of RC specimen beams was analyzed by conducting a four-point bending fatigue loading test under constant amplitude load, and the original database was established. A deep belief network (DBN) model for fatigue performance prediction of RC beams was established for the first time by using deep learning with massively parallel processing and self-learning characteristics. After training the RBM layer by layer and extracting feature information from complex data, the DBN model parameters were adaptively adjusted using a particle swarm optimization algorithm. The results of the fatigue performance prediction of RC beams based on the PSO-DBN model were compared with those of the single DBN model and the BP model. The comparative analysis of the simulation model predictions verified the feasibility and accuracy of the fatigue performance prediction of RC beams based on the PSO-DBN model and provided new ideas for engineering structure evaluation and fatigue durability research.

2. Experimental Program

2.1. Specimens

Six RC specimen beams were designed for this test, and the design strength grade of concrete for the specimen beams was divided into three types, C35, C40, and C60 (150 mm cubic compressive strengths are 35 MPa, 40 MPa, and 60 MPa at the age of 28 days), two of each type, one of which was used for a static loading test and the remaining one for a fatigue loading test. The cross-sectional dimensions of the specimen beam were 150 mm × 200 mm, with a total length of 600 mm, and the tensile reinforcement was CRB600H (a new type of cold-rolled ribbed steel bar developed in China in recent years) with high strength and high ductility. The specific arrangement of the reinforcement is shown in Figure 1. Figure 2 shows the specimen beams before and after concreting.

2.2. Material Properties

The mechanical properties of materials were tested, as shown in Figure 3. According to the Standard for Test Methods of Physical and Mechanical Properties of Concrete (GB/T 50081-2019) [38], three standard cubic specimens were reserved for each of the three types of concrete with different design strengths during the casting of the specimen beams, and after the specimens were cured for 28 days, the average values of their cubic compressive strengths were tested to be 36.9 MPa, 55 MPa, and 63.7 MPa, respectively. The concrete mix ratios are shown in Table 1. According to the Metal Axial Tensile Test Method (GB/T 228.1-2010) [39], three pieces of each type of reinforcement were reserved for material mechanical property tests. The average yield strength and ultimate tensile strength of CRB600H reinforcements with a nominal diameter of 12 mm were 619 MPa and 671 MPa, respectively, and the average yield strength and ultimate tensile strength of HRB400 reinforcements with a nominal diameter of 8 mm were 467 MPa and 560 MPa, respectively.

2.3. Test Loading Device and Loading System

After finishing the maintenance of the specimen beams, the PMS-500 hydraulic pulsation testing machine produced by China Jinan Times Assay Testing Machine Co., Ltd. was used for test loading, and the loading device is shown in Figure 4a. The test included a static loading test and fatigue loading test, using vertical force control loading and four-point bending cycle constant amplitude loading, respectively. The purpose of the static loading test was to determine the static ultimate bearing capacity of the RC specimen beams and then determine the upper limit value of the fatigue load of the RC specimen beams, and the static test loading is shown in Figure 4b. Before the official start of the static loading test, it was preloaded to 10 kN to promote sufficient contact between the couplings. After checking that the channels of each measurement point were normal, the loading was carried out in increments of 10 kN per level of load, and the load increment per level was appropriately reduced when the specimen beams were close to damage. The indicators, such as mid-span deflection, the strain of tensile reinforcement, and concrete strain in the compression zone at the top of the beam, were measured during the test loading. With the increasing load, the tensile reinforcement at the bottom of the RC-1, RC-3, and RC-5 specimen beams yielded, and the concrete at the top of the beams was crushed. Finally, the static ultimate bearing capacities of the three specimen beams were 205 kN, 232 kN, and 243 kN, respectively.
The fatigue test used four-point bending equal amplitude load, and the fatigue test loading is shown in Figure 4c. According to the results of the static loading test, the upper limit of fatigue load for RC-2, RC-4, and RC-6 specimen beams was taken as 40 kN, 70 kN, and 90 kN, and the lower limit of fatigue load was taken as 10 kN, and the fatigue lives of the three specimen beams were 3.59 million, 2.76 million, and 3.28 million times, respectively. The test loading scheme is shown in Table 2.

2.4. Measurement Point Arrangement and Data Acquisition

Before the RC specimen beams were cast, reinforcement strain gauges were arranged on the tensile reinforcement in the middle of the span of the beam to measure the strain of the tensile reinforcement in the specimen beams. Before the test, two concrete strain gauges were arranged at the top of the span of the specimen beams to measure the strain in the concrete of the specimen beams. Displacement sensors were arranged at the bottom of the span of the specimen beams and the top surface of the beams above the support to record the vertical displacement and the support settlement, respectively. During the loading process of the fatigue test, the DH3820 high-speed static strain testing and analysis system produced by China Jiangsu Donghua Testing Technology Co., Ltd. was used to automatically collect the deflection, strain on the tensile reinforcement, and strain on the concrete top of the beams during the test.

2.5. Test Results and Analysis

Figure 5 shows the fatigue performance of RC-2, RC-4, and RC-6 specimen beams under different load cycles. The relationship curves between the maximum crack width and fatigue life ratio n/N of the specimen beams are shown in Figure 5a. From Figure 5a, it can be seen that the crack development process of the specimen beams shows a "three-stage" pattern. The first stage is the crack derivation stage, in which the crack width is about 0–0.175 mm, the second stage is the stable crack development stage, in which the crack width is about 0.175–0.25 mm, and the third stage is the fatigue damage stage, in which the crack width expands sharply with the increase of the number of cycles, and finally leads to the fatigue damage of the specimen beams.
The curves of the bottom tensile reinforcement strain and top compressive zone concrete strain versus fatigue life ratio n/N for the specimen beams loaded statically to the corresponding upper fatigue limit are shown in Figure 5b,c. It can be seen from the figures that the strains in the bottom tensile reinforcement and the concrete strains in the top compressive zone increase rapidly when the specimen beams are loaded to the corresponding fatigue upper limit within the first 10% of the cycles of fatigue life. Within 10–70% of the fatigue life, the strains develop steadily. After that, additional cyclic loading leads to rapid strain development.
The mid-span deflection versus fatigue life ratio n/N of the specimen beams when statically loaded to the corresponding upper fatigue limit is shown in Figure 5d. From Figure 5d, it can be seen that the change in the mid-span deflection of the specimen beams also shows a three-stage pattern. In the first stage, the mid-span deflection increases significantly. In the second stage, the mid-span deflection of the specimen beams changes relatively smoothly. After that, the mid-span deflection increases rapidly as the number of cycles continues to increase.

3. Method

3.1. Principle of the DBN

The deep belief network is a neural network model consisting of multiple stacked restricted Boltzmann machines, as shown in Figure 6. The DBN training model mainly consists of two processes: pre-training and reverse fine-tuning. The pre-training process is a top-down independent and unsupervised learning process, where the output vector of the previous layer of the RBM network is used as the input vector of the next layer of the RBM network, and the deep feature information of the input vector data is extracted layer by layer to realize the training of the model layer by layer, to obtain the initialized network parameters. Then, the BP neural network is established in the last layer of DBN, and the output vector of the RBM network is used as the input vector of BP. The BP algorithm adjusts network parameters such as weights and biases, and to finish training the entire DBN, the error between the actual output and the desired output is propagated backward layer by layer.
A single RBM is made up of a visible layer and a hidden layer. Figure 6 depicts the layout of a network made up of a three-layer RBM, where h stands for the hidden layer, v for the visible layer, and W for the link weight between the two. Neurons in the same layer of the RBM are independent of one another, whereas connection weights connect neurons in adjacent layers. Binary values 0 and 1, respectively, reflect the inactive and active states of neurons.
The RBM is a thermodynamics-based energy model whose energy function can be expressed as:
E ( v , h | θ ) = i = 1 n j = 1 m v i w i j h j i = 1 n a i v i j = 1 m b j h j
where vi is the state of neuron i in the visible layer, ai is the offset corresponding to vi, hj is the state of neuron j in the hidden layer, bj is the offset corresponding to hj, wij is the connection weight of neurons i and j, θ =(wij,ai,bj) is the RBM parameter, m is the number of neurons in the hidden layer, and n is the number of neurons in the visible layer.
When the parameters in the RBM model are determined, the following joint probability distribution of (v,h) can be obtained:
p ( v , h | θ ) = 1 Z ( θ ) exp ( E ( v , h | θ ) )
where Z ( θ ) = v h exp ( E ( v , h | θ ) ) is the normalization factor.
If the number of training samples is N, the parameter θ can be obtained by learning the maximum log-likelihood function of the samples:
θ = arg max θ L ( θ ) = arg max θ n = 1 N log p ( v n | θ )
where p ( v | θ ) = 1 Z ( θ ) h exp ( E ( v , h | θ ) ) is the likelihood function of the data v.
The difficult normalizing factor Z(θ) computation is typically approximated using sampling techniques, such as Gibbs [40]. Hinton developed the contrastive divergence (CD) algorithm, which allows the jth neuron in the hidden layer to be determined from the state of the neuron in the visible layer and the ith neuron in the visible layer to be reconstructed from the hidden layer using the activation probabilities listed in Equations (4) and (5):
p ( h j = 1 | v , θ ) = σ b j + i v i w i j
p ( v i = 1 | h , θ ) = σ a i + j w i j h j
where σ = 1 1 + exp ( x ) is the sigmoid activation function.
The maximum value of the log-likelihood function can be solved by the stochastic gradient ascent method. The amount of variation of parameters such as RBM weights and bias can be calculated as follows:
Δ w i j = ε v i h j d a t a v i h j r e c o n s t r u c t e d Δ a i = ε v i d a t a v i r e c o n s t r u c t e d Δ b j = ε h j d a t a h j r e c o n s t r u c t e d
where d a t a is the distribution defined by the original data model, r e c o n s t r u c t e d is the distribution defined by the reconstructed model, and ε is the learning rate.

3.2. Parameter Optimization based on the PSO Algorithm

The particle swarm optimization (PSO) originated from the study of the foraging behavior of bird flocks. The basic idea is to find the optimal solution through collaboration and information sharing among different individuals in a swarm. PSO treats each individual in a swarm as a particle without volume and mass in a multidimensional search space. The PSO algorithm first initializes a set of particles in the multidimensional search space and then iteratively updates its speed and direction according to its optimal value and the global optimal value of the population to output the optimal solution [41].
Set a population X = (X1, X2,···Xn) consisting of n particles, and in the D-dimensional search space, each particle ni in this population has its velocity vector Vi = (Vi1, Vi2,···ViD)T and position vector Xi = (xi1, xi2···xiD)T, and the particle’s superiority concerning the target is evaluated by its corresponding individual fitness value, which leads to the individual optimal solution Pi = (Pi1, Pi2···PiD)T when the particle flies in the D-dimensional search space. In turn, the individual optimal solution Pi = (Pi1, Pi2···PiD)T is obtained. The particle, when flying in the D-dimensional search space, will combine the flight experience of other particles and its previous flight state to adjust the next position and velocity, thus outputting the current global optimal solution Pg = (Pg1, Pg···PgD)T and obtaining the optimal solution by k iterations. The velocity and position of the particle are updated as follows:
V i d k + 1 = ω V i d k + c 1 r 1 P i d k X i d k + c 2 r 2 P g d k X i d k
X i d k + 1 = X i d k + V i d k + 1
where c1 and c2 are acceleration values, which are used to adjust the maximum step of individual and group optimal positions, r1 and r2 are inertia factors, distributed between [0,1], ω is the inertia weight, which can be used to balance the global search ability and local search ability, the velocity and position of particles are generally limited to the interval: V max , V max and X max , X max , to avoid blindly searching for particles.

3.3. Fatigue Performance Prediction of RC Beams based on the PSO-DBN Model

The flow chart of the fatigue performance prediction of RC beams based on the PSO-DBN model is shown in Figure 7. The specific steps of the prediction model are as follows:
In the first step, the data collected during the fatigue loading test are organized and pre-processed. The data are normalized by normalizing the original data to ensure that the data are relatively undistorted. After the normalization is completed, the data set is divided into two mutually exclusive sets, one of which is used as the training data set and the other as the test data set. After that, DBN initialization is completed.
In the second step, the initialization of particle position and velocity is completed, the particle fitness value is calculated, and then the particle position and velocity are updated. When the fitness value satisfies the set condition or the number of iterations is equal to M, the second step of PSO optimization is finished. Otherwise, the process of calculating the particle fitness value and updating the position and velocity of the particles will be repeated until the determination condition is satisfied.
In the third step, the DBN model structure parameters after PSO optimization are used to calculate the test data set by the optimal DBN structure and complete the prediction of the fatigue performance of the RC specimen beams.

3.4. Test Data Pre-Process

The data in this section were selected from the loading time range data of the RC specimen beams fatigue test. For the detailed data, see Table A1 in Appendix A. A total of 300 original data samples were selected as the input data for the PSO-DBN model based on 100-time course data for each of the different damage stages of the fatigue loading tests of RC-2, RC-4, and RC-6 specimen beams. In this study, based on the 300 experimental data samples, 70% of them were randomly selected as training sets, and the remaining 30% of data samples were used as test sets.
Based on the a priori knowledge of fatigue loading of RC beams, four relevant factors were identified as the input vectors of the model, namely: load amplitude (kN), concrete strength (MPa), static loading values at different damage stages (kN), and fatigue life ratio(n/N). Since the crack width time-range data were not available for the time being, the three mechanical property indices of concrete strain (με), tensile reinforcement strain (με), and mid-span deflection (mm) at the top of the RC specimen beams were used as the output vectors of the model in this paper.
Since the magnitudes of different variables were different, the data were normalized and mapped to the interval [0,1] before the formal experiments to prevent their variability from affecting the modeling effect, reduce the computational complexity, and accelerate the convergence speed. In this paper, the normalization method was used to linearly transform the original data, as in Equation (9):
x i = x i x min x max x min
where x i is the normalized value of the sample data, x i is the original value of the input variable, and x max and x min are the maximum and minimum values of the original data, respectively.

3.5. Evaluation Indicators

In this study, the model’s performance was assessed using the root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and coefficient of determination (R2). The performance of the prediction increases with decreasing RMSE, MAE, and MAPE. R2 measures the percentage of the dependent variable’s total variation that the independent variable explains through the regression connection. It is generally accepted that the closer R2 is to 1, the better the regression relationship fits the data. The formula for calculating the above valuation metrics is shown below:
R M S E = 1 N i = 1 N ( y i y ¯ i ) 2
M A E = 1 N i = 1 N y i y ¯ i
M A P E = 1 N i = 1 N y i y ¯ i y i
R 2 = 1 i = 1 N ( y i y ¯ i ) 2 i = 1 N ( y i y ¯ ) 2
where y i is the actual sample value, y ¯ i is the model predicted value, and N is the number of samples.

3.6. Model Parameter Setting

The selection of optimal parameters was achieved by using a longitudinal comparison method. Any three different test sample data sets were selected to compare the effect of the different number of hidden layers on model prediction. Each test sample set contains 90 data samples. The number of hidden layer nodes was set to 105, and the effect of different hidden layer choices on model prediction was analyzed, as shown in Table 3.
As can be seen from Table 1, when the number of hidden layers is less than four, the model error gradually decreases as the number of hidden layers increases, which indicates that the model prediction accuracy keeps improving. When the number of hidden layers is greater than four, the model error keeps increasing as the number of hidden layers increases, which is because too many hidden layers make the training of the model complicated and lead to the overfitting phenomenon. Therefore, the optimal number of hidden layers for this model is set to four.
The proper number of hidden layer nodes was discussed and decided upon when the number of hidden layers was established. The research discusses the implications of the prediction when there are 65–125 nodes, as indicated in Table 4. When there are fewer than 105 nodes, the model error gradually decreases as there are more nodes. When there are more than 105 nodes, the model error gradually rises with the number of nodes. Consequently, 105 nodes are needed for the best effect to be realized.
The parameters of the deep belief network model were automatically adjusted using the PSO algorithm, and the parameters of the model were output. The number of DBN hidden layer nodes was automatically output as (115, 129, 109, 105) after PSO completes the parameter optimization and converges. The number of particle swarm is 20, acceleration factor c1 = c2 = 1.49, learning rate is 0.01. The specific parameters of the model are taken as shown in Table 5.

4. Results and Discussion

4.1. Prediction Capability of the PSO–DBN Model

In this section, the predicted results of the PSO-DBN model are evaluated. The comparison between the test values and the predicted values of the three output vectors (mid-span deflection, tensile reinforcement strain, and top compressive zone concrete strain) in the model is shown in Figure 8, including both the training and test data sets. The comparison shown in Figure 8 shows that the predicted values of mid-span deflection, tensile reinforcement strain, and concrete strain in the top compression zone of the RC specimens output from the model are in good agreement with the test values. The errors between the predicted and tested values for both the training and test datasets are relatively small, with the mean absolute percentage errors (MAPE) of 0.075, 0.111, and 0.124 for the test dataset part and 0.067, 0.075, and 0.071 for the training dataset part for the three output vectors, respectively. Error-values indicate that the PSO-DBN model proposed in this paper predicts well the mechanical properties of RC specimen beams under cyclic loading at different damage stages.
The regression plots of the two parts of the training set and the test set are shown in Figure 9. From Figure 9, it can be seen that the PSO-DBN model has a better prediction ability. For the training set, the coefficients of determination (R2) of the three output vectors are 0.983, 0.991, and 0.993, respectively. For the test set, the coefficients of determination (R2) of the three output vectors are 0.979, 0.986, and 0.989, respectively. The coefficients of determination of the three output vectors remain above 0.97 for both the training and test parts. Therefore, it is feasible to use the PSO-DBN model to predict the mechanical properties of RC beams under cyclic loading at different damage stages with high accuracy and low error. It can be applied to develop a numerical tool to evaluate the deterioration performance of RC structures.
To illustrate the accuracy of the model in this paper in predicting the fatigue performance of RC beams, the model predicts the mid-span deflection, tensile reinforcement strain, and concrete strain in the compression zone of RC-2 specimen beam under different fatigue life ratios and compares them with the test values under the same load level (25 kN). As shown in Figure 10, the predicted values of the model are compared with the test values for a high degree of compliance. The predicted values develop rapidly at the fatigue life ratio of 0–0.15, and then the development rate slows down relatively around the fatigue life ratio of 0.15–0.7, which is consistent with the trend of the test results.

4.2. Models Comparison and Analysis

To highlight the efficiency of the PSO-DBN model, the prediction results of the PSO-DBN model, single DBN model, and BP model are compared.
As shown in Figure 11, the correlation values of the two model algorithms in the training part (Figure 11a,c) and the testing part (Figure 11b,d) are determined considering RMSE, MAE, MAPE, and R2 as evaluation metrics. As can be seen in Figure 11, the PSO-DBN model is more accurate than the single DBN model and BP model, as evidenced by the reduction in the error values of RMSE, MAE, and MAPE, and the improvement in prediction accuracy is more pronounced in the training part than in the testing part. Considering the coefficient of determination (R2) as the fitted regression error criterion, the PSO-DBN model has R2 values closer to 1 compared with the DBN model without optimization and BP model and shows some superiority in both the training and test sets.
For comparison purposes, Table 6 and Table 7 show the exact values of the four error criteria when using the PSO-DBN and DBN models, respectively, where the numbers 1, 2, and 3 correspond to mid-span deflection, reinforcement strain, and concrete strain, respectively. Focusing on the test section, the average increase in RMSE, MAE, MAPE, and R2 reached 53.7%, 59.6%, 63.3%, and 6.0%, respectively. Thus, the use of PSO to adjust the weights and biases of DBN greatly improved the accuracy of the predictions.

5. Conclusion

In this paper, the fatigue performance of RC specimen beams was analyzed by conducting a four-point bending fatigue loading test under constant amplitude load, and the original database was established. A deep belief network (DBN) model for fatigue performance prediction of RC beams was established for the first time by using deep learning with massively parallel processing and self-learning characteristics. After training RBM layer by layer and extracting feature information from complex data, the fatigue performance prediction model of RC beams based on PSO-DBN was established by adaptively adjusting DBN model parameters using a particle swarm optimization (PSO) algorithm. The conclusions of this study are as follows.
  • Under the action of constantamplitude four-point bending cyclic loading, the mid-span deflection, tensile reinforcement strain, and concrete strain in the compression zone at the top of the beam of RC specimen beams with CRB600H for tensile reinforcement showed a three-stage trend at different damage stages with different static loading, i.e., rapid development at the initial stage, stable and slow development at the middle stage, and rapid development at the later stage until the fatigue fracture of the reinforcement.
  • The PSO-DBN model describes the complex nonlinear mapping relationship between the RC specimen beams and their material properties, load magnitude, and other factors and accurately predicts and reflects the real process of fatigue damage evolution of the RC specimen beams. By collecting the static loading time data of RC specimen beams at different damage stages during the fatigue loading test, a database containing 300 samples was established and used to train a DBN model. The parameters of the DBN model were adjusted by using PSO to establish the PSO-DBN model. Four evaluation metrics, namely root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and coefficient of determination (R2), were used to evaluate the errors between the predicted values of the PSO-DBN model and test values. The prediction results of the PSO-DBN model showed high reliability, and in the three output vectors of the test section, the coefficient of determination (R2) reached 0.979, 0.986, and 0.989, respectively.
  • The prediction performance of the model on the development process of mid-span deflection, reinforcement strain, and concrete strain in the compressive zone under cyclic loading of the specimen beams was analyzed by using an RC-2 specimen beam as an example. The results showed that the predicted values of mid-span deflection, strain in the tensile reinforcement, and concrete strain in the compression zone of the specimen beam under static load (25 kN) at different fatigue life ratios do not differ significantly from the tested values, and the model prediction of the development trend is consistent with the test results.
  • The PSO-DBN model was compared with the single DBN model and BP model, and the comparison showed that the prediction performance of the PSO-DBN model is better, and the accuracy of RMSE, MAE, MAPE, and R2 are improved to different degrees. Focusing on the test set of the PSO-DBN model and the single DBN model, the average increases in RMSE, MAE, MAPE, and R2 reached 53.7%, 59.6%, 63.3%, and 6.0%, respectively. This indicated that the PSO-DBN model could predict the fatigue performance of RC specimen beams more efficiently and accurately.

Author Contributions

Conceptualization, L.S. and L.W.; methodology, L.S. and L.W.; software, L.S. and L.W.; validation, L.S., L.W. and H.S.; formal analysis, L.S. and H.S.; investigation, C.C.; resources, Z.Y.; data curation, L.W.; writing—original draft preparation, L.S. and L.W.; writing—review and editing, L.S.; visualization, C.C. and L.W.; supervision, L.S. and Z.Y.; project administration, L.S. and L.W.; funding acquisition, L.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded in part by the National Natural Science Foundation of China, grant numbers 51778631, 52078492, and U1934217, in part by the Scientific Research Project of Shuohuang Railway Development Co., Ltd, grant number SHGF-18-50, and in part by the Major Research Project of China Railway Group Limited, grant number 2020-Major-2.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

Table A1. The experimental data sets.
Table A1. The experimental data sets.
No.Load
Amplitude
(kN)
Concrete Strength
(MPa)
Static
Loading
Values
(kN)
Fatigue Life Ratio(n/N)Concrete Strain
(με)
Tensile
Reinforcement Strain
(με)
Mid-Span Deflection
(mm)
16055350.73−660.207917.46590.392024
28063.7150.3−348.898555.14270.217485
36055100.55−325.581475.92220.189237
43036.9250.35−355.138483.82920.311466
58063.7500−208.525313.3280.096914
68063.7850.38−1080.041461.950.666317
76055550.91−1076.231516.5390.594662
83036.9200.28−266.424408.79690.208433
96055250.0914−262.274323.95810.193996
106055450.64−691.86968.3670.414213
116055200−63.307552.002580.053224
126055250.46−447.674522.910.253336
133036.9350.42−469.747727.03750.451553
146055200.46−425.065495.16950.22686
153036.9400.7−805.2811190.1680.61478
163036.9350.28−406.872636.48120.424396
178063.7650.53−880.1181298.9790.539176
188063.7800.69−1216.691653.2030.715272
198063.7400.61−622.718982.88780.396408
206055350.1825−379.845438.06520.246948
213036.9200.07−159.305300.12940.173507
228063.7350.46−503.693801.03380.334273
236055400.365−497.416596.92820.306398
246055400.46−560.724687.3820.334973
258063.7600−330.364447.41820.125569
268063.7400.23−401.207675.13160.318897
273036.9250.28−313.223450.1940.303701
286055600.91−1130.491648.270.647523
296055600.365−664.729933.7710.445335
308063.7750.914−1416.271994.3670.880539
316055450.46−587.855755.79120.365869
328063.7800−564.464747.52510.221702
333036.9200.77−520.255724.45020.276353
343036.9100−14.442531.047870.037254
356055500.1825−438.63634.19090.326397
368063.7550.23−587.363863.50340.396788
373036.9300.139−243.594439.84480.327171
383036.900.42−146.709222.50970.116433
398063.7100.61−421.279768.71260.235399
403036.900.49−176.977250.97020.122254
416055400.1825−379.845447.71790.264636
423036.9150.56−331.381509.70250.20048
438063.7800.46−1060.821450.6980.65326
446055450.55−669.251855.25990.38785
453036.9200.56−406.139571.79820.243361
468063.700.83−577.03913.66630.263571
473036.9100.35−172.779320.82790.153683
486055550.73−836.5631236.1050.508962
498063.7450.15−406.738684.77630.334508
503036.9350.35−444.129690.8150.436039
518063.7600.46−787.0621134.6570.492507
528063.7900.23−1033.621442.9190.669016
536055100.73−420.543575.39090.2222
543036.9200.7−482.986675.29110.268582
556055300.64−587.855717.87630.345753
568063.7650.914−1205.561799.6560.771727
578063.7700.46−923.9481286.3040.552216
588063.7650.61−933.4461342.0310.575364
596055450.1825−416.021511.54720.293324
603036.9350.07−250.819489.00390.381702
616055450.73−737.081049.820.444983
628063.7900.83−1586.062224.260.932645
633036.9250.07−192.125349.28850.268775
648063.7400−123.614168.06450.075977
653036.9300.21−301.82499.35320.342692
663036.9300.56−530.029724.45020.38732
676055100.365−271.318380.98990.154072
683036.900.21−83.833173.35060.098968
698063.7800.914−1484.712137.9450.947851
703036.9400.28−441.995827.94310.50999
718063.7550.3−629.749935.25460.417464
726055650.2737−637.5971033.8630.484995
733036.9300.07−222.633403.62230.307766
746055250.55−497.416604.27640.268725
753036.9400.07−306.925670.11640.461478
763036.9400.35−502.542866.75290.523577
773036.9150.7−403.571600.25870.219882
783036.9300.35−392.646592.49680.352401
796055600.1825−560.724793.66260.412368
806055600.2737−614.987897.60980.425548
818063.7750.61−1053.911501.6460.663472
828063.7200.23−328.464526.51230.230504
838063.7650−382.39530.42210.146402
843036.9400.013−269.66597.67140.447885
856055250.1825−325.581391.78760.209384
863036.9250.49−390.063615.78270.33668
878063.7400.76−751.2551150.3140.43518
883036.9350.49−532.615783.95860.470952
898063.7900.015−865.4261230.8360.599245
903036.9150.07−135.765256.14490.130619
913036.9150.77−447.818626.1320.231531
926055350.55−574.289763.71910.339269
936055650.64−917.9591359.4150.583906
943036.9400.139−330.215701.16430.484757
958063.7250.38−435.185658.94220.274556
966055700.55−972.2221391.7210.616983
976055600−307.494504.24240.24312
986055700.64−1003.881518.4090.645558
996055700.365−822.9971224.3650.57302
1008063.7500.08−438.25705.57550.34238
1013036.9300.28−355.392532.98840.352404
1026055150.64−434.109585.01460.24433
1036055150.91−624.031838.21560.323451
1046055700.1825−678.2951070.560.548837
1053036.9200.63−448.065602.84610.255001
1068063.7450.61−663.8091030.7940.414616
1078063.7650.15−673.632959.34370.453897
1086055550−239.664404.1940.196857
1093036.9250.56−448.263662.35450.336683
1106055550.2737−560.724829.18610.385882
1113036.9150.49−287.127473.47990.194658
1123036.9350.56−588.502835.7050.484542
1133036.9300.42−441.557628.71930.364041
1148063.7600.15−598.368893.87880.412404
1153036.9100.63−282.219421.73350.176972
1168063.7350.3−423.016700.57840.316189
1178063.7550−261.917377.17770.109933
1183036.900.77−298.063437.25740.145537
1198063.7300.23−360.04590.46560.279872
1203036.900.139−74.5224150.06470.091206
1218063.7200.61−504.851847.01760.277024
1226055150.365−339.147417.70190.184982
1236055200.64−483.85648.90180.26422
1243036.9300.013−171.396341.52650.292246
1253036.9350.63−625.755879.68950.49425
1268063.700.914−620.7841100.2290.348846
1278063.7200.08−260.09461.13660.19692
1288063.7900.76−1517.72087.1330.865453
1298063.7900−673.995899.16770.312398
1308063.7450.015−308.294613.01970.29058
1313036.900.56-190.948256.14490.126135
1323036.9200.21−217.519362.22510.200671
1338063.7850.46−1125.171573.5650.692161
1346055150.55−388.889530.7510.228942
1353036.9250.7−550.725739.97410.359972
1366055250.82−637.597843.94050.354446
1373036.9150.21−184.665326.00260.157789
1383036.9250.21−268.976401.03490.293999
1398063.7200.15−297.01507.37620.217584
1406055350.64−614.987827.02660.367848
1416055650.91−1284.241721.10.720159
1423036.9400.49−646.9261029.7540.564327
1436055600.73−899.8711313.5730.555243
1448063.7200.53−463.83743.37570.256352
1453036.9200.42−329.307489.00390.2259
1463036.9100.7−333.441465.7180.188612
1478063.7300.83−710.0911102.3110.41166
1488063.7300.61−556.948915.74630.331556
1498063.7500.3−579.093829.94710.388912
1508063.7600.83−1044.131514.1410.621715
1513036.900.63−221.22287.19280.130013
1528063.7900.08−913.2841283.4510.625089
1533036.9400−111.316393.2730.352791
1543036.9350.139−299.727553.68690.399166
1556055600.64−850.1291223.1620.517877
1568063.7100.23−264.031462.56890.194059
1576055650.82−1125.971585.470.652035
1588063.7600.914−1123.441719.8410.719911
1598063.700.3−243.391465.60750.155043
1606055550.1825−501.938725.20990.368299
1618063.7150.08−260.019440.33740.189057
1626055450.2737−461.24606.49390.308701
1638063.7150.15−291.468478.60430.207145
1646055600.55−818.4751105.620.506877
1653036.9300.63−585.923768.43470.395082
1668063.7400.15−351.977640.05010.305977
1676055350.82−773.2561053.110.427178
1686055400.73−696.382963.25080.4185
1698063.7850.08−858.5191163.7710.578445
1703036.900.35−118.762212.16040.108668
1718063.7800.08−822.8991055.2710.518868
1728063.7750.76−1220.731661.0990.715156
1736055300.2737−406.977473.67570.244651
1746055550.46−646.641964.8740.429834
1756055500.55−691.86937.22050.429704
1766055200.0914−253.23282.62270.171915
1778063.7900.15−951.5751371.1540.648344
1788063.7200.76−601.937960.2320.333863
1793036.9100.77−366.052522.63910.194437
1803036.9100.21−137.858256.14490.134281
1813036.9200−35.874582.794310.07841
1828063.7550.53−745.9691077.1710.458812
1833036.9150.139−159.051287.19280.150027
1843036.9100.139−119.225238.03360.122632
1858063.7750.83−1331.491780.6970.748748
1868063.7700.08−705.165964.20660.461785
187605500.0914−144.703226.08370.105498
1888063.7550.015−441.049716.8110.355436
1898063.7350.914−812.7221392.6080.533239
1906055500.46−601.421864.86910.394546
1916055250.365−406.977473.16840.229157
1928063.7350.08−284.91565.04050.246409
1938063.7700.61−1005.991417.0610.621988
1948063.7450.69−718.4971062.6860.424956
1958063.700.53−319.963628.24520.183467
1966055650−379.845644.97410.298164
1976055150.73−479.328621.17580.26411
1986055100.0914−189.922254.3750.127698
1998063.7200.46−431.013698.72710.251184
200605500.82−483.85655.67030.235172
2018063.7550.38−684.439995.85040.43038
2023036.9150.42−249.858429.49550.184953
2036055350.365−474.806573.72420.290911
2046055650.1825−619.509929.87230.467416
2058063.7850.914−1558.622267.2041.004806
2066055400−131.137117.55660.095419
2078063.7200.69−541.768864.55920.295112
2083036.9150.63−340.703551.09960.208239
2093036.9350−71.5098261.31950.292439
210605500.2737−189.922284.88370.120883
2113036.9250.013−157.2292.36740.245483
2128063.7800.15−847.5131123.8310.539548
2138063.7600.3−674.9371011.8730.456336
2146055500.0914−361.757584.44930.311015
2158063.7400.914−853.8091464.4460.551451
2163036.9250−45.4206108.66750.152333
2178063.7500.76−860.7681257.2740.497423
2188063.7250.015−221.873453.24680.199628
2196055150.1825−289.406327.27710.165199
2203036.9250.42−369.113556.27430.321162
2218063.7600.61−869.1061251.0550.539027
2226055150.82−560.724734.22490.297077
2233036.9300−47.9772170.76330.214616
2248063.7350.69−633.598963.67510.365276
2256055100.1825−230.62313.1750.145281
2266055150.2737−302.972376.98970.178389
2276055300.0914−298.45360.64110.216083
2283036.900−11.65047.7619660.007774
2296055250.73−569.767744.47180.312684
2303036.9250.63−483.184716.68820.342501
2313036.9150.28−210.283336.35190.167488
2323036.9150−23.992354.333760.037457
2333036.9200.013−129.026253.55760.161864
2348063.700.69−410.211735.08060.206719
2353036.9350.21−346.317592.49680.414693
2363036.9250.139−229.379377.7490.278475
2376055700.2737−741.6021165.5510.564225
2386055200.1825−316.537327.87140.187304
2396055200.55−438.63576.53590.244439
2408063.7850.015−814.7641107.9670.550017
2416055350.2737−420.543510.40220.27772
2428063.7250.69−591.067909.28620.315891
2436055650.73−976.7441449.8980.612478
2448063.7850.23−978.8631315.2420.612036
2453036.900.013−48.9041119.01680.069858
2463036.900.07−62.8757139.71540.081504
2473036.9100.49−230.99372.57440.169207
2483036.9200.49−359.571522.63910.23754
2493036.9300.49−485.801677.87840.383442
2506055600.0914−497.416743.9210.392581
2513036.900.7−263.138346.70120.135834
2523036.9350.013−206.59460.54330.352595
2533036.9100.28−149.493300.12940.149805
2546055500.365−538.114810.60550.365968
2553036.9100.013−77.3182173.35060.093525
256605500.1825−176.357253.21550.114289
2578063.7450−153.762215.98240.091605
2586055450.82−813.9531180.9140.488939
2598063.7450.83−860.7081314.6180.49214
2606055700.46−913.4371346.4870.5994
261605500.46−239.664348.17670.149454
2628063.7800.76−1302.841752.0720.75146
2636055300.55−547.158641.01740.310592
2646055400.64−651.163863.79650.396525
2656055100.2737−257.752358.38010.151878
2668063.7200.38−398.196631.75320.243432
2676055300−90.439389.294290.066604
2686055350.0914−307.494361.20640.231552
2698063.7150.76−545.795923.47850.318252
2708063.7250.83−699.0891076.7160.385663
2716055500.64−732.5581032.1530.449484
2723036.9100.42−189.083336.35190.163386
2736055350−90.4393103.41090.071123
2746055100.82−501.938697.48390.270548
2758063.7500.69−800.6031136.0970.471592
276605500.91−547.158737.06560.281322
2773036.900.28−104.79186.28720.102846
2788063.7350.53−535.146845.68070.342024
2796055550.64−795.8661141.1580.486981
2806055150−49.741646.915370.044319
2818063.7900.61−1350.861879.8320.813765
2828063.7150.23−318.818496.1460.212317
2838063.7150.015−194.387408.4450.176141
2843036.9400.21−395.415750.32340.500284
2858063.7550.69−863.5651193.5740.50016
2863036.9150.35−242.891380.33640.179131
2873036.9400.56−684.1831055.6270.589556
2883036.9300.7−637.152817.59380.402844
2896055200.365−384.367431.84750.204882
2906055300.46−506.46586.73930.288611
2918063.7700.69−1079.831458.5130.624568
2923036.9150.013−96.1786214.74770.107333
2936055700.73−1053.621577.1650.702711
2943036.9400.63−709.8051164.2950.599259
2953036.9100.56−263.59385.5110.169207
2963036.9100.07−105.254206.98580.11487
2973036.9200.139−198.898318.24060.190971
2983036.9200.35−306.021457.9560.216198
2993036.9350.7−695.62926.26130.507831
3003036.9400.42−556.114934.02330.544925

References

  1. Niu, D.; Miao, Y. Experimental study on fatigue performance of corroded highway bridges based on vehicle loading. China Civ. Eng. J. 2018, 51, 112308. [Google Scholar]
  2. Nor, N.M.; Saliah, S.N.M.; Hashim, K.A. Fatigue damage assessment of reinforced concrete beam using average frequency and rise angle value of acoustic emission signal. Int. J. Struct. Integr. 2020, 11, 633–643. [Google Scholar] [CrossRef]
  3. Liu, F.; Zhou, J.; Yan, L. Study of stiffness and bearing capacity degradation of reinforced concrete beams under constant-amplitude fatigue. PLoS ONE 2018, 13, e0192797. [Google Scholar] [CrossRef] [PubMed]
  4. Liang, J.; Ding, Z.; Li, J. Analytical method for fatigue process of concrete structures. J. Build. Struct. 2017, 38, 149–157. [Google Scholar]
  5. Huang, B.-T.; Li, Q.-H.; Xu, S.-L.; Zhang, L. Static and fatigue performance of reinforced concrete beam strengthened with strain-hardening fiber-reinforced cementitious composite. Eng. Struct. 2019, 199, 109576. [Google Scholar] [CrossRef]
  6. Chen, Z.; Huang, P. Experimental Study on Fatigue Performance of CFRP-RC Beams Under Variable Amplitude Overloads. In Proceedings of the International Conference on Theoretical, Applied and Experimental Mechanics, Cyprus, Greece, 17–20 June 2018; pp. 152–154. [Google Scholar]
  7. Zhao, Y.; Wang, Z.; Xie, Y.; Wang, L. Experimental study on flexural behavior and fatigue life of steel fiber reinforced concrete beams. Concrete 2018, 5, 42–50. [Google Scholar]
  8. Wang, Y.; Li, J. Experimental study on stochastic responses of reinforced concrete beams under fatigue loading. Int. J. Fatigue 2021, 151, 106347. [Google Scholar] [CrossRef]
  9. Yang, J.; Zhang, T.; Sun, Q. Experimental study on flexural fatigue properties of reinforced concrete beams after salt freezing. Adv. Mater. Sci. Eng. 2020, 2020, 1032317. [Google Scholar] [CrossRef]
  10. Wang, X.; Zhou, C.; Ai, J.; Petrů, M.; Liu, Y. Numerical investigation for the fatigue performance of reinforced concrete beams strengthened with external prestressed HFRP sheet. Constr. Build. Mater. 2020, 237, 117601. [Google Scholar] [CrossRef]
  11. Banjara, N.K.; Ramanjaneyulu, K.; Sasmal, S.; Srinivas, V. Flexural fatigue performance of plain and fibre reinforced concrete. Trans. Indian Inst. Met. 2016, 69, 373–377. [Google Scholar] [CrossRef]
  12. Zhao, G.Q.; Li, K.; Zhang, L. Finite Element Analysis of Fatigue Performance for HRBF500 Reinforced Concrete Beams. J. Zhengzhou Univ. Eng. Sci. 2010, 31, 10–13. [Google Scholar]
  13. Jin, L.; Lan, Y.; Zhang, R.; Du, X. Numerical analysis of the mechanical behavior of the impact-damaged RC beams strengthened with CFRP. Compos. Struct. 2021, 274, 114353. [Google Scholar] [CrossRef]
  14. Qingfeng, H.; Jiang, L.; Leiyang, Y.; Jiawei, M. Numerical Simulation on Impact Test of CFRP Strengthened Reinforced Concrete Beams. Front. Mater. 2020, 7, 252. [Google Scholar] [CrossRef]
  15. Dobromil, P.; Jan, C.; Radomir, P. Material model for finite element modelling of fatigue crack growth in concrete. Procedia Eng. 2010, 2, 203–212. [Google Scholar] [CrossRef] [Green Version]
  16. Nguyen, H.Q.; Ly, H.-B.; Tran, V.Q.; Nguyen, T.-A.; Le, T.-T.; Pham, B.T. Optimization of artificial intelligence system by evolutionary algorithm for prediction of axial capacity of rectangular concrete filled steel tubes under compression. Materials 2020, 13, 1205. [Google Scholar] [CrossRef]
  17. Nguyen, Q.H.; Ly, H.-B.; Le, T.-T.; Nguyen, T.-A.; Phan, V.-H.; Tran, V.Q.; Pham, B.T. Parametric investigation of particle swarm optimization to improve the performance of the adaptive neuro-fuzzy inference system in determining the buckling capacity of circular opening steel beams. Materials 2020, 13, 2210. [Google Scholar] [CrossRef]
  18. Dao, D.V.; Ly, H.-B.; Trinh, S.H.; Le, T.-T.; Pham, B.T. Artificial intelligence approaches for prediction of compressive strength of geopolymer concrete. Materials 2019, 12, 983. [Google Scholar] [CrossRef]
  19. Ly, H.-B.; Pham, B.T.; Dao, D.V.; Le, V.M.; Le, L.M.; Le, T.-T. Improvement of ANFIS model for prediction of compressive strength of manufactured sand concrete. Appl. Sci. 2019, 9, 3841. [Google Scholar] [CrossRef]
  20. Dao, D.V.; Ly, H.-B.; Vu, H.-L.T.; Le, T.-T.; Pham, B.T. Investigation and optimization of the C-ANN structure in predicting the compressive strength of foamed concrete. Materials 2020, 13, 1072. [Google Scholar] [CrossRef]
  21. Santarsiero, G.; Mishra, M.; Singh, M.K.; Masi, A. Structural health monitoring of exterior beam–column subassemblies through detailed numerical modelling and using various machine learning techniques. Mach. Learn. Appl. 2021, 6, 100190. [Google Scholar] [CrossRef]
  22. Salimbahrami, S.R.; Shakeri, R. Experimental investigation and comparative machine-learning prediction of compressive strength of recycled aggregate concrete. Soft Comput. 2021, 7, 242. [Google Scholar] [CrossRef]
  23. Liu, D.; Cui, P. Forecasting for safety vibration velocity of freshly-made concrete based on BP neural network. J. Saf. Environ. 2014, 14, 43–46. [Google Scholar]
  24. Asteris, P.G.; Armaghani, D.J.; Hatzigeorgiou, G.D.; Karayannis, C.G.; Pilakoutas, K. Predicting the shear strength of reinforced concrete beams using Artificial Neural Networks. Comput. Concr. Int. J. 2019, 24, 469–488. [Google Scholar]
  25. Cong, Q.; Yu, W. Integrated soft sensor with wavelet neural network and adaptive weighted fusion for water quality estimation in wastewater treatment process. Measurement 2018, 124, 436–446. [Google Scholar] [CrossRef]
  26. Sahoo, S.; Mahapatra, T.R. ANN Modeling to study strength loss of Fly Ash Concrete against Long term Sulphate Attack. Mater. Today Proc. 2018, 5, 24595–24604. [Google Scholar] [CrossRef]
  27. Onyari, E.; Ikotun, B. Prediction of compressive and flexural strengths of a modified zeolite additive mortar using artificial neural network. Constr. Build. Mater. 2018, 187, 1232–1241. [Google Scholar] [CrossRef]
  28. Zhang, X.; Gong, J.G.; Xuan, F.Z. A deep learning based life prediction method for components under creep, fatigue and creep-fatigue conditions. Int. J. Fatigue 2021, 148, 106236. [Google Scholar] [CrossRef]
  29. Yang, J.; Kang, G.; Liu, Y.; Kan, Q. A novel method of multiaxial fatigue life prediction based on deep learning. Int. J. Fatigue 2021, 151, 106356. [Google Scholar] [CrossRef]
  30. Hinton, G.E.; Osindero, S.; Teh, Y.-W. A fast learning algorithm for deep belief nets. Neural Comput. 2006, 18, 1527–1554. [Google Scholar] [CrossRef]
  31. Qiu, F.; Zhang, B.; Guo, J. A deep learning approach for VM workload prediction in the cloud. In Proceedings of the 2016 17th IEEE/ACIS International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD), Shanghai, China, 30 May–1 June 2016; pp. 319–324. [Google Scholar]
  32. Wang, W.; He, Z.; Han, Z.; Qian, Y. Landslides Susceptibility Assessment Based on Deep Belief Network. J. Northeast. Univ. 2020, 41, 7. [Google Scholar]
  33. Chen, T.; Zhong, Z.; Niu, R.; Liu, T.; Chen, S. Mapping Landslide Susceptibility Based on Deep Belief Network. Geomat. Inf. Sci. Wuhan Univ. 2020, 45, 1809–1817. [Google Scholar]
  34. Xu, X.; Zhang, Y.; Li, X.; Li, Z.; Zhang, J. Study on damage identification method of reinforced concrete beam based on acoustic emission and deep belief nets. J. Build. Struct. 2018, 39, 400–407. [Google Scholar]
  35. Assad, A.; Deep, K. A hybrid harmony search and simulated annealing algorithm for continuous optimization. Inf. Sci. 2018, 450, 246–266. [Google Scholar] [CrossRef]
  36. Rani, S.; Suri, B.; Goyal, R. On the effectiveness of using elitist genetic algorithm in mutation testing. Symmetry 2019, 11, 1145. [Google Scholar] [CrossRef]
  37. Gong, D.; Sun, J.; Miao, Z. A set-based genetic algorithm for interval many-objective optimization problems. IEEE Trans. Evol. Comput. 2016, 22, 47–60. [Google Scholar] [CrossRef]
  38. China National Standard. In Standard for Test Methods of Concrete Physical and Mechanical Properties(GB/T 50081−2019); China Architecture & Building Press: Beijing, China, 2019.
  39. China National Standard. In Tensile Testing of Metallic Materials(GB/T 228.1−2010); Standards Press of China: Beijing, China, 2010.
  40. Jiang, H.; Pan, Z.; Liu, N.; You, X.; Deng, T. Gibbs-sampling-based CRE bias optimization algorithm for ultradense networks. IEEE Trans. Veh. Technol. 2016, 66, 1334–1350. [Google Scholar] [CrossRef]
  41. Liu, H.; Song, W.; Li, M.; Kudreyko, A.; Zio, E. Fractional Lévy stable motion: Finite difference iterative forecasting model. Chaos Solitons Fractals 2020, 133, 109632. [Google Scholar] [CrossRef]
Figure 1. Details of test specimens (unit: mm): (a) Geometry; (b) Midspan section.
Figure 1. Details of test specimens (unit: mm): (a) Geometry; (b) Midspan section.
Materials 15 06349 g001
Figure 2. Specimen beams: (a) Before concreting; (b) After concreting.
Figure 2. Specimen beams: (a) Before concreting; (b) After concreting.
Materials 15 06349 g002
Figure 3. Material performance testing: (a,b) physical and mechanical properties test of concrete; and (c) material mechanical properties test of reinforcement.
Figure 3. Material performance testing: (a,b) physical and mechanical properties test of concrete; and (c) material mechanical properties test of reinforcement.
Materials 15 06349 g003
Figure 4. Test loading process: (a) test loading device; (b) specimen beams static loading test; (c) specimen beams fatigue loading test.
Figure 4. Test loading process: (a) test loading device; (b) specimen beams static loading test; (c) specimen beams fatigue loading test.
Materials 15 06349 g004
Figure 5. Fatigue performance of RC specimen beams under different numbers of cyclic loading: (a) maximum crack width; (b) tensile reinforcement strain; (c) concrete strain; (d) mid-span deflection.
Figure 5. Fatigue performance of RC specimen beams under different numbers of cyclic loading: (a) maximum crack width; (b) tensile reinforcement strain; (c) concrete strain; (d) mid-span deflection.
Materials 15 06349 g005
Figure 6. A prediction model based on the deep belief network (DBN).
Figure 6. A prediction model based on the deep belief network (DBN).
Materials 15 06349 g006
Figure 7. Fatigue performance prediction model of RC beams based on the PSO-DBN.
Figure 7. Fatigue performance prediction model of RC beams based on the PSO-DBN.
Materials 15 06349 g007
Figure 8. Comparison between predicted and experimental values of PSO-DBN model: (a) mid-span deflection prediction for training data; (b) mid-span deflection prediction for test data; (c) reinforcement strain prediction for training data; (d) reinforcement strain prediction for test data; (e) concrete strain prediction for training data; and (f) concrete strain prediction for test data.
Figure 8. Comparison between predicted and experimental values of PSO-DBN model: (a) mid-span deflection prediction for training data; (b) mid-span deflection prediction for test data; (c) reinforcement strain prediction for training data; (d) reinforcement strain prediction for test data; (e) concrete strain prediction for training data; and (f) concrete strain prediction for test data.
Materials 15 06349 g008aMaterials 15 06349 g008b
Figure 9. Correlation analysis between test and predicted values of fatigue performance of RC specimen beams: (a) mid-span deflection prediction for training data; (b) mid-span deflection prediction for test data; (c) reinforcement strain prediction for training data; (d) reinforcement strain prediction for test data; (e) concrete strain prediction for training data; and (f) concrete strain prediction for test data.
Figure 9. Correlation analysis between test and predicted values of fatigue performance of RC specimen beams: (a) mid-span deflection prediction for training data; (b) mid-span deflection prediction for test data; (c) reinforcement strain prediction for training data; (d) reinforcement strain prediction for test data; (e) concrete strain prediction for training data; and (f) concrete strain prediction for test data.
Materials 15 06349 g009aMaterials 15 06349 g009b
Figure 10. Analysis of fatigue loading process of RC-2 specimen beam: (a) mid-span deflection; (b) reinforcement strain; (c) concrete strain.
Figure 10. Analysis of fatigue loading process of RC-2 specimen beam: (a) mid-span deflection; (b) reinforcement strain; (c) concrete strain.
Materials 15 06349 g010
Figure 11. Models Comparison and Analysis:(a) RMSE and MAE for training data; (b) RMSE and MAE for testing data; (c) MAPE and R2 for training data; and (d) MAPE and R2 for testing data.
Figure 11. Models Comparison and Analysis:(a) RMSE and MAE for training data; (b) RMSE and MAE for testing data; (c) MAPE and R2 for training data; and (d) MAPE and R2 for testing data.
Materials 15 06349 g011aMaterials 15 06349 g011b
Table 1. Mixture ratio and type of concrete.
Table 1. Mixture ratio and type of concrete.
Design StrengthCement
(kg/m3)
Fly Ash (kg/m3)Sand
(kg/m3)
Rocks (kg/m3)Water (kg/m3)Additives (kg/m3)Admixture (kg/m3)
C35260/73411011607.8112
C403502083581017877.4100
C603656571311751284.9565
Table 2. Test design loading scheme.
Table 2. Test design loading scheme.
Specimen NumberDesign StrengthLoad Range (kN)Load Level
Pmax/Pu
Mode Failure
PminPmax
RC-1C351.0Static damage
RC-2C3510400.20Fatigue damage
RC-3C401.0Static damage
RC-4C4010700.30Fatigue damage
RC-5C601.0Static damage
RC-6C6010900.27Fatigue damage
Table 3. The effect of the choice of the number of hidden layers on the prediction effect of the model.
Table 3. The effect of the choice of the number of hidden layers on the prediction effect of the model.
The Number of Hidden LayersThe Test Set
1-MAPE
The Test Set
2-MAPE
The Test Set
3-MAPE
20.1280.1230.122
30.1260.1220.118
40.1160.1150.114
50.1180.1160.116
60.1220.1200.121
70.1230.1230.123
80.1290.1280.126
Table 4. The effect of the number of nodes in the hidden layer on the prediction effect of the model.
Table 4. The effect of the number of nodes in the hidden layer on the prediction effect of the model.
The Number of NeuronsThe Test Set
1-MAPE
The Test Set
2-MAPE
The Test Set
3-MAPE
650.1250.1260.124
1000.1230.1190.118
1050.1150.1170.113
1100.1180.1210.116
1150.1230.1220.118
1250.1280.1230.124
Table 5. PSO-DBN model parameters.
Table 5. PSO-DBN model parameters.
DescriptionSymbolValue
The number of neurons in the input layer-4
The number of neurons in the output layer-3
The number of RBMs-4
Iteration number of each RBM-100
The number of neurons in the first hidden layerh1115
The number of neurons in the second hidden layerh2129
The number of neurons in the third hidden layerh3109
The number of neurons in the fourth hidden layerh4105
The learning rate of the DBNη0.01
The momentum of the DBNα0.5
The acceleration factor of PSOc1,c21.49
The iteration number of PSOM100
The inertia weight of PSOw0.9
The population factor of PSOW20
Table 6. Comparison of evaluation metrics between PSO-DBN model and single DBN model (a).
Table 6. Comparison of evaluation metrics between PSO-DBN model and single DBN model (a).
DataModelRMSE-1RMSE-2RMSE-3MAE-1MAE-2MAE-3
TrainingPSO-DBN0.02441.46126.4310.01531.00120.472
DBN0.05994.18673.2800.04570.76656.783
%Gain+59.4+56.0+64.0+65.7+56.2+60.6
TestingPSO-DBN0.02754.41634.7760.01836.95625.922
DBN0.053107.64894.0960.04380.36374.360
%Gain+48.6+49.5+63.0+59.6+54.0+65.1
Table 7. Comparison of evaluation metrics between PSO-DBN model and single DBN model (b).
Table 7. Comparison of evaluation metrics between PSO-DBN model and single DBN model (b).
DataModelMAPE-1MAPE-2MAPE-3R2-1R2-2R2-3
TrainingPSO-DBN0.0670.0750.0710.9830.9910.993
DBN0.2150.1850.2220.8980.9520.944
%Gain+68.9+59.8+68.2+9.5+4.1+5.2
TestingPSO-DBN0.0750.1110.1240.9790.9860.989
DBN0.2210.2500.3900.9210.9460.921
%Gain+66.1+55.6+68.2+6.3+4.2+7.4
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Song, L.; Wang, L.; Sun, H.; Cui, C.; Yu, Z. Fatigue Performance Prediction of RC Beams Based on Optimized Machine Learning Technology. Materials 2022, 15, 6349. https://doi.org/10.3390/ma15186349

AMA Style

Song L, Wang L, Sun H, Cui C, Yu Z. Fatigue Performance Prediction of RC Beams Based on Optimized Machine Learning Technology. Materials. 2022; 15(18):6349. https://doi.org/10.3390/ma15186349

Chicago/Turabian Style

Song, Li, Lian Wang, Hongshuo Sun, Chenxing Cui, and Zhiwu Yu. 2022. "Fatigue Performance Prediction of RC Beams Based on Optimized Machine Learning Technology" Materials 15, no. 18: 6349. https://doi.org/10.3390/ma15186349

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop