• Computers and Concrete
• /
• v.32 no.2
• /
• pp.149-163
• /
• 2023

One of the main design parameters traditionally utilized in projects of geotechnical engineering is the uniaxial compressive strength. The present paper employed three artificial intelligence methods, i.e., the stochastic fractal search (SFS), the multi-verse optimization (MVO), and the vortex search algorithm (VSA), in order to determine the compressive strength of concrete (CSC). For the same reason, 1030 concrete specimens were subjected to compressive strength tests. According to the obtained laboratory results, the fly ash, cement, water, slag, coarse aggregates, fine aggregates, and SP were subjected to tests as the input parameters of the model in order to decide the optimum input configuration for the estimation of the compressive strength. The performance was evaluated by employing three criteria, i.e., the root mean square error (RMSE), mean absolute error (MAE), and the determination coefficient (R²). The evaluation of the error criteria and the determination coefficient obtained from the above three techniques indicates that the SFS-MLP technique outperformed the MVO-MLP and VSA-MLP methods. The developed artificial neural network models exhibit higher amounts of errors and lower correlation coefficients in comparison with other models. Nonetheless, the use of the stochastic fractal search algorithm has resulted in considerable enhancement in precision and accuracy of the evaluations conducted through the artificial neural network and has enhanced its performance. According to the results, the utilized SFS-MLP technique showed a better performance in the estimation of the compressive strength of concrete (R²=0.99932 and 0.99942, and RMSE=0.32611 and 0.24922). The novelty of our study is the use of a large dataset composed of 1030 entries and optimization of the learning scheme of the neural prediction model via a data distribution of a 20:80 testing-to-training ratio.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.16 no.5
• /
• pp.148-160
• /
• 2012

This study proposed more reasonable prediction models on compressive strength and carbonation of concrete structure and developed a more effective tunnel safety diagnosis and maintenance method through field application of the proposed prediction models. For this study, the Seoul Metro's Line 1 through Line 4 were selected as target structures because they were built more than 30 years ago and have accumulated numerous diagnosis and maintenance data for about 15 years. As a result of the analysis of compressive strength and carbonation, we were able to draw prediction models with accuracy of more than 80% and confirmed the prediction model's reliability by comparing it with the existing models. We've also confirmed field suitability of the prediction models by applying field, the average error of an estimate on compressive strength and carbonation depth was about 20%, which showed an accuracy of more than 80%. We developed a more effective maintenance method using durability prediction Map before field inspection. With the durability prediction Map, diagnostic engineers and structure managers can easily detect the vulnerable points, which might have failed to reach the standard of designed strength or have a high probability of corrosion due to carbonation, therefore, it is expected to make it possible for them to diagnose and maintain tunnels more effectively and efficiently.

• Structural Engineering and Mechanics
• /
• v.53 no.1
• /
• pp.41-55
• /
• 2015

In this paper, a model for the evaluation of shear strength of fibre reinforced polymer (FRP)-reinforced concrete beams is given. The survey of literature indicates that the FRP reinforced beams tested with shear span to depth ratio less than or equal to 1.0 is limited. In this study, eight concrete beams reinforced with GFRP rebars without stirrups are cast and tested over shear span to depth ratio of 0.5 and 1.75. The concrete compressive strength is varied from 40.6 to 65.3 MPa. The longitudinal reinforcement ratio is varied from 1.16 to 1.75. The experimental shear strength and load-deflection response of the beams are determined and reported in this paper. A model is proposed for the prediction of shear strength of beams reinforced with FRP bars. The proposed model accounts for compressive strength of concrete, modulus of FRP rebar, longitudinal reinforcement ratio, shear span to depth ratio and size effect of beams. The shear strength of FRP reinforced concrete beams predicted using the proposed model is found to be in better agreement with the corresponding test data when compared with the shear strength predicted using the eleven models published in the literature. Design example of FRP reinforced concrete beam is also given in the appendix.

• Geomechanics and Engineering
• /
• v.23 no.1
• /
• pp.61-69
• /
• 2020

The uniaxial compressive strength (UCS) of rock is a basic parameter in underground engineering design. The disadvantages of this commonly employed laboratory testing method are untimely testing, difficulty in performing core testing of broken rock mass and long and complicated onsite testing processes. Therefore, the development of a fast and simple in situ rock UCS testing method for field use is urgent. In this study, a multi-function digital rock drilling and testing system and a digital core bit dedicated to the system are independently developed and employed in digital drilling tests on rock specimens with different strengths. The energy analysis is performed during rock cutting to estimate the energy consumed by the drill bit to remove a unit volume of rock. Two quantitative relationship models of energy analysis-based core drilling parameters (ECD) and rock UCS (ECD-UCS models) are established in this manuscript by the methods of regression analysis and support vector machine (SVM). The predictive abilities of the two models are comparatively analysed. The results show that the mean value of relative difference between the predicted rock UCS values and the UCS values measured by the laboratory uniaxial compression test in the prediction set are 3.76 MPa and 4.30 MPa, respectively, and the standard deviations are 2.08 MPa and 4.14 MPa, respectively. The regression analysis-based ECD-UCS model has a more stable predictive ability. The energy analysis-based rock drilling method for the prediction of UCS is proposed. This method realized the quick and convenient in situ test of rock UCS.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.16 no.6
• /
• pp.533-542
• /
• 2014

In this study TBM disc cutter prediction models including Gehring, CSM and NTNU models were investigated and the characteristics of the models were examined. The influence of penetration, uniaxial compressive strength and abrasiveness index on the models was analyzed. The life of disc cutter linearly increases with penetration per revolution and decreases with increasing uniaxial compressive strength of rocks. As the abrasiveness index, CAI, increases, the life of disc cutter in Gehring and CSM model decreases. On the contrary, the life of disc cutter life in NTNU model decreases with increasing CLI. Also, comparisons of predicted disc life were made between models using actual job site data.

• Journal of the Korea Concrete Institute
• /
• v.27 no.2
• /
• pp.169-176
• /
• 2015

Since the size effect law announced currently has been based on the normal weight concrete, for light weight concrete having different fracture characteristics, its application is questionable. Accordingly, in this study, a model equation to predict the effect of dried unit weight of the concrete on size effect of its compressive strength was developed and a database using existing research results was created. After determining the experimental constants of prediction models of Ba${\check{z}}$ant based on nonlinear fracture mechanics, Kim and Eo, and this study using the database, their results are mutually compared. Finally, it was found that the prediction model of this study considered dried unit weight of concrete predicted well the test results for light weight concrete than that of the models of Ba${\check{z}}$ant and Kim and Eo.

• Advances in concrete construction
• /
• v.13 no.2
• /
• pp.123-132
• /
• 2022

This paper aims to adapt Multilinear regression (MLR) to predict the strength and toughness of SIFCON containing various pozzolanic materials. Slurry Infiltrated Fibrous Concrete (SIFCON) is one of the most common terms used in concrete manufacturing, known for its benefits such as high ductility, toughness and high ultimate strength. Assessment of compressive strength (CS.), flexural strength (F.S.), splitting tensile strength (STS), dynamic elasticity modulus (DME) and impact energy (I.E.) using the experimental approach is too costly. It is time-consuming, and a slight error can lead to a repeat of the test and, to solve this, alternative methods are used to predict the strength and toughness properties of SIFCON. In the present study, the experimentally investigated SIFCON data about various mix proportions are used to predict the strength and toughness properties using regression analysis-multilinear regression (MLR) models. The input parameters used in regression models are cement, fibre, fly ash, Metakaolin, fine aggregate, blast furnace slag, bottom ash, water-cement ratio, and the strength and toughness properties of SIFCON at 28 days is the output parameter. The models are developed and validated using data obtained from the experimental investigation. The investigations were done on 36 SIFCON mixes, and specimens were cast and tested after 28 days of curing. The MLR model yields correlation between predicted and actual values of the compressive strength (C.S.), flexural strength, splitting tensile strength, dynamic modulus of elasticity and impact energy. R-squared values for the relationship between observed and predicted compressive strength are 0.9548, flexural strength 0.9058, split tensile strength 0.9047, dynamic modulus of elasticity 0.8611 for impact energy 0.8366. This examination shows that the MLR model can predict the strength and toughness properties of SIFCON.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2021.11a
• /
• pp.30-31
• /
• 2021

This study aims to build a deep learning model that can predict the value of concrete mixing properties according to a given concrete strength value. A model was created for a total of 1,291 concrete data, including 8 characteristics related to concrete mixing elements and environment, and the compressive strength of concrete. As the deep learning model, DNN-3L-256N, which showed the best performance on the prior study, was used. The average value for each characteristic of the data set was used as the initial input value. In results, in the case of 'curing temperature', which had a narrow range of values in the existing data set, showed the lowest error rate with less than 1% error based on MAE. The highest error rate with an error of 12 to 14% for fly and bfs.

• Computers and Concrete
• /
• v.23 no.6
• /
• pp.377-398
• /
• 2019

Behavior of RC beam-column joint is very complex as the composite material behaves differently in elastic and inelastic range. The approaches generally used for predicting joint shear strength are either based on theoretical, strut-and-tie or empirical methods. These approaches are incapable of predicting the accurate response of the joint for entire range of loading. In the present study a new generalized RC beam-column joint shear strength model based on hybrid approach i.e. combined strut-and-tie and empirical approach has been proposed. The contribution of governing parameters affecting the joint shear strength under compression has been derived from compressive strut approach whereas; the governing parameters active under tension has been extracted from empirical approach. The proposed model is applicable for various conditions such as, joints reinforced either with or without shear reinforcement, joints with wide beam or wide column, joints with transverse beams and slab, joints reinforced with X-bars, different anchorage of beam bar, and column subjected to various axial loading conditions. The joint shear strength prediction of the proposed model has been compared with 435 experimental results and with eleven popular models from literature. In comparison to other eleven models the prediction of the proposed model is found closest to the experimental results. Moreover, from statistical analysis of the results, the proposed model has the least coefficient of variation. The proposed model is simple in application and can be effectively used by designers.

• Journal of the Korea Concrete Institute
• /
• v.16 no.6 s.84
• /
• pp.813-822
• /
• 2004

For the prediction of the shear strength of reinforced concrete members subjected to axial force, this paper presents a truss model, Transformation Angle Truss Model (TATM), that can predict the shear behavior of reinforced concrete members subjected to combined actions of shear, axial force, and bending moment. In TATM, as axial compressive stress increases, crack angle decreases and concrete contribution due to the shear resistance of concrete along the crack direction increases in order to consider the effect of the axial force. To verify if the prediction results of TATM have an accuracy and reliability for the shear strength of reinforced concrete members subjected to axial forces, the shear test results of a total of 67 RC members subjected to axial force reported in the technical literatures were collected and compared with TATM and existing analytical models(MCFT RA-STM and FA-STM). As a result of comparing with experimental and theoretical results, the test results was better predicted by TATM with 0.94 in average value of $\tau_{test}/\tau_{ana}$. and $11.2\%$ in coefficient of variation than other truss models. And theoretical results obtained from TATM were not effect by steel capacity ratio, axial force, shear span-to-depth ratio, and compressive steel ratio.