Yazar "Kim, Sanghun" seçeneğine göre listele
Listeleniyor 1 - 12 / 12
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe CO2 Emission Optimization of Concrete-Filled Steel Tubular Rectangular Stub Columns Using Metaheuristic Algorithms(Mdpi, 2021) Cakiroglu, Celal; Islam, Kamrul; Bekdas, Gebrail; Kim, Sanghun; Geem, Zong WooConcrete-filled steel tubular (CFST) columns have been assiduously investigated experimentally and numerically due to the superior structural performance they exhibit. To obtain the best possible performance from CFST columns while reducing the environmental impact, the use of optimization algorithms is indispensable. Metaheuristic optimization techniques provide the designers of CFST members with a very efficient set of tools to obtain design combinations that perform well under external loading and have a low carbon footprint at the same time. That is why metaheuristic algorithms are more applicable in civil engineering due to their high efficiency. A large number of formulas for the prediction of the axial ultimate load-carrying capacity Nu of CFST columns are available in design codes. However, a limitation of the usage of these design formulas is that most of these formulas are only applicable for narrow ranges of design variables. In this study a newly developed set of equations with a wide range of applicability that calculates Nu in case of rectangular cross-sections is applied. In order to optimize the cross-sectional dimensions, two different metaheuristic algorithms are used, and their performances are compared. The reduction in CO2 emission is demonstrated as a function of cross-sectional dimensions while considering certain structural performance requirements. The outcome of the more recently developed social spider algorithm is compared to the outcome of the well-established harmony search technique. The objective of optimization was to minimize CO2 emissions associated with the fabrication of CFST stub columns. The effects of varying the wall thickness as well as the concrete compressive strength on CO2 emissions are visualized by using two different optimization techniques.Öğe Explainable ensemble learning models for the rheological properties of self-compacting concrete(MDPI-Multidisciplinary Digital Publishing Institute, 2022) Çakıroğlu, Celal; Bekdaş, Gebrail; Kim, Sanghun; Geem, Zong WooSelf-compacting concrete (SCC) has been developed as a type of concrete capable of filling narrow gaps in highly reinforced areas of a mold without internal or external vibration. Bleeding and segregation in SCC can be prevented by the addition of superplasticizers. Due to these favorable properties, SCC has been adopted worldwide. The workability of SCC is closely related to its yield stress and plastic viscosity levels. Therefore, the accurate prediction of yield stress and plastic viscosity of SCC has certain advantages. Predictions of the shear stress and plastic viscosity of SCC is presented in the current study using four different ensemble machine learning techniques: Light Gradient Boosting Machine (LightGBM), Extreme Gradient Boosting (XGBoost), random forest, and Categorical Gradient Boosting (CatBoost). A new database containing the results of slump flow, V-funnel, and L-Box tests with the corresponding shear stress and plastic viscosity values was curated from the literature to develop these ensemble learning models. The performances of these algorithms were compared using state-of-the-art statistical measures of accuracy. Afterward, the output of these ensemble learning algorithms was interpreted with the help of SHapley Additive exPlanations (SHAP) analysis and individual conditional expectation (ICE) plots. Each input variable's effect on the predictions of the model and their interdependencies have been illustrated. Highly accurate predictions could be achieved with a coefficient of determination greater than 0.96 for both shear stress and plastic viscosity.Öğe Interpretable machine learning algorithms to predict the axial capacity of frp-reinforced concrete columns(MDPI-Multidisciplinary Digital Publishing Institute, 2022) Çakıroğlu, Celal; Islam, Kamrul; Bekdaş, Gebrail; Kim, Sanghun; Geem, Zong WooFiber-reinforced polymer (FRP) rebars are increasingly being used as an alternative to steel rebars in reinforced concrete (RC) members due to their excellent corrosion resistance capability and enhanced mechanical properties. Extensive research works have been performed in the last two decades to develop predictive models, codes, and guidelines to estimate the axial load-carrying capacity of FRP-RC columns. This study utilizes the power of artificial intelligence and develops an alternative approach to predict the axial capacity of FRP-RC columns more accurately using data-driven machine learning (ML) algorithms. A database of 117 tests of axially loaded FRP-RC columns is collected from the literature. The geometric and material properties, column shape and slenderness ratio, reinforcement details, and FRP types are used as the input variables, while the load-carrying capacity is used as the output response to develop the ML models. Furthermore, the input-output relationship of the ML model is explained through feature importance analysis and the SHapely Additive exPlanations (SHAP) approach. Eight ML models, namely, Kernel Ridge Regression, Lasso Regression, Support Vector Machine, Gradient Boosting Machine, Adaptive Boosting, Random Forest, Categorical Gradient Boosting, and Extreme Gradient Boosting, are used in this study for capacity prediction, and their relative performances are compared to identify the best-performing ML model. Finally, predictive equations are proposed using the harmony search optimization and the model interpretations obtained through the SHAP algorithm.Öğe Metaheuristic optimization of laminated composite plates with Cut-Outs(MDPI-Multidisciplinary Digital Publishing Institute, 2021) Çakıroğlu, Celal; Islam, Kamrul; Bekdaş, Gebrail; Kim, Sanghun; Geem, Zong WooThe stacking sequence optimization of laminated composite plates while maximizing the structural performance or minimizing the weight is a subject investigated extensively in the literature. Meanwhile, research on the optimization of laminates with cut-outs has been relatively limited. Cut-outs being an indispensable feature of structural components, this paper concentrates on the stacking sequence optimization of composite laminates in the presence of circular cut-outs. The buckling load of a laminate is used as a metric to quantify the structural performance. Here the laminates are modeled as carbon fiber-reinforced composites using the finite element analysis software, ABAQUS. For the optimization, the widely used harmony search algorithm is applied. In terms of design variables, ply thickness, and fiber orientation angles of the plies are used as continuously changing variables. In addition to the stacking sequence, another geometric variable to consider is the aspect ratio (ratio of the length of the longer sides to the length of the shorter sides of the plate) of the rectangular laminates. The optimization is carried out for three different aspect ratios. It is shown that, by using dispersed stacking sequences instead of the commonly used 0 & DEG;/& PLUSMN;45 & DEG;/& PLUSMN;90 & DEG; fiber angle stacks, significantly higher buckling loads can be achieved. Furthermore, changing the cut-out geometry is found to have a significant effect on the structural performance.Öğe Neural Network Predictive Models for Alkali-Activated Concrete Carbon Emission Using Metaheuristic Optimization Algorithms(2024) Aydın, Yaren; Çakıroğlu, Celal; Bekdaş, Gebrail; Işıkdağ, Ümit; Kim, Sanghun; Hong, Junhee; Geem, Zong WooDue to environmental impacts and the need for energy efficiency, the cement industry aims to make more durable and sustainable materials with less energy requirements without compromising mechanical properties based on UN Sustainable Development Goals 9 and 11. Carbon dioxide (CO2) emission into the atmosphere is mostly the result of human-induced activities and causes dangerous environmental impacts by increasing the average temperature of the earth. Since the production of ordinary Portland cement (PC) is a major contributor to CO2 emissions, this study proposes alkali-activated binders as an alternative to reduce the environmental impact of ordinary Portland cement production. The dataset required for the training processes of these algorithms was created using Mendeley as a data-gathering instrument. Some of the most efficient state-of-the-art meta-heuristic optimization algorithms were applied to obtain the optimal neural network architecture with the highest performance. These neural network models were applied in the prediction of carbon emissions. The accuracy of these models was measured using statistical measures such as the mean squared error (MSE) and coefficient of determination (R2). The results show that carbon emissions associated with the production of alkali-activated concrete can be predicted with high accuracy using state-of-the-art machine learning techniques. In this study, in which the binders produced by the alkali activation method were evaluated for their usability as a binder material to replace Portland cement, it is concluded that the most successful hyperparameter optimization algorithm for this study is the genetic algorithm (GA) with accurate mean squared error (MSE = 161.17) and coefficient of determination (R2 = 0.90) values in the datasets.Öğe Optimal dimensioning of retaining walls using explainable ensemble learning algorithms(MDPI-Multidisciplinary Digital Publishing Institute, 2022) Bekdaş, Gebrail; Çakıroğlu, Celal; Kim, Sanghun; Geem, Zong WooThis paper develops predictive models for optimal dimensions that minimize the construction cost associated with reinforced concrete retaining walls. Random Forest, Extreme Gradient Boosting (XGBoost), Categorical Gradient Boosting (CatBoost), and Light Gradient Boosting Machine (LightGBM) algorithms were applied to obtain the predictive models. Predictive models were trained using a comprehensive dataset, which was generated using the Harmony Search (HS) algorithm. Each data sample in this database consists of a unique combination of the soil density, friction angle, ultimate bearing pressure, surcharge, the unit cost of concrete, and six different dimensions that describe an optimal retaining wall geometry. The influence of these design features on the optimal dimensioning and their interdependence are explained and visualized using the SHapley Additive exPlanations (SHAP) algorithm. The prediction accuracy of the used ensemble learning methods is evaluated with different metrics of accuracy such as the coefficient of determination, root mean square error, and mean absolute error. Comparing predicted and actual optimal dimensions on a test set showed that an R-2 score of 0.99 could be achieved. In terms of computational speed, the LightGBM algorithm was found to be the fastest, with an average execution speed of 6.17 s for the training and testing of the model. On the other hand, the highest accuracy could be achieved by the CatBoost algorithm. The availability of open-source machine learning algorithms and high-quality datasets makes it possible for designers to supplement traditional design procedures with newly developed machine learning techniques. The novel methodology proposed in this paper aims at producing larger datasets, thereby increasing the applicability and accuracy of machine learning algorithms in relation to optimal dimensioning of structures.Öğe Optimal Dimensions of Post-Tensioned Concrete Cylindrical Walls Using Harmony Search and Ensemble Learning with SHAP(Mdpi, 2023) Bekdas, Gebrail; Cakiroglu, Celal; Kim, Sanghun; Geem, Zong WooThe optimal design of prestressed concrete cylindrical walls is greatly beneficial for economic and environmental impact. However, the lack of the available big enough datasets for the training of robust machine learning models is one of the factors that prevents wide adoption of machine learning techniques in structural design. The current study demonstrates the application of the well-established harmony search methodology to create a large database of optimal design configurations. The unit costs of concrete and steel used in the construction, the specific weight of the stored fluid, and the height of the cylindrical wall are the input variables whereas the optimum thicknesses of the wall with and without post-tensioning are the output variables. Based on this database, some of the most efficient ensemble learning techniques like the Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), Categorical Gradient Boosting (CatBoost) and Random Forest algorithms have been trained. An R-2 score greater than 0.98 could be achieved by all of the ensemble learning models. Furthermore, the impacts of different input features on the predictions of different machine learning models have been analyzed using the SHapley Additive exPlanations (SHAP) methodology. The height of the cylindrical wall was found to have the greatest impact on the optimal wall thickness, followed by the specific weight of the stored fluid. Also, with the help of individual conditional expectation (ICE) plots the variations of predictive model outputs with respect to each input feature have been visualized. By using the genetic programming methodology, predictive equations have been obtained for the optimal wall thickness.Öğe Optimal dimensions of post-tensioned concrete cylindrical walls using harmony search and snsemble learning with SHAP(2023) Çakıroğlu, Celal; Bektaş, Gebrail; Geem, Zong Woo; Kim, SanghunThe optimal design of prestressed concrete cylindrical walls is greatly beneficial for economic and environmental impact. However, the lack of the available big enough datasets for the training of robust machine learning models is one of the factors that prevents wide adoption of machine learning techniques in structural design. The current study demonstrates the application of the well-established harmony search methodology to create a large database of optimal design configurations. The unit costs of concrete and steel used in the construction, the specific weight of the stored fluid, and the height of the cylindrical wall are the input variables whereas the optimum thicknesses of the wall with and without post-tensioning are the output variables. Based on this database, some of the most efficient ensemble learning techniques like the Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), Categorical Gradient Boosting (CatBoost) and Random Forest algorithms have been trained. An R2 score greater than 0.98 could be achieved by all of the ensemble learning models. Furthermore, the impacts of different input features on the predictions of different machine learning models have been analyzed using the SHapley Additive exPlanations (SHAP) methodology. The height of the cylindrical wall was found to have the greatest impact on the optimal wall thickness, followed by the specific weight of the stored fluid. Also, with the help of individual conditional expectation (ICE) plots the variations of predictive model outputs with respect to each input feature have been visualized. By using the genetic programming methodology, predictive equations have been obtained for the optimal wall thickness.Öğe Optimisation of shear and lateral-torsional buckling of steel plate girders using meta-heuristic algorithms(Mdpi, 2020) Çakıroğlu, Celal; Bekdaş, Gebrail; Kim, Sanghun; Geem, Zong WooThe shear buckling of web plates and lateral-torsional buckling are among the major failure modes of plate girders. The importance of the lateral-torsional buckling capacity of plate girders was further evidenced when several plate girders of a bridge in Edmonton, Alberta, Canada failed in 2015, because insufficient bracing led to the lateral buckling of the plate girders. In this study, we focus on the optimisation of the cross-sections of plate girders using a well-known and extremely efficient meta-heuristic optimisation algorithm called the harmony search algorithm. The objective of this optimisation is to design the cross-sections of the plate girders with the minimum area that satisfies requirements, such as the lateral-torsional buckling load and ultimate shear stress. The base geometry, material properties, applied load and boundary conditions were taken from an experimental study and optimised. It was revealed that the same amount of load-carrying capacity demonstrated by this model can be achieved with a cross-sectional area 16% smaller than that of the original specimen. Furthermore, the slenderness of the web plate was found to have a decisive effect on the cost-efficiency of the plate girder design.Öğe Optimization and predictive modeling of reinforced concrete circular columns(MDPI-Multidisciplinary Digital Publishing Institute, 2022) Bekdaş, Gebrail; Çakıroğlu, Celal; Kim, Sanghun; Geem, Zong WooMetaheuristic optimization techniques are widely applied in the optimal design of structural members. This paper presents the application of the harmony search algorithm to the optimal dimensioning of reinforced concrete circular columns. For the objective of optimization, the total cost of steel and concrete associated with the construction process were selected. The selected variables of optimization include the diameter of the column, the total cross-sectional area of steel, the unit costs of steel and concrete used in the construction, the total length of the column, and applied axial force and the bending moment acting on the column. By using the minimum allowable dimensions as the constraints of optimization, 3125 different data samples were generated where each data sample is an optimal design configuration. Based on the generated dataset, the SHapley Additive exPlanations (SHAP) algorithm was applied in combination with ensemble learning predictive models to determine the impact of each design variable on the model predictions. The relationships between the design variables and the objective function were visualized using the design of experiments methodology. Applying state-of-the-art statistical accuracy measures such as the coefficient of determination, the predictive models were demonstrated to be highly accurate. The current study demonstrates a novel technique for generating large datasets for the development of data-driven machine learning models. This new methodology can enhance the availability of large datasets, thereby facilitating the application of high-performance machine learning predictive models for optimal structural design.Öğe Optimum design of cylindrical walls using ensemble learning methods(MDPI-Multidisciplinary Digital Publishing Institute, 2022) Bekdaş, Gebrail; Çakıroğlu, Celal; Islam, Kamrul; Kim, Sanghun; Geem, Zong WooThe optimum cost of the structure design is one of the major goals of structural engineers. The availability of large datasets with preoptimized structural configurations can facilitate the process of optimum design significantly. The current study uses a dataset of 7744 optimum design configurations for a cylindrical water tank. Each of them was obtained by using the harmony search algorithm. The database used contains unique combinations of height, radius, total cost, material unit cost, and corresponding wall thickness that minimize the total cost. It was used to create ensemble learning models such as Random Forest, Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), and Categorical Gradient Boosting (CatBoost). Generated machine learning models were able to predict the optimum wall thickness corresponding to new data with high accuracy. Using SHapely Additive exPlanations (SHAP), the height of a cylindrical wall was found to have the greatest impact on the optimum wall thickness followed by radius and the ratio of concrete unit cost to steel unit cost.Öğe Performance of optimum tuned PID controller with different feedback strategies on active-controlled structures(MDPI-Multidisciplinary Digital Publishing Institute, 2021) Ulusoy, Serdar; Bekdaş, Gebrail; Niğdeli, Sinan Melih; Kim, Sanghun; Geem, Zong WooIn this study, multi-story structures with different combinations (on each floor and only the first floor) of active tendon control systems driven by a proportional-integral-derivative (PID) controller were actively controlled. The PID parameters, K-p (proportional gain), T-d (derivative gain), and T-i (integral gain) for each structure, were optimally tuned by using both the harmony search algorithm (HS) and flower pollination algorithm (FPA), which are metaheuristic algorithms. In two different active-controlled structures, which are formed according to the position of the PID, the structural responses under near-fault records defined in FEMA P-695 are examined to determine the appropriate feedback which was applied for displacement, velocity, acceleration, and total acceleration. The performance of the different feedback strategies on these two active-controlled structures is evaluated. As a result, the acceleration feedback is suitable for all combinations of the active control system with a PID controller. The HS algorithm outperforms the optimum results found according to the FPA.
