Browsing by Author "Çiftçioglu, AO"

Now showing 1 - 4 of 4
  • No Thumbnail Available
    Item
    Wind Load Design of Hangar-Type Closed Steel Structures with Different Roof Pitches Using Abaqus CAE Software
    Çiftçioglu, AO; Yildizel, SA; Yildirim, MS; Dogan, E
    Structures convert the kinetic energy available in the air into potential energy which is in the form of pressure and suction forces reducing or fully stopping its motion. The potential impact of the wind depends on the geometric properties and pertinacity of a building, the angle of the wind flow, its strength and velocity. Design gains importance for tall buildings against the impact of the resonance along with the force based on pressure. Relevant calculations are made in Turkey based on the TS 498 Wind Load Velocity Criterion and this standard is currently being updated. This study develops the wind load design of hangar-type closed steel structures with different roof pitches using Abaqus CAE software.
  • No Thumbnail Available
    Item
    A Study on the Performance of Grey Wolf Optimizer
    Çiftçioglu, AO
    In recent years there is an enormous increase in the emergence of non-deterministic search methods. The effective way of animals in problem-solving (like discovering the shortest path to the food source) has been examined by scientists and swarm intelligence has become a research field that imitates the behaviour of animals in swarm. The moth-flame optimization (MFO) algorithm, salp swarm algorithm (SSA), firefly algorithm (FFA), bat (BAT) algorithm, cuckoo search (CS) algorithm, genetic algorithm (GA), and grey wolf optimizer (GWO) are some of the swarm intelligence based non-deterministic methods. In the present study, the seven methods above are investigated separately. Five mathematical functions are resolved individually by these seven methods. Each algorithm is run 30 times in each benchmark function. The performances of these optimization methods are evaluated and compared within each function individually. Performances of algorithms over convergence are compared by plotting convergence rate graph and boxplots of methods for each function. Considering most of the functions, GWO is observed to be stronger than other algorithms.
  • No Thumbnail Available
    Item
    Grey wolf optimizer integrated within boosting algorithm: Application in mechanical properties prediction of ultra high-performance concrete including carbon nanotubes
    Çiftçioglu, AO; Kazemi, F; Shafighfard, T
    Nowadays, the construction industry has increasingly recognized the superior performance characteristics of ultra high-performance concrete (UHPC). Known for its exceptional durability and high tensile strength, UHPC material is revolutionizing structure standards subjected to extreme environmental conditions and heavy loads. This paper explores the enhancement of UHPC with nano- and micromaterials, employing advanced machinelearning (ML) techniques to optimize the prediction of mechanical properties. Moreover, by introducing the novel extreme gradient boosting (XGBoost) improved by grey wolf optimizer (GWO) algorithm, this research represents the first integration of ML with GWO in UHPC research, significantly enhancing the accuracy of predictions for key properties such as compressive, tensile, and flexural strengths. The study investigates the impact of nanotechnology on UHPC, specifically how carbon nanotubes (CNTs) and microscale reinforcements contribute to advances in strength, durability, and resilience. These enhancements are pivotal in addressing limitations of traditional concrete, especially in high-demand construction environments. The proposed GWOXGB model has demonstrated a remarkable ability to achieve R2 values of 98.4% and 94.8% for UHPC with nanomaterial and micromaterial, respectively, indicating very high level of accuracy in predicting mechanical properties. This model also had one of the lowest error values, demonstrating its precision and ability to minimize prediction errors. This approach significantly facilitates the testing and development of UHPC by automating the accurate determination of its mechanical properties, thereby reducing the reliance on costly and time-consuming experimental methods. Highlighting the transformative potential of combining ML with engineering science, this study offers promising avenues for innovations in construction practices.
  • No Thumbnail Available
    Item
    RAGN-R: A multi-subject ensemble machine-learning method for estimating mechanical properties of advanced structural materials
    Kazemi, F; Çiftçioglu, AO; Shafighfard, T; Asgarkhani, N; Jankowski, R
    The utilization of advanced structural materials, such as preplaced aggregate concrete (PAC), fiber-reinforced concrete (FRC), and FRC beams has revolutionized the field of civil engineering. These materials exhibit enhanced mechanical properties compared to traditional construction materials, offering engineers unprecedented opportunities to optimize the design, construction, and performance of structures and infrastructures. This formal description elucidates the inherent mechanical properties of PAC, FRC, and FRC beams, explores their diverse applications in civil engineering projects. This research aims to propose a surrogate multi-subject ensemble machine-learning (ML) method (named RAGN-R) for estimating mechanical properties of aforementioned advanced materials. The proposed learning approach, RAGN-R, integrates Random forest, Adaptive boosting, and GradieNt boosting techniques, employing a Ridge regression framework for stacking the ensemble. For this purpose, three experimental dataset have been prepared to determine the capability of RAGN-R and the results of the study have been compared with six well-known ML models. It is noteworthy that the proposed RAGN-R has the ability of self-optimizing the hyperparameters, which facilitate the adoptability of the model with engineering problems. Moreover, three datasets have been investigated to show the ability of the RAGN-R for diverse problems. Different performance evaluation metrics have been conducted to present results and compare ML models, which confirms the highest performance of RAGN-R (i.e., 97.7% accuracy) in handling complex relationships and improving overall prediction accuracy.