Fiber-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.

/pdf/interpretable-machine-learning-algorithms-to-predict-the-2nfy8huw.pdf

Interpretable Machine Learning Algorithms to Predict the Axial Capacity of FRP-Reinforced Concrete Columns

This 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 R2 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.

/pdf/optimal-dimensioning-of-retaining-walls-using-explainable-22k6mc2z.pdf

Optimal Dimensioning of Retaining Walls Using Explainable Ensemble Learning Algorithms

This study focuses on tuned liquid dampers (TLDs) using liquids with different characteristics optimized with the adaptive harmony search algorithm (AHS). TLDs utilize the characteristic features of the liquid to absorb the dynamic forces entering the structure and benefit from the sloshing movement and the spring stiffness created by the liquid mass. TLDs have been optimized to investigate the effect of liquid characteristics on the control by analyzing various liquids. For optimization, the memory consideration ratio (HMCR) and fret width (FW) values were adapted from the classical harmony search (HS) algorithm parameters. The TLDs were used on three types of structure models, such as single-story, 10, and 40 stories. The contribution of the liquid characteristics to the damping performance was investigated by optimizing the minimum displacement under seismic excitation. According to the results, it was understood that the liquid density and kinematic viscosity do not affect single-story structures alone. However, two characteristic features should be evaluated together. As the structure mass increases, the viscosity and density become more prominent.

/pdf/optimization-of-tuned-liquid-damper-including-different-26664bjb.pdf

Optimization of Tuned Liquid Damper Including Different Liquids for Lateral Displacement Control of Single and Multi-Story Structures

The Mosaic of Metaheuristic Algorithms in Structural Optimization

Metaheuristic 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.

/pdf/optimization-and-predictive-modeling-of-reinforced-concrete-2onz8hvk.pdf

Optimization and Predictive Modeling of Reinforced Concrete Circular Columns

In this study, the tuned liquid damper (TLD) device was optimized by the harmony search (HS) and adaptive harmony search algorithms (AHS). Using the harmony search algorithm, seismic excitations were directed at single and ten-story structures, and TLD parameters were optimized to minimize building movement. To improve design parameters, the optimization process was repeated by adapting the design factors of the harmony search algorithm. For this purpose, both the harmony memory consideration ratio (HMCR) and fret width (FW) were gradually reduced by providing an initial value, and optimum algorithm parameters were obtained. As a result of both optimizations, in a critical seismic analysis, the displacements of the adaptive harmony search showed smaller means and standard deviations than those of the classical harmony search.

/pdf/adaptive-harmony-search-for-tuned-liquid-damper-optimization-3uyxmoon.pdf

Adaptive Harmony Search for Tuned Liquid Damper Optimization under Seismic Excitation

In this study, a seismic isolator placed on the base of a structure was optimized under various earthquake records using an adaptive harmony search algorithm (AHS). As known, the base-isolation systems with very low stiffness provide a rigid response of superstructure, so it was assumed that the structure is rigid and the base-isolated structure can be considered as a single-degree of freedom structure. By using this assumption, an optimization method that is independent of structural properties but specific to the chosen earthquake excitation set is proposed. By taking three different damping ratio limits and isolator displacement limits, the isolator period and damping ratio were optimized so that the acceleration of the structure was minimized for nine cases. In the critical seismic analysis performed with optimum isolator parameters, the results obtained for different damping ratios and isolator periods were compared. From the results, it is determined that isolators with low damping ratios require more ductility, and as the damping ratio increases, further restriction of the movement of the isolator increases the control efficiency. Thus, it is revealed that increasing the ductility of the isolator is effective in reducing the total acceleration in the structure.

/pdf/optimization-of-seismic-base-isolation-system-using-adaptive-2xmv39kj.pdf

Optimization of Seismic Base Isolation System Using Adaptive Harmony Search Algorithm

In this study, a truss system is optimized with the Harmony Search Algorithm (HS) inspired by the process of searching for the best harmony. It is compared with the Adaptive Harmony Search Algorithm whose initial values of parameters are defined and gradually decrease. The optimization process is aimed to optimize the volume by calculating the optimum volume of the elements of the truss system. In design, the effect of design factors specific to the harmony search algorithm on optimization was investigated. For this purpose, optimization was performed with different harmony memory consideration ratio (HMCR) values. By adapting the harmony search algorithm, optimum results are obtained for the values that gradually decrease according to the HMCR and fret width (FW) values. Optimum volume function with the lowest standard deviation values was obtained with adaptive harmony search optimization in the mean of the optimum value found for both optimizations. 

Comparison of Classical and Adaptive Parameter Setting for Harmony Search on a Structural Optimization Problem

This study focuses on the performance analysis of optimum tuned liquid dampers (TLDs) under the different live loads of three different structure models, designed as both single and multi-story, under earthquake excitations. For this purpose, single, ten, and forty story structure models have been created and tuned liquid damping devices that contain liquids of different densities and viscosities such as acetone, mercury, and seawater are placed on the structure. For the analysis conducted under earthquake excitations, optimum damping device parameters were previously obtained with the Adaptive Harmony Search Algorithm (AHS), and minimizing the movement of the structure was aimed. The effect of the damping device on the control performance was investigated under increasing and decreasing live loads for the uncertain mass of the structure because of variable actions. When the results are examined, it has been determined that the increase in the story number of the structure will less affect the displacement reduction performance of TLDs for the structure under uncertain variable loading.

Ayla Ocak

Papers

Optimization of Tuned Liquid Damper Including Different Liquids for Lateral Displacement Control of Single and Multi-Story Structures

Adaptive Harmony Search for Tuned Liquid Damper Optimization under Seismic Excitation

Optimization of Seismic Base Isolation System Using Adaptive Harmony Search Algorithm

Comparison of Classical and Adaptive Parameter Setting for Harmony Search on a Structural Optimization Problem

Robustness evaluation of optimum tuned liquid dampers for uncertain variable loading of structures