Structural Engineering Problems Research Articles

Application of Artificial Neural Network for Prediction of Story Drift in High-Rise Buildings

The difficulties arising from growing population expansion and extensive urbanization are driving up the demand for high-rise structures. Comprehending the symmetry and composition of structural components correspondingly is becoming more and more important in light of current developments in computer technology, novel materials, and unique structures. The structural responses of multistory buildings predicted by Artificial Neural Networks (ANN) briefly outlined in a systematic manner. This review enhances the application of ANN for the prediction of story drift in multistory buildings with greater level accuracy. A multistory structure might fall because of the intense movement of the ground, resulting in fatalities and economic loss. As a result, it is imperative that the multistory structure be appropriately constructed to minimize the earthquake hazards. Based on a literature analysis, it can be concluded that artificial neural networks are widely used in simulating and predicting story drift in high-rise building systems. The objective of the current investigation is to present a comprehensive overview of the many approaches and uses of artificial neural network modeling research in the domain of structural engineering. Accordingly, the articles were categorized in this way based on author, year, algorithm used, software application and description of findings and conclusions. Here, the most influential parameter (input vector) is also rigorously reviewed. This information may be taken into account when creating a neural network model that operates more effectively. The review's conclusions might prove beneficial for structural and/or civil engineering problems. The approaches described here assist the structural practitioner in understanding the limitations and advantages of ANN in comparison to other traditional mathematical modelling. This review contributes a body of knowledge for the ANN modelling approach to forecast the story drift evaluation within the field of structural engineering.

Advancing ultimate bond stress–slip model of UHPC structures through a novel hybrid machine learning approach

The significance of Ultimate Bond Stress-Slip (UBS-S) in reinforced Ultra-High Performance Concrete (UHPC) structures cannot be overstated, as it directly affects their load-carrying capacity, structural integrity, and long-term performance. A comprehensive analysis of the UHPC-Parallel Micro Element System (UHPC-PMES), including 144 specimens, evaluated the computational efficiency of the proposed UBS-S model. To this end, the most critical settings of steel bar diameter, concrete cover, bond length, and UHPC compressive strength (db,c,lb,fUHPC′) were directed to create this parametric research. Applying the hybrid approach of three different optimization techniques, namely Physics-Informed Neural Networks (PINN), Genetic Algorithms (GA), and Multiple Linear Regression (MLR), this study predicted the UBS-S at the interface of UHPC and steel bars. It formulated the hyper-parameters effect values (a,m,β,G). The presented research used these algorithms to solve an inverse problem in structural engineering. Comparing the results obtained from PINN, GA, and MLR demonstrated that machine learning techniques and the proposed PMES model could effectively and accurately investigate the ultimate bond stress-slip for reinforced UHPC structures.

Using the Grey Wolf Aquila Synergistic Algorithm for Design Problems in Structural Engineering.

The Aquila Optimizer (AO) is a metaheuristic algorithm that is inspired by the hunting behavior of the Aquila bird. The AO approach has been proven to perform effectively on a range of benchmark optimization issues. However, the AO algorithm may suffer from limited exploration ability in specific situations. To increase the exploration ability of the AO algorithm, this work offers a hybrid approach that employs the alpha position of the Grey Wolf Optimizer (GWO) to drive the search process of the AO algorithm. At the same time, we applied the quasi-opposition-based learning (QOBL) strategy in each phase of the Aquila Optimizer algorithm. This strategy develops quasi-oppositional solutions to current solutions. The quasi-oppositional solutions are then utilized to direct the search phase of the AO algorithm. The GWO method is also notable for its resistance to noise. This means that it can perform effectively even when the objective function is noisy. The AO algorithm, on the other hand, may be sensitive to noise. By integrating the GWO approach into the AO algorithm, we can strengthen its robustness to noise, and hence, improve its performance in real-world issues. In order to evaluate the effectiveness of the technique, the algorithm was benchmarked on 23 well-known test functions and CEC2017 test functions and compared with other popular metaheuristic algorithms. The findings demonstrate that our proposed method has excellent efficacy. Finally, it was applied to five practical engineering issues, and the results showed that the technique is suitable for tough problems with uncertain search spaces.

A general framework of high-performance machine learning algorithms: application in structural mechanics

Data-driven models utilizing powerful artificial intelligence (AI) algorithms have been implemented over the past two decades in different fields of simulation-based engineering science. Most numerical procedures involve processing data sets developed from physical or numerical experiments to create closed-form formulae to predict the corresponding systems’ mechanical response. Efficient AI methodologies that will allow the development and use of accurate predictive models for solving computational intensive engineering problems remain an open issue. In this research work, high-performance machine learning (ML) algorithms are proposed for modeling structural mechanics-related problems, which are implemented in parallel and distributed computing environments to address extremely computationally demanding problems. Four machine learning algorithms are proposed in this work and their performance is investigated in three different structural engineering problems. According to the parametric investigation of the prediction accuracy, the extreme gradient boosting with extended hyper-parameter optimization (XGBoost-HYT-CV) was found to be more efficient regarding the generalization errors deriving a 4.54% residual error for all test cases considered. Furthermore, a comprehensive statistical analysis of the residual errors and a sensitivity analysis of the predictors concerning the target variable are reported. Overall, the proposed models were found to outperform the existing ML methods, where in one case the residual error was decreased by 3-fold. Furthermore, the proposed algorithms demonstrated the generic characteristic of the proposed ML framework for structural mechanics problems.

State of art for hybrid mixed finite element formulation in non-linear analysis of structures

Since the late 80’s the Structural Analysis Research Group of the Instituto Superior Técnico (IST) has been involved in the development of non-conventional finite element formulations in order to overcome some of the limitations associated with the use of the CFE method and to develop high performance numerical tools for the analysis of structural engineering problems. Several alternative models for the linear and non-linear structural analysis have been developed using hybrid and mixed models techniques. These works are summarized in this paper, in which their past and future applications of this formulation in non-linear analysis of structures are fully detailed.

A multi-strut model for the hysteresis behavior and strength assessment of masonry-infilled steel frames with openings under in-plane lateral loading

The analysis of masonry-infilled frames (MIFs) with openings is a complex problem in structural engineering. Previous studies have shown that when an infill panel has an opening, it behaves as four separate subpanels around the opening. Therefore, it can be conceived that a perforated masonry infill panel behaves as four inclined compression-only equivalent struts in each direction. This paper presents a multi-strut model (MSM) using the equivalent strut concept for the nonlinear analysis of masonry-infilled steel frames (MISFs) with openings. The MSM consists of two inelastic side pier struts connected to the surrounding frame joints using rigid trusses representing the subpanels above and below the opening. The model incorporates an envelope curve and a hysteretic law to define the initial stiffness, peak strength, and post-peak strength behavior of the inclined struts under reversal loads. The MSM was used to analyze the lateral load behavior of tested MISFs with openings found in the literature. It was shown that the proposed model is efficient and can predict the stiffness and capacity of MISFs with openings under both monotonic and cyclic lateral loadings.

The Efficiency of Using Machine Learning Techniques in Fiber-Reinforced-Polymer Applications in Structural Engineering

Sustainable solutions in the building construction industry have emerged as a new method for retrofitting applications in the last two decades. Fiber-reinforced polymers (FRPs) have garnered much attention among researchers for improving reinforced concrete (RC) structures. The existing design guidelines for FRP-strengthened RC members were developed using empirical methods that are based on specific databases, limiting the accuracy of the predicted results. Therefore, the use of innovative and efficient prediction tools to predict the behavior of FRP-strengthened RC members has become essential. During the last few years, efforts have been progressively focused on the use of machine learning (ML) as a feasible and effective technique for solving various structural engineering problems. Its capability to predict the behavior of complex nonlinear structural systems while considering a wide range of parameters offers a distinctive opportunity to make the behavior of RC members more predictable and accurate. This paper aims to evaluate the current state of using various ML algorithms in RC members strengthened with FRP to enable researchers to determine the capabilities of current solutions as well as to find research gaps to carry out more research to bridge revealed knowledge and practice gaps. Scopus databases were searched using predefined standards. The search revealed ninety-six articles published between 2016 and 2023. Consequently, these articles were analyzed for ML applications in the field of FRP retrofitting, including flexural and shear strengthening of RC beams, flexural strengthening of slabs, confinement and compressive strength of columns, and FRP bond strength. The results reveal that 32% of the reviewed studies focused on the application of ML techniques to the flexural and shear strengthening of RC beams, 32% on the confinement and compressive strength of columns, 6.5% on the flexural strengthening of slabs, 22% on FRP bond strength, 6.5% on materials, and 1% on beam–column joints. This research also revealed that the application of various ML algorithms has shown a significant improvement in resistance prediction accuracy as compared with the existing empirical solutions. Supervised learning techniques were the most favorable learning method due to their good generalization, interpretability, adaptability, and predictive efficiency. In addition, the selection of suitable ML algorithms and optimization techniques is found to be mainly dictated by the nature of the problem and the characteristics of the dataset. Nonetheless, selecting the most appropriate ML model and optimization algorithm for each specific application remains a challenge, given that each algorithm is developed with different principles and methodologies.

Ultra-High-Performance Concrete (UHPC): A state-of-the-art review of material behavior, structural applications and future

Ultra-high-performance concrete commonly known as UHPC is rising curiosity among structural engineers all over. Though early research on this material dates back to a couple of decades, some initial knowledge about this material, its behavior, and its properties is largely limited to a few research circles in a handful of advanced countries. This paper introduces UHPC as a material, elaborates on its ingredients, and describes its properties. A detailed review of available research literature about UHPC is made. The contributions made by several researchers have been discussed in detail. Following this, the structural behavior and strength of the material are reviewed comprehensively. Comparisons are made between conventional concrete and UHPC with respect to their properties, stress-strain relation, cracking behavior, compressive, tensile, and shear strengths. A detailed evaluation is made of the enhanced properties of the material with respect to its durability and long-term performance. The resistance of this material to moisture permeability, chloride ingress, and chemical attacks is understood. The impact resistance and energy absorption characteristics of the material are compared with conventional concrete. The study documented the structural applications of UHPC as well as the potential applications in the field of civil engineering. Finally, the authors enlisted the impacts of this new material (UHPC) on the future direction of structural engineering and the innovative solutions it can provide to structural engineering problems

Multi-Objective Optimization Design of FRP Reinforced Flat Slabs under Punching Shear by Using NGBoost-Based Surrogate Model

Multi-objective optimization problems (MOPs) in structural engineering arise as a significant challenge in achieving a balance between prediction accuracy and efficiency of the surrogate models, which are conventionally adopted as mechanics-driven models or numerical models. Data-driven models, such as machine learning models, can be instrumental in resolving intricate structural engineering issues that cannot be tackled through mechanics-driven models. This study aims to address the challenges of multi-objective optimization punching shear design of fiber-reinforced polymer (FRP) reinforced flat slabs by using a data-driven surrogate model. Firstly, this study employs an advanced machine learning model, namely Natural Gradient Boosting (NGBoost), to predict the punching shear resistance of FRP reinforced flat slabs. The comparisons with other machine learning models, design provisions and empirical theory models illustrate that the NGBoost model has higher accuracy in predicting the punching shear resistance. Additionally, the NGBoost model is explained with Shapley Additive Explanation (SHAP), revealing that the slab’s effective depth is the primary factor affecting the punching shear resistance. Then, the formulated NGBoost model is adopted as a surrogate model in conjunction with the Non-Dominated Sorting Genetic Algorithm-II (NSGA-II) algorithm for multi-objective optimization design of FRP reinforced flat slabs subjected to punching shear. Through a case study, it is demonstrated that the Pareto-optimal set of the punching shear resistance and cost of the FRP reinforced flat slabs can be successfully obtained. By discussing the effects of design parameter changes on the results, it is also shown that increasing the slab’s effective depth is a relatively effective way to achieve higher punching shear resistance of FRP reinforced flat slabs.

Sparse polynomial chaos expansion for nonlinear finite element simulations with random material properties

AbstractThis contribution deals with the uncertainty quantification for applied nonlinear structural engineering problems, including high stochastic dimensions. A finite element problem with different material models is investigated. The efficiency, accuracy and convergence of sparse PCE are studied numerically and compared with Monte‐Carlo Simulation (MCS) for non‐linear structural analysis including elasto‐plastic and damage models. In both models, the Young's modulus is considered as random fields discretised by Karhunen Loeve Expansion (KLE). In the provided studies, sparse PCE converges fast and is highly efficient for linear elastic and elasto‐plastic material models. However, sparse PCE loses its effectiveness and exhibits lower accuracy for the damage material model.

Unified Flexural Resistance Design Method and Evaluation Frame for the B-Regions of RC Flexural Members—Theory and Application

The load and resistance factor design (LRFD) method is normally used to design B-regions of reinforced concrete (RC) flexural members. The design includes many checks corresponding to different limit states. The LRFD method requires many loop calculation steps in the design, demonstrating its relative inefficiency. It cannot be applied to compare limit states directly and quantitatively. Different design limit states are separated and isolated. How to improve the analytical calculation efficiency of the LRFD method and to realize direct and quantitative comparisons between limit states are very important problems in structural engineering. This paper presents an innovative unified flexural resistance design (UFRD) method and a unified flexural resistance evaluation (UFRE) frame to solve these problems to some extent. The main contents include the unified flexural resistance (UFR) principles, formulas for the unified flexural resistance design (UFRD) method, the operation procedure to facilitate its usage, the UFRE framework to compare limit states, and three examples. The results show that the UFRD method can provide the same design outcomes as the LRFD one. However, UFRD calculations are simpler, requiring at most 20% of the calculation steps of the LRFD method. The UFRE frame can make different limit states compare with each other directly and quantitatively, which cannot be realized by the LRFD method. It helps expose some potential and insufficient flexural resistance hazards for some limit states, such as the only 10% relative strength reservation of one example. Thus, the UFRD method and the UFRE frame supplement and develop the LRFD method to some degree. The simplicity and practicality of the approach and the frame make them appropriate for many applications.

Applications and Potential of Artificial Intelligence in Structural Engineering

This review article discusses the various applications of artificial intelligence (AI) in structural engineering. The paper presents a comprehensive overview of the latest research in AI and its use in prediction, risk analysis, decision-making, resources optimization, classification and selection, construction, maintenance, and management of structural engineering problems. The review highlights the significant benefits of using AI, including improved accuracy, efficiency, and cost-effectiveness, compared to conventional methods. It also discusses the limitations and challenges associated with AI in structural engineering and future directions for research. The paper concludes that AI has enormous potential in structural engineering and will transform the field in the coming years. Key Words: artificial intelligence, structural engineering, prediction, resources optimization, construction, maintenance.

Use of Real Coded Genetic Algorithm as a Pre-Dimensioning Tool for Prestressed Concrete Beams

In project practice, the search for optimal solutions is based on the traditional process of trial and error, which consumes much time and does not guarantee that solutions found are the optimal solutions for the problem. Many studies have been developed in recent years with the aim of solving problems in various fields of structural engineering with the aid of intelligent algorithms; however, when it comes to the optimization of structural designs, the approaches considered by the authors involve a large number of variables and constraints, making the implementation of optimization techniques difficult and consuming significant processing time. This research aims to evaluate the efficiency of intelligent algorithms when associated with structural optimization approaches that are simpler to implement. Therefore, a Genetic Algorithm in Real Coding was built to serve as an auxiliary tool for pre-dimensioning prestressed concrete beams. With this, the problem becomes simpler to implement, as it depends on a smaller number of variables, leading to less processing time consumption. Simulations were performed to calibrate the Genetic Algorithm and find the optimal solution later. The solution found by the algorithm was compared with the real solution of a project that had already gone through a traditional optimization process. Even in these circumstances, the proposed Genetic Algorithm was able to find, in 210 s, a more economical solution. Our studies found that even with more straightforward approaches, intelligent algorithms can help in the search for optimal solutions to structural engineering problems; in addition, using real coding in fact proved to be a great strategy due to the nature of the problem, making the implementation of the algorithm simpler and ensuring answers with little processing time.

Deep learning model for prediction of non-linear cyclic hysteresis of seismic isolation devices: Full-scale experimental validation

Recurrent neural networks predicting the non-linear hysteresis behavior of specific triple pendulum bearing (TPB) and lead rubber bearing (LRB) seismic isolation devices, are developed, and tested in this paper. Experimental datasets of full-scale isolators were derived from a shake-table test program of a five-story building specimen, performed at the Hyogo Engineering Research Center of Miki, Japan. Data measured during several table motions of different amplitudes, frequency content, and durations, are appropriately processed to construct a substantial TPB/LRB dataset of 158/55 samples. A comprehensive framework is proposed to process the data, to optimize the network architecture, and to train, validate, and to test the machine learning models. The comparisons with reference experimental data showed that developed models could predict the two-dimensional hysteresis behavior of studied isolators with a very good accuracy. The R2 value of the TPB/LRB model was of 0.83/0.96 on new unseen dataset. The illustration of representative predictions showed the power of a single model to capture different and irregular hysteresis patterns, performed by a typical isolator under variable and realistic loading conditions and that can't be described by conventional analytical models. The generalization capability of these surrogate models on such a substantial data, revealed the benefit of applying machine learning to solve complex structural engineering problems.

Seismic Response Analysis of Vertically Irregular RC Building with MR Dampers

Seismic hazard mitigation has become one of the emerging problems in structural engineering therefore; many researchers have been studied on seismic performance of reinforced concrete structures having vertically regular and irregular. It is noted that vertically irregular structures are more prone to hazards due to earthquake. The recent emerging trend is to use active, passive, semi-active controls that aid in keeping the responses of structures in permissible. The most dynamically varied semi-active damper is a Magneto-Rheological (MR) damper which is used in this study to control the seismic responses. In the present study, four different models with and without MR dampers of which three models are vertically irregular in height are subjected to four realistic ground motions. The governing equation of motion for various building models is solved numerically by Newmark’s step-by-step integration method. The dynamic behaviour of semiactive MR damper has been predicted by modified Bouc-Wen model. This study employed the Lyapunov direct approach as a control algorithm for stability analysis and design of MR controller. The responses of various building models are simulated through MATLABâcomputing software. The study observed that considerable reduction in seismic responses of models with MR dampers is obtained as compared to respective building models without control. It is also observed that responses of building models are influenced by the vertical irregularities in mass and stiffness of building models.

Revisitation of elastic buckling of circular rings: Some analytical and numerical issues

Elastic buckling of circular rings under external pressure is one of the oldest and most ubiquitous problems in structural engineering and the first meaningful solution to it was proposed by Levy (1884). This basic solution, based on the Euler–Bernoulli beam model, can be found in the vast majority of textbooks on stability. However, there are a number of subtleties in the formulation of the problem which need to be taken into account so and one should always be well aware of the underlying analytical fundamentals, especially in the current world which strongly relies on numerical analyses. Therefore, starting from the general nonlinear kinematics of the problem, the buckling load is revisited by means of a direct energy approach and on the basis of an in-depth discussion of the classic kinematic hypotheses. Possibly different outcomes are pointed out and remarked. Finally, some inconsistencies which unexpectedly arise in pursuing the solution by means of commercial FE packages are presented and commented.

UHPC-PINN-parallel micro element system for the local bond stress–slip model subjected to monotonic loading

The bond stress between Ultra-High-Performance Concrete (UHPC) and steel bars is an essential issue in designing reinforced UHPC structures. Although, to date, considerable studies have been directed on the Local Bond Stress-Slip (LBSS) relationship, not enough progress has been achieved in the parametric investigation of the Parallel Micro Element System (PMES) on the LBSS relationship at the steel bar-UHPC interface. Therefore, the presented research applied the two-parameter Weibull Cumulative Distribution Function (WCDF) and conducted a series of monotonic pullout tests on 144 specimens considering the most influential parameters on the LBSS, such as concrete cover (c), UHPC compressive strength (fUHPC′), steel bar diameter (db), and bond length (lb). The Finite Element Method (FEM), ABAQUS, calibrated the experimental results to provide a reliable database. The Genetic Algorithm (GA) and the Physics-Informed Neural Network (PINN) were adopted to solve the inverse problem of this research, predict the results, and check the accuracy of the PMES. It can be concluded that PINN's application, compared to GA, leads to more precise results. Observing the specimens R16C1L2F155 and R18C1L3F155, for instance, PINN provided more consistent values for the friction coefficient (β) [0.2549, 0.8165] and the LBSS relationship stiffness (G) [453.1250, 249.9757], while GA failed to acquire not only β and G but also the hyper-parameters of (a) [0.3016, 0.3088] and (m) [0.3035, 0.3635], respectively. This research developed the application of GA and PINN to solve an inverse problem in structural engineering.

Bias-variance tradeoff in machine learning: Theoretical formulation and implications to structural engineering applications

When machine learning (ML) is used to solve practical problems in structural engineering (SE), the associated data may not reflect the overall distribution of the population. With this in mind, a common practice in prior ML-SE studies is to optimize the model such that the testing error is minimized. However, this strategy can lead to a situation where the testing error is significantly greater than the error obtained during training, which is a clear sign of overfitting. Therefore, simply selecting a model with the highest testing performance does not necessarily guarantee generalizability. To address this challenge, this study aims to provide some practical guidelines that consider both the model complexity and generalizability. The proposed solution is fairly straightforward to implement. The generalized bias-variance decomposition is first introduced, followed by a discussion of different error variation patterns (bell-shaped, double-descent, and monotonically decreasing). A mathematical formulation that establishes a relationship between the model complexity and generalizability via an objective function is then presented to facilitate the model parameter selection. Subsequently, four numerical experiments are conducted to apply the formulation within the structural engineering domain. Finally, the proposed model selection approach is compared with existing methods to demonstrate its advantages and limitations.

Automated classification of linear bifurcation buckling eigenmodes in thin-walled cylindrical shell structures

Computational problems in structural engineering are growing ever larger and solutions must increasingly be based on correspondingly large datasets obtained from detailed parametric sweeps. However, the acquisition of computational datasets of useful size is also becoming increasingly unfeasible without extensive use of automation. In computational shell buckling studies, particularly those of thin-walled shells under complex loading conditions, an important qualitative piece of information is the class of buckling mode which reveals the dominant destabilising membrane stress components. Unfortunately, the diversity of geometries that can be encountered in computational shell buckling studies is truly vast, and there is currently no way to rapidly assess the buckling mode without laborious direct human observation of the model output.This paper presents an automated classification tool for linear bifurcation buckling eigenmodes in cylindrical shells such as those found as wind turbine support towers, chimneys, silos, tanks, piles and pipelines. It is based on a convolutional neural network implemented using the PyTorch machine learning framework. The adopted network architecture is based on those widely adopted for image classification and recognition tasks, chosen based on a stratified five-fold cross-validation exercise. The network is trained on a purposefully generated basic dataset of 13,392 linear bifurcation buckling eigenmodes modes encoded as chromatic signatures in .jpg images (enhanced to 25,726 by transformations). An example parametric sweep of a cylindrical shell under unsymmetrical wind loading illustrates the performance of the classifier. A GitHub repository offers Python scripts and instructions on how to download the dataset and trained network.

Neural network prediction of web-crushing strength of i-shaped reinforced concrete beams

Introduction. Under the combined action of shear force and bending moment web-shear cracking takes place in the thin webs of reinforced concrete beams. The strength of the concrete struts between the cracks depends on the size of the web, concrete strength, parameters of the transverse and longitudinal reinforcement, and the shear span-to-depth ratio. Plane stress constitutive models for reinforced concrete are difficult to implement. For this reason, building codes employ empirical formulas for web-crushing strength which were obtained from the analysis of the existing experimental data. Using machine learning tools — artificial neural networks (ANN) — can serve as a solution that allows to take into account the impact of structural and loading parameters more accurately. The analyzed experimental base included the test results of 77 beams. The input layer of the ANN consisted of 7 independent variables, the output — of 1 dependent, and both linear and nonlinear functions were considered as activation functions.
 
 Materials and methods. The article studies I-shaped concrete beams with basalt fiber reinforced polymer and steel transverse reinforcement tested by authors. To ensure the statistical significance of independent variables in all indicators, the database of other authors was also considered. Neural networks were developed using STATISTICA software package. In the first stage, the input and output variables were normalized. The accuracy of the ANN model prediction was compared with the accuracy of regression models. In the last stage, the calculations were performed without normalizing the variables.
 
 Results. Artificial neural networks prediction has high accuracy. The relative error of prediction was 28.6 % for the regression method and 10.9 % for the ANN. For calculations without preliminary normalization relative error of prediction was 6.6 %.
 
 Conclusions. The results of research and other similar studies suggest ANN to be a promising tool for solving intractable problems of structural engineering.