Prediction Of Structural Performance Research Articles

Predictive Modeling of Structural Performance Using Machine Learning: A Comprehensive Review

Predicting structural performance is a critical aspect of civil engineering, ensuring the safety, efficiency, and durability of buildings and infrastructure. Traditional methods, such as finite element analysis and empirical modeling, often fall short in addressing the complexities of modern structural systems. The advent of machine learning (ML) has revolutionized this domain by offering data-driven approaches capable of handling non-linear relationships and large datasets, enhancing the accuracy and efficiency of structural performance predictions. This review paper examines the applications of ML techniques, including Artificial Neural Networks (ANN), Support Vector Regression (SVR), Random Forest (RF), Decision Tree Regression (DTR), and hybrid models, in predicting structural metrics such as load-bearing capacity, deflection, durability, and seismic performance. The paper synthesizes findings from recent studies, highlighting key achievements and challenges, such as limited real-world validation, the need for hybrid approaches, and barriers to integrating ML into engineering workflows. By identifying critical research gaps and proposing future directions, this review aims to provide a comprehensive framework for advancing ML applications in structural engineering. The findings emphasize the transformative potential of ML to optimize design processes, enhance safety, and promote sustainable practices in civil engineering projects.

Light Weight Deflectometer Configurations for Testing Thin Asphalt Pavements

The light weight deflectometer (LWD) is used to determine fundamental soil properties such as deflection and modulus. Recently, the LWD has been investigated as a potential structural evaluation tool for thin flexible pavements to generate input data for structural performance prediction models of thin flexible pavements. The falling weight deflectometer (FWD) can be used for the evaluation of thin flexible pavements; however, the relatively high cost and logistics burden, which also may render FWD unsuitable for many local agencies, are disadvantageous for the rapid assessment of dispersed rural roadways. The LWD is a potential alternative; however, since various LWD models are available and multiple testing configurations are possible, this study focused on determining which LWD configuration(s) produce structural responses that most closely correspond to FWD. Ten different LWD combinations of base plate diameter, drop weight, and drop height were tested on two full-scale thin asphalt test sections under closely comparable temperature and moisture conditions. The LWD’s ability to indicate structural conditions was compared with that of the FWD through impulse stiffness modulus. The distribution of stress within the pavements was measured during testing. In addition, the LWD’s repeatability was assessed and found to be similar to that of the FWD. As a result, the small 5.91-in. plate diameter, 22-in. drop height, and largest available hammer weights for both LWD models obtained good repeatability and the closest correspondence to FWD when tested on thin flexible pavements.

Embedding Prior Knowledge into Data-Driven Structural Performance Prediction to Extrapolate from Training Domains

Embedding Prior Knowledge into Data-Driven Structural Performance Prediction to Extrapolate from Training Domains

Uncertainty and bias in generic ground motion sets used for PBEE

Probabilistic assessment of seismic structural reliability relies on the simulation of structural behavior using nonlinear dynamic analysis of numerical models. This typically involves analyzing samples of ground motions that are assumed to represent specific populations of signals recorded at the site of interest. While the literature extensively discusses uncertainties arising from material properties, modeling randomness, and ground motion record-to-record (RTR) variability, there are two factors that are often overlooked: (1) variability in ground motion sets and (2) the interaction between ground motion RTR variability and other sources of uncertainty. Ideally, the ground motion records selected for probabilistic analyses should reflect the seismic hazard characteristics at the site of interest. However, in practice, researchers and practitioners often rely on pre-collected generic ground motion sets (a.k.a. standardized sets of ground motions). This study aims to address this gap by quantifying the variability and bias in dispersion and median values within different generic ground motion sets. To achieve this, the incremental dynamic analysis is performed on a series of nonlinear single-degree-of-freedom systems. The results are summarized in terms of inter- and intra-set variability, and are contrasted with other techniques based on resampling plans belonging to the bootstrap family. The findings will contribute to the development of improved framework for selecting and incorporating ground motion records in probabilistic seismic assessments, enabling more reliable predictions of structural performance.

FRAMED: An AutoML Approach for Structural Performance Prediction of Bicycle Frames

This paper presents a data-driven analysis of the structural performance of 4500 community-designed bicycle frames. We introduce FRAMED – a parametric dataset of bicycle frames based on bicycles designed by bicycle practitioners from across the world. To support our data-driven approach, we also provide a dataset of structural performance values such as weight, displacements under load, and safety factors for all the bicycle frame designs. Our structural simulations are validated against results from physical experiments on real bicycle frames. By exploring a diverse design space of frame design parameters and a set of ten competing design objectives, we present a data-driven approach to analyze the structural performance of bicycle frames. Through our analysis, we highlight overall trends in bicycle frame designs created by community members and study several bicycle frames under different loading conditions. We then undertake a systematic search for optimal performance and feasibility-predictive Machine Learning models, applying a state-of-the-art Automated Machine Learning framework. We demonstrate that the proposed AutoML models outperform commonly used models such as Neural Networks and XGBoost, which we tune using Bayesian hyperparameter optimization. This work aims to simultaneously serve researchers focusing on bicycle design as well as researchers focusing on the development of data-driven design algorithms, such as surrogate models and Deep Generative Models. The dataset and code are provided at http://decode.mit.edu/projects/framed/ . • We introduce a dataset of 4500 bicycle frames inspired by bicycles designed by community members using the BikeCAD software. For each frame, we provide ten structural performance indicators evaluating the frame's performance under three load cases (in-plane, transverse, and eccentric loading). Indicators consist of seven deflections, two safety factors, and a weight value and are calculated through Finite Element Analysis. • We validate our Finite Element Analysis framework through a mesh convergence study and verify the accuracy of our simulation results against physical testing of bicycle frames. • We identify optimal surrogate models which predict the performance and feasibility of frames. Surrogates are selected using an AutoML framework which automates the selection of algorithms, architectures, hyperparameters, and instantiations for optimal model performance. AutoML models achieved a coefficient of determination of 0.605 in structural performance prediction and an F1 score of 0.915 in feasibility classification. • We validate our proposed AutoML framework against several common models (Neural Networks, XGBoost, etc.) which we optimize using Bayesian hyperparameter tuning. The proposed AutoML models attain the best coefficient of determination and mean absolute error among all methods tested in regression and the best F1 score, precision, recall, accuracy, and ROC AUC in classification.

Neutronics Assessment of the Spatial Distributions of the Nuclear Loads on the DEMO Divertor ITER-like Targets: Comparison between the WCLL and HCPB Blanket

The Plasma Facing Components (PFCs) of the divertor target contribute to the fundamental functions of heat removal and particle exhaust during fusion operation, being subjected to a very hostile and complex loading environment characterized by intense particles bombardment, high heat fluxes (HHF), varying stresses loads and a significant neutron irradiation. The development of a well-designed divertor target, which represents a crucial step in the realization of DEMO, needs the assessment of all these loads as accurately as possible, to provide pivotal data and indications for the design and structural performance prediction of the PFCs. In a particular way, this study is fully devoted to the comprehension of the distributions on the divertor target of the main nuclear loads due to neutron irradiation, performed for the first time using an extremely detailed approach. This work has been carried-out considering the latest configuration of the DEMO reactor, including the updated design of the divertor and ITER-Like PFCs geometry, varying the blanket layout (Water Cooled Lithium Lead—WCLL and Helium Cooled Pebble Bed—HCPB), thus evaluating the impact of the different blanket concept on the above-mentioned distributions. Neutronics analyses have been performed with MCNP5 Monte Carlo code and JEFF3.3 nuclear data libraries. 3D DEMO MCNP models have been created, focusing in particular on a thorough representation of the divertor and PFCs, allowing for the assessment of the distributions of the main nuclear loads: radiation damage (dpa/FPY), He-production rate (appm/FPY) and nuclear heating density (W/cm3) and total nuclear power deposition (MW). These results are presented by means of 2D maps and plots for each PFCs sub-component both for WCLL and HCPB blanket case: W-monoblocks, Cu-interlayers\CuCrZr-pipe and PFC-CB (Cassette Body) supports made of Eurofer steel.

Structural performance prediction based on the digital twin model: A battery bracket example

Battery bracket for new energy commercial vehicles is subjected to variable loads and battery temperature changes both during the design road test phase and in-service operation. Therefore, their structural performance must be evaluated in real-time for reliability design and health monitoring. With the rapid development of industrial digitization, the digital twin has become an indispensable technology. This paper proposes a digital twin approach for predictive monitoring of the performance of mechanical structures. Taking the structural performance for the battery bracket of new energy commercial vehicles as an example, this paper builds a unit-level digital twin model—DTMAR. It comprises the numerical model, NN-RSR model, and hybrid machine learning model. The results reveal that the DTMAR model can efficiently and accurately calculate and predict the structural performance. This can not only provide constructive guidance for optimal design of the next generation product structure, but also aid in evaluating the structural reliability of the battery bracket of new energy commercial vehicles and improve their driving safety.

Methodology for multidimensional approximation of current velocity fields around offshore aquaculture installations

Consistent growth of the marine aquaculture industry over the past decades calls for potential deployments of large-scale aquaculture structures to produce finfish, shellfish and macroalgae in varying inshore and offshore environments. Numerical simulations for engineering design applications become more challenging with increase of scale since current velocity fields are no longer uniform, complicating accurate hydrodynamic load calculations. Horizontal and vertical velocity profiles in this case are spatially (depth and particular location within the deployment site) and temporary (date and time) dependent. Thus, proper representation of the current velocity field in numerical models becomes crucial for accurate predictions of structural performance of aquaculture installations. In this paper, an advanced multidimensional approximation method based on discrete current velocity data is formulated. The approach implies presenting the continuous current velocity function as a superposition of weighted radial basis functions extended by a linear polynomial. To address overfitting issues, the thin plate regularization is applied in the method. The approximation is then constrained in order to fit the velocity values on the domain boundaries. The method is implemented in finite element software Hydro-FE and its performance is compared to other approximation methods on the example of a kelp grow line deployed at the Wood Island research site, Maine, USA. It was found that the difference between regular (mean or linearly interpolated) velocity profiles and the velocity profiles approximated with the radial basis function method can reach up to 34–38 % in terms of grow line mooring tensions, and 6–18% in terms of grow line displacement.

An elastoplastic damage constitutive model for hybrid steel-polypropylene fiber reinforced concrete

An appropriate constitutive model to characterize the mechanical behavior of fiber reinforced concrete (FRC) plays a critical role in accurate prediction of structural performance. In this study, an elastoplastic damage constitutive model is developed for hybrid steel-polypropylene fiber reinforced concrete (HFRC), which can predict the versatile mechanical responses of HFRC as a result of various hybrid fiber inclusions. For plasticity growth, the Willam-Warnke five-parameter failure envelope is modified in the effective stress space to govern the yield condition of HFRC, while a nonlinear Drucker-Prager type plastic potential function is proposed to control the non-associated plastic flow. Regarding the damage evolution, the damage criterion and evolution laws are established in the Cauchy stress space, in which the plastic hardening and unilateral behaviors are taken into consideration. Upon two parallel integration algorithms to describe the development of plasticity and damage, as well as the identification of fiber-dependent parameters, the proposed model is numerically implemented and then validated by the experimental results from available research. The comparisons between the numerical predictions and the test results at both material scale and structural scale solidly demonstrate the capacity of the model in capturing the main features of HFRC when subjected to different loading paths.

Rapid prediction of the ultimate moment of flush endplate connections at elevated temperatures through an artificial neural network

Steel connections are vulnerable structural components, and their failure can lead to the collapse of the whole structure in a fire. Hence, it is crucial to predict the connection behaviors accurately to enable a precise prediction of structural performance in the fire. This paper proposes an efficient approach for predicting the ultimate moment of flush endplate connections at elevated temperatures using the finite element method and an Artificial Neural Network (ANN). Firstly, a finite element model considering both geometric and material nonlinearities is developed and verified with existing experimental results conducted by others. Based on the validated finite element model, an ABAQUS plug-in, namely COinFIRE, is implemented in ABAQUS to facilitate the parametric study. Consequently, 886 finite models are generated and used as training data for developing the ANN model. Accordingly, the endplate thickness, the pitch of bolts, the bolt row distance, the outer edge distance, the gage distance, the bolt diameter, the number of bolt rows, and the material properties at elevated temperatures are considered input variables. Meanwhile, the output is the ultimate moment. The obtained results show that the proposed ANN model can accurately predict the ultimate moment of flush endplate connections (R2 = 1.0, 1.0, 1.0, RMSE = 0.811 kN.m, 1.627 kN.m, 1.256 kN.m, and a20-index = 1.0, 1.0, 1.0 for the training, testing and validation datasets, respectively). Finally, closed-form formulas and a graphical user interface (GUI) tool are developed to evaluate the ultimate moment of flush endplate connections at elevated temperatures for practical use. The proposed approach shows great promise as a powerful design tool for steel connections at elevated temperatures.

Real-time multiscale prediction of structural performance in material extrusion additive manufacturing

The material extrusion additive manufacturing (AM) has been extensively used in fabricating structures with complex geometries. However, geometric defects often exist in an AM structure, which could compromise its final performance. In this paper, a real-time multiscale performance evaluation method is developed for material extrusion-based honeycomb structures. The representative cell boundary is extracted from three-dimensional (3D) point clouds obtained via an in-situ monitoring approach. The cell boundary is then used to generate the digital twin of the unit cell of the printed layer based on the finite element (FE) method. A physics-based multiscale modeling approach called mechanics of structure genome (MSG) is then employed to predict the effective material properties of the printed layer and plate stiffness matrix of the final structure. The proposed approach provides a highly efficient way to predict the real-time performance of the as-manufactured products. Moreover, the numerical example shows that the geometric defects could result in complex mechanical behaviors in the defected parts, which cannot be captured by the conventional approaches based on the shape deviations. The numerical results are validated by the three-point bending tests. The proposed method can be used in the closed-loop control of material extrusion-based manufacturing systems.

Framed: An Automl Approach for Structural Performance Prediction of Bicycle Frames

Framed: An Automl Approach for Structural Performance Prediction of Bicycle Frames

Machine learning in predicting mechanical behavior of additively manufactured parts

Although applications of additive manufacturing (AM) have been significantly increased in recent years, its broad application in several industries is still under progress. AM also known as three-dimensional (3D) printing is layer by layer manufacturing process which can be used for fabrication of geometrically complex customized functional end-use products. Since AM processing parameters have significant effects on the performance of the printed parts, it is necessary to tune these parameters which is a difficult task. Today, different artificial intelligence techniques have been utilized to optimize AM parameters and predict mechanical behavior of 3D-printed components. In the present study, applications of machine learning (ML) in prediction of structural performance and fracture of additively manufactured components has been presented. This study first outlines an overview of ML and then summarizes its applications in AM. The main part of this review, focuses on applications of ML in prediction of mechanical behavior and fracture of 3D-printed parts. To this aim, previous research works which investigated application of ML in characterization of polymeric and metallic 3D-printed parts have been reviewed and discussed. Moreover, the review and analysis indicate limitations, challenges, and perspectives for industrial applications of ML in the field of AM. Considering advantages of ML increase in applications of ML in optimization of 3D printing parameters, prediction of mechanical performance, and evaluation of 3D-printed products is expected.

Collapse Simulation of CFRP Retrofitted Column using Hybrid Testing Technique

The performance of Reinforced Concrete (RC) structures and their collapse safety experiencing moderate to high-magnitude seismic excitations might not be fully understood. Factors related to the high levels of uncertainties in ground motion content, structural modelling and design uncertainties, or to the empirical nature of design codes and standards affect the collapse assessment of RC structures. Therefore, experimental test researches are very important in verifying and improving the accuracy of structural performance predictions and highlighting the actual behavior during seismic excitation. In this paper, the test results of a full-length damaged RC column retrofitted with Carbon Fiber Reinforced Polymer (CFRP) sheets and then retested using pseudo-dynamic experimental testing approach through the state-of-the-art hybrid simulation testing facility, referred to Multi-Axis Substructure Testing (MAST) system at Swinburne University of Technology are presented. A comparative collapse assessment of the initial and retrofitted column showed more ductile column response reaching higher drift ratios. In addition, the experimental test results displayed clearly the effectiveness of CFRP sheets under increasing intensity of earthquake demands in restoring the collapse resistance of the retrofitted RC column by altering concrete failure type.

Mathematical Formulation for Prediction of Structural Performance of Green Geopolymer Concrete Beams and Columns

In order to minimise the environmental impact of the waste and its disposal processes and to reduce global warming by cement production, this study has focused on the demand. This research looked at how waste wood ash can be used to make geopolymer concrete beams and columns, which can be used to replace traditional reinforced concrete elements in the construction industry. Waste wood ash, a waste produced in the nearby hotel and manufacturing factories by burning the waste wood collected in the woodworking industry and throwing the ash to land causing considerable environmental pollution. Geopolymer is an innovative inorganically friendly, alkaline-based binding agent that stimulates the source material of aluminosilicate (such as metakaolin, fly ash and GGBS). For three types of concretes (30 percent WWA – 70 percent Fly ash Geo-polymer concrete, Fly ash Geo-polymer concrete, and Reinforced Cement Concrete), the mathematical formula for the behaviour of beams in deflection, ductility factor, flexural strength, and columns in load carrying ability, stress strain behaviour, and load-deflection behaviours were investigated in this study. The results showed that the inclusion of waste wood ash into geopolymer concrete contributed to the 42 and 28 percent increase of the capacity for beam and column. Furthermore, by replacing waste wood ash, the compotation of structural elements was improved in their rigidity and ductility. The derived mathematical equations were most suitable for the prediction of forecast.

Behavior Deterioration and Microstructure Change of Polyvinyl Alcohol Fiber-Reinforced Cementitious Composite (PVA-ECC) after Exposure to Elevated Temperatures.

In the case of fire, explosive spalling often occurs in cementitious composites due to dense microstructure and high pore-pressure. Polymer fibers were proved to be effective in mitigating such behavior. However, deterioration of these fiber-reinforced cementitious composites inevitably occurs, which is vital for the prediction of structural performance and prevention of catastrophic disaster. This paper concentrates on the behavior and mechanism of the deterioration of polyvinyl alcohol fiber-reinforced engineered cementitious composite (PVA-ECC) after exposure to elevated temperatures. Surface change, cracking, and spalling behavior of the cubic specimens were observed at room temperature, and after exposure to 200 °C, 400 °C, 600 °C, 800 °C, and 1200 °C. Losses in specimen weight and compressive strength were evaluated. Test results indicated that explosive spalling behavior was effectively prevented with 2.0 vol% polyvinyl alcohol fiber although the strength monotonically decreased with heating temperature. X-ray diffraction curves showed that the calcium hydroxide initially decomposed in the range of 400–600 °C, and finished beyond 600 °C, while calcium silicate hydrate began at around 400 °C and completely decomposed at approximately 800 °C. Micrographs implied a reduction in fiber diameter at 200 °C, exhibiting apparent needle-like channels beyond 400 °C. When the temperature was increased to 600 °C and above, the dents were gradually filled with newly produced substance due to the synergistic effect of thermal expansion, volume expansion of chemical reactions, and pore structure coarsening

Prediction of Structural Performance of Vinyl Ester Polymer Concrete Using FEM Elasto-Plastic Model.

This paper presents the methodology for predicting the mechanical performance of structural elements made of polymer concrete (PC). A vinyl ester polymer concrete composition and the results of experimental studies to determine the basic mechanical properties of the material are presented. Following the strategy for sustainable development in the building industry, the material cost of polymer concrete was lowered by reducing the consumption of raw materials and the partial replacing of the microfiller fraction with recycled waste products—calcium fly ash. An accurate computational model enabling stress analysis is a convenient way to verify the suitability of PC as a construction material in structural applications. Due to difficulty in deriving an accurate analytical formula, numerical approximation (finite element method) was used as a method for solving the problem. Constitutive modeling of PC is a very important aspect of the strength calculations and here it was done within the framework of elasto-plasticity. Numerical evaluation of the static bearing capacity of PC manhole covers is shown as an example of the proposed FEM methodology. The results of computer simulations were compared with laboratory tests. Finally, the adequacy of the numerical modeling for testing new construction and material improvements is discussed. The study showed that the concrete damaged plasticity material model can be effectively used for the description of PC mechanical behavior.

Reliability analysis of a reinforced concrete bridge under moving loads

This study presents a reliability analysis procedure for a reinforced concrete bridge exposed to different moving loads. Bridges are one of the important part of transportation infrastructure systems. As bridges age, structural weakening due to heavy traffic and aggressive environmental factors lead to an increase in repair frequency and decrease in load carrying capacity. Therefore, bridges require periodic maintenance and repair in order to function and be reliable throughout their lifetimes. In other words, condition and safety of the bridges must be monitored at regular time intervals to avoid the disadvantages of deterioration. Otherwise, sudden collapse of a bridge may lead to irreversible loss of life and property. Therefore, the importance of the structural assessment of bridges is rapidly increasing in developed countries. In this study, reliability analysis which is one of the structural performance prediction method is applied to a reinforced concrete bridge subjected to the different moving loads. The aim of this study is to observe the safety of the bridge for the effect of the increasing traffic factor over the years.

Computational platform for probabilistic optimum monitoring planning for effective and efficient service life management

Over the past decades, significant advances have been accomplished in developing SHM techniques to detect the existing damages in deteriorating structures and maintenance techniques to extend the service life of these structures. The application of SHM can lead to more accurate damage detection. By using the information obtained from SHM, the uncertainties associated with structural performance assessment and prediction can be reduced. If the advanced SHM techniques are optimally integrated in life-cycle management, the efficiency and effectiveness of service life management of deteriorating structures can be maximized. In this paper, a computational platform for optimum monitoring planning based on multi-objective optimization (MOPT) and decision making is presented. The main components integrated in this computational platform are (a) formulation of objectives for optimum monitoring planning; (b) MOPT and decision making for application of the best monitoring plan; and (c) updating the damage propagation and structural performance prediction. The objectives for optimum monitoring planning are formulated considering the availability of monitoring data, damage detection, maintenance, service life and life-cycle cost. Through the MOPT and decision making, the best monitoring plan is determined. The updating process integrates the information obtained from monitoring to improve the accuracy and reduce the uncertainty associated with the damage occurrence and propagation prediction and monitoring planning.

Life-cycle reliability-based design and reliability updating of reinforced concrete shield tunnels in coastal regions

In this paper, a novel approach for life-cycle reliability-based design and reliability updating of reinforced concrete (RC) shield tunnels in coastal regions has been developed. For new shield tunnels, a reliability-based durability design criterion to determine the concrete cover and concrete quality for preventing the deterioration of structural performance during the structural service life is proposed. For existing shield tunnels, an approach to update structural reliability is established based on Sequential Monte Carlo Simulation. Epistemic uncertainty associated with the structural performance prediction of existing shield tunnels can be reduced compared with that of new shield tunnels. In illustrative examples, durability design factors and maximum water to cement ratio for designing RC segment satisfying the target durability reliability are investigated. For existing RC shield tunnels, the effect of the observational deformation of segmental lining on the time-dependent failure probability is discussed.