High-fidelity Finite Element Model Research Articles

Machine learning techniques for predicting the peak response of reinforced concrete beam subjected to impact loading

To meet the growing need for resilient structures in seismic and high-impact zones, accurate prediction of the response of reinforced concrete (RC) beams under impact loads is essential. Traditional methods, such as experimental testing and high fidelity finite element models, are often time consuming and resource intensive. To address these challenges, this study investigates various ensemble and non-ensemble machine learning techniques—including support vector machine, gaussian process regression (GPR), k-nearest neighbor (KNN), gene expression programming, random forest, decision tree, boosted tree, adaptive boosting tree, gradient boosting algorithm, stochastic gradient descent, and artificial neural network—for predicting the peak response of RC beams under impact loads. A set of 145 experimental data points from 12 different sources is used to train and evaluate these machine learning models. Key parameters in the data include beam width and depth, span, reinforcement ratios, concrete strength, steel yield strength, deflection, and impact characteristics. Except for KNN, all models showed satisfactory generalization capabilities with R2 values over 0.8. Statistical errors such as RMSE, a-10 index, MAE, and a-20 index are within acceptable limits. The GPR model is the most effective with R2 value of 0.95. Moreover, Shapely analysis identified beam depth, impact velocity, and beam breadth as critical factors. Overall, this study demonstrates the efficacy of machine learning in accurately predicting the behavior of RC structures under impact loads, providing valuable tools for civil engineers in design and analysis.

Low-velocity impact (LVI) and compression after impact (CAI) of Double-Double composite laminates

Tailorability is a key advantage of fiber-reinforced composites over other material systems. While tailoring a single isolated laminate is relatively simple, challenges arise when designing larger integrated components while ensuring compatibility between laminates and avoiding sharp changes in local stiffness. The innovative Double-Double (DD) laminate design method simplifies the optimization and processing of laminates by incorporating 4-ply building blocks consisting of +ϕ, −ϕ, +ψ, and −ψ ply orientations. As a relatively new concept, DD laminate design requires careful assessment to ensure its performance is equivalent to that of conventional designs. The current study compares impact damage tolerance of quadriaxial (QUAD) laminates consisting of 0°, 90°, and ±45° ply orientations with equivalent DD laminates under Low-Velocity Impact (LVI) and Compression After Impact (CAI) loadings. To this end, a validated three-dimensional high-fidelity finite element model capable of capturing fiber breakage, splitting, kinking, as well as matrix cracking and delamination, was used. A computer tool was developed to identify equivalent DD laminates and to find the best stacking sequence for achieving layup homogenization. Three equivalent DD laminates were selected for the [0/45/90/−45]4s. The first laminate had an equal in-plane stiffness [A] matrix ([67.5/–22.5/22.5/−67.5]8T), the second laminate had an equal flexural stiffness [D] matrix ([64.5/−17/17/−64.5]8T), and the third laminate ([65.5/−18.5/18.5/−65.5]8T) had a similar [D] matrix while keeping the difference between each element of [A] matrices below 10 %. The results indicate that the QUAD laminates can be replaced by equivalent DD without compromising impact damage tolerance while benefiting from the improved design and manufacturing ease of the DD laminate configuration.

A parametric approach for creating finite element models from point clouds for steel superstructure components in steel girder bridges

Finite element (FE) analysis is widely used to evaluate the conditions of existing bridges. However, the creation of high-fidelity FE models (FEM) for existing bridges is often a challenging task. In view of the ability of 3D laser scanning to capture detailed geometrical information, this study proposes a direct scan-to-FEM strategy specifically for steel bridge superstructure and hence to facilitate the creation of high-fidelity FE models for existing steel girder bridges. Parametric representations are proposed to address partial occlusions in the point clouds as well as to reinforce the connectivity and alignment between the superstructure elements. Mesh generation algorithms are developed to automatically convert the parametric representations into all-hexahedron FE mesh assembly that can be used directly for FE analysis without any mesh verification or editing. Experimental validations are carried out to evaluate the effectiveness of proposed method in term of both mesh quality and analysis accuracy.

Progressive bearing failure analysis and strength prediction method for the initial assembly and tensile process of composite bolted joints

Composite bolted joints have extensive application in aerospace and marine equipment. Failure analysis and strength prediction are crucial steps in the design of composite bolted joints. This study proposes a method that can accurately analyze the progressive bearing failure and static strength of composite bolted joints under different initial assembly parameters. This method first obtains an accurate initial assembly state based on the high-fidelity finite element model with detailed thread structure. Then it uses this as an input to analyze the progressive bearing failure process and the bearing strength value. Using the above method, the formation and evolution mechanism of the tightening yielding phenomenon during the assembly process are analyzed. And the influence of the two key assembly parameters, washer type and initial bolt tension, on the bearing strength of composite bolted joints is analyzed. The experimental results verify the effectiveness and accuracy of the simulation methods and results.

Multi-objective optimization design for anti-fatigue lightweight of dump truck carriage combined with machine learning

As urbanization continues to accelerate, dump trucks assume an increasingly important role in the transportation and construction of infrastructure. The carriage represents a critical structural assembly of dump trucks. One of the primary failure modes of the carriage is weld fatigue failure, which frequently gives rise to the problem of weld fatigue cracking during transportation. To increase the fatigue life of welds and enhance the degree of structural lightweight of a heavy dump truck carriage, a method for anti-fatigue lightweight design based on machine learning and multi-objective optimization is proposed. A high-fidelity finite element model of the carriage is established for static simulation analysis of the typical conditions. Based on the virtual reliability simulation test of the dump truck and the equivalent structural stress method, the fatigue life of the critical welds in the carriage is calculated. The important part thicknesses are selected as design variables through the comprehensive contribution analysis method. The maximum displacement and maximum stress under the dangerous condition are considered as constraints. The mass of the carriage and the minimum fatigue life of the critical welds are considered as optimization objectives. The GA-XGBoost machine learning approximation models (GA-XGBoost-MLAM) and NSGA-II algorithm are employed for multi-objective optimization design of the carriage. The entropy weighted TOPSIS method is utilized for multi-objective decision-making of Pareto solutions. The design after optimization and decision-making shows that, while satisfying the requirements of static structural performance, the minimum fatigue life mileage of the critical welds of the carriage is increased by 157,570 km, representing an increase of 36.58%. Additionally, the mass of the carriage is reduced by 295.69 kg, representing a decrease of 9.47%. Therefore, the proposed design method achieves a good effect in the anti-fatigue lightweight of dump truck carriage.

Cyclic Behavior of Concrete-Filled Tube Columns with Bidirectional Moment Connections Considering the Local Slenderness Effect

In this research, the cyclic behavior of concrete-filled thin tube (CFTT) columns with bidirectional moment connections was numerically studied within the context of thin-walled structures. Novel considerations in the design of CFTT columns with slenderness sections are proposed through a parametric study. A total of 70 high-fidelity finite element (FE) models are developed using ANSYS software v2022 calibrated from experimental research using similar 3D joint configurations. Furthermore, a comparison of different width-to-thickness ratios in columns was considered. The results showed that the models with a high slenderness ratio reached a stable cyclic behavior until 0.03 rad of drift, and a flexural strength of 0.8 Mp was reached for 4% of the drift ratio according to the Seismic Provisions. However, this effect slightly decreased the strength and the dissipated energy of the moment connection in comparison to columns with a high ductility ratio. Moreover, an evaluation of concrete damages shows concrete cracked for cyclic loads higher than 3% of drift. Finally, the joint configurations studied can achieve a good performance, avoiding brittle failure mechanisms and ensuring the plastic hinges in the beams.

Computational engine for finite element digital twins of structural dynamics via motion data

Typical structural health monitoring systems employ limited numbers of sensors capable of measuring discrete local behaviours. However, practical challenges arise as these sensor arrays cannot cover all local areas of interest. To address this challenge, this article introduces a novel method for twinning structural dynamic behaviour by constructing a finite-element-model-based digital twin, enabling the observation of non-sensor positions crucial for downstream tasks. The approach utilises streaming monitoring data, e.g., displacement and acceleration, as external dynamic loads to reproduce the dynamic response of the entire physical structure. Subsequently, the dynamic behaviour of specific non-sensor locations can be extracted from the digital twin. The method is formulated as a local-global-local procedure. To validate the proposed approach, two virtual experiments were conducted on: 1) a simply supported Euler-Bernoulli beam subjected to static loads and 2) a high-fidelity finite element model of a composite bridge carrying dynamic traffic loads. The results demonstrate remarkable accuracy in reproducing both global and local behaviours, facilitating observations at non-sensor positions for downstream estimations.

Spatiotemporally nonlocal homogenization method for viscoelastic porous metamaterial structures

The underlying size-dependent mechanism of viscoelastic porous metamaterial structures is still unclear, and the solution to their response using the high-fidelity numerical method is time-consuming, even prohibitive. This paper explores how the size-dependent effect of metamaterial structures emerges from the underlying nonlocal interactions in metamaterial microstructures, and then develops an accurate but efficient homogenization method for analyzing their size-dependent and time-dependent mechanical behaviors of viscoelastic porous metamaterial structures. By accounting for the nonlocal interaction between different representative volume elements and the history-dependent effect, a spatiotemporally nonlocal discrete element model is developed to analyze the mechanical behavior of viscoelastic porous metamaterial structures. With the help of the spatiotemporally nonlocal discrete element model, a spatiotemporally nonlocal homogenization method is proposed to accurately yet efficiently assess the size-dependent mechanical behavior of the porous metamaterial structure. The remarkable potential of the spatiotemporally nonlocal homogenization method is illustrated by simulating the stress relaxation of a viscoelastic porous metamaterial structure under axial strain. Results indicate that the size-dependent phenomenon can be significant, and the spatiotemporally nonlocal homogenization model can efficiently capture the size-dependent relaxation stress evaluated by the high-fidelity finite element model without loss of accuracy.

Hygrothermal effects on ballistic behavior of toughened CFRP laminates

High-speed impact resistance is a critical consideration in the design and assessment of composites for aerospace applications during their service life, especially in the context of hygrothermal aging. This study, combining experiment and numerical simulation, investigates hygrothermal effects on the high-speed impact performance of a toughened T700/EH301 carbon/epoxy composite material. Three distinct hygrothermal conditions, namely Unaged, Aged, and Redried, are investigated. Experimental results reveal an apparent improvement in ballistic limit after hygrothermal aging (170.2 m/s, a notable increase of 11.2%), still maintaining an excellent performance even after redrying (158.5 m/s, an increase of 3.2%). Inspired by the micrographic analysis, a high-fidelity finite element model, referred to as HFEDCM (High Fidelity Explicit Discrete Crack Model), has been developed. This model not only explains the positive effects of hygrothermal aging on high-speed impact performance but also accurately captures the macroscopic damage morphology of composite laminates. It has been quantitatively and qualitatively demonstrated that the enhancement of the fracture toughness of the “fiber bundle” plays a critical role in improving the high-speed impact resistance of composites under hygrothermal aging.

Physics-informed surrogate modeling for a damaged rotating shaft

In the online structural health monitoring framework, surrogate modeling aims to approximate the predictions of the underlying high-fidelity model to facilitate fast simulations. Generally, conventional machine learning techniques based on purely data-driven approaches often behave as black boxes, producing predictions that may lack physical consistency. In an effort to overcome these limitations, we proposed a neural network mapping constrained by the known physical relationships of an engineering model, which can be considered as a physics-informed inductive bias. This approach aims to enhance computational efficiency while reducing the required number of training examples. The physics-informed term guides the training process to yield predictions that exhibit greater physical consistency while maintaining high performance. We here propose a surrogate modeling approach relying on a Physics-Informed Neural Network (PINN) scheme, which is designed to enable robust and real-time monitoring of ballistic impact damage on a rotor shaft. At first, a high fidelity finite element model is developed to simulate the response of the rotor shaft under various operational conditions. As a first reduction step, substructuring is used to decrease the dimensionality of the system's matrices, thereby reducing the computational cost. For comparison and validation, an alternative model based on the Timoshenko beam theory is utilized. Both models examine the effects of impact-induced vibration loads due to ballistic occurrences. The PINN model receives the external load as input and predicts the system's response. The system matrices extracted from both models constitute the physics-informed term. This term is incorporated into the PINN architecture in order to steer the learning process towards more physically consistent and reliable predictions regarding the state and the response of the structure. The algorithm's performance is evaluated using response data derived from both models.

Surface temperature field real-time reconstruction of hot forging die based on 1DCNN

Accurate reconstruction of the die surface temperature field during the hot forging process is crucial for ensuring optimal control of die temperature and high-quality metal forgings. In this study, we propose a novel method that combines hot forging process parameters with a deep neural network to reconstruct the temperature field on the die surface in real-time. Firstly, a high-fidelity finite element model of the hot forging process is created and validated with a forging experiment. Secondly, temperature observation points are initially selected based on the simulated temperature distribution of the forging die, and the correlation coefficient method is used to optimize the layout of observation points to improve reconstruction accuracy and efficiency. Subsequently, multiple high-quality simulations with varying process parameters are conducted, generating a comprehensive dataset. The 1DCNN algorithm is then presented to produce a mapping from the input to the surface temperature field. Finally, the effectiveness of the surface temperature field reconstruction method is verified. The results indicate that the overall performance of the proposed 1DCNN algorithm is better than the traditional regression algorithms. In terms of reconstruction accuracy, it achieves a prediction accuracy within an absolute error of 3 °C for the outer surface points and within a relative error of 5 % for the cavity region. Regarding reconstruction efficiency, it requires an approximate execution time of 0.064 s for predicting a new sample. This paper presents a method for real-time reconstruction of the die surface temperature field, leveraging hot forging process parameters and temperature data from observation points. This approach enables reliable online monitoring of transient temperature variations on the hot forging die surface, offering valuable support for operational control and maintenance.

Comparative analysis of rolling contact fatigue life in a wind turbine pitch bearing with different modeling approaches

The load calculation for wind turbines usually does not take pitch bearings into account, and the pitch-bearing internal load distribution is calculated in a decoupled process. In the current study, a coupled wind turbine load simulation considering the pitch bearings is proposed. Internal pitch-bearing load distribution and roller contact fatigue life are compared to a decoupled approach using a high-fidelity finite-element model. Internal pitch-bearing load distribution can be an essential variable in almost all major pitch-bearing failures. Furthermore, rolling contact fatigue is one of the major types of pitch-bearing failure that can represent the bearing service life. The results show that the flexibility of the surroundings has an effect on the contact forces. As a result, the finite element model-based life calculation of the NREL 5 MW turbine predicts a longer rolling contact fatigue life for the bearing than the multi-body system-based one. For the IWT7.5, sample load cases indicate that this behavior is different, leading to the conclusion that the life of the bearing is significantly affected by the stiffness of the bearing and its surrounding components.

Combined effect of random porosity and surface defect on fatigue lifetime of additively manufactured micro-sized Ti6Al4V components: An investigation based on numerical analysis and machine learning approach

Surface defects and internal porosities resulting from the additive manufacturing process contribute to a scatter in fatigue lifetime, increasing uncertainty in applications such as aerospace engineering. This study proposes a combined approach of machine learning and finite element modelling to explore the interaction between random porosity distribution and high surface roughness on the fatigue lifetime of micro-sized additively manufactured parts. To achieve this, a high-fidelity finite element model is constructed with randomly distributed porosities and surface defects. A fatigue lifetime estimation model, developed using a combination of continuum damage mechanics and the theory of critical distance, is then applied to predict lifetime using effective stress extracted from simulations. A total of 1200 simulations are conducted using these models to examine the interaction effects of subsurface porosity and surface defects on localized stress distribution and fatigue lifetime. Subsequently, two machine learning algorithms, Support Vector Regression and Kernel Ridge Regression, are employed to estimate fatigue lifetime, accounting for the effects of randomly distributed subsurface porosities and surface defects. Eighty percent of the developed finite element models are employed to construct an input dataset, incorporating geometrical features of pores, and to train the algorithms, while the remaining 20% are reserved for verification purposes. The observed results demonstrate that the combined effect of random porosity and surface defects increases fatigue scatter behaviour, leading to greater uncertainty in lifetime prediction. Furthermore, the machine learning algorithms exhibit promising prediction performance.

Effect of three-dimensional space on progressive collapse resistance of reinforced concrete frames under various column removal scenarios

In this study, high fidelity finite element (FE) model was adopted to investigate the progressive collapse resistance of reinforced concrete (RC) frame structures under various column removal scenarios. The validity of the numerical model was verified by comparing with the existing experimental results. Afterwards, to facilitate the research on possible simplified boundary conditions served for further various column removal scenarios, the effect of multi-span continuous beams was firstly evaluated and the simplified methods were compared. Then, the effect of various column removal scenarios addressing the three-dimensional effect on the progressive collapse resistance of RC frame structures with and without slab was systematically evaluated. The analysis results showed that the continuous span beams increased the internal compressive force of the beam section at compressive arch action (CAA) stage, which effectively improved the CAA strength of the substructure. For simple modeling, the use of horizontal constraints at the end of the beam could replace the continuous span beams. The presence of transverse beams and slab provided additional load transfer paths for collapse resistance. The compressive membrane action (CMA) and tensile membrane action (TMA) of the slab significantly improved the progressive collapse resistance of the substructure in the small and large deformation stages, respectively. The location of the removed column significantly affected the contribution ratio of transverse beams and slab to the progressive collapse resistance of the substructure, which was quantitatively evaluated for practical design.

Compression after impact (CAI) failure mechanisms and damage evolution in large composite laminates: High-fidelity simulation and experimental study

This study focuses on developing and validating a high-fidelity finite element model for predicting damage evolution and residual strength in fiber-reinforced composite panels. Impact and compression after impact (CAI) tests were conducted at both barely visible impact damage (BVID) and clearly visible impact damage (CVID) levels. The ASTM D7137 standard 100 mm × 150 mm CAI coupons were inadequate to cover the range of experimental studies required for model validation. Therefore, larger 254 mm × 304.8 mm laminates were investigated under two CAI testing conditions: one a scaled-up version of ASTM standard coupon, and the other with additional anti-buckling support plates to reduce unsupported areas to 127 mm × 177.8 mm. The model captured inter- and intra-laminar failure modes, including fiber breakage, splitting, kinking, pull-out, and crushing as well as matrix cracking, delamination, and their interactions. This was achieved by cohesive zone modeling technique and enhancement of the LaRC05 failure criteria through modeling the fiber damage evolution and utilizing an efficient search algorithm to determine the matrix fracture plane and fiber kink band angle. This study underscores the efficacy of the high-fidelity modeling approach in accurately predicting both impact damage and CAI strength in typical aircraft impact damage scenarios. Additionally, it provides insights into complex CAI failure mechanisms and energy release associated with various damage modes and highlights the effect of global buckling on the failure behavior and compressive strength of composite laminates. Furthermore, it shows that the proposed fixture with support plates is suitable for testing a broader range of impact scenarios without experiencing global buckling.

Risk assessment and prevention for typical railway bridge pier under rockfall impact

Railway bridges in mountainous regions are susceptible to damage and collapse due to rockfall impacts, posing a risk to train operations. This research specifically focuses on risk assessment of a typical railway bridge pier considering correlations between the collision parameters. Firstly, a comprehensive analysis of rockfall trajectories is conducted using the tool ROCFALL. Statistical features and correlation of impact parameters at the pier location are obtained and further characterized based on functional statistical models. Additionally, a high-fidelity finite element (FE) model of a two-span simply supported box girder with three piers is developed using LS-DYNA. Multiple structural dynamic responses including impact force, pier displacement, and internal energy of the directly impacted pier, are evaluated for both general rockfalls and potential boulders. The study examines the effects of individual impact parameters on the internal energy and displacement responses at the pier top, highlighting their significance in relation to load-bearing capacity and operational safety. Machine learning techniques are utilized to perform a comprehensive evaluation of risk states, wherein an Artificial Neural Network (ANN) model is employed for accurate prediction of risk indexes, and the K-means algorithm is applied to classify pier risk levels into minor, moderate, and severe. Furthermore, Monte Carlo (MC) simulations are employed to calculate the probabilities associated with different risk levels. A composite multilayer prevention measure is suggested to enhance the performance of a rockfall-prone bridge. It is proven that this measure can rapidly reduce associated vibrations and effectively mitigate structural damage resulting from the collision.

A Bayesian Multi-fidelity Neural Network to Predict Nonlinear Frequency Backbone Curves

Abstract The use of structural mechanics models during the design process often leads to the development of models of varying fidelity. Often low-fidelity models are efficient to simulate but lack accuracy, while the high-fidelity counterparts are accurate with less efficiency. This paper presents a multi-fidelity surrogate modeling approach that combines the accuracy of a high-fidelity finite element model with the efficiency of a low-fidelity model to train an even faster surrogate model that parameterizes the design space of interest. The objective of these models is to predict the nonlinear frequency backbone curves of the Tribomechadynamics Research Challenge benchmark structure which exhibits simultaneous nonlinearities from frictional contact and geometric nonlinearity. The surrogate model consists of an ensemble of neural networks that learn the mapping between low and high-fidelity data through nonlinear transformations. Bayesian neural networks are used to assess the surrogate model's uncertainty. Once trained, the multi-fidelity neural network is used to perform sensitivity analysis to assess the influence of the design parameters on the predicted backbone curves. Additionally, Bayesian calibration is performed to update the input parameter distributions to correlate the model parameters to the collection of experimentally measured backbone curves.

Error approximation and bias correction in dynamic problems using a recurrent neural network/finite element hybrid model

This work proposes a hybrid modeling framework based on recurrent neural networks (RNNs) and the finite element (FE) method to approximate model discrepancies in time-dependent, multi-fidelity problems, and use the trained hybrid models to perform bias correction of the low-fidelity models. The hybrid model uses FE basis functions as a spatial basis and RNNs for the approximation of the time dependencies of the FE basis' degrees of freedom. The training data sets consist of sparse, non-uniformly sampled snapshots of the discrepancy function, pre-computed from trajectory data of low- and high-fidelity dynamic FE models. To account for data sparsity and prevent overfitting, data upsampling and local weighting factors are employed, to instigate a trade-off between physically conforming model behavior and neural network regression. The proposed hybrid modeling methodology is showcased in three highly non-trivial engineering test-cases, all featuring transient FE models, namely, heat diffusion out of a heat sink, eddy-currents in a quadrupole magnet, and sound wave propagation in a cavity. The results show that the proposed hybrid model is capable of approximating model discrepancies to a high degree of accuracy and accordingly correct low-fidelity models.

Experimental tests and numerical simulations of miniature buckling-restrained braces using iron-shape memory alloy bar

Miniature buckling-restrained braces (BRBs) have a concise configuration and clear working mechanism. Owing to stable and full hysteresis, miniature BRBs are emerging as favorable energy-dissipating fuse elements in seismic applications. This paper proposes using iron-shape memory alloy (FeSMA) in lieu of conventional steel as the yielding core and focuses on the cyclic behavior. Four specimens, corresponding to four loading protocols, were fabricated and tested. According to experimental data, the miniature FeSMA BRBs exhibited full hysteresis, which is characterized by large damping, noticeable isotropic and kinematic hardening behavior and appreciable post-yield stiffness ratio. The cumulative plastic deformations meet the requirement by ANSI/AISC 341-16. Further, high fidelity finite element models, which explicitly considered damage and fracture rules, were established for numerical simulations. Good agreement can be found between the experimental data and numerical simulations, in terms of failure modes, fracture locations and global hysteresis.

Improving accuracy in parametric reduced-order models for classical guitars through data-driven discrepancy modeling

Recently developed high-fidelity finite element (FE) models represent a state-of-the-art approach for gaining a deeper understanding of the vibrational behavior of musical instruments. They can also be used as virtual prototypes. However, certain analyses, such as optimization or parameter identification, necessitate numerous model evaluations, resulting in long computation times when utilizing the FE model. Projection-based parametric model order reduction (PMOR) proves to be a powerful tool for enhancing the computational efficiency of FE models while retaining parameter dependencies. Despite their advantages, projection-based methods often require complete system matrices, which may have limited accessibility. Consequently, a systematic discrepancy is introduced in the reduced-order model compared to the original model. This contribution introduces a discrepancy modeling method designed to approximate the parameter-dependent effect of a radiating boundary condition in an FE model of a classical guitar that cannot be exported from the commercial FE software Abaqus. To achieve this, a projection-based reduced-order model is augmented by a data-driven model that captures the error in the approximation of eigenfrequencies and eigenmodes. Artificial neural networks account for the data-driven discrepancy models. This methodology offers significant computational savings and improved accuracy, making it highly suitable for far-reaching parametric studies and iterative processes. The combination of PMOR and neural networks demonstrate greater accuracy than using either approach alone. This paper extends our prior research presented in the proceedings of Forum Acusticum 2023, offering a more comprehensive examination and additional insights.