Mechanics-based Model Research Articles

Physics-guided machine learning frameworks for fatigue life prediction of AM materials

Introducing random defects is a type of the dominant causes of fatigue scatter of additive manufacturing (AM) materials. The fracture mechanics-based models oversimplify the fatigue damage process and ignore the key influencing factors. Given data-driven nature of machine learning (ML), sufficient samples are required to train, which generally cannot be met in practice. In view of this, this work develops two physics-guided machine learning frameworks which combine physics-based models and ML algorithms, for improving the prediction ability. In the proposed framework, ML models can consider those factors ignored by physics-based models and physics-based models can ensure the consistency with physical results. Finally, fatigue data of three AM materials are used for model evaluation and comparison. The results show that, physics-based models lack sufficient explanatory power for scatter of fatigue life. Moreover, compared with the purely data-driven methods, the proposed framework can maintain high accuracy and alleviate the over fitting phenomenon under insufficient fatigue data sources.

Experimental, analytical and numerical investigation on the capacity of composite glulam beams with holes

Holes in timber beams can significantly affect their bearing capacity. This paper presents an extensive experimental investigation of the effect of circular holes with and without plywood reinforcement on glulam joists behaviour. The tests consider the variability of hole position, number and diameter. In the second step, a finite element model based on fracture mechanics was developed and validated against the experimental force–displacement curves of thirteen configurations. The model reproduces crack initiation and propagation through the adoption of cohesive contact layers. The satisfactory agreement with the experimental data has been the base of extensive parametric analyses considering multiple beam and hole geometry selections and two load arrangements at the upper and lower side of the beam. Finally, the results of the parametric analyses, initially used for a qualitative understanding of the structural behaviour, are used for calibrating probabilistic capacity models of the capacity of simply-supported beams with circular holes. The mechanics-based probabilistic model calculates the capacity as the product between the analytical capacity associated with the reduced cross-section and an adimensional correction factor. The factor is expressed as a linear combination of a set of explanatory variables selected after a step-wise deletion process.

Warping caused by surface elasticity in a nanowire under torsion

By combining molecular statics simulations and continuum mechanics-based modeling, we show in this paper that the torsion of a $\ensuremath{\langle}001\ensuremath{\rangle}$ single-crystal copper nanowire with a circular cross section gives rise to a warp, i.e., to a displacement field along the wire axis that renders the cross section nonplanar. This behavior, which is in apparent contradiction with what is predicted by continuum mechanics for an isotropic cylinder, can be well explained if we take into account the elastic response of the wire lateral surface. The latter is characterized by the anisotropy of the surface elastic constants and, more specifically in the case of torsion, by the surface shear constant ${C}_{44}^{S}$ whose strength as a function of the local orientation of the lateral surface is estimated independently from atomistic calculations on slabs presenting different vicinal surfaces. The solution of the torsion problem is then obtained by adopting a semi-inverse method in the framework of the finite strain theory in linear elasticity with Gurtin-Murdoch boundary conditions linking surface stress and bulk stress. It is shown that such an approach is well suited to explain quantitatively the warp obtained in our atomistic simulations and to prove the preponderant role played by the surface elastic constant ${C}_{44}^{S}$.

Hydrostatic, strike-slip and normal stress true triaxial hydrofracturing testing of Blanco Mera Granite: breakdown pressure and tensile strength assessment

We have designed and built a versatile testing device to perform hydraulic fracturing experiments under true triaxial conditions. The device, based on a stiff biaxial frame that can be installed in a servocontrolled press, can accommodate cube rock samples of up to 150 mm-edge. Using a low-permeability rock known as Blanco Mera granite, we have performed a series of tests across a range of confining pressures including hydrostatic, normal, and strike-slip regimes. We have verified the applicability of two simple fracture mechanics-based models for the interpretation of experimental results, and we have determined the value of tensile strength of the rock from the injection curves recorded. The orientation of the hydraulically-triggered fractures with respect to the applied stress has also been analyzed. Although the models proposed by Rummel and Abou-Sayed provided reasonably satisfactory results, especially for hydrostatic and strike-slip tests, the presence of heterogeneities and defects in the rock matrix may have a strong influence on the fracture behavior and, therefore, affect the interpretation of hydrofracturing tests.

Investigating the Effect of Bending on the Seismic Performance of Hollow-Core Flooring

Even if precast pre-stressed hollow-core (PPHC) slabs are usually designed as simply supported elements, continuity with the supporting beam may exist when constructed together with a reinforced concrete topping and continuity reinforcing bars. During an earthquake (and possibly other lateral load), this continuity may result in bending moments being induced close to the supports as the buildings sway laterally. The response of precast floors to earthquake-induced demands has been addressed by past research. However, further investigation is required to improve understanding of several aspects of precast floor behaviour either revealed or emphasized by recent earthquakes in New Zealand. This paper proposes a mechanics-based modelling approach for the analysis of PPHC slab-to-beam seating connections. The model has been calibrated against existing test data to predict the failure of a PPHC slab under negative bending moments. The numerical outcomes allow comparison of the moment–drift response, principal tensile stresses, and crack progression during loading. The developed modelling approach will allow future studies to exhaustively investigate all aspects of precast floor behaviour by varying the properties and geometry of the PPHC seating connection.

Fatigue Reliability Analysis of Submarine Pipelines Using the Bayesian Approach

A fracture mechanics-based fatigue reliability analysis of a submarine pipeline is investigated using the Bayesian approach. The proposed framework enables the estimation of the reliability level of submarine pipelines based on limited experimental data. Bayesian updating method and Markov Chain Monte Carlo simulation are used to estimate the posterior distribution of the parameters of a fracture mechanics-based fatigue model regarding different sources of uncertainties. Failure load cycle distribution and the reliability-based performance assessment of API 5L X56 submarine pipelines as a case study are estimated for three different cases. In addition, the impact of different parameters, including the stress ratio, maximum load, uncertainties of stress range and initial crack size, corrosion-enhanced factor, and also the correlation between material parameters on the reliability of the investigated submarine pipeline has been indicated through a sensitivity study. The applied approach in this study may be used for uncertainty modelling and fatigue reliability-based performance assessment of different types of submarine pipelines for maintenance and periodic inspection planning.

Integrating automated machine learning and interpretability analysis in architecture, engineering and construction industry: A case of identifying failure modes of reinforced concrete shear walls

Machine learning (ML) has been recognized by researchers in the architecture, engineering, and construction (AEC) industry but undermined in practice by (i) complex processes relying on data expertise and (ii) untrustworthy ‘black box’ models. As a result, ML results of complex non-linear AEC problems, such as the failure mechanism of reinforced concrete (RC) shear walls, are not comparable with empirical and mechanics-based models. This paper aims to integrate automated ML (AutoML) and interpretability analysis to study the failure mechanism of RC shear walls. In this study, we collected a dataset of 351 comprehensive samples for the failure mode identification of RC shear walls. First, the AutoML model trained using the dataset outperformed a set of conventional ML methods in terms of the F1 accuracy score. Then, three model-agnostic interpretability analysis methods confirmed the trustworthiness of the AutoML model. The contribution of this paper is three-fold. First, AutoML sheds light on the automatic identification of failure modes of RC shear walls. Second, the interpretability analysis can validate ‘black-box’ ML models against long-established domain knowledge in solving non-linear AEC problems. Third, for AEC industrial practitioners, the whole process is automatic, accurate, less reliant on data expertise, and interpretable.

Flexible Needle Bending Model for Spinal Injection Procedures.

An in situ needle manipulation technique used by physicians when performing spinal injections is modeled to study its effect on needle shape and needle tip position. A mechanics-based model is proposed and solved using finite element method. A test setup is presented to mimic the needle manipulation motion. Tissue phantoms made from plastisol as well as porcine skeletal muscle samples are used to evaluate the model accuracy against medical images. The effect of different compression models as well as model parameters on model accuracy is studied, and the effect of needle-tissue interaction on the needle remote center of motion is examined. With the correct combination of compression model and model parameters, the model simulation is able to predict needle tip position within submillimeter accuracy.

Efficient boosting-based algorithms for shear strength prediction of squat RC walls

Reinforced concrete shear walls have been considered as an effective structural system due to their optimal cost and great behavior in resisting lateral loads. For the slender type of these walls, failure modes are mainly related to flexure, while for the squat type with height-to-length ratios less than two, shear is the dominant factor. Thus, accurate estimation of shear strength for squat shear walls is necessary for design applications and can also be complex due to the various effective parameters. In order to address this issue, first a comprehensive dataset with 558 samples of squat shear walls is conducted, and three hybrid models consisting of genetic algorithms and boosting-based ensemble learning methods, i.e., XGBoost, CatBoost, and LightGBM, are used for estimation of shear strength. The results showed high prediction accuracy, with a coefficient of determination of at least 98.6% for all three models. Genetic algorithm has been proven to be an effective method for tuning boosting-based algorithms compared to manual testing. In addition, the results of the algorithms are compared to their default hyperparameters and other conventional regression Models. Also, multicollinearity and principal component analysis (PCA) were studied. Furthermore, the performance of three tuned models is compared with that of a mechanics-based semi-empirical model and other genetic programming (GP)-based models. Finally, parametric and sensitivity analyses were performed, to demonstrate the ability of the models to identify the most critical parameters with significant influence on shear strength.

Development of rutting forecasting models for distinct asphalt pavement structures in RIOH testing track using different approaches

Rutting is one of the main exacerbation phenomena of the asphalt pavements, which seriously affects road safety and the quality of service. The predicted models are indispensable for preventing and controlling the damage caused by the exacerbation of pavement management systems. This paper is centered on developing more appropriate and exact rutting forecasting models for estimating the rutting performance of distinct asphalt pavement structures on the basis of the field data set of the Research Institute of Highway (RIOH) testing track. First of all, the critical influential factors giving rise to the rutting evolution are identified through analyzing the field data set of the RIOH testing track. After that, the framework of the rutting forecasting models based on the artificial neural networks (ANNs) coupled with genetic algorithm (GA) (GA-ANNs-based rutting forecasting models), the frameworks of the experiential and mechanics-based experiential rutting forecasting models are established according to the implicit relationship between the rutting and the key influential factors. Then, the unknown parameters of the experiential and mechanics-based experiential rutting forecasting models for different asphalt pavement structures are designed by the multivariate nonlinear optimization algorithm. Finally, the applicability of the GA-ANNs-based rutting forecasting models, the experiential and mechanics-based experiential rutting forecasting models are discussed and illustrated through analyzing their forecasting effects. The result illustrates that the proposed rutting forecasting models have better applicability for different asphalt pavement structures (i.e., the semi-rigid base pavement structure, the rigid base pavement structure, the flexible base pavement structure, and the full depth AC pavement structure).

Classification of failure modes of pipelines containing longitudinal surface cracks using mechanics-based and machine learning models

This paper applies the mechanics-based approach and five machine learning algorithms to classify the failure mode (leak or rupture) of steel oil and gas pipelines containing longitudinally oriented surface cracks. The mechanics-based approach compares the nominal hoop stress remote from the surface crack at failure and the remote nominal hoop stress to cause unstable longitudinal propagation of the through-wall crack to predict the failure mode. The employed machine learning algorithms consist of three single learning algorithms, namely naïve Bayes, support vector machine and decision tree; and two ensemble learning algorithms, namely random forest and gradient boosting. The classification accuracy of the mechanics-based approach and machine learning algorithms is evaluated based on 250 full-scale burst tests of pipe specimens collected from the open literature. The analysis results reveal that the mechanics-based approach leads to highly biased classifications: many leaks erroneously classified as ruptures. The machine learning algorithms lead to markedly improved accuracy. The random forest and gradient boosting models result in the classification accuracy of over 95% for ruptures and leaks, with the accuracy of the decision tree and support vector machine models somewhat lower. This study demonstrates the value of employing machine learning models to improve the integrity management practice of oil and gas pipelines.

Data-Driven Shear Strength Prediction of FRP-Reinforced Concrete Beams without Stirrups Based on Machine Learning Methods

Due to the intrinsic complexity, there has been no widely accepted mechanics-based estimation model of the shear performance of Fiber-Reinforced Polymer (FRP)-reinforced concrete beams. Capitalizing on a large amount of previous experimental data, data-driven machine learning (ML) models could be potentially suitable for addressing this problem. In this paper, four existing shear design provisions are reviewed and four typical ML models are analyzed. The accuracy of codified methods and ML models are compared and analyzed based on our established extensive database of FRP-reinforced concrete beams with rectangular cross sections. A series of artificially selected features considering the shear-carrying mechanisms of FRP-reinforced beams are incorporated into the proposed ML models to show their influence on the model validity. Bayesian optimization is utilized to automatically tune the hyperparameters of different ML models. Compared to the most satisfying codified predictions from CSA S806, the best ML model, XGBoost, can provide more accurate and consistent predictions for the database, with R2 enhanced by 15% and the MAE and RMSE reduced by 59% and 52%, respectively. With the selected features based on domain knowledge, the performance of ML models is further enhanced, shown by the most important features being the added ones. With outstanding performance on a large database and singular test, the ML approaches have great potential in guiding the shear design of FRP-reinforced concrete.

Experimental and Numerical Study of the Influence of Pre-Existing Impact Damage on the Low-Velocity Impact Response of CFRP Panels.

This paper presents an experimental and numerical investigation on the influence of pre-existing impact damage on the low-velocity impact response of Carbon Fiber Reinforced Polymer (CFRP). A continuum damage mechanics-based material model was developed by defining a user-defined material model in Abaqus/Explicit. The model employed the action plane strength of Puck for the damage initiation criterion together with a strain-based progressive damage model. Initial finite element simulations at the single-element level demonstrated the validity and capability of the damage model. More complex models were used to simulate tensile specimens, coupon specimens, and skin panels subjected to low-velocity impacts, being validated against experimental data at each stage. The effect of non-central impact location showed higher impact peak forces and bigger damage areas for impacts closer to panel boundaries. The presence of pre-existing damage close to the impact region leading to interfering delamination areas produced severe changes in the mechanical response, lowering the impact resistance on the panel for the second impact, while for non-interfering impacts, the results of the second impact were similar to the impact of a pristine specimen.

Bond deterioration of corroded reinforcements in SFRC: Experiments and 3D laser scanning

The bond deterioration behavior of reinforcements in steel fiber-reinforced concrete (SFRC) subjected to chloride-induced corrosion has not yet been fully elucidated. This study investigates the corrosion characterization, resistivity, corrosion-induced cracking, and associated effects on the bond performance of reinforcing bars in SFRC with milled-cut steel fibers (M fibers), referred to as MCSFRC. Pull-out tests were performed following galvanostatic corrosion, with the M fiber, concrete cover, and corrosion level being the main variables. A closer look at the bar corrosion was conducted using 3D laser scanning technology, which allowed for estimating the spatial-domain corrosion pattern and bond index degradation that characterize the rib-concrete interaction. The results indicated that, with the increase in corrosion levels, the bond index decreased exponentially as a result of pits propagating from the rib root to the rib peak. The bond strength of the MCSFRC specimens deteriorated less than that of the control counterparts and hook-end fiber-reinforced specimens in the literature. It is further revealed that the crack opening and bond index reduction are the bond deterioration mechanisms under corrosion, reducing the contact area between the bar and concrete through the ribs. On that basis, a mechanics-based model for bond strength deterioration was developed. The proposed model considered the two aforementioned effects, as well as the fiber reinforcement effect on crack suppression, showing good agreement with the experimental results.

NUMERICAL INVESTIGATION OF HYDRAULIC FRACTURE PROPAGATION IN HETEROGENEOUS ROCKS USING THE DAMAGE MECHANICS MODEL

In this study, we use a three-dimensional damage mechanics-based finite element model to investigate hydraulic fracture propagation in heterogeneous rocks. In the proposed finite element model, the Weibull random distribution function is adopted to characterize the rock mechanical heterogeneity of the elastic modulus. This damage mechanics model can well reflect the damage evolution of hydraulic fracture propagation in heterogeneous rocks under tensile and compression states. The fully coupled fluid-solid system of the hydraulic fracturing problem is numerically solved by the Newton-Raphson iterative method. Our numerical results show that key factors including the degree of mechanical heterogeneity, tensile strength, elastic modulus, and natural fractures have a significant impact on hydraulic fracture propagation in heterogeneous rocks. Both the damage zone and the first principal stress contour present a random distribution due to the mechanical heterogeneity. In a naturally fractured heterogeneous reservoir, the damage zone presents a banded distribution, which has the same dip angle of natural fractures. Some natural fractures away from the injection point are also partially opened under the combined action shear and normal stress field. This investigation provides new insight into the formation mechanisms of complex fracture networks in heterogeneous rocks.

Probabilistic prediction of joint shear strength using Gaussian process regression with anisotropic compound kernel

Shear strength of reinforced concrete (RC) beam-column joints is inevitably affected by various uncertainties existing in complex shear behaviors. However, traditional mechanics-based and data-driven methods fail to reasonably quantify these uncertainties. In order to overcome this limitation, a probabilistic prediction model for shear strength of RC beam-column joints based on the Gaussian process regression (GPR) with anisotropic compound kernel was proposed. Firstly, the GPR algorithm was employed to establish the probabilistic shear strength model of RC beam-column joints. Then the anisotropic compound kernel capturing complex variation characteristics of shear strength of RC beam-column joints was designed based on the sum of single kernels and the automatic relevance determination function. Meanwhile, the characteristic length-scales of the designed kernel were used to analyze the feature importance of shear strength of RC beam-column joints. Moreover, the anisotropic compound kernel was integrated to the GPR algorithm and the hyper-parameter optimization of the developed GPR was performed based on the maximum likelihood estimation. Finally, the developed GPR was verified by comparing with traditional models and 237 sets of experimental data. Analysis results show that the proposed kernel can effectively improve the prediction accuracy and generalization performance of the GPR. Comparing with traditional data-driven methods, the developed GPR not only has satisfactory estimation and generalization ability, but also describes probabilistic characteristics of shear strength of RC beam-column joints. Mechanics-based models were further calibrated based on the confidence interval of the developed GPR.

Nonlinear dynamics of system with combined rolling–sliding contact and clearance

Rolling–sliding contacts are found in a variety of systems, such as gears, drum brakes, and tire pavements. Such systems inherently have multiple nonlinearities such as kinematic, contact, and friction nonlinearities. Further, most of these systems have clearance between components which causes excessive vibration and noise. The combination of clearance with other nonlinearities makes the system dynamically interesting. It is important to understand the dynamic behavior of such systems under various operating conditions. Hence, this article focuses on theoretically investigating the transient and steady-state responses of a cam–follower system with rolling–sliding contact and clearance as an exemplary case. A contact mechanics-based model for the same has been developed for this purpose. The transient behavior of the system is examined based on energy and time-varying frequency contents. The domain of attractors, frequency response plots, and phase portraits are deployed to analyze the effect of initial conditions and excitation speed on the steady-state behavior, quantitatively and qualitatively. The steady-state solutions were classified into various branches based on periodicity and phase portraits. In addition, parametric analyses of the effects of loading, damping, and friction on the response have been conducted. Finally, the system is examined for start-up and shut-down characteristics with a constant rate of change in excitation frequency.

Comparison and correlation between polymer modified asphalt binders and mastics in high- and intermediate-temperature rheological behaviors

Traditional evaluation parameters for multiple stress creep recovery (MSCR) test and linear amplitude sweep (LAS) test are difficult to deeply and theoretically characterize the high- and intermediate-temperature rheological behaviors of polymer modified asphalt (PMA) binders and mastics. In this paper, one neat asphalt and four types of PMA binders and mastics are selected (i.e., the rubber modified ones, SBS modified ones, thermoplastic elastomer modified ones and SBS/Rubber modified ones). Two new analysis methods, i.e. Burgers model and damage mechanics-based crack growth model, are applied in analyzing the high- and intermediate-temperature rheological behaviors of PMA materials respectively. The Burgers model results indicate adding filler decreases the viscous deformation, and leads to lower non-recoverable compliance in the MSCR test. The polymer modification transfers the delayed elastic deformation to elastic deformation, thus causing more excellent rutting resistance. The damage mechanics-based crack growth model analysis reveals that adding filler enhances the crack growth rate and crack length, thus sacrificing the fatigue resistance. By contrast, the polymer modification enhances the fatigue resistance regardless of polymer types. Notably, a two-stage pattern based on crack growth rate and energy release rate was further proposed to identify the crack initiation and propagation of PMA materials. Specifically, 30% rubber modified asphalt binders and mastics perform best in both rutting and fatigue resistance. Moreover, PMA binders and mastics have strong correlation based on the recovery percentage and delayed elastic deformation. The correlation of fatigue life between PMA binders and mastics is very significant, however there was no obvious correlation in terms of crack length.

A Validated Three-Dimensional, Heterogenous Finite Element Model of the Rotator Cuff and The Effects of Collagen Orientation.

Continuum mechanics-based finite element models of the shoulder aim to quantify the mechanical environment of the joint to aid in clinical decision-making for rotator cuff injury and disease. These models allow for the evaluation of the internal loading of the shoulder, which cannot be measured in-vivo. This study uses human cadaveric rotator cuff samples with surface tendon strain estimates, to validate a heterogeneous finite element model of the supraspinatus-infraspinatus complex during various load configurations. The computational model was considered validated when the absolute difference in average maximum principal strain for the articular and bursal sides for each load condition estimated by the model was no greater than 3% compared to that measured in the biomechanical study. The model can predict the strains for varying infraspinatus loads allowing for the study of load sharing between these two tightly coordinated tendons. The future goal is to use the modularity of this validated model to study the initiation and propagation of rotator cuff tear and other rotator cuff pathologies to ultimately improve care for rotator cuff tear patients.

A mechanics-based approach to realize high-force capacity electroadhesives for robots.

Materials with electroprogrammable stiffness and adhesion can enhance the performance of robotic systems, but achieving large changes in stiffness and adhesive forces in real time is an ongoing challenge. Electroadhesive clutches can rapidly adhere high stiffness elements, although their low force capacities and high activation voltages have limited their applications. A major challenge in realizing stronger electroadhesive clutches is that current parallel plate models poorly predict clutch force capacity and cannot be used to design better devices. Here, we use a fracture mechanics framework to understand the relationship between clutch design and force capacity. We demonstrate and verify a mechanics-based model that predicts clutch performance across multiple geometries and applied voltages. On the basis of this approach, we build a clutch with 63 times the force capacity per unit electrostatic force of state-of-the-art electroadhesive clutches. Last, we demonstrate the ability of our electroadhesives to increase the load capacity of a soft, pneumatic finger by a factor of 27 times compared with a finger without an electroadhesive.