Virtual Test Data Research Articles

[formula omitted]-space physics-informed neural network (k-PINN) for compressed spectral mapping and efficient inversion of vibrations in thin composite laminates

The vibrational response of structural components carries valuable information about their underlying mechanical properties, health status and operational conditions. This underscores the need for the development of efficient physics-based inversion algorithms which, given a limited set of sensing data points and in the presence of measurement noise, can reconstruct the response at locations where measurement data is not available and/or identify the unknown mechanical properties. Addressing this challenge, Physics-Informed Neural Networks (PINNs) have emerged as a promising approach. PINNs seamlessly integrate governing equations into their architecture and have gained significant interest in solving inversion problems. In the context of learning and inversion of multimodal, multiscale vibrational responses, this paper introduces a novel spectral extension of PINNs, utilizing Fourier basis functions in the wavenumber domain, commonly known as k-space. The proposed method, referred to as k-space PINN (k-PINN), offers a robust framework for adjusting complexity and wavenumber composition of the response. Notably, the spectral formulation of k-PINN, coupled with the generally sparse representation of vibrations in k-space, facilitate efficient reconstruction and learning of broadband vibrations and alleviate the spectral bias associated with standard PINN. Additionally, the spectral solution space introduced by k-PINN substantially reduces the computational cost associated with computing physics-informed loss terms. We evaluate the effectiveness of the proposed methodology on reconstructing the bending vibrational mode shapes of a thin composite laminate and identifying its effective bending stiffness coefficients. Mode shapes are initially obtained from finite element simulation, and virtual test data with added noise are generated for evaluation purposes. It is shown that the proposed k-PINN methodology outperforms the standard PINN in terms of both learning and computational efficiency. The performance of k-PINN is further demonstrated by show casing its capability in learning different selections of symmetric, anti-symmetric and asymmetric mode shapes.

Fusion of Multiple Data Sources for Vehicle Crashworthiness Prediction Using CycleGAN and Temporal Convolutional Networks

Abstract Computer-aided engineering (CAE) models play a pivotal role in predicting crashworthiness of vehicle designs. While CAE models continue to advance in fidelity, an inherent discrepancy between CAE model predictions and the responses of physical tests remains inevitable. Machine learning (ML) models have shown promising potential in improving the prediction accuracy of CAE models. Nevertheless, the scarcity of vehicle crash data poses a significant challenge for the training of such models. This paper overcomes these challenges by fusing multiple data sources from two different types of vehicles. More specifically, the Cycle-Consistent Generative Adversarial Neural networks (CycleGAN) are first employed to translate features of time-series test data from one domain to another using cycle consistency loss. Such a translation allows for the generation of synthetic crash test data for the second vehicle type by leveraging existing tests from both vehicle types. In parallel, an initial temporal convolutional network (TCN) model is trained using CAE simulation data and physical test data of the first vehicle type. This pre-trained TCN model is then fine-tuned using three sources of data from the second vehicle type, namely the CAE data, test data, and the augmented virtual test data generated using CycleGAN. Through these data fusion, the crashworthiness prediction accuracy of the second vehicle type can be improved. The proposed method is demonstrated by a real-world case study with a small-size SUV and a medium-size SUV. Results show enhancements in the predictive performance of the medium-size SUV model.

Validation of an Abaqus impact wave propagation model

Structural Health Monitoring (SHM) is an important topic in aircraft industry to reduce the maintenance costs and increase the availability of an aircraft fleet through automating the determination of the health of the aircraft structure and systems using on-site sensors, like piezoelectric (PZT) and optic fibre Bragg grating (FBG) sensors that are often applied in SHM sensor networks. A finite element model for wave propagation due to an impact is very desirable to enable the design of optimal sensor network configurations for SHM applications, to obtain a better understanding of wave propagation (especially in composite structures) and also to generate simulated (virtual) test data for SHM algorithm development. Therefore, an Abaqus explicit finite element (FE) model was developed for wave propagation due to impacts. To validate such a model, a series of impact tests were performed with a drop tower on an aluminium clamped square panel, at different impact energy levels and different impact locations. On the panel, eight FBG and six PZT sensors were installed. These experimental results were used to validate the FE model, in which the impacts were simulated and the computed PZT and FBG sensor responses were compared against the experimental results. The paper will present the impact test setup, the finite element model and the validation results. The main focus of the paper will be the sensitivity analysis, identifying the model parameters that significantly affect the impact sensor responses and their optimal settings. A very good correlation between the numerical and experimental results was obtained. In near future work, the model will be extended to a (thermoplastic) composite panel, for which experimental results on a flat composite panel already have been collected, and a stiffened composite panel representative of an actual horizontal tail plane are planned.

Mechanical Biomarkers in Bone Using Image-Based Finite Element Analysis

The purpose of this review is to summarize insights gained by finite element (FE) model-based mechanical biomarkers of bone for in vivo assessment of bone development and adaptation, fracture risk, and fracture healing. Muscle-driven FE models have been used to establish correlations between prenatal strains and morphological development. Postnatal ontogenetic studies have identified potential origins of bone fracture riskand quantified the mechanical environment during stereotypical locomotion and in response to increased loading. FE-based virtual mechanical tests have been used to assess fracture healing with higher fidelity than the current clinical standard; here, virtual torsion test data was a better predictor of torsional rigidity than morphometric measures or radiographic scores. Virtual mechanical biomarkers of strength have also been used to deepen the insights from both preclinical and clinical studies with predictions of strength of union at different stages of healing and reliable predictions of time to healing. Image-based FE models allow for noninvasive measurement of mechanical biomarkers in bone and have emerged as powerful tools for translational research on bone. More work to develop nonirradiating imaging techniques and validate models of bone during particularly dynamic phases (e.g., during growth and the callus region during fracture healing) will allow for continued progress in our understanding of how bone responds along the lifespan.

Virtual Test Based on Far Side Occupant Protection

Virtual testing is getting more and more attention in the field of vehicle detection. In order to improve the evaluation accuracy of virtual test data under the condition of far side occupant protection, the paper studies the evaluation method of remote virtual test data accuracy. First, the dynamic responses of 11 simulation and test data are determined by analyzing the basic data under remote conditions. Then, based on CORA and ISO_ The measurement method of TS18571 is used to evaluate the correlation, and the error analysis method is used to analyze the responses with differences in the correlation evaluation results. Finally, the conclusion of the precision evaluation of CORA and ISO measurement methods is obtained through experimental research.

An Investigation on the Influence of Operation Experience on Virtual Hazard Perception Using Wearable Eye Tracking Technology.

Poor electrical hazard recognition is a widespread issue in the production industry. Hazard perception has impacts workers’ hazard recognition, causing them to experience unanticipated hazard exposure and suffer catastrophic injuries. To improve the factors of affecting hazard perception, the current study examined hazard recognition as an everyday visual search task. A comparative test was carried out combining the advantages and disadvantages of the two test methods. It was confirmed that the virtual image test data can replace the real image test data and demonstrate superior flexible settings performance, so the virtual image test method is used. A hazard perception test method based on wearable eye tracking technology was proposed to analyze the eye-tracking data (i.e., fixation, count, search duration, mean fixation duration, eye tracking, and hazard recognition performance feedback) were compared between experts in the field of electrical safety: skilled workers with at least five years of work experience and workers who had been on the job for less than a year. It was found that experts had a better hazard recognition accuracy and missed detection rate than other workers. Experts’ hazards research track was more concised and paid less attention time. This advantage is most obvious in complex risk environments. The findings also suggest that workers who have different working years was not obvious visual search patterns other than the search duration. As can be seen the work experience is not an absolute factor in improving hazard perception. The present research will be useful to understand the influence of working years on hazard perception and provide a theoretical basis for corresponding training.

A new method for the in vivo identification of degenerated material property ranges of the human eye: feasibility analysis based on synthetic data

This paper proposes a new method for in vivo and almost real-time identification of biomechanical properties of the human cornea based on non-contact tonometer data. Further goal is to demonstrate the method’s functionality based on synthetic data serving as reference. For this purpose, a finite element model of the human eye is constructed to synthetically generate full-field displacements from different data sets with keratoconus-like degradations. Then, a new approach based on the equilibrium gap method combined with a mechanical morphing approach is proposed and used to identify the material parameters from virtual test data sets. In a further step, random absolute noise is added to the virtual test data to investigate the sensitivity of the new approach to noise. As a result, the proposed method shows a relevant accuracy in identifying material parameters based on full-field displacements. At the same time, the method turns out to work almost in real time (order of a few minutes on a regular workstation) and is thus much faster than inverse problems solved by typical forward approaches. On the other hand, the method shows a noticeable sensitivity to rather small noise amplitudes rendering the method not accurate enough for the precise identification of individual parameter values. However, analysis show that the accuracy is sufficient for the identification of property ranges which might be related to diseased tissues. Thereby, the proposed approach turns out promising with view to diagnostic purposes.

Establishing prediction master curve of dynamic modulus of asphalt mixture considering randomness of aggregate morphology

The objective of this paper is to study the influence of aggregate morphology on dynamic modulus of asphalt mixture and establish a method to predict dynamic modulus of asphalt mixture at different temperatures and loading frequencies based on aggregate morphology and asphalt mortar material parameters. The Digital image processing (DIP) method is used to obtain statistical laws of the fractal dimension D and aspect ratio lw of aggregate section profiles, and the midpoint displacement method is used to generate virtual random aggregate sections. Based on the 2P2S1D constitutive model using the WLF time temperature equivalence, master curves of dynamic modulus of asphalt mortar at different reference temperatures are obtained and converted into the Prony series. Interface Transition Zones (ITZs) between asphalt mortars and aggregates are modeled by the Coulomb friction model, and dynamic moduli of asphalt mixture calculated by this model are in the range of calculation results by the tie model and the frictionless model. A total of 100 virtual specimens including random aggregates, asphalt mortars, ITZs and air voids are generated. Under different temperature and frequency combinations, extensive virtual tests of dynamic modulus of asphalt mixture are carried out on the virtual specimens. Based on the virtual test data of dynamic modulus, prediction master curves of dynamic modulus of asphalt mixture are obtained. The results show that aggregate morphology has a significant effect on dynamic moduli of asphalt mixture, and this effect has a clear correlation with the distribution laws of aggregate morphology indexes. By comparing with the test data of different literature, the probabilistic prediction method of dynamic modulus of asphalt mixture established in this paper is feasible.

Multisensor Data Fusion in Testability Evaluation of Equipment

The multisensor data fusion method has been extensively utilized in many practical applications involving testability evaluation. Due to the flexibility and effectiveness of Dempster–Shafer evidence theory in modeling and processing uncertain information, this theory has been widely used in various fields of multisensor data fusion method. However, it may lead to wrong results when fusing conflicting multisensor data. In order to deal with this problem, a testability evaluation method of equipment based on multisensor data fusion method is proposed. First, a novel multisensor data fusion method, based on the improvement of Dempster–Shafer evidence theory via the Lance distance and the belief entropy, is proposed. Next, based on the analysis of testability multisensor data, such as testability virtual test data, testability test data of replaceable unit, and testability growth test data, the corresponding prior distribution conversion schemes of testability multisensor data are formulated according to their different characteristics. Finally, the testability evaluation method of equipment based on the multisensor data fusion method is proposed. The result of experiment illustrated that the proposed method is feasible and effective in handling the conflicting evidence; besides, the accuracy of fusion of the proposed method is higher and the result of evaluation is more reliable than other testability evaluation methods, which shows that the basic probability assignment of the true target is 94.71%.

Three-dimensional virtual dynamometer to measure the process of overcoming obstacles by disc cultivator

The process of interaction between elastically fastened tillage tools and obstacles is investigated by means of computer applications that simulate the dynamics of the movement of 3d-models (MDB) created in the CAD. Design of the virtual stand and dynamometer block, implemented in the SolidWorks CAD system and the application for modeling the movement dynamics of SolidWorks Motion are described. The virtual stand enables to simultaneously capture all the components of traction resistance vector (RX, RY, RZ). Measured volumetric power characteristics of the process of stump overcoming by a section of a modular disk cultivator and the forces arising on the spring of a safety mechanism are given as an example of the virtual stand use. The discrepancy between experimental data and modeling data at the entrance to the stump is 8 … 10%. In the future, virtual test data can be used to optimize the structures of tillage tools and safety mechanisms, as well as to reliably simulate disturbances in modeling the working process of the machine-tractor unit as a whole.

Computationally efficient necking prediction using neural networks trained on virtual test data

The onset of localized necking under monotonic and non-monotonic loading can be well-predicted by the imperfection-based approach proposed by [1] (MK). However, a large number of virtual imperfections has to be investigated for an accurate necking prediction, making the MK approach computationally expensive and hence preventing the industrial application for full-scale vehicle models. To overcome these issues, a computationally efficient neural network (NN) model is proposed for replacing the MK model in the present work. An extended version of the MK model has been implemented into a User Material for an explicit crash solver. The model continuously computes the “distance to localized necking” as an important engineering quantity. Single shell element simulations are utilized for creating a comprehensive virtual test database for monotonic and non-monotonic loading for a 22MnB5 grade in an as-delivered state. A simple feed-forward NN model, featuring only one hidden layer, is trained and tested against the virtual data, where invariants of the stress and plastic strain tensors represent the input features of the NN and the “distance to necking” represents the output value of the NN. For comparison of the computational cost, the NN architecture has also been implemented in a User Material Routine for shell elements. The predictions of the NN and the MK model are in good agreement, where, due to the simple mathematical structure of the NN, the computational cost of the NN is significantly lower than for the MK implementation.

A novel GREY‒ASMAA model for reliability growth evaluation in the large civil aircraft test flight phase

PurposeThe purpose of this paper is to present a novel GREY‒ASMAA model for reliability growth evaluation in the large civil aircraft test flight phase.Design/methodology/approachAs limited data are collected during the large civil aircraft test flight phase, which are not enough to meet the requirements of the ASMAA model for reliability growth, four basic GM(1, 1) models, even grey model, original difference grey model, even difference grey model and discrete grey model, are presented. Then both forward and backward grey models GM(1,1) are built to forecast and obtain virtual test data on left and right sides. Then the ASMAA model for reliability growth evaluation can be built based on original and virtual test data.FindingsAiming at the background of poor information data during the large civil aircraft test flight phase, first, a novel GREY‒ASMAA model, which was combined by the grey model GM(1,1) with the ASMAA model, has been put forward in this paper.Practical implicationsThe GREY‒ASMAA model for reliability growth evaluation can be used to solve the problem of reliability growth evaluation with poor information data during the large civil aircraft test flight phase, and it has been used in reliability evaluation of C919 at the test flight stage.Originality/valueThis paper presents two new definitions of forward grey model GM(1,1) and backward grey model GM(1,1), as well as a novel GREY‒ASMAA model for reliability growth evaluation of large civil aircraft during test flight phase.

Necking-induced fracture prediction using an artificial neural network trained on virtual test data

The imperfection-based necking model by Marciniak and Kuczyński (MK) is frequently used for predicting the onset of localized necking under proportional and non-proportional loading, which can be considered a lower limit for the occurrence of fracture in a vehicle body structure subjected to crash loading. A large number of virtual imperfection lines at different orientation angles have to be analysed simultaneously in order to find the critical imperfection causing necking under arbitrary loading. This, and the continuous computation of a “distance to necking” quantity, representing a crucial output quantity for the simulation engineer, makes the model computationally expensive and limits industrial use in full-scale vehicle crash simulations.In this work, an extended MK model is used for creating a virtual test data base under proportional and non-proportional loading for training of a computationally more efficient simple feed-forward neural network (NN). Both models are implemented in a User Material routine of an explicit crash code, where the predictions of the NN are in good agreement with the predictions of the MK reference model, however at a significantly reduced computational cost. Besides a pure numerical validation study, an experimental validation study has been performed, imposing biaxial tension loading followed by plane strain tension loading until necking using a special punch test apparatus. Whereas MK and NN are in good agreement with the experimental observations, the agreement of classical necking models, applied in conjunction with a linear damage accumulation (forming severity) concept was less accurate.

On the synergy between physical and virtual sheet metal testing: calibration of anisotropic yield functions using a microstructure-based plasticity model

This paper scrutinizes in detail the synergy between physical and virtual testing of the mechanical response of sheet metal. In this context, physical testing refers to the usage of physical samples onto which mechanical tests are conducted, while virtual testing refers to multi-scale plasticity simulations onto a model representation of the metallic microstructure derived from microstructural measurement. An extensive experimental campaign was conducted to capture the plastic material response of mild steel sheet in the first quadrant of the biaxial stress space, i.e. for tensile stresses along the Rolling Direction (RD) and Transverse Direction (TD). The experimental data was acquired using state-of-the-art stress-controlled material tests enabling to probe the material up to large plastic strains in the first quadrant of stress space. Identical stress paths were simulated using the Virtual Experimentation Framework (VEF) software suite adopting the ALAMEL multi-scale plasticity model. The ALAMEL model was calibrated solely on the basis of the initial crystallographic texture of the material and a reference hardening curve obtained through a uniaxial tensile test in the rolling direction. The predictive accuracy of the ALAMEL model to reproduce the experimentally acquired material response has been thoroughly assessed. To avoid any bias in the assessment due to extrapolation of the material behavior, predictions were limited to the pre-necking regime of the material. The predictions of the VEF show good agreement with the physical test results. Subsequently, the experimental and virtual test data were used to calibrate Hill’s quadratic yield criterion and the Yld2000-2d yield function. The calibration accuracy of these yield criteria is thoroughly assessed by comparing theoretical predictions and experimental data. In addition, the calibrated yield functions are used to simulate a hydraulic bulge test and the results are compared with experimental observations. It is shown that the adopted virtual material testing procedure has reached a sufficient level of maturity to potentially serve as a viable alternative for physical material testing of steel sheet.

Parameter distribution characteristics of material fatigue life using improved bootstrap method

The bootstrap method is mostly used to estimate statistical characteristics of small sample data. However, the limitations of the bootstrap method itself lead to a reduction in the reliability of small-sample estimates. In this article, an improved bootstrap method is developed to address this problem. In the statistically significant error range (the sample average error and the limit error of sampling) of the original single sample data, expanding the virtual test data that obey two distributions to overcome the limitations of the bootstrap method itself. This article compares and analyses these two methods through the case; the result indicates that the improved bootstrap method can enhance the reliability of the estimation results without changing its probability distribution. We also discussed how to reduce the fluctuation of the improved bootstrap method. And the effectiveness and feasibility of this improved method are discussed in the analysis of fatigue life test data.

A novel approach to temperature-dependent total emissivity estimation based on isothermal cooling

This paper proposes a novel approach to estimate the temperature-dependent total emissivity based on the isothermal cooling process. The realizability of the ideal isothermal cooling phase has been verified through the finite-element model, even for specimens with low thermal conductivity. During the cooling process, the temperature difference of the specimen can be less than 0.8%. Moreover, based on the temperature curve of isothermal cooling, nonlinear total emissivity can be calculated using the inverse method over a broad temperature range, e.g., 500–1200 K, and the relative errors will be less than 8.83% with virtual test data. For a real-world application, we introduce the Savitzky-Golay (SG) filter to weaken the effect of noise on computation precision and successfully gain the smooth temperature-dependent total emissivity.

Design of an improved predictive LTR for rollover warning systems

Rollover warning system predicts whether the vehicle has a risk of rollover in a period of time based on the current driving state of the vehicle. Most previous studies on rollover evaluation index have only used vehicle’s current states and often ignored the judgment of the overall tendency of the motion. In the calculation of the rollover early warning algorithm, it is assumed that the vehicle input was known. Based on the above reasons, this paper introduces the derivative of LTR into the rollover index, using an improved predictive LTR (IPLTR) as the rollover index, and realized the real-time calculation through an 8-DOF nonlinear vehicle model. The vehicle rollover prediction system is achieved through the preview driver model and the rollover prediction index calculation model. To achieve a proper preview time for the driver model, a comprehensive evaluation formula that considers the trajectory deviation, the steering wheel rotation speed and lateral acceleration is used. Finally, an IPLTR is developed, which utilizes a driver model to predict the steering angle and several other sensor signals available from the electronic stability control system as inputs. The new IPLTR index can provide a time-advanced measure of rollover propensity and its effectiveness is verified by virtual tests and experimental test data. Simulation results by double lane change (DLC) and “Highway Merge” road maneuvers show that the proposed rollover warning algorithm is rather satisfying. Experimental test by pylon course slalom maneuver also proves its effectiveness and has more advantages of such a system in rollover prevention.

Error assessment in post-necking strain hardening behaviour identification of mild steel sheet

The information hidden in the diffuse neck of a tensile test on a thin metal sheet can be extracted using a special case of the non-linear virtual fields method yielding the so-called post-necking strain hardening behaviour. The method, however, requires a number of assumptions which are scrutinized in this paper. To eliminate experimental errors which could potentially hamper the assessment, virtual test data (i.e. strain fields at different load steps) is generated using a FE model of the tensile test. The identification strategy is then used to retrieve the reference strain hardening behaviour used in the FE simulation. This approach is used to study the necessity of incorporating rate-dependent plasticity in the identification procedure. Additionally, the necessary plane stress condition in the diffuse neck is studied.

Thermal Conductivity of Cement Concrete Using Discrete Element Method

Consideration on the traditional experiment was time-consuming and required a major investment in human and material resource, even leaded to a great level of error in the process of experiment. The new test method of thermal conductivity of cement concrete based on discrete element method was presented. The cylinder-shaped specimen was created via the mass graduation of the cement concrete, and make sure the numerical model was identical to the truth experiment. The thermal micro-properties composed of density,specific heat, coefficient of linear thermal expansion and thermal resistance were conducted. The applied constant heat flux and constant temperature as the boundaries were investigated via the herein developed model. The results testified that the virtual test data was nearly identical with the analytical values for both different boundaries, and it confirmed that the new numerical model using discrete element method is feasible and reliable. It also supplied a new method on thermal properties study.

Testability demonstration with component level data from virtual and physical tests

Testability demonstration plays an important role in assuring the testability capability, which can decrease the fault diagnosis time and accelerate the maintenance actions. However, testability demonstration test with classical planning method has the problems of large fault sample size, high test cost, and long test period. A new testability demonstration planning method is proposed, which takes the component level data from both virtual and physical demonstration tests as prior information. Owing to the limitations of the testability modeling technology and relevant programming tools, the virtual testability prototype of the system level cannot be established and the virtual testability test data is not totally credible. So a data conversion method based on the information entropy theory is proposed to convert the component level virtual and physical test data into equivalent system test data, in which the data credibility is taken into consideration. The equivalent system test data is then used to get the prior probability density function of the testability indexes with an empirical Bayesian method. Then, a testability demonstration planning method of Bayesian posterior risk criteria is presented. Finally, the fault detection rate demonstration tests of a flight control system and a heating controller are taken as examples to verify the proposed method. The results show that the introduction of prior test data can effectively decrease the sample size and the credibility of the virtual testability test data can affect the test plan.