Training Response Research Articles

Dynamic response of high-speed trains under uneven temperature fields of slab tracks on cable-stayed bridges

The temperature deformation of a long-span cable-stayed bridge (LSCSB) could significantly deteriorate the smoothness of rails, thereby affecting the dynamic behavior of high-speed trains (HSTs). In this paper, based on the thermal-fluid-structure coupling theory and vehicle-track coupling dynamics theory, an analysis method was proposed to explore the influence of temperature deformation of LSCSBs on the dynamic behavior of HSTs. The temperature deformation of LSCSBs within a day was analyzed, and the dynamic behavior of HSTs passing through the LSCSBs at different times was revealed. The results showed that on a summer day, the temperature deformation of LSCSBs varied significantly and rapidly, and the variation experienced a significant increase of 23.46% within just 2 h. Due to the influence of structural temperature deformation, the wheel-rail interaction, train safety, and passenger comfort changed noticeably. The construction temperature of the bridge and track decisively affected the dynamic behavior of the train. Therefore, it was crucial to construct the structure under suitable temperature conditions to prevent further exacerbation of the dynamic behavior of HSTs due to violent temperature differentials. In summary, the method proposed in this paper could contribute to a comprehensive and in-depth understanding of the operational characteristics of HSTs on LSCSBs.

O97 RESPONSE TO AEROBIC TRAINING MAY DIFFER BETWEEN ELDERLY MEN AND WOMEN WITH TREATED HYPERTENSION

Background and objective. Hypertension incidence increases with aging, with expressive enhancement in women after the menopause due to the important physiological changes induced by the estrogen levels decline. Aerobic training is recommended for lowering blood pressure (BP) in men and women with hypertension. However, the possible gender-related differences in training responses have been poorly investigated, especially in elderly patients. Therefore, the objective of this study was to analyze responses to aerobic training in elderly men and women with treated hypertension. Methods: Thirty-one elderly hypertensives (15 women, 68±6ys and 16 men, 66±7ys) who were taking regular antihypertensive medication for at least 4 months participated in the study. They underwent 10 weeks of an aerobic training designed in accordance with the guidelines’ recommendations for hypertension (i.e. 3 session/week, cycle ergometer, 30-45 min, moderate intensity). Before and after the training period, peak oxygen uptake (VO2peak, cardiopulmonary exercise test), BP (auscultatory), systemic hemodynamics (CO2 rebreathing technique) and vascular function (brachial doppler ultrasound) were assessed. In each group (men and women), the training effects were evaluated using paired t-tests or Wilcoxon tests, considering p<0.05 as significant. Results: Aerobic training increased aerobic capacity in both sexes (men = 20.8±4.2 vs. 21.8±4.8 ml.kg-1.min-1, p=0.04 and women = 15.3±2.6 vs. 16.4±2.3 ml.kg-1.min-1, p=0.01). In men, systolic BP and systemic vascular resistance decreased (129±13 vs. 125±12 mmHg, p=0.003 and 21.0±3.8 vs. 18.9±4.1 u, p=0.02), while flow mediated dilation increased (5.5±2.9 vs. 7.2±3.1%, p=0.01) after the training period. On the other hand, in women, none of these variables changed with the training (124±15 vs. 126±12 mmHg, p=0.92; 23.2±5.5 vs. 23.5±5.6 u, p=0.80; and 4.2±3.2 vs. 5.4±3.5%, p=0.10). The other variables assessed did not change in either group. Conclusions: The proposed aerobic training program increased aerobic capacity in both male and female elderly hypertensives, but improved BP, systemic vascular resistance, and endothelial function only in males. Key-words: aerobic training, gender, blood pressure, hypertension. Financial Support: CAPES-0001, CNPq-302309/2022-5, and FAPESP- 2022/12605-3.

Mitigating crosswind response of a high-speed train passing the end of windbreak walls

The sudden change in wind speed at the end of a windbreak wall poses a significant threat to train safety, as it amplifies train’s responses. This study aims to investigate countermeasures to reduce the maximum unsteady crosswind response of high-speed trains at the end of windbreaks. Wind flow around windbreak walls was systematically examined by particle image velocimetry (PIV), considering both constant and varying porosities. Subsequently, a gust model was proposed to evaluate the wind distributions behind windbreak walls, validated by experimental results. Aerodynamic forces on high-speed train, incorporating windbreak walls with diverse porosities, were assessed by integrating the gust model and the quasi-steady theory. Ultimately, the unsteady crosswind responses were computed employing multibody simulations (MBS), accounting for both solid windbreak walls and those with varying porosities. A noteworthy finding was the reduction of approximately 25 % in the maximum unsteady crosswind response when windbreak porosity varied from 0 % to 80 %, thus affirming the efficacy of windbreaks featuring varying porosities in diminishing unsteady crosswind response. Overall, this study provides valuable insights into the design of windbreaks for high-speed trains, which can improve their safety and performance in strong wind regions.

PSXII-1 Comparative analysis of the training response of Raramuri Criollo and Angus Cattle to virtual fencing

PSXII-1 Comparative analysis of the training response of Raramuri Criollo and Angus Cattle to virtual fencing

Effects of Altitude Training and Sea-Level Training Programs on the Physiological and Performance of Elite Chinese Male Marathon Runners

Background and Aims: Examining the differences between sea-level training programs and altitude offers important new perspectives on how environmental influences affect physiological adaptations and athletic performance. By being aware of these variations, coaches and athletes can tailor training plans to the unique conditions of altitude, which may improve performance in terms of endurance, oxygen use, and general athletic ability. Thus, this study aimed to investigate the effects of altitude training on physiological adaptation and performance in Chinese elite male marathon runners compared to plains training. By comparing the results of altitude training with those of sea-level training, this study sought to determine whether altitude training has a significant advantage in terms of physiological adaptation and performance enhancement for the Chinese elite male marathoner. In addition, exploring the variability of altitude training responses will help to inform an in-depth understanding of training strategies affecting elite marathoners and help to optimize performance enhancement methods in endurance sports. Methodology: Thirty-two Chinese elite male marathon runners will be recruited and subjected to a randomized controlled trial (RCT) in which they will be assigned to either a high altitude or plains training group. SPSS was used to perform independent samples t-tests and paired samples t-tests on the data of technical indicators. Both groups will receive structured and standardized training for a predetermined period, with the altitude of the high-altitude training set at 2366 m for 8 weeks. Results: (1) In the experimental group, the differences in red blood cells (RBC), maximal heart rate (HRmax), maximal oxygen uptake (VO2max), and full marathon performance before the altitude experiment compared to after the altitude experiment were significant (P=0.001); the differences in maximal anaerobic power and hemoglobin (HB) before the experiment were substantial (P<0.005); and the differences in maximal heart rate (HRmax) between the 28 days of the experiment and the post-experiment were substantial. (HRmax) difference was substantial (P=0.001). (2) The red blood cells (RBC), hemoglobin (HB), and maximal oxygen uptake (VO2max) of the experimental group were significantly higher than those of the control group compared with the control group (P < 0.005) Conclusion: The results of the study showed that eight weeks of altitude training at an altitude of 2,366 meters above sea level could improve the athletes' performance in the marathon and aerobic capacity, and could be used in the training of marathon runners.

Simulation on passenger evacuation of metro train fire in the tunnel

The metro tunnel space is relatively closed, it is difficult to rescue after the metro train fire accident, which causes heavy casualties and property losses. Studying the safe evacuation of passengers under fire smoke spread has important significance. The numerical simulation methods are used to analyze the reliability of passenger evacuation in metro train fire scenarios. Firstly, the required safety evacuation time (RSET) calculation methods in short-distance and long-distance tunnels are developed combining the pre-action time and train response time. Then, a passenger evacuation model in metro trains is established in Pathfinder software to simulate the evacuation process under different passenger loads, carriage layouts, evacuation strategies, and train locations. The simulation results show that the RSET is affected by passenger load, seat arrangement, evacuation strategies and train locaitons. The critical point calculation method effectively finds out the critical point of the train location, which guides emergency evacuation strategies in fire scenarios.

Study on the misalignment deformation threshold of girder end of curved bridge with ballasted track

When a train traverses a curved bridge, it inherently tends to shift laterally outward. To counteract this propensity, the strategy of track superelevation is employed. After earthquake, curved bridges may exhibit diverse degrees of residual deformation. The precise failure mechanisms that compromise train safety and passenger comfort during transit over curved bridges, particularly under the combined influence of residual deformation, centrifugal force, and track superelevation, remain not entirely comprehended. A comprehensive system integrating trains with curved track-bridges is developed to examine the impact of girder end misalignments on residual track irregularities. The analysis concentrated on the dynamic time-history responses of trains traversing bridges featuring girder misalignments, exploring a range of amplitudes, train velocities, and curve radii. The findings revealed significant variances in the dynamic train responses, especially in lateral acceleration and the Nadal index, when contrasting centripetal and centrifugal girder side misalignments. Furthermore, it is discerned that lateral misalignment at the girder ends markedly affects train lateral acceleration and the Nadal index, with a relatively minor influence on other dynamic response metrics. Finally, the thresholds for misalignment deformation were summarized, for curve radii of 800, 1000, and 1200 meters, the comfort deformation thresholds at 100 km/h are determined to be 23, 23, and 20 mm respectively, while the safety deformation thresholds are found to be 26, 30, and 38 mm respectively. At a speed of 120 km/h, the comfort deformation thresholds for the same curve radii are 18, 18, and 15 mm, with the safety deformation thresholds at 12, 15, and 21 mm.

Trained immunity is regulated by T cell-induced CD40-TRAF6 signaling

Trained immunity is characterized by histone modifications and metabolic changes in innate immune cells following exposure to inflammatory signals, leading to heightened responsiveness to secondary stimuli. Although our understanding of the molecular regulation of trained immunity has increased, the role of adaptive immune cells herein remains largely unknown. Here, we show that Tcells modulate trained immunity via cluster of differentiation 40-tissue necrosis factor receptor-associated factor 6 (CD40-TRAF6) signaling. CD40-TRAF6 inhibition modulates functional, transcriptomic, and metabolic reprogramming and modifies histone 3 lysine 4 trimethylation associated with trained immunity. Besides invitro studies, we reveal that single-nucleotide polymorphisms in the proximity of CD40 are linked to trained immunity responses invivo and that combining CD40-TRAF6 inhibition with cytotoxic T lymphocyte antigen 4-immunoglobulin (CTLA4-Ig)-mediated co-stimulatory blockade induces long-term graft acceptance in a murine heart transplantation model. Combined, our results reveal that trained immunity is modulated by CD40-TRAF6 signaling between myeloid and adaptive immune cells and that this can be leveraged for therapeutic purposes.

Progression and perceptual responses to blood flow restriction resistance training among people with multiple sclerosis.

Resistance exercise can attenuate muscular impairments associated with multiple sclerosis (MS), and blood flow restriction (BFR) may provide a viable alternative to prescribing heavy training loads. The purpose of this investigation was to examine the progression of upper and lower body low-load (30% of one-repetition maximum [1RM]) resistance training (RT) with BFR applied intermittently during the exercise intervals (RT + BFR) versus volume-matched heavy-load (65% of 1RM) RT. Men and women with MS (n = 16) were randomly assigned to low-load RT + BFR (applied intermittently) or heavy-load RT and completed 12weeks (2 × /week) of RT that consisted of bilateral chest press, seated row, shoulder press, leg press, leg extension, and leg curl exercises. Exercise load, tonnage, and rating of perceived exertion were assessed at baseline and every 6 weeks. Training load increased to a greater extent and sometimes earlier for RT + BFR (57.7-106.3%) than heavy-load RT (42.3-54.3%) during chest press, seated row, and leg curl exercises, while there were similar increases (63.5-101.1%) for shoulder press, leg extension, and leg press exercises. Exercise tonnage was greater across all exercises for RT + BFR than heavy-load RT, although tonnage only increased during the chest press (70.7-80.0%) and leg extension (89.1%) exercises. Perceptions of exertion (4.8-7.2 au) and compliance (97.9-99.0%) were similar for both interventions. The training-induced increases in load, high compliance, and moderate levels of exertion suggested that RT + BFR and heavy-load RT are viable interventions among people with MS. RT + BFR may be a preferred modality if heavy loads are not well tolerated and/or to promote early-phase training responses.

Digital Twin of Space Environment: Development, Challenges, Applications, and Future Outlook

This paper explores and discusses the revolutionary applications of digital twin technology in space environments and its profound impact on future space exploration activities. Originating from a proposal by the National Aeronautics and Space Administration (NASA) in 2002, digital twin technology aims to enhance the safety and reliability of space missions by creating precise virtual models. As the technology has evolved, its applications have successfully expanded beyond aerospace to include Industry 4.0, healthcare, and urban management, demonstrating remarkable cross-industry adaptability and broad impact. In space applications, digital twin technology can not only improve spacecraft design and maintenance processes but also enhance the efficiency of mission planning and execution. It plays a crucial role in astronaut training and emergency response as well. Particularly in extreme space conditions, this technology provides real-time monitoring and fault prediction, significantly enhancing mission safety and success rates. However, despite its recognized potential, the implementation of digital twins in space environments faces numerous challenges, including data transmission delays, model accuracy, and the design of user–system interactions. In the future, as artificial intelligence (AI) and machine learning (ML) technologies become mature and integrated, the digital twin will play a more central role in space missions, especially in remote operations, complex system management, and deep space exploration. This article is to overview key technical features, application examples, and challenges of digital twin technology, aiming to provide a comprehensive reference framework for researchers and developers while inspiring further in-depth studies and innovative applications.

RORα negatively regulates BCG-induced trained immunity

Trained immunity is a long-lasting change in the responsiveness of innate immune cells, leading to a stronger response upon an unrelated secondary challenge. Epigenetic, transcriptional, and metabolic reprogramming contribute to the development of trained immunity. By investigating the impact of gene variants on trained immunity responses after Bacillus Calmette–Guérin (BCG) vaccination, we identified a strong association between polymorphisms in the RORA gene and BCG-induced trained immunity in PBMCs isolated from healthy human donors. RORα, encoded by the RORA gene in humans, is a nuclear receptor and a transcription factor, regulating genes involved in circadian rhythm, inflammation, cholesterol, and lipid metabolism. We found that natural RORα agonists in the circulation negatively correlate with the strength of trained immunity responses after BCG vaccination. Moreover, pharmacological inhibition of RORα in human PBMCs led to higher cytokine production capacity and boosted trained immunity induction by BCG. Blocking RORα activity also resulted in morphological changes and increased ROS and lactate production of BCG-trained cells. Blocking lactate dehydrogenase A (LDHA) and glycolysis with sodium oxamate reduced the cytokine production capacity of cells trained with a combination of BCG and the RORα agonist. In conclusion, this study highlights the potential role of RORα in trained immunity, and its impact on human vaccination and diseases should be further investigated.

N of 1: Optimizing Methodology for the Detection of Individual Response Variation in Resistance Training.

Most resistance training research focuses on inference from average intervention effects from observed group-level change scores (i.e., mean change of group A vs group B). However, many practitioners are more interested in training responses (i.e., causal effects of an intervention) on the individual level (i.e., causal effect of intervention A vs intervention B for individual X). To properly examine individual response variation, multiple confounding sources of variation (e.g., random sampling variability, measurement error, biological variability) must be addressed. Novel study designs where participants complete both interventions and at least one intervention twice can be leveraged to account for these sources of variation (i.e., n of 1 trials). Specifically, the appropriate statistical methods can separate variability into the signal (i.e., participant-by-training interaction) versus the noise (i.e., within-participant variance). This distinction can allow researchers to detect evidence of individual response variation. If evidence of individual response variation exists, researchers can explore predictors of the more favorable intervention, potentially improving exercise prescription. This review outlines the methodology necessary to explore individual response variation to resistance training, predict favorable interventions, and the limitations thereof.

Assessment of Food Safety Knowledge Among Fresh Produce Growers and Extension Workers Before and After Training in Nepal

Access to safe food, including safe fresh produce, is essential to ensure healthy food and diets for the households and communities. However, there is limited awareness of food safety at household levels in Nepal. Food safety educational programs bridge the knowledge gap in addressing food safety. A pre- and post-assessment study was conducted during food safety trainings in five cities of Nepal to understand and evaluate the effectiveness of the trainings. Data from 102 fresh produce growers and 100 extension workers was collected through pre- and post- training responses. Questionnaires featuring 15 food safety statements were administered to the participants before and after the training. Then subsequent comparative analysis was done using the Wilcoxon signed ranks test, Mann Whitney test and Kruskal-Wallis test. The scores across subgroups: gender, age, respondent type and location were evaluated to test the significant differences. We found that the median knowledge score (indication of the knowledge or awareness level) elevated from 9.00 to 11.00 after the training. Our analysis suggested a statistically significant increase in food safety knowledge following the trainings (p<0.001). We found significant differences in the pre- and post-knowledge scores by gender (p<0.001), respondent type (p=0.005) and location (p<0.001). Our findings suggest that the targeted trainings to growers and extension workers could be an effective tool in enhancing food safety knowledge.

Visceral fat mass dynamics in a 2-year randomized STrength versus ENdurance training trial in people with obesity.

To compare the effectiveness of strength versus endurance training on reducing visceral fat in individuals with obesity. For the STrength versus ENdurance (STEN) 24-month randomized clinical trial, we assigned 239 participants with abdominal obesity to either strength or endurance training (two to three times a week, 60 min/training session) in addition to standard nutritional counselling to promote a healthy diet. Changes in abdominal visceral adipose tissue (VAT) area quantified by magnetic resonance imaging after 12 months were defined as a primary endpoint. Participants (aged 44 years, 74% women, body mass index: 37 kg/m2, mean VAT volume: 4050 cm3) had an approximately 50% retention rate and a 30% good training programme adherence at 12 months. There was no difference between strength and endurance training in VAT volume dynamics after 12 and 24 months (p = .13). Only in the good adherence group did we find a trend for reduced VAT volume in both training regimens. Independently of the exercise programme, there was a continuous trend for moderate loss of abdominal subcutaneous AT volume, body fat mass, body mass index and improved parameters of insulin sensitivity. Although parameters of physical fitness improved upon both exercise interventions, the dynamics of resting energy expenditure, glucose and lipid metabolism parameters were not different between the intervention groups and did not significantly improve during the 2-year trial (p > .05). Despite heterogeneous individual training responses, strength and endurance training neither affected VAT volume nor key secondary endpoints differently.

Dynamic characteristics of electromechanical coupling of body-suspended drive system for high-speed trains under wheel polygonal wear

Wheel polygonal wear can trigger intense wheel–rail interactions in railway vehicles, posing a significant threat to traffic safety. However, the mechanism and interaction of the influence of wheel polygonal wear on the electromechanical coupling characteristics of high-speed trains remain unclear. Therefore, this paper proposes a joint simulation modelling method that takes into account the integration of the electrical subsystem with the complete mechanical subsystem. A comprehensive analysis is conducted on the dynamic response of the high-speed train’s body-suspended drive system under a wide range of frequency excitations arising from harmonic torque, gear meshing, and wheel polygon wear. The results indicate that the electrical system generates a substantial amount of harmonic frequency components, resulting in a significant increase in the vibration of the train. The wheel polygon wear causes a broadening of the low-frequency resonance band, revealing a relatively prominent observation interval for the dynamic characteristics. Concurrently, it is discovered that the 23rd-order wheel polygon wear excites the system’s resonance frequency, and as the wavelength of the wheel polygon increases, the acceleration vibrations intensify progressively. Research reveals first the interaction of electrical system and mechanical system under wheel polygon wear, especially in the resonance gain characteristics under polygons and harmonic torques.

Validation of a scaled dynamic test system for simulating a high‐speed train passing bridges under seismic excitation

AbstractThis study presents the validation of a dynamic test system to simulate a high‐speed train passing bridges under seismic excitation. The system comprises scaled models of a CRH380A high‐speed train and an 11‐span simply supported bridge on a shake table array. This innovative apparatus combines seismic loading with the moving train load to replicate train‐track‐bridge interaction (TTBI) during earthquakes. It allows investigation of various train speeds and seismic excitations, providing invaluable insights into TTBI. First, the detailed similarity design principle based on the equations of motion was discussed, and its applicability to the wheel‐rail contact relationship was verified. Then, the dynamic characteristics of the scaled model were identified, and the impact of the error between the scaled model and the theoretical model on the TTBI response was assessed. Furthermore, the comparison of the dynamic test model and the numerical simulation in acceleration responses validated the accuracy of the rigid‐flexible coupling model method for the actual TTBI system. Test cases without external excitation, with simple harmonic excitation and with seismic excitation were conducted on the dynamic test system. Results showed that the influence of track irregularity and running speed on train response aligns with the classical train‐bridge interaction theory. The successful implementation of this test system marks a significant advance in understanding TTBI mechanics and has significant implications for seismic safety enhancement of railway infrastructure.

High-throughput screen identifies non inflammatory small molecule inducers of trained immunity

Trained immunity is characterized by epigenetic and metabolic reprogramming in response to specific stimuli. This rewiring can result in increased cytokine and effector responses to pathogenic challenges, providing nonspecific protection against disease. It may also improve immune responses to established immunotherapeutics and vaccines. Despite its promise for next-generation therapeutic design, most current understanding and experimentation is conducted with complex and heterogeneous biologically derived molecules, such as β-glucan or the Bacillus Calmette-Guérin (BCG) vaccine. This limited collection of training compounds also limits the study of the genes most involved in training responses as each molecule has both training and nontraining effects. Small molecules with tunable pharmacokinetics and delivery modalities would both assist in the study of trained immunity and its future applications. To identify small molecule inducers of trained immunity, we screened a library of 2,000 drugs and drug-like compounds. Identification of well-defined compounds can improve our understanding of innate immune memory and broaden the scope of its clinical applications. We identified over two dozen small molecules in several chemical classes that induce a training phenotype in the absence of initial immune activation-a current limitation of reported inducers of training. A surprising result was the identification of glucocorticoids, traditionally considered immunosuppressive, providing an unprecedented link between glucocorticoids and trained innate immunity. We chose seven of these top candidates to characterize and establish training activity in vivo. In this work, we expand the number of compounds known to induce trained immunity, creating alternative avenues for studying and applying innate immune training.

Seasonal variation in BCG-induced trained immunity

The Bacille Calmette-Guerin (BCG) vaccine is a well-established inducer of innate immune memory (also termed trained immunity), causing increased cytokine production upon heterologous secondary stimulation. Innate immune responses are known to be influenced by season, but whether seasons impact induction of trained immunity is not known. To explore the influence of season on innate immune memory induced by the BCG vaccine, we vaccinated healthy volunteers with BCG either during winter or spring. Three months later, we measured the ex vivo cytokine responses against heterologous stimuli, analyzed gene expressions and epigenetic signatures of the immune cells, and compared these with the baseline before vaccination. BCG vaccination during winter induced a stronger increase in the production of pro-inflammatory cytokines by peripheral blood mononuclear cells (PBMCs) upon stimulation with different bacterial and fungal stimuli, compared to BCG vaccination in spring. In contrast, winter BCG vaccination resulted in lower IFNγ release in PBMCs compared to spring BCG vaccination. Furthermore, NK cells of the winter-vaccinated people had a greater pro-inflammatory cytokine and IFNγ production capacity upon heterologous stimulation. BCG had only minor effects on the transcriptome of monocytes 3 months later. In contrast, we identified season-dependent epigenetic changes in monocytes and NK cells induced by vaccination, partly explaining the higher immune cell reactivity in the winter BCG vaccination group. These results suggest that BCG vaccination during winter is more prone to induce a robust trained immunity response by activating and reprogramming the immune cells, especially NK cells. (Dutch clinical trial registry no. NL58219.091.16)

Research on control parameters of high-speed maglev train under stochastic track irregularities

During operational phases, maglev trains encounter track irregularities, presenting a formidable challenge to the integrity of their control systems owing to the inherently stochastic nature, a challenge accentuated in the context of high-speed maglev trains. This study aims to comprehensively examine control parameters within the context of stochastic excitation resulting from track irregularities. Through the development of a rigorous stochastic response analysis model utilizing the pseudo excitation method, the investigation delves into the dynamics of an EMS-type high-speed maglev train-rigid track system. Evaluation indices, derived from the power spectral density of the maglev train's response, are established to quantify both control stability and operational stability. Subsequently, based on these indices, a recommended range of PD control parameters is proposed, accounting for the requirements stipulated by the evaluation metrics. The study elucidates that variation in PD control parameters exert discernible effects on the system's stability, primarily altering the nature frequency and damping ratio. Specifically, control stability demonstrates a positive correlation with increasing PD control parameters, while operational stability exhibits a nuanced relationship—initially bolstered by escalating proportional coefficients but subsequently tempered, albeit gradually, with heightened derivative coefficients. The delineated range of values for PD control parameters is meticulously determined, considering pertinent physical parameters of the electromagnetic system, such as mass, static suspension current, and static suspension gap, alongside factors like the Sperling indicator and suspension gap variation.

Novel gradient-enhanced Bayesian neural networks for uncertainty propagation

Uncertainty propagation (UP) is crucial for assessing the impact of input uncertainty on structural responses, holding significant importance in engineering applications. However, achieving accurate and efficient UP remains challenging, especially for highly nonlinear structures. Bayesian neural networks (BNN) have gained attention for addressing UP issues, yet current BNN models only utilize input samples and corresponding structural responses for training. However, incorporating gradients of structural responses with respect to input samples provides valuable information. This study proposes a novel approach called gradient-enhanced Bayesian neural networks (GEBNN) to tackle UP tasks. In the GEBNN, a modified evidence lower bound (MELBO) loss is developed to consider both structural responses and gradient information simultaneously. This includes disparities between actual and predicted responses, as well as disparities between actual and predicted derivatives. Additionally, a gradient screening strategy based on the marginal probability density functions (PDFs) of input samples is established to identify significant derivative data for GEBNN training. Once the GEBNN is configured, it is utilized to replace the computationally intensive finite element model to efficiently provide UP results. Various applications, including nonlinear numerical examples, and mechanical, civil, and aeronautical structures, are presented to demonstrate the effectiveness of the GEBNN. The results show that the GEBNN significantly enhances the computational accuracy of the BNN in solving UP tasks.