ABSTRACT A multifunctional electrode design is proposed for structural power composites (SPCs) by integrating high‐mass‐loading Zn‐ion hybrid supercapacitor (ZHSC) components with stainless steel (SS) mesh current collectors. A combined wet–dry processing route was developed to overcome fabrication challenges in thick electrodes, such as compositional inhomogeneity and drying‐induced cracking. The resulting freestanding activated carbon (AC) films were thermally laminated onto SS mesh to form ACSS electrodes with strong interfacial adhesion and improved electrical conductivity. A complementary ZnSS electrode was fabricated by calendering Zn foil onto SS mesh, enhancing mechanical robustness and electrical properties. The electrodes were assembled on opposite sides of an aramid nanofiber/glass fiber (ANFGF) composite electrolyte and subsequently protonated to enable strong interfacial bonding through the ANF network. The resulting ZHSC‐SPC device demonstrates capacitive charge storage, with areal capacitance scaling linearly with electrode stacking while retaining gravimetric performance. The device delivers a specific capacitance of 72.2 F g −1 , areal capacitance of 1284.3 mF cm −2 , energy density of 516 µWh cm −2 , and power density of 6.2 mW cm −2 , with 76% retention after 500 cycles. Mechanical tests confirm high performance, with a tensile strength of 114 MPa and an elastic modulus of 2.8 GPa, demonstrating suitability for structural energy storage applications.

Sandwich composite floor panels are widely used in aircraft interior structures because of their lightweight and high stiffness-to-weight characteristics. However, the guided-wave response of such panels is strongly influenced by their multilayer configuration, honeycomb core, attenuation behavior, sensor-path geometry, and excitation frequency. In this study, an active guided-wave-based Structural Health Monitoring (SHM) configuration was experimentally evaluated on an aerospace-type sandwich composite floor panel using a piezoelectric (PZT) sensor network. The specimen consisted of glass fiber reinforced polyetherimide (GFR-PEI) face sheets and a phenolic-coated aramid honeycomb core. Controlled bonded patch-induced surface perturbations were sequentially applied over 25 predefined panel regions to introduce repeatable local mass-loading and damping changes. Guided-wave measurements were performed using an Acellent ScanGenie system over a frequency range of 75–600 kHz with 25 kHz increments and twelve directed actuator–receiver paths. The results showed that the measured Damage Index (DI) response depends strongly on excitation frequency, sensing path, and perturbation location. The 400–450 kHz range produced relatively higher DI values under the tested configuration, and 425 kHz yielded the highest mean DI among valid measurements. However, the valid sensing coverage at 425 kHz was only 50%; therefore, this frequency was not interpreted as the most robust overall monitoring frequency. Lower frequencies around 100–150 kHz provided full sensing coverage while maintaining relatively high DI values. Frequencies above 550 kHz showed reduced measurement reliability due to increased attenuation and poor usable signal response. Overall, the study provides a comparative sensitivity assessment of a guided-wave-based PZT network on a sandwich composite floor panel under controlled bonded patch-induced perturbations, rather than a direct validation of realistic internal sandwich-panel damage mechanisms.

Structural Health Monitoring (SHM) systems are now mature and are being widely considered for implementation on commercial aircraft. SHM systems can provide immense benefits for commercial aircraft including providing savings in planned maintenance, increasing aircraft flying time by minimizing aircraft downtime and streamlining operational logistics. The pathway towards implementation will however need to address some key challenges including the availability of limited in-service data and the need to prove the viability of SHM systems compared to traditional inspection methods. The SHM system used to replace a conventional NDI method will need to demonstrate that the damage detection capability of the SHM system is equal to or better than that of the existing NDI method. In order to meet these requirements, it is imperative that the SHM system have the ability for robust and accurate damage detection in the aircraft flight environments. The SMART Layer PZT based SHM system can help address the challenges associated with SHM system certification and implementation. The SMART Layers utilize PZT transducers that are now widely used for crack detection and structural health monitoring in aerospace applications. The PZT SHM system is based on traditional NDI-UT physics but utilizes a different wave propagation mode (Guided waves). SMART Layers are thin, flexible, optimally designed PZT sensor networks that are integrated with the structure for maximum energy transmission and easy integration. In addition, the SMART Layers are designed to be installed in hidden and hard to access locations on the aircraft potentially replacing Low or High Frequency Eddy Current (LFEC, HFEC) and enabling structural inspection without structural disassembly provide robust data in multi-environments for any type of application provide “area coverage” for crack detection irrespective of the crack orientation and location in the area of concern minimize false calls that can lead to unscheduled maintenance enable sensor placement such that the area can still be validated using traditional NDI methods if needed In addition, the PZT SHM system has been tested over the years and can leverage historical data from previous aircraft installations precluding validation using alternate NDI methods. This presentation will discuss the advantages of the PZT SHM system in “actively” providing structural inspection information using multiple diagnostic methods developed for the sensor networks. This paper will demonstrate that the SMART Layer PZT system inherently provides multiple, independent diagnostic methods that operate simultaneously within a single structurally integrated sensor network. Using the same PZT array, several physics-distinct approaches can be utilized including guided-wave interrogation, impedance methods, scattering-based time-of-flight analysis, and damage imaging to characterize crack initiation and growth at designated inspection areas. Each diagnostic method is governed by a different physical mechanism, providing orthogonal and cross-validating indicators of structural degradation. These methods enable accurate detection of cracks in various flight environments. Historical and recently developed test data from multiple sources including coupon, flight tests and other non-commercial aircraft usage of the SHM system will be discussed to demonstrate a pathway towards the system certification while minimizing any false calls.

This presentation discusses the evolution of SMART Layer, based SHM system from a targeted inspection aid to a key enabler of IVHM and CBM strategies. Lessons learned from fielded rotorcraft applications are discussed, along with a practical path for integrating the SHM system components into both sustainment programs and future aircraft designs. The role of automation, data management, and health-state awareness in supporting aircraft readiness and lifecycle optimization is will also be discussed.

Transverse matrix cracking is the predominant failure mode in the early stages of progressive degradation, making its accurate identification crucial for ensuring the safety of carbon fiber reinforced plastics (CFRP) composite structures. However, when models trained on guided wave sensor data from one structure are applied to predict crack density in another structure with a different ply orientation, conventional deep learning methods face significant challenges due to the sensitivity of this regression task to feature scale and ply orientation. To address these challenges, we propose a domain adaptive relational graph convolutional network (DA-RGCN) model specifically designed for crack density prediction, leveraging deep domain adaptation to transfer identification knowledge learned from one laminate to another with completely different ply orientation. First, crack damage-related features are automatically extracted from sensor signals by capturing their temporal relationships during guided wave propagation. These features, along with geometric information from sensor networks, are embedded as node features within a graph structure, allowing for the learning of enhanced feature representations through the fusion of information from neighboring nodes. Subsequently, the fused features are utilized to identify crack density along each path by measuring its spatial distance from two reference states (baseline and saturation). Additionally, we employ a representation subspace distance based on principal angles to minimize distribution discrepancies between features without altering their scales. As a result, combined with the physical guidance from the damage index model, the extracted features achieve domain invariance, significantly enhancing the cross-structural generalization of the DA-RGCN. To validate the model’s capability for cross-structural identification of transverse matrix cracks, we designed transfer tasks between two layups using the CFRP Composites Dataset. The results indicate that the proposed DA-RGCN achieves an average root mean square error of 1.1841 in crack density identification, demonstrating the lowest error compared to other deep transfer learning-based and purely physics-based methods.

Advances in computing and machine learning methods have led to a rapid rise in artificial intelligence (AI) research and applications in many fields. AI research benefitted from advances in computation hardware, collection and distribution of large data sets, and proliferation of software techniques. AI techniques include machine learning for provable results, deep learning for data exploration, reinforcement learning for control, and active learning for adaptive systems. Likewise, AI algorithms can handle large amounts of data, construct unknown representations, and provide a direct link between data and classification for decision making. These unmatched capabilities have been seen as a path to solving hard engineering problems, including that of structural health monitoring (SHM). SHM consists of automating the condition assessment task of civil, health, mechanical, and aerospace systems using measurements obtained from temporary or permanently installed sensors. Often, the systems of interest are geometrically large and/or technically complex, which complicates the development and application of physics-based methods. It follows that AI is seen as a key potential contributor enabling SHM in field applications for data-driven analysis. As with many research endeavors, many concepts using AI for SHM have been explored in the literature. Nevertheless, very few AI methods have been deployed in the context of SHM, which may be due to the lack of available data supporting their capabilities, limited integrated AI-SHM systems capable of providing results to users and operators with decision-making capabilities, or certification of AI methods for safety-critical applications. The objective of this Roadmap publication is to discuss the integration of AI at the system level enabling SHM, including associated challenges and opportunities such as those found in common metrics of concern (e.g., transparency, interpretability, explainability, security, certifiability, etc.), with a particular focus on providing a path to research and development efforts that could yield impactful field applications. The overview of available methods and directions will provide the readers with applicability of AI for certain SHM designs (software), availability of common data sets for further AI comparisons (data), and lessons learned in implementation (hardware).

Background: Several artificial intelligence–enhanced electrocardiogram (AI-ECG) models have shown promise in detecting left ventricular systolic dysfunction (LVSD), but their head-to-head agreement and performance have not been independently compared within the same cohort. Objectives: This study aimed to compare the performance of published AI-ECG models for LVSD detection in a standardized external cohort and evaluate the field’s transparency and reproducibility. Methods: We systematically reviewed AI-ECG models predicting LVSD and assessed the risk of bias. Authors were invited to share models for external validation in a well-phenotyped registry of patients undergoing routine clinical cardiac magnetic resonance imaging with cardiologist-adjudicated reports and paired ECGs. Model performance was evaluated in all consecutive patients and a lower-complexity subgroup with 15% LVSD prevalence. Results: We identified 35 studies describing 51 models, reporting high (area under the receiver-operating characteristic curve [AUROC] >0.80) or excellent (AUROC >0.90) performance. The risk of bias is high and primarily attributed to the limited description of development and validation cohort characteristics, as well as the lack of independent external validation. Four groups (from Korea, the United States, Taiwan, and the Netherlands) shared models for independent testing. AUROCs ranged from 0.83 to 0.93 in all patients (n = 1,203; mean age 59 ± 15 years; 450 [35%] female) and from 0.87 to 0.96 in the lower-complexity subset. Performance remained consistent across subgroups, with slight decreases in ECGs showing wide QRS complexes or atrial fibrillation. Conclusions: In this first-in-kind independent validation and head-to-head comparison study, AI-ECG for LVSD detection demonstrated strong performance despite training on disparate populations. However, the limited-availability of models hinders independent validation.

Background/Objectives: To develop and validate a model system using deep learning algorithms for the automatic detection of type A aortic dissection (AD), and differentiate it from normal and type B AD patients. Methods: In this retrospective study, a deep learning model is developed, based on aortic computed tomography angiography (CTA) scans of 498 patients using training, validation and test sets of 398, 50 and 50 patients, respectively. An independent test set of 316 patients is used to validate and evaluate its performance. Results: Our model comprises two components. The first one is an objection detection model, which can identify the aorta from CTA. The second one is a dissection classification model, which can automatically detect the presence of aortic dissection and determine its type based on Stanford classification. Overall, the sensitivity and specificity for Type A AD were 0.969 and 0.982, for Type B AD were 0.946 and 0.996 and for normal cases were 0.988 and 1.000, respectively. The average processing time per CTA scan was 7.9 ± 2.8 s. (mean ± standard deviation). Conclusions: This deep learning automatic model can accurately and quickly detect type A AD patients, and could serve as an imaging triage in an emergency setting and facilitate early intervention and surgery to decrease the mortality rates of type A AD patients.

Deep learning analysis of electrocardiography (ECG) may predict cardiovascular outcomes. We present a novel multi-task deep learning model, the ECG-MACE, which predicts the one-year first-ever major adverse cardiovascular events (MACE) using 2,821,889 standard 12-lead ECGs, including training (n = 984,895), validation (n = 422,061), and test (n = 1,414,933) sets, from Chang Gung Memorial Hospital database in Taiwan. Data from another independent medical center (n = 113,224) was retrieved for external validation. The model’s performance achieves AUROCs of 0.90 for heart failure (HF), 0.85 for myocardial infarction (MI), 0.76 for ischemic stroke (IS), and 0.89 for mortality. Furthermore, it outperforms the Framingham risk score at 5-year MACEs and 10-year mortality prediction. Over 10-year follow-ups, the model-predicted-positive group exhibits significantly higher MACE incidences than the model-predicted-negative group (relative incidence ratio: HF: 15.28; MI: 7.87; IS: 4.74; mortality: 13.18). Using solely ECGs, ECG-MACE effectively predicts one-year events and exhibits long-term anticipation. It provides potential applications in preventive medicine.

Rheumatoid arthritis (RA) is distinguished by the presence of modified bone microarchitecture, also known as 'texture,' in the periarticular regions. The radiographic detection of such alterations in RA can be challenging. To train and to validate a deep learning model to quantitatively produce periarticular texture features directly from radiography and predict the diagnosis of early RA without human reading. Two kinds of deep learning models were compared for diagnostic performance. Anterior-posterior bilateral hands radiographs of 891 early RA (within one year of initial diagnosis) and 1237 non-RA patients were split into a training set (64%), a validation set (16%), and a test set (20%). The second, third, and fourth distal metacarpal areas were segmented for the Deep Texture Encoding Network (Deep-TEN; texture-based) and residual network-50 (ResNet-50; texture and structure-based) models to predict the probability of RA. The area under the curve of the receiver operating characteristics curve for RA was 0.69 for the Deep-TEN model and 0.73 for the ResNet-50 model. The positive predictive values of a high texture score to classify RA using the Deep-TEN and ResNet-50 models were 0.64 and 0.67, respectively. High mean texture scores were associated with age- and sex-adjusted odds ratios (ORs) with 95% confidence interval (CI) for RA of 3.42 (2.59-4.50) and 4.30 (3.26-5.69) using the Deep-TEN and ResNet-50 models, respectively. The moderate and high RA risk groups determined by the Deep-TEN model were associated with adjusted ORs (95% CIs) of 2.48 (1.78-3.47) and 4.39 (3.11-6.20) for RA, respectively, and those using the ResNet-50 model were 2.17 (1.55-3.04) and 6.91 (4.83-9.90), respectively. Fully automated quantitative assessment for periarticular texture by deep learning models can help in the classification of early RA.

BACKGROUND:The stroke risk in patients with subclinical atrial fibrillation (AF) is underestimated. By identifying patients at high risk of embolic stroke, health-care professionals can make more informed decisions regarding anticoagulation treatment to prevent stroke. The main aim of this study was to forecast the risk of AF both retrospectively and prospectively. METHODS:The research used a dataset of patients who had received a standard 12-lead electrocardiogram (ECG) at the seven branches of Chang Gung Memorial Hospital between October 2007 and December 2019. Using convolutional neural network (CNN) ECG models, the study classified the risk of AF development both retrospectively and prospectively in 1,776,968 patients by analyzing their 12-lead ECG. The study also examined the risk of stroke, hospitalization for heart failure (HF), myocardial infarction (MI), and death among patients with predicted AF versus that of those with normal sinus rhythm. RESULTS:The CNN models could be used to accurately diagnose AF, assess the risk of past AF episodes, and predict the risk of future AF episodes with high accuracy, as shown by areas under the receiver operating characteristic curve of 0.99, 0.86, and 0.85, respectively. Patients who were estimated to have had past AF or predicted to have future AF were at a higher risk of developing stroke, HF hospitalization, MI, and mortality. The ECGs of patients with predicted AF tended to exhibit lower R-wave amplitudes and flattened T waves. Additionally, we observed that the QRS complexes in leads V1, aVL, and aVR were highly weighted in predicting AF in the CNN models. CONCLUSIONS:The CNN models were effective for estimating the past and future risk of AF by analyzing 12-lead ECG. Patients with predicted AF had a higher risk of developing stroke, hospitalization for HF, MI, and death. By using this AF prediction model, physicians may be able to identify patients who should be screened for AF and taking action to prevent stroke and manage cardiovascular risk.

It is crucial to accurately monitor the performance, health, and lifespan of lithium-ion batteries to ensure reliable, efficient, and on-demand delivery of stored electrical energy for hybrid and electric vehicle technologies. This paper presents a method to monitor the state of charge (SoC) and state of health (SoH) of lithium-ion batteries by utilizing ultrasonic guided wave propagation signals. The lithium-ion battery is modeled as a time- varying single-degree-of-freedom vibrating system and the time-varying nature of the system is captured by successive utilization of the time-invariant autoregressive models. The proposed technique focuses on extracting the time-dependent natural frequencies and damping ratios of the lithium-ion batteries from the ultrasonic guided wave signals through the use of autoregressive model-based modal analysis techniques. The extracted natural frequencies and damping ratios are then correlated with the battery SoC and SoH to assess the battery conditions accurately. Lengthwise signals exhibit a decreasing trend in natural frequencies as the SoC increases, whereas, the thickness direction signals show the opposite trend, frequencies increase with rising SoC. This method elucidates the potential of ultrasonic guided wave-based real-time monitoring of battery SoC and SoH in its life cycle.

This study evaluated the performance of a deep learning model trained to detect scaphoid fractures in radiographs when various perturbations were added to the images. The datasets were modified by applying Gaussian noise; via blurring, JPEG compression, contrast-limited adaptive histogram equalization (CLAHE), and resizing; and the addition of geometric offsets. Model accuracy declined as the severity of perturbations increased; however, the extent of performance decline varied according to the type of perturbation. The model demonstrated greater resistance to color perturbations than to grayscale perturbations, but the application of Gaussian blur exerted a considerably negative impact on model performance. CLAHE increased the false-positive rate. There was a strong linear relationship between image quality and model performance; the model performed better on higher-quality images. We also found that geometric offset or pixel value rescaling did not affect the performance of the deep learning model. Resolution was the most significant factor influencing model performance; localizing the region of interest may minimize any decrease in accuracy. Overall, this study provides insights into the robustness of deep learning models that detect small fractures, such as scaphoid fractures, in radiographs subjected to various perturbations. The findings could guide the development of more accurate models for medical image analysis. The cumulative insights gained from this study will contribute to the design of more accurate and robust deep learning models tailored for medical image analysis, especially under conditions of varying image quality.

Inspired by avian species, "Fly-by-Feel" introduces a data-driven, physics-aware method for aircraft flight state awareness. In contrast to conventional approaches that rely on discrete, specialized sensors, this novel technique uses hybrid modeling algorithms to process distributed, multimodal sensor data and infer global aerodynamic state. Signals from an ultra-low-profile sensor network mounted on the wing(s) are continuously recorded by high-precision data-acquisition electronics, capturing the static and dynamic stresses at multiple locations. This uniquely rich data set is used to train inference models and generate real-time estimates of safety-critical flight conditions, key piloting variables, and aerodynamic metrics. A modeling architecture based on Fourier neural operators (FNOs) is designed to facilitate "learning" fluid-structure interaction and frequency-domain features. By incorporating a physical awareness lacking in purely data-driven approaches and the discretization-invariance missing from conventional neural-network-based approaches, such models are particularly well-suited for describing flight mechanics. Fly-by-Feel's application is demonstrated through wind-tunnel testing of a commercial unmanned aerial system (UAS) retrofitted with sensing hardware. When its accuracy, safety, and robustness metrics are compared to competing systems, the Fly-by-Feel system's outputs match or exceed the capabilities of the UAS's built-in systems under a vast majority of conditions across the flight envelope while also comparing remarkably well to wind-tunnel measurements. A sensitivity study of models trained on restricted datasets further underscores the advantages of the bio-inspired system's unique multimodal sensing and multi-local to global properties.

This study investigates the interaction between ultrasonic guided waves ( GW) propagation and bond quality of induction-welded thermoplastic composite joints. Using ultrasonic piezoelectric transducers, the proposed robust structural health monitoring (SHM) method established a strong correlation between ultrasonic signals and the health status of joint. Three joint batches were manufactured with varying shear strength by adjusting welding parameters. Batch 1 had superior properties, while batches 2 and 3 were intermediate and lower, respectively. Transducers were placed on opposite sides of the joint overlap, operating at 350 kHz and 450 kHz. Time of flight (ToF) measurements, indicating wave group velocity, showed a direct correlation with weld-line stiffness and wave propagation velocity. These findings significantly contribute to the understanding of the interaction between mechanical properties and guided wave propagation in induction-welded thermoplastic composite joints, providing valuable insights for monitoring the performance of such joints in engineering applications.

Purpose Osteoporosis, affecting over 200 million individuals, often remains unrecognized and untreated, increasing the risk of fractures in older adults. Osteoporosis is typically diagnosed with bone mineral density (BMD) measured by dual-energy X-ray absorptiometry (DXA). This study aims to develop DeepDXA-Hand, a deep learning model using the efficient CNN-based deep learning architecture, for opportunistic osteoporosis screening from hand radiographs. Methods DeepDXA-Hand utilizes a CNN-based, HarDNet, approach to predict BMD non-invasively. A total of 10,351 hand radiographs and DXA pairs were used for model training and validation. The model's interpretability was enhanced using GradCAM for hotspot analysis to determine the model's attention areas. Results The predicted and ground truth BMD were significantly correlated with a correlation coefficient of 0.745. For binary classification of osteoporosis, DeepDXA-Hand demonstrated a sensitivity of 0.73, specificity of 0.83, and accuracy of 0.80, indicating its clinical potential. The model mainly focused on the carpal bones, such as the capitate, trapezoid, hamate, triquetrum, and the head of the second metacarpal bone, suggesting these areas provide radiological features for inferring BMD. Conclusion DeepDXA-Hand shows potential for the early detection of osteoporosis with high sensitivity and specificity. Further studies should explore its utility in predicting fracture risks. Mini abstract Osteoporosis affects millions and often goes undetected and untreated. DeepDXA-Hand, a HarDNet-based deep learning model, predicted bone mineral density with a correlation of 0.745 and classified osteoporosis with 0.80 accuracy. This model enhances early detection and has significant clinical potential as osteoporosis opportunistic screening tool.

BackgroundElectrocardiogram (ECG) abnormalities have demonstrated potential as prognostic indicators of patient survival. However, the traditional statistical approach is constrained by structured data input, limiting its ability to fully leverage the predictive value of ECG data in prognostic modeling.MethodsThis study aims to introduce and evaluate a deep-learning model to simultaneously handle censored data and unstructured ECG data for survival analysis. We herein introduce a novel deep neural network called ECG-surv, which includes a feature extraction neural network and a time-to-event analysis neural network. The proposed model is specifically designed to predict the time to 1-year mortality by extracting and analyzing unique features from 12-lead ECG data. ECG-surv was evaluated using both an independent test set and an external set, which were collected using different ECG devices.ResultsThe performance of ECG-surv surpassed that of the Cox proportional model, which included demographics and ECG waveform parameters, in predicting 1-year all-cause mortality, with a significantly higher concordance index (C-index) in ECG-surv than in the Cox model using both the independent test set (0.860 [95% CI: 0.859- 0.861] vs. 0.796 [95% CI: 0.791- 0.800]) and the external test set (0.813 [95% CI: 0.807- 0.814] vs. 0.764 [95% CI: 0.755- 0.770]). ECG-surv also demonstrated exceptional predictive ability for cardiovascular death (C-index of 0.891 [95% CI: 0.890- 0.893]), outperforming the Framingham risk Cox model (C-index of 0.734 [95% CI: 0.715-0.752]).ConclusionECG-surv effectively utilized unstructured ECG data in a survival analysis. It outperformed traditional statistical approaches in predicting 1-year all-cause mortality and cardiovascular death, which makes it a valuable tool for predicting patient survival.