Search Results (77)

The Qingshankou Formation (K₂qn) represents a key interval for lacustrine shale oil accumulation in the Songliao Basin. However, the spatial heterogeneity of organic-rich shales and their controlling mechanisms remain poorly constrained. Here, we investigate the Qijia–Gulong and Sanzhao sags by integrating drilling, well-log, geochemical, and mineralogical data to systematically evaluate source rock characteristics and their dominant controls. Based on well-log data from 442 wells, total organic carbon (TOC) was continuously predicted using an improved ΔlogR method. In addition, mineral compositions and lithofacies distributions were quantitatively characterized for representative wells in the eastern and western sags by combining X-ray diffraction (XRD) data with a deep residual shrinkage network (DRSN) model. The results reveal a dual depocenter pattern within K₂qn across the study area. The Qijia–Gulong Sag is characterized by thicker mudstone successions (30–600 m), higher sedimentation rates, and stronger stratigraphic continuity, whereas the Sanzhao Sag exhibits comparatively thinner deposits (30–300 m). Significant differences are also observed in organic matter type and thermal maturity: the Qijia–Gulong Sag is dominated by Type II₁ kerogen with higher maturity (R_o = 1.0%–1.5%), while the Sanzhao Sag mainly contains Type I kerogen with relatively lower maturity (R_o = 0.8%–1.3%). Despite this, TOC values in the Sanzhao Sag are markedly higher than those in the Qijia–Gulong Sag, with average values of 3.34% and 2.19%, respectively. These differences reflect the coupled control of palaeoenvironmental conditions and terrigenous input on organic matter enrichment. Elevated salinity and enhanced water-column stratification in the Sanzhao Sag promoted the development of reducing conditions favorable for organic matter preservation, resulting in higher TOC contents. In contrast, although the Qijia–Gulong Sag experienced high sedimentation rates and developed thick shale sequences, strong terrigenous input and dilution effects limited organic matter enrichment, while simultaneously leading to higher thermal maturity. Consequently, two distinct enrichment modes are identified in the study area: a “high-salinity stratification–efficient preservation” mode and a “high maturity–thick shale development” mode, which together govern the spatial heterogeneity of shale oil resources. Full article

This study proposes a stress field data-driven prediction method that combines a finite element thermo-mechanical coupling model with a multi-machine learning framework. This method takes the inversion of stress based on the shrinkage behavior of deposition layers as the core logic, extracts the node displacement shrinkage during the cooling to solidification process of the melt pool in the thermal coupling simulation as the key feature input, and constructs extreme gradient boosting (XGBoost), Gaussian process regression (GPR), and deep convolutional neural network (DCNN) models, respectively, to achieve accurate prediction of nodal effect stress and triaxial stress in the laser directed energy deposition (L-DED) node process. The experimental results show that the XGBoost algorithm performs the best in various stress prediction indicators, and its generated stress distribution cloud map is highly consistent with the thermal coupling simulation results, suggesting a strong correlation between deposition layer shrinkage behavior and the stress field under the investigated conditions. In addition, compared to traditional finite element simulations, this method significantly improves computational efficiency while ensuring prediction accuracy, providing a new approach for rapid assessment of residual stresses. Full article

Quantitative trading systems require predictive models that simultaneously deliver accurate forecasts, calibrated uncertainty quantification, and actionable risk measures. This paper proposes an information-theoretic semiparametric regression framework combining a convolutional neural network–Transformer (CNN–Transformer) network for nonlinear temporal dependencies with a Gaussian process (GP) prior for residual autocorrelation and calibrated predictive distributions. Three theoretical results are established: an identifiability theorem guarantees joint recoverability of the nonparametric and GP components; a consistency theorem showing that the penalised maximum likelihood estimator converges at a rate $n^{-1/(2+d_{\mathrm{eff}})}$; and a coverage theorem proving asymptotic nominal coverage of the GP’s credible intervals. The framework enables an entropy-regulated trading module where predictive differential entropy informs position sizing via an uncertainty-penalised Kelly criterion, Kullback–Leibler divergence quantifies model uncertainty, and CVaR-constrained optimisation controls the tail risk. Simulations show the method outperforms the CNN, long short-term memory (LSTM), Transformer, XGBoost, random forest, least absolute shrinkage and selection operator (LASSO), and standard GP regression approaches. Backtesting on four Chinese A-share stocks yielded annualised returns of 15.9–22.4% with Sharpe ratios of 0.49–0.62, maximum drawdowns below 15%, and daily 95% CVaR reductions of 28–31% relative to a full-Kelly baseline, confirming both predictive accuracy and risk management effectiveness. Full article

This study examines how assessments of coastal extreme sea levels depend on the separability and reconstructability of the astronomical tide in hourly sea-level records. Using a global tide-gauge network, it proposes an ensemble-learning correction framework that integrates a physical-baseline threshold with multi-criteria consistency testing to determine whether machine-learning enhancement is genuinely effective across stations and time windows. The analysis uses hourly records from 528 UHSLC tide gauges, with 31-day short sequences used to reconstruct 180-day sea-level variability. Taking the physical tidal model as the baseline, residuals are corrected using Extremely Randomized Trees, Random Forest, and Gradient Boosting. To avoid false improvement driven solely by error reduction, a hierarchical decision framework is established. Baseline model quality is first screened using NSE and the coefficient of determination, after which mathematical artefacts are identified through diagnostics of peak suppression and variance shrinkage. A five-level classification is then derived from the convergent evidence of twelve performance metrics and four statistical significance tests. The results show a consistent global pattern across all three algorithms. Approximately 57% of stations meet the criterion for genuine improvement, whereas about 42% are associated with an unreliable physical baseline, indicating that the dominant source of failure arises not from the ensemble-learning algorithms themselves, but from spatially varying limitations in the underlying physical baseline. Spatially, the credibility of machine-learning correction is strongly conditioned by baseline quality: stations with effective correction are more continuous along the eastern North Atlantic and European coasts, whereas stations with ineffective correction are more concentrated in the Gulf of Mexico, the Caribbean, and the marginal seas and archipelagic regions of the western Pacific. These results indicate that the observed spatial heterogeneity primarily reflects geographically varying physical and dynamical conditions that control baseline reliability and residual learnability, rather than a standalone difference in the intrinsic capability of ensemble learning itself. Full article

As a key technology for wireless signal identification, modulation recognition plays an important role in the fields of unmanned aerial vehicle (UAV) communications, low-altitude spectrum management, etc. However, the accuracy of modulation recognition often cannot be guaranteed in scenarios with serious noise interference when a few samples are available. In this paper, we propose an intelligent modulation recognition method for UAV signals based on small-sample augmentation and soft threshold denoising. We first propose a new dual-driven dataset expansion method by combining the UAV air–ground channel propagation model with the received data samples. Then, we construct a background learning-based long short-term memory (BL-LSTM) model to extract the environmental background features embedded in the UAV signal, including Line-of-Sight (LoS) state, multi-scale fading parameters and Doppler shift characteristics. We integrate environmental background information into the data training model and optimize the authenticity of data distribution. As a result, the model adaptability can be enhanced. Finally, we construct a deep residual shrinkage network based on the soft threshold function (STF-DRSN). By leveraging the capability of the soft threshold that resists noise interference, we integrate it into each residual block of the deep residual shrinkage network. Simulation results show that compared with the state of the art, our method can improve the modulation recognition accuracy of UAV signals in small-sample scenarios. Full article

To investigate how the content of silica fume (SF) influences the performance of foam concrete (FC) after high-temperature exposure and the underlying mechanisms, this study prepared standard FC cube specimens with SF contents of 0%, 0.15%, 0.2%, 0.25%, and 0.3%. The working properties of the material at room temperature were systematically tested, and the mass loss, residual compressive strength, failure mode, microstructure and acoustic emission (AE) data at different temperatures (100 °C, 200 °C, 300 °C and 400 °C) were analyzed. The test results indicate that increasing the SF content reduces the fluidity of the fresh paste yet significantly enhances the compressive strength and lowers the water absorption of FC at room temperature. After high-temperature exposure, the effect of SF exhibits a dual character: at 200 °C and below, SF effectively mitigates the performance degradation of FC. However, when the temperature reaches 300–400 °C, specimens with an excessively high SF content (e.g., 0.3%) experience rapidly built-up internal steam pressure that cannot escape in time, which triggers the formation and propagation of a microcrack network and leads to a sharp drop in strength. Based on AE detection and scanning electron microscopy (SEM) image analysis, the failure process of silica fume foam concrete (SFFC) proceeds through three stages: free water evaporation at low temperatures, dehydration shrinkage of the C-S-H gel at medium temperatures, and finally, structural failure marked by the collapse of the C-S-H gel network at high temperatures. This study indicates that an SF content of 0.25% allows FC to achieve an optimal balance between mechanical properties and high-temperature stability. The findings provide a theoretical basis for optimizing FC mix proportions and enhancing fire prevention design. Full article

Camber in high-speed railway prestressed concrete (PC) girders increases with service time and affects profile control, ride comfort, and durability; reliable long-term midspan camber prediction is therefore required. Building on established hybrid physics–data modeling and discrepancy-correction ideas, we present a monitoring-oriented two-layer strategy for long-term camber prediction. In the physics layer, a physics-informed neural network (PINN) is formulated in a quasi-static, stage-aware manner to capture the physics-consistent low-frequency trend governed by creep, shrinkage, prestress loss, and staged loading. In the data layer, an XGBoost model learns a bounded, measurement-level residual correction from monitoring features to account for additional effects not explicitly represented in the physics layer, without altering the underlying physics-driven trend. The approach is evaluated using monitoring data from five 1:4 scaled specimens of a 24 m post-tensioned simply supported box girder and is compared against a theoretical calculation and a standalone PINN. Across prediction stages and specimens, the proposed strategy reproduces the measured camber evolution more closely than the benchmarks while preserving physically plausible trend behavior and yielding more consistent errors among girders. These results indicate that, under the present scaled-specimen and independently calibrated setting, a stage-aware physics baseline combined with bounded residual correction can provide closer agreement with the observed camber evolution than the benchmark models under sparse-monitoring conditions. Its engineering applicability can be repeatedly demonstrated across girders with different construction-condition combinations after girder-wise calibration. Full article

High-power ultrasound has been widely used to regulate the anti-digestibility of starch-based products, including the formation of resistant starch (RS-V) in amylopectin assemblies. This can contribute to the attenuation of postprandial hyperglycemia. However, the mechanisms by which high-power ultrasound modulates RS-V remain to be elucidated. Therefore, nano-scale Euryale ferox amylopectin (EFA) ternary assemblies were constructed under high-power ultrasound. All EFA assemblies exhibited ternary self-assembly peaks and V-type crystallinity. Combined chemometric analyses revealed that, with increasing ultrasound power, the rising self-assembly sites within B2 and C chains promoted the increase in self-assembly index but decreased semicrystalline lamellae thickness and structural fractal dimension. Consequently, a compact and ordered molecular cross-linking network was formed, contributing to increases in residual crystallinity, molecular weight, short-range order, and molecular density. This resulted in the shrinkage of digestion channel structures and optimization of the molecular gel network. As a result, the reduction in hydrolysis sites with increasing ultrasound power led to increased RS-V content (22.66–60.17%), causing a decline in the estimated glycemic index. The EFA–lauric acid–lactoglobulin assemblies prepared under 600 W ultrasound were the optimal composition and exhibited enhanced anti-digestibility relative to amylopectin assemblies derived from staple crops such as white waxy maize. The present investigation not only serves as a valuable supplement for studying the precise regulation mechanisms of nano-scale amylopectin RS-V, but also provides critical theoretical guidance for the development of foods aimed at preventing hyperglycemia. Full article

To address the challenges of fault identification in renewable energy plant outgoing lines within “double-high” power systems, this paper proposes a novel parallel dual-channel method that fuses time-series signals and images. On one hand, the fault current signals from the renewable energy plant outgoing lines are acquired and fed into a constructed Multi-scale Adaptive Residual Shrinkage Network (MARS-Net) for one-dimensional temporal feature extraction. On the other hand, one-dimensional fault data is transformed into two-dimensional images via a Relative Angle Matrix (RAM). The generated 2D image data is then input into a network incorporating Dynamic Convolution (D-Conv) and a Transformer-enhanced MobileNetV3 (DT-MobileNetV3) for spatial feature extraction. Finally, feature fusion of the one-dimensional and two-dimensional information is performed to achieve fault type identification. To comprehensively evaluate the method’s performance, this paper designs experiments including noise interference tests, multi-network comparative experiments, ablation studies, comparisons of different 2D transformation methods and data loss. The results demonstrate that the proposed method possesses significant advantages in terms of identification accuracy, noise immunity, data loss tolerance, and generalization capability. Full article

To address the durability deficiencies and limited service life of concrete structures exposed to complex service environments such as chloride attack in marine and underground engineering, this study employs fly ash (FA) and ground granulated blast-furnace slag (GGBS), typical eco-friendly materials, as functional mineral admixtures to systematically investigate the effects of their combined incorporation on the mechanical properties, durability, drying shrinkage, and microstructural characteristics of concrete. The objective is to develop a concrete material that achieves high durability while maintaining structural safety and service performance, with the additional benefit of improved resource utilization efficiency. Single-factor tests were first conducted to determine the sensitivity ranges of FA and GGBS within 10–30% for slump, compressive strength, chloride migration coefficient (RCM), and drying shrinkage. Subsequently, response surface methodology (RSM) was employed to establish quadratic regression models using FA and GGBS as independent variables and compressive strength, RCM, and drying shrinkage as response indicators. The models exhibited high fitting accuracy, and their reliability was validated through analysis of variance (ANOVA), residual analysis, and predictive performance indices. Multi-objective optimization based on the desirability function identified the optimal mix proportion as FA = 14.8% and SL = 29.3%, yielding predicted values of 56.2 MPa for 28-day compressive strength, 6.03 × 10⁻¹² m²/s for RCM, and 639 με for 90-day drying shrinkage. Microstructural analysis using SEM and MIP further revealed that the binary-blended system promotes the formation of a dense C–S–H/C–A–S–H gel network, refines pore-size distribution, and reduces pore connectivity, thereby improving long-term mechanical and durability performance. The findings provide quantitative guidance for designing high-durability, environmentally friendly concrete suitable for marine and underground engineering applications. Full article

With the ongoing miniaturization of smart devices, fine-grained hand gesture recognition using millimeter-wave radar has attracted increasing attention, yet practical deployment remains challenging in continuous-gesture segmentation, robust feature extraction, and reliable classification. This paper presents an end-to-end fine-grained gesture recognition framework based on frequency modulated continuous wave(FMCW) millimeter-wave radar, including gesture design, data acquisition, feature construction, and neural network-based classification. Ten gesture types are recorded (eight valid gestures and two return-to-neutral gestures); for classification, the two return-to-neutral gesture types are merged into a single invalid class, yielding a nine-class task. A sliding-window segmentation method is developed using short-time Fourier transformation(STFT)-based Doppler-time representations, and a dataset of 4050 labeled samples is collected. Multiple signal classification(MUSIC)-based super-resolution estimation is adopted to construct range–time and angle–time representations, and instance-wise normalization is applied to Doppler and range features to mitigate inter-individual variability without test leakage. For recognition, a variable-channel deep residual shrinkage network (DRSN) is employed to improve robustness to noise, supporting single-, dual-, and triple-channel feature inputs. Results under both subject-dependent evaluation with repeated random splits and subject-independent leave one subject out(LOSO) cross-validation show that DRSN architecture consistently outperforms the RefineNet-based baseline, and the triple-channel configuration achieves the best performance (98.88% accuracy). Overall, the variable-channel design enables flexible feature selection to meet diverse application requirements. Full article

To address the computational redundancy, inadequate multi-scale feature capture, and poor noise robustness of traditional deep networks used for bearing vibration and acoustic signal feature extraction, this paper proposes a fault diagnosis method based on Depthwise Separable Atrous Convolution (DSAC) and Acoustic Spatial Pyramid Pooling (ASPP). First, the Continuous Wavelet Transform (CWT) is applied to the vibration and acoustic signals to convert them into time–frequency representations. The vibration CWT is then fed into a multi-scale feature extraction module to obtain preliminary vibration features, whereas the acoustic CWT is processed by a Deep Residual Shrinkage Network (DRSN). The two feature streams are concatenated in a feature fusion module and subsequently fed into the DSAC and ASPP modules, which together expand the effective receptive field and aggregate multi-scale contextual information. Finally, global pooling followed by a classifier outputs the bearing fault category, enabling high-precision bearing fault identification. Experimental results show that, under both clean data and multiple low signal-to-noise ratio (SNR) noise conditions, the proposed DSAC-ASPP method achieves higher accuracy and lower variance than baselines such as ResNet, VGG, and MobileNet, while requiring fewer parameters and FLOPs and exhibiting superior robustness and deployability. Full article

To overcome the challenge of degraded classification accuracy in automatic modulation recognition under low signal-to-noise ratio (SNR) conditions, this paper introduces an end-to-end framework utilizing a Deep Residual Shrinkage Long Short-Term Memory (DRSLSTM) network. The proposed DRSLSTM architecture synergistically integrates two dedicated components: a deep residual shrinkage module, specifically designed for I/Q signals to perform simultaneous denoising and spatial feature extraction, and a Long Short-Term Memory (LSTM) network that captures long-range temporal dependencies from the refined feature sequences. Extensive simulations on a public dataset show that the DRSLSTM model achieves a recognition accuracy of 51.19% at −8 dB SNR, an improvement of 3.36 percentage points over the CLDNN baseline, and consistently surpasses six benchmark models at SNR levels above 0 dB. Moreover, it exhibits higher average recognition accuracy across a wide range of SNR conditions. The experimental results validate the overall superiority of the proposed approach. Full article

Soil desiccation cracks are common in farmland under dry conditions, which can alter soil water movement by providing preferential flow paths and thus affect water and fertilizer use efficiency. Understanding the mechanism of soil shrinkage and cracking is of great significance for optimizing field management by crack utilization or prevention. The behavior of soil shrinkage and cracking was monitored during drying experiments and analyzed with the help of a digital image processing method. The results showed that during shrinkage, the changes in soil height and equivalent diameter with water content differed significantly. The height change consisted of a rapid decline stage and a residual stage, while the equivalent diameter had a stable stage before the rapid decline stage. The VG-Peng model was suitable to fit the soil shrinkage characteristic curves, and the curves revealed that the soil shrinkage contained structural shrinkage, proportional shrinkage, residual shrinkage, and zero shrinkage stages. According to the changes in evaporation intensity, soil water evaporation could be divided into three stages: stable stage, declining stage, and residual stage. Cracks first formed in the defect areas and edge areas of the soil, and they mainly propagated in the stable evaporation stage. Crack development was dominated by an increase in crack length during the early cracking stage, while the propagation of crack width played a major role during the later stage. At the end of drying, the contribution ratio of crack length and width to the crack area was approximately 30% and 70%, respectively. The box-counting fractal dimension of the stabilized cracks was approximately 1.65, indicating that the crack network had significant self-similarity. The experimental results were used to implement the discrete element method to model the process of soil shrinkage and cracking. The models could effectively simulate the variation characteristics of soil height and equivalent diameter during shrinkage, as well as the variation characteristics of crack ratio and length density during cracking, with acceptable relative errors. In particular, the modeled morphology of the crack network was highly similar to the experimental observation. Our results provide new insights into the characterization and simulation of soil desiccation cracks, which will be conducive to understanding crack evolution and soil water movement in farmland. Full article

In this study, xanthan gum (XG)-modified coal-fly-ash-based cementitious materials were synthesized to realize the resource utilization of coal fly ash and to develop a low-carbon emission cementitious sealing material that can substitute cement-based sealing material to prevent coal fires. The optimal formulation for coal-fly-ash-based mining cementitious sealing material was developed using response surface methodology based on Box–Behnken Design. The optimized formulation was obtained with a coal fly ash-to-precursor ratio of 0.65, alkali-activator modulus of 1.4, and alkali-activator dosage of 7.5%. Under the optimal conditions, the initial and final setting time were 26 min and 31 min, respectively, fluidity was 245 mm, and the 7-day compressive strength approached 36.60 MPa, but there were still thermal shrinkage and cracking phenomena after heating. XG was then introduced to improve the thermal shrinkage and cracking of coal-fly-ash-based cementitious materials. Incorporating 1 wt.‰ XG was found to decrease the fluidity while maintaining the setting time and increasing the 1-day and 7-day compressive strength by 15.44% and 1.97%, respectively. The results demonstrated that the gels generated by XG cross-linking and coordinating with Al³⁺/Ca²⁺ were interspersed in the original C(N)-A-S-H gel network, which not only made the 1 wt.‰ XG modified coal-fly-ash-based cementitious material show minor expansion at ambient temperatures, but also improved the residual compressive strength, thermal shrinkage resistance and cracking resistance in comparison to unmodified cementitious material. However, due to the viscosity of XG and the coordination of Al³⁺ and non-terminal carboxyl groups in XG breaking the gel network, XG incorporation should not exceed 1 wt.‰ as the compressive strength and fluidity are decreased. Full article