Search Results (3,647)

This study numerically investigated the applicability of the Forchheimer and Izbash equations for describing non-Darcian flow in three-dimensional (3D) rough-walled rock fractures. High-precision flow simulations were conducted on 52 synthetic 3D rough-walled fractures with varied apertures and surface roughness, generated from 56 combinations after excluding four cases with surface contact, under different hydraulic gradients. The simulation results captured transverse flow, back flow, and non-uniform streamlines on horizontal planes, which cannot be observed in conventional two-dimensional (2D) fracture models. The total eddy volume ratio negatively correlated with the aperture and positively correlated with roughness, and the 3D fractures exhibited a much smaller eddy volume ratio than the 2D fractures. Both equations provided excellent fits to the simulated data, with coefficients of determination R² > 0.996. Notably, the Forchheimer coefficients showed strong and monotonic correlations with the aperture and roughness and are therefore predictable and characterizable, whereas the Izbash coefficients showed weak and non-monotonic correlations. Since non-negligible prediction errors occurred at low Reynolds numbers when the equations were fitted over the entire flow range, a piecewise fitting strategy was proposed, which reduced the prediction errors of both equations to within 5% across the full range and quantitatively divided the flow into the Darcy, weak inertial, and strong inertial regimes. Double-parameter equations relating the critical Reynolds numbers to the aperture and roughness were then established, allowing the flow regime to be predicted directly from the geometric parameters without additional simulation. These findings facilitate reasonable flow regime division and accurate full-range flow characterization in rock fractures. Full article

Pressure-dependent pore-volume deformation is a critical but poorly constrained variable in dynamic reserve assessment for ultra-deep fractured-vuggy carbonate reservoirs, where fractures, dissolution pores, and vugs respond differently to effective-stress loading. In this work, a pore-structure-aware evaluation strategy was developed by integrating high-temperature and high-pressure volumetric measurements with data-driven regression. Twelve carbonate core plugs from the Ordovician Yijianfang and Yingshan formations of the Fuman Oilfield were selected to represent matrix-pore, dissolution-pore, fracture-vug, and fracture-dominated pore systems. Stepwise net-pressure experiments were performed under simulated reservoir conditions, and pore-volume compressibility (Cp) was calculated from corrected pore-volume changes. Measured C_p values reveal a distinct stress-sensitive response: C_p declines sharply during the low-net-pressure stage and then tends toward a quasi-stable level as net pressure increases, indicating progressive closure of mechanically compliant fractures, narrow throats, and weakly supported dissolution pores. Although porosity is positively associated with Cp, samples with comparable porosity display markedly different compressibility values, confirming that pore-space geometry and fracture-related compliance must be considered. Eight representative regression algorithms were then compared, using net pressure, porosity, permeability, initial pore volume, surface porosity, temperature, and a pore-structure index as model inputs. To further assess model generalization to completely unseen core plugs, additional core-ID-based leave-one-core-out (LOCO) validation was performed for k-nearest neighbors and AdaBoost. Under this grouped validation, k-nearest neighbors yielded an RMSE of 13.5978 × 10⁻⁴ MPa⁻¹ and an R² of 0.8408, whereas AdaBoost achieved an RMSE of 10.0160 × 10⁻⁴ MPa⁻¹ and an R² of 0.9136, indicating greater cross-core robustness of AdaBoost. Permutation-importance analysis of the split-specific KNN model indicated that net pressure, porosity, surface porosity, and pore-structure index made the largest predictive contributions within that model. Moreover, the predicted normalized Cp values reproduced the experimentally observed decreasing trend with increasing net pressure, supporting the physical consistency of the k-nearest neighbors predictions. The proposed experimental–machine learning framework offers a pressure-dependent method for estimating pore-volume compressibility within the geological and petrophysical domain represented by the investigated Fuman Oilfield cores, and provides more representative inputs for material-balance analysis, dynamic reserve evaluation, and production adjustment. Full article

An understanding of vortex formation in stirred tank reactors is of great importance, as in some processes, vortices are necessary for a chemical reaction to take place at all or to be accelerated, whilst in other processes, vortex formation is undesirable and can cause high mechanical stresses on the stirrer shaft, sealing and motor drive unit, or may cause undesirable surface aeration or foaming. To gain a better understanding of vortex formation, various baffle systems with different geometries are being experimentally investigated concerning power consumption and the resulting vortices on a single-stage Rushton turbine setup. The shapes of the resulting vortices are described mathematically in terms of vortex depth, width and volume, and the stirring systems prone to vortex formation are simulated using a CFD model based on the Lattice Boltzmann method. The CFD data obtained are compared, validated and verified against the experimental results in order to ultimately be able to fully describe, model and predict vortex formation through simulation. Furthermore, based on the detailed CFD data, vortex formation can be directly correlated with the swirl number, offering a mechanistic characterization method for the vortex shape in various mixing vessel configurations. Full article

Background: This study aimed to quantitatively assess structural changes in the hippocampus, amygdala, entorhinal cortex, lateral ventricles, precuneus, and posterior cingulate using deep-learning-based FastSurfer morphometry and evaluate their contribution to classification across the Alzheimer’s disease (AD) spectrum. Methods: Three-dimensional T1-weighted MRI data from the AD Neuroimaging Initiative (ADNI) were analyzed using FastSurfer. The study included 791 participants (425 female and 366 male): 365 cognitively normal (CN), 273 with mild cognitive impairment (MCI), and 153 with AD. Morphometric measures included hippocampal and amygdala volumes, lateral ventricular volume, entorhinal cortical thickness and surface area, and precuneus and posterior cingulate thickness and folding index. Volumetric measures were normalized to estimated total intracranial volume (eTIV). Principal morphometric comparisons were additionally adjusted for age, sex, and education. Six machine-learning classifiers were evaluated using participant-level training (n = 632) and held-out test (n = 159) sets, with an additional MMSE-ablation analysis. Results: Hippocampal and amygdala volumes were significantly lower and lateral ventricular volume was significantly higher in AD than in CN (all p < 0.001), with these differences persisting after eTIV normalization and remaining significant after adjustment for age, sex, and education (all adjusted p < 0.001). MCI generally showed intermediate volumetric values between CN and AD. Mini-Mental State Examination (MMSE) scores correlated positively with total hippocampal (r = 0.495) and amygdala (r = 0.462) volumes and negatively with lateral ventricular volume (r = −0.256) (all p < 0.001). For three-class CN/MCI/AD classification, Logistic Regression with ElasticNet achieved a mean macro-F1 score of 0.706 ± 0.017 in five-fold cross-validation and a held-out macro-F1 score of 0.705 with MMSE included. Excluding MMSE reduced the held-out macro-F1 score to 0.531 and ROC-AUC to 0.735, compared with 0.857 when MMSE was included. Performance was higher for binary CN-versus-AD classification, with a held-out test macro-F1 score of 0.919 and ROC-AUC of 0.990. Conclusions: FastSurfer-based morphometry demonstrated medial temporal atrophy and lateral ventricular enlargement across the CN–MCI–AD spectrum, with the principal volumetric differences remaining robust after adjustment for age, sex, and education. The intermediate morphometric profile of MCI and the associations between medial temporal volumes and cognitive performance support the relevance of these structural measures. Machine-learning performance improved substantially when MMSE was incorporated, indicating that the combined models reflect integrated morphometric, demographic, genetic, and cognitive information rather than morphometry alone. External validation in independent cohorts is required before clinical application. Full article

Objectives: Non-contrast computed tomography (NCCT) is used before retrograde intrarenal surgery (RIRS) to describe stone size, location, and complexity, but the best way to summarize stone burden remains unsettled. Therefore, we compared multidimensional NCCT-derived burden metrics with cumulative stone diameter (CSD) with respect to their ability to predict stone-free rate (SFR) and operative time after RIRS. Methods: We reviewed 352 RIRSs performed by one surgical team from April 2024 to May 2026. The primary SFR endpoint was defined as no residual fragments or a maximum residual fragment size of ≤4 mm, with 1-month computed tomography (CT) used as the reference imaging modality. Strict 0 mm CT stone-free status was analyzed as a sensitivity endpoint. The candidate predictors included one-dimensional, two-dimensional, estimated three-dimensional, spatial-distribution, attenuation, and distribution-complexity metrics extracted from preoperative NCCT. The cohort was split into training (n = 247) and validation (n = 105) sets. Results: The main model retained area-equivalent diameter, staghorn stone, and stone volume distribution entropy, with a validation area under the receiver operating characteristic curve (AUC) of 0.779. A simpler CSD model combining CSD and staghorn stone achieved a validation AUC of 0.761, without a significant AUC difference (p = 0.201). In repeated patient-level full-pipeline validation, the median validation AUC was 0.761. Under the strict 0 mm CT endpoint, the validation AUC was 0.670. Entropy showed lower selection stability in sensitivity analyses. Reclassification favored the main model (validation continuous net reclassification improvement [NRI] 0.460, p = 0.007; integrated discrimination improvement [IDI] 0.038, p = 0.003). Regarding operative time, area-, volume-, and surface-area-based metrics performed better than CSD in regard to both the whole cohort and stone-free episodes. Conclusions: NCCT-derived burden metrics can refine preoperative assessment before RIRS, but their incremental value over CSD for binary SFR prediction was modest. CSD remains a practical approximation for SFR prediction, whereas multidimensional metrics better reflect operative workload. Full article

Seismic sensors deployed in land acquisition record wavefields that are strongly distorted by near-surface heterogeneity, which introduces trace-specific, frequency-dependent phase perturbations that persist even after advanced time processing. These distortions are more pronounced for dense single sensor acquisition, where individual sensor coupling, local site conditions, and receiver-level near-surface variability drive phase behavior that differs from one sensor to the next. Conventional processing relies primarily on surface-consistent deconvolution, which targets long- to mid-wavelength phase variability under the highly simplified assumption of surface consistency, equalizing large-scale trends and taming variability through overdetermination across many sensors. However, this approximation is inherently unable to correct localized, non-surface-consistent phase distortions, and its effectiveness further degrades when such effects dominate, as is often the case for modern high-density single-sensor data. A separate and equally important limitation is that conventional workflows provide no direct, quantitative per sensor measure of phase reliability, that is, trace-to-trace phase coherence. Phase quality is therefore assessed only indirectly, typically through amplitude behavior or visual inspection, leaving residual phase disorder largely undiagnosed. We introduce phase variance as a seismic quality-control attribute for seismic sensor-recorded data, by treating seismic phases as circular random variables and analyzing local trace ensembles using circular statistics. This data-driven measure quantifies localized phase dispersion without phase unwrapping, enabling analysis of local phase trends and sensor-to-sensor fluctuations without global assumptions or wavelet models. Phase variance provides frequency-by-frequency classification of the data, ranging from coherent signal behavior to fully randomized, noise-dominated phase. Synthetic tests confirm that phase variance reliably captures imposed phase perturbations and their frequency dependence. Application of phase variance analysis to field prestack land data shows that conventional processing reduces phase variability primarily in the low-to-intermediate frequency range and struggles within the noise cone, while the highest and lowest frequencies often show little improvement in phase coherence. Phase variance operates automatically over the full prestack volume, from shallow to deep, and frequency by frequency, providing a consistent, human-independent metric for defining effective bandwidth based on phase coherence and supporting phase-sensitive workflows such as AVO, migration, and full-waveform inversion. Full article

Waste-to-hydrogen technology requires the involvement of new material development and performance evaluation for additive manufacturing (AM). Metal 3D printing is a very good possible alternative and is delivering results visible in the circular-economy environment. By using the AM technique, complex operations are avoided when combining the individual components, which is a typical disadvantage in obtaining composite materials; such samples are prepared with only one operation from the starting melt, which is chemically defined. Pyrolysis rotary kiln sealing rings are to be upgraded with several different microstructure coatings in order to improve the surface performance. The surface topology is aimed to be fine, dense and smooth. Also, the target characteristics are a low friction coefficient and a high hardness value, suggesting enhanced wear resistance. For elevated temperatures, 900 °C is selected for cobalt-based superalloy Stellite types with particle reinforcement. Two possibilities for advanced materials production are proposed with the Directed Energy Deposition Plasma Arc (DED-arc) and Laser Directed Energy Deposition (DED-LB). The shell of the rotary kiln sealing ring is made of stainless steel as the base, with the coating overlaid. Selected mixtures in powder form with defined composition are applied. For the DED-arc, commercially available Stellite 6 (Deloro Stellite^® 6) and 20 vol% WC particles with a grain size of 63–150 µm were employed. For the DED-LB, we employed TRIBALOY^® T-800 (Kennametal Stellite^TM) with 25 vol% TiC and a mesh size of −100/+325 (particle diameter between 45 and 150 µm). After the representative samples were metallurgically bonded with the base stainless steel, the relevant properties were obtained. Manufactured samples are compared in terms of microstructures and mechanical properties. Analysis of structure: Intermetallic carbides that formed on the cobalt basis make the composite harder and increase the plasticity in a defined direction. The hypoeutectic structures of Stellite 6 + 20% WC consist of dendrite and interdendrite eutectic. It is observed that with an increase in WC volume fraction, the size of the dendrites becomes finer, and the amount of eutectic structure is increased. For the TRIBALOY^® T-800 with TiC, we obtained relatively smaller grain sizes. The roughness values for the tested samples with WC were initially Ra = 0.8 µm, increasing up to Ra = 3.44 µm after the wear test, whereas for the TiC, they were slightly lower. Microhardness testing revealed increased values compared to the base stainless steels. Advanced sensor analysis with acoustic emission (AE) and electrical contact resistance (ECR) also showed the properties of the new materials. Customizable coatings with tailored properties were deposited by DED-arc and DED-LB. From the tests performed, a new technological procedure for the production of novel pyrolysis rotary kiln sealing rings is proposed. The microhardness, roughness, microstructure and abrasive wear-resistant response of the metallic composite material were examined in order to characterize the stable multiphase system. Full article

This article investigates the numerical simulation of heat-and-aerodynamic cycles in the ventilated air gap of a multilayer composite wall façade system in a hot climate using ANSYS 19/2 Fluent. Standard normative techniques rely on averaged, stationary boundary conditions and account for neither the height-wise inequality of solar exposure nor the dependence of air density and viscosity on barometric pressure and temperature, resulting in significant errors in predicting the actual heating of such structures. The model was calibrated on the authors’ own full-scale, in situ measurements of temperature, air speed and solar exposure in the ventilated gap of a nine-storey building, from which linear height-dependent surface-temperature relations were derived and used as boundary conditions for 3D models of façades 25 and 60 m tall. Thirty-two finite-volume experiments were performed under free convection (Boussinesq approximation), varying gap width (5 and 10 cm), inlet width (20 and 40 cm), barometric pressure (690 and 770 mmHg) and external air temperature (20 and 40 °C). Façade height proved the dominant factor (air speed up to 1.8 times higher, temperature 3–12.1 °C higher), followed by gap width (speed lower by 1.7 times, temperature by 3–5 °C), whereas pressure and inlet width altered the results by no more than 6%. Discrepancies with the standard calculation reached 10 °C in temperature and a two-fold difference in flow speed, confirming the need for verified CFD simulation when designing ventilated composite wall façades in hot climates. Full article

Long-distance free-surface water-conveyance tunnels require a spatially continuous representation of transient water depth, yet in practice flow is monitored at only a few sections. This study develops a physics-informed neural network (PINN) that reconstructs the transient water-depth field of unsteady free-surface flow in such tunnels from two- or three-point sensors. The one-dimensional Saint-Venant equations, closed with a Darcy–Weisbach steady friction term in hydraulic-radius form, are embedded as a soft constraint in the training loss, so that sparse depth observations are combined with the governing conservation laws to recover the field at unobserved interior and downstream locations. High-resolution finite-volume (FVM) solutions of a 500 m circular tunnel under a flood-rise scenario provide the reference data. The PINN reduces the relative L₂ error at unobserved sections to approximately one-quarter of that of an otherwise identical, physics-free ANN (2.50% versus 10.05% at an interior section; 4.77% versus 13.69% at an extrapolation section). Runs repeated with different random seeds confirm statistical stability at zero noise, while revealing that a minority of trainings at 10% noise converge to spurious solutions. Sensor-placement experiments, including layouts anchored at the true domain boundaries (x = 0 and 500 m), show that boundary anchoring—particularly of the upstream boundary—governs both accuracy and noise robustness: boundary-anchored two-sensor layouts remain accurate in most runs under 10–20% observation noise, whereas interior-only layouts degrade sharply. The method is presented as an offline reconstruction tool; its extension to streaming data assimilation is discussed as future work. Full article

Low-temperature sintering flue gas contains both nitrogen oxides (NO_x) and carbon monoxide (CO), requiring bifunctional catalysts for concurrent pollutant abatement. Herein, CuSmTi composite oxides with an identical nominal composition were synthesized via impregnation, mechanical grinding, and sol–gel methods to examine the effects of preparation route on their structure, surface properties, and catalytic performance in coupled NH₃-SCR and CO oxidation. CuSmTi-SG exhibited the best performance, achieving >40% NO_x conversion at 125 °C, complete NO_x and CO conversion at 200 °C, and nearly 100% N₂ selectivity over 100–300 °C, with stable performance over 24 h. It possessed a higher surface area (131.4 m² g⁻¹), pore volume (0.230 cm³ g⁻¹), and smaller TiO₂ crystallite size (8.5 nm) than the other catalysts. Spectroscopic analyses showed that sol–gel synthesis altered the surface electronic states of Cu and Sm species, resulting in a higher Cu⁺ fraction and greater amounts of medium-to-strong Lewis acid sites and labile surface oxygen species. In situ DRIFTS indicated that CO oxidation proceeded predominantly via a Mars–van Krevelen mechanism over Cu⁺ sites, whereas NH₃-SCR mainly followed an Eley–Rideal pathway. CO and NH₃ preferentially interacted with different surface sites, resulting in limited mutual inhibition. These results demonstrate that the preparation route can modify the structure and surface chemistry of CuSmTi catalysts without changing their nominal composition, thereby affecting their performance in low-temperature NO_x and CO abatement. Full article

Surface-water networks evolve as depressions fill, spill, and connect, whereas conventional digital elevation model (DEM) routing often treats flats and depressions through preprocessing or complete filling. This study introduces selective depression filling (SDF), a terrain-based framework that generates routing states characterized by depression extent, network connectivity, and geometric fill-to-spill volume. SDF aggregates 8-connected DEM cells with identical elevations into irregular cells (ICs), preserving the lateral geometry of flats, lakes, and channels, and uses a catchment-to-destination area (C/D) ratio to control depression-filling levels. IC-D8 (IC-based single-flow-direction routing scheme) and IC-MFD (IC-based multiple-flow-direction routing scheme) routing are evaluated across four contrasting landscapes. Both methods avoid ambiguous cellwise routing across flat and low-gradient features. Depression filling changes connectivity and contributing-area propagation as depressions transition from internal termini to widened flow-path components at spillover. Filling depths define DEM-derived topographic accommodation space rather than available water-storage capacity. Differences in C/D ratios define a Contributing-Area Sensitivity Metric (CSM), a terrain-based routing contrast rather than flood probability. Because the C/D ratio and filling states are derived from terrain data without independent hydrological forcing, the uncoupled SDF is a terrain-morphological analysis framework, not a hydrological model; hydrologic interpretation and regional transferability require independent calibration, validation, and testing across regions. Full article

Background/Objectives: First NIBUT, Average NIBUT, and tear meniscus height (TMH) are routinely used to characterize tear film stability and tear volume at individual visits, whereas their longitudinal behavior across the hospital-attending population is less well characterized. Aggregating routine examinations over time may provide a continuous hospital-level view of ocular surface status and enable the short-term forecasting of expected trajectories. We therefore evaluated a benchmark-first framework for monthly aggregated adult ocular surface indicators. Methods: This retrospective time-series study used de-identified adult eye-level Keratograph 5M records from July 2018 to November 2023. After cleaning, 31,492 records from 13,749 patients and 15,334 examination occasions were aggregated across 65 calendar months (64 observed months; March 2020 had no eligible records). Seven benchmark models and two exploratory deep learning comparators were evaluated using eight rolling-origin 3-month test windows. Results: Linear trend had the lowest mean origin-level macro-normalized RMSE (0.875; 95% bootstrap CI, 0.497–1.421), followed by simple exponential smoothing (0.907) and SARIMA(1,0,0)(1,0,0,12) (0.940). The paired difference between linear trend and simple exponential smoothing was small and did not show clear superiority (mean difference, −0.032; 95% bootstrap CI, −0.217 to 0.135; p = 0.789). The model with the lowest pooled error differed by target, while patient-month and sample-size-weighted sensitivity analyses gave a similar overall benchmark pattern. The exploratory LSTM and Transformer did not show a consistent advantage over the leading simple models. Conclusions: In this short hospital-level monthly series, simple forecasting models remained competitive, while no single model showed consistent superiority across forecast origins and sensitivity analyses. By extending ocular surface assessment from isolated examinations to longitudinal hospital-level trajectories, this framework provides a methodological basis for monitoring temporal changes in tear film stability and tear volume and for future quality monitoring, clinical, and epidemiological applications. Full article

The performance of Peltier-driven atmospheric water generation (AWG) systems depends strongly on the surface geometry and wetting behavior of the condensation structure. Triply periodic minimal surfaces (TPMS) provide high surface area-to-volume ratio and geometric tunability, but their effectiveness as three-dimensional condensation structures requires experimental evaluation. In this study, five additively manufactured TPMS geometries, Gyroid, Diamond, Lidinoid, SplitP, and Schwarz, were evaluated in a Peltier-driven AWG setup under controlled laboratory conditions. The measured water collection response varied among the tested TPMS geometries, which showed different condensation, retention, and collection trends. Water collection was measured with and without surface treatment, while the monitored surface temperature remained below the calculated dew point during testing. Without surface treatment, total water collection ranged from approximately 0.9 to 1.4 g, whereas surface-treated specimens collected approximately 0.6 to 1.2 g. The specimens with surface treatment exhibited predominantly discrete droplets rather than the film-wise morphology observed without surface treatment, but the total water collection did not increase consistently. Gyroid and Lidinoid showed slight increases with surface treatment, while SplitP, Diamond, and Schwarz showed reductions. Water collection also did not scale directly with calculated TPMS surface area, which suggests that effective air exposure, droplet retention, drainage, and coating uniformity contributed strongly to the observed performance. These findings provide experimental insights into additively manufactured TPMS geometry and surface treatment conditions for Peltier-driven AWG. Full article

Rice husk, an abundant agro-industrial by-product rich in SiO₂, represents a promising precursor for the sustainable synthesis of zeolites. In this study, rice husk ash was used to synthesize faujasite-type X and faujasite-type Y, and their performance for Hg²⁺ removal from aqueous solutions was comparatively evaluated. X-ray diffraction confirmed the successful formation of the faujasite structures, while physicochemical characterization revealed differences in pore structure and surface chemistry. FAU-type X exhibited higher Hg²⁺ removal than FAU-type Y, consistent with the combined influence of its lower Si/Al ratio, higher framework charge density and ion-exchange capacity, as well as its larger pore volume and average pore diameter. Based on its higher Hg²⁺ removal, FAU-type X was selected for a comprehensive evaluation of its adsorption performance and applicability under environmentally relevant conditions. The pseudo-second-order model best described adsorption kinetics for both zeolites, whereas thermodynamic analyses indicated that the adsorption process was spontaneous and endothermic. Optimal adsorption conditions for FAU-type X were achieved at pH 6.8, using an adsorbent dosage of 0.75 g L⁻¹, a contact time of 24 h, and an initial Hg²⁺ concentration of 1 mg L⁻¹. Equilibrium data were best fitted by the Sips isotherm model, indicating adsorption on a heterogeneous surface with a maximum adsorption capacity of 83.14 mg g⁻¹. FAU-type X retained appreciable adsorption performance after four regeneration cycles, although Hg²⁺ removal efficiency decreased in Caquetá River water because of competition from coexisting metal ions. To assess the environmental implications of the treated water beyond Hg²⁺ removal efficiency, ecotoxicological assays demonstrated the sensitivity of Daphnia magna to residual Hg²⁺ concentrations, whereas reductions in Escherichia coliforms were mainly attributed to the adsorption process. In addition, Lactuca sativa seedlings exhibited approximately 50% inhibition of elongation after treatment. Overall, these findings demonstrate the potential of rice husk-derived faujasite-type X as a sustainable adsorbent for Hg²⁺ removal, while highlighting the need for complementary treatment strategies to ensure the environmentally safe discharge of water and its agricultural reuse. Full article

Polyimide membranes are generally limited by the permeability–selectivity trade-off during gas separation, which significantly restricts their further development and practical applications. Carbon molecular sieve membranes derived from polyimide (PI) precursors via pyrolysis have attracted considerable attention due to their excellent molecular sieving capability. In this study, polyaniline (PANI) was introduced to modify the polyimide matrix, and a series of PI/PANI precursor membranes with different PANI loadings were fabricated, followed by pyrolysis to obtain the corresponding CMS/PANI membranes. The effects of PANI incorporation on the structural evolution and gas separation performance of the membranes were systematically investigated. The results demonstrated that PANI incorporation effectively increased the fractional free volume (FFV) of the precursor membranes and promoted the formation of microporous structures during pyrolysis, resulting in enhanced BET surface area and a more developed microporous network in CMS/PANI membranes. Compared with the pristine PI-derived CMS membrane, CMS/PANI membranes exhibited significantly improved gas permeability while maintaining comparable gas pair selectivity. Among the prepared membranes, CMS/PANI-20 achieved the optimal overall separation performance, with H₂ and CO₂ permeabilities of 2012 and 754 Barrer, respectively, and H₂/CH₄ and CO₂/CH₄ selectivities of 253 and 95, respectively, exceeding the corresponding Robeson upper bounds. Full article