1663 papers
This paper investigates how substrate heat treatment influences the tribological performance of +Ni and DLC coatings applied to tool steels, aiming to mitigate surface degradation and extend the service life of components used in plastic injection molds and forming tools.
This study demonstrates that applying Bayesian Kalman filtering as a preprocessing step significantly enhances the accuracy of deep learning models (GRU and LSTM) for short-term air quality forecasting across multiple Canadian cities, with optimal model configurations varying by pollutant type and local conditions.
This paper introduces and experimentally validates a high-order optical neural network framework under partially coherent illumination that overcomes the nonlinearity limitation of traditional diffractive systems, achieving performance comparable to electrical neural networks with significantly lower computational complexity and superior robustness.
This study demonstrates that increasing ultrasonic frequency and cutting speed during the machining of ultrasonic rolling-strengthened 7075-T6 aluminum alloy reduces cutting forces and surface roughness while enhancing compressive residual stress, ultimately yielding a finer, more uniform surface with fewer defects.
This study develops and validates metaheuristic-optimized machine learning models, specifically PSO-Cat and BO-Cat, which significantly outperform traditional methods in predicting the shear strength of circular concrete-filled steel tubes while utilizing SHAP and PDP analyses to interpret key influencing factors and providing a practical GUI for engineering application.
This study presents a validated analytical model that accurately predicts cure-induced spring-in in thin-walled curved composite parts by explicitly incorporating tool–part interaction, including interfacial friction and shear-lag deformation, to account for thickness-dependent dimensional deviations.
This paper proposes a Consistency-Gated Cascade framework that dynamically fuses a high-precision coordinate regression branch with a robust detection branch via a parameter-free consistency gate, achieving superior accuracy and stability for tungsten electrode tip localization in challenging K-TIG welding environments where standalone methods fail.
This paper proposes a synergistic multi-time-scale scheduling strategy integrating wind, solar, hydro, and pumped storage with improved Model Predictive Control and precise hydraulic coupling constraints to significantly reduce renewable energy curtailment and enhance system reliability.
This paper presents a four-parameter gas optical fiber sensor based on surface plasmon resonance that utilizes a 4D-type coreless fiber substrate and specific sensing materials to simultaneously detect methane, relative humidity, hydrogen, and hydrogen sulfide with high sensitivity and without spatiotemporal deviations.
This paper proposes a layered Edge-AI predictive maintenance architecture that integrates Industrial IoT sensor data with Digital Twin modeling and ensemble machine learning to achieve 97.43% detection accuracy, significantly reducing false negatives and downtime while enabling low-latency inference on edge devices.