Ironmaking process(IP)is indispensable to modern iron and steel industry,where real-time monitoring is crucial for achieving high molten iron quality(MIQ)with low energy consumption.While neural network-based models s...Ironmaking process(IP)is indispensable to modern iron and steel industry,where real-time monitoring is crucial for achieving high molten iron quality(MIQ)with low energy consumption.While neural network-based models show some promising results,they are generally limited by non-negligible drawbacks such as interpretability issues of feature learning.To address these issues,we propose a novel concept based on the shallow-to-deep correlation network representation regression(Sh-to-De CNRR).Our approach,shallow correlation network representation regression(ShCNRR),combines neural network and canonical correlation analysis thoughts to generate explainable features via shallow correlation network representation(CNR).A twin inverse network is then derived to obtain the explicit model output,leveraging the shallow CNR.To capture deeper nonlinear information,we extend ShCNRR into a hierarchical deep correlation network representation regression(DeCNRR)model that features stacked neural networks,enabling us to learn deeper CNR from process data.The feasibility and advantages of our proposals are validated by theoretical derivations and practical IP cases,which contain one MIQ regression and three MIQ-related fault detection tasks.The results reveal that highly fused statistical and neural network models yield superior monitoring performance compared to current state-of-the-art models,while statistical tests verify the convincing feature mining.展开更多
With the increasing complexity of industrial automation,planetary gearboxes play a vital role in largescale equipment transmission systems,directly impacting operational efficiency and safety.Traditional maintenance s...With the increasing complexity of industrial automation,planetary gearboxes play a vital role in largescale equipment transmission systems,directly impacting operational efficiency and safety.Traditional maintenance strategies often struggle to accurately predict the degradation process of equipment,leading to excessive maintenance costs or potential failure risks.However,existing prediction methods based on statistical models are difficult to adapt to nonlinear degradation processes.To address these challenges,this study proposes a novel condition-based maintenance framework for planetary gearboxes.A comprehensive full-lifecycle degradation experiment was conducted to collect raw vibration signals,which were then processed using a temporal convolutional network autoencoder with multi-scale perception capability to extract deep temporal degradation features,enabling the collaborative extraction of longperiod meshing frequencies and short-term impact features from the vibration signals.Kernel principal component analysis was employed to fuse and normalize these features,enhancing the characterization of degradation progression.A nonlinear Wiener process was used to model the degradation trajectory,with a threshold decay function introduced to dynamically adjust maintenance strategies,and model parameters optimized through maximum likelihood estimation.Meanwhile,the maintenance strategy was optimized to minimize costs per unit time,determining the optimal maintenance timing and preventive maintenance threshold.The comprehensive indicator of degradation trends extracted by this method reaches 0.756,which is 41.2%higher than that of traditional time-domain features;the dynamic threshold strategy reduces the maintenance cost per unit time to 55.56,which is 8.9%better than that of the static threshold optimization.Experimental results demonstrate significant reductions in maintenance costs while enhancing system reliability and safety.This study realizes the organic integration of deep learning and reliability theory in the maintenance of planetary gearboxes,provides an interpretable solution for the predictive maintenance of complex mechanical systems,and promotes the development of condition-based maintenance strategies for planetary gearboxes.展开更多
Fault diagnosis in industrial process is essential for ensuring production safety and efficiency.However,existing methods exhibit limited capability in recognizing hard samples and struggle to maintain consistency in ...Fault diagnosis in industrial process is essential for ensuring production safety and efficiency.However,existing methods exhibit limited capability in recognizing hard samples and struggle to maintain consistency in feature distributions across domains,resulting in suboptimal performance and robustness.Therefore,this paper proposes a fault diagnosis neural network for hard sample mining and domain adaptive(SmdaNet).First,the method uses deep belief networks(DBN)to build a diagnostic model.Hard samples are mined based on the loss values,dividing the data set into hard and easy samples.Second,elastic weight consolidation(EWC)is used to train the model on hard samples,effectively preventing information forgetting.Finally,the feature space domain adaptation is introduced to optimize the feature space by minimizing the Kullback–Leibler divergence of the feature distributions.Experimental results show that the proposed SmdaNet method outperforms existing approaches in terms of classification accuracy,robustness and interpretability on the penicillin simulation and Tennessee Eastman process datasets.展开更多
Complex industrial processes present typical uncertainty due to fluctuations in the composition of raw materials and frequently changing operating conditions.This poses three challenges for precise fault diagnosis,inc...Complex industrial processes present typical uncertainty due to fluctuations in the composition of raw materials and frequently changing operating conditions.This poses three challenges for precise fault diagnosis,including random noise interference,less distinguishability between multi-class faults,and the new fault emerging.To address these issues,this study formulates fault diagnosis in uncertain industrial processes as a multilevel refined fault diagnosis problem.A hierarchical stochastic network approach is proposed to refine fault diagnosis of multiclass faults.This method considers the augmentation of fault categories as naturally following a hierarchical structure.At each hierarchical stage,stochastic network methods are designed according to the sources of uncertainty.For fault feature extraction,a doubly stochastic attention-based variational graph autoencoder is introduced to suppress noise during the messagepassing process,ensuring the extraction of high-quality fault features and providing the provision of differentiated information.Subsequently,multiple stochastic configuration networks are deployed to realize multi-level fault diagnosis from coarse to fine granularity via a hierarchical structure rather than treating all faults equally.This approach effectively enhances the precision of multi-class fault diagnosis and ensures its robust generalization capability.Finally,the feasibility and effectiveness of the proposed method are validated using two industrial processes.The results demonstrate that the proposed method can effectively suppress the random noise interference and adapt to the emergence of small samples and imbalanced extreme fault-type data,achieving a satisfactory fault diagnosis performance.展开更多
In the electroslag remelting(ESR)process,it mainly relies on thermal experiments or analysis via mechanistic models to realize the physical fields simulation of the electromagnetic field and temperature field coupled ...In the electroslag remelting(ESR)process,it mainly relies on thermal experiments or analysis via mechanistic models to realize the physical fields simulation of the electromagnetic field and temperature field coupled transfer,which has the limitations of high cost,a large amount of calculating data and high computing power requirements.A novel network based on physics-informed neural network(PINN)was designed to realize the fast and high-fidelity prediction of the distribution of electromagnetic field and temperature field in ESR process.The physical laws were combined with the deep learning network through PINN,and physical constraints were embedded to achieve effective solution of partial differential equations(PDEs).PINN was used to minimize the loss function consisting of data error,physical information error and boundary condition error.The physical laws and boundary condition constraints in the ESR process were considered to maintain high PDE solution accuracy under different spatial and temporal resolutions.Automatic differentiation(Autodiff)technique and gradient descent algorithm were used to optimize the network parameters.The experimental results show that compared with the mechanistic models,PINN can effectively replace thermal experiments to realize the physical field simulation of ESR process with only a few experimental data,which can avoid the disadvantages of pure data-driven network simulation that requires a large amount of training data.Moreover,the solution of PINN has good physical interpretability and reliability of simulation results.For simulating electromagnetic field and temperature field distribution,the training time of the network is only 140 and 203 s,and the regression indicators of root mean square error can reach 12.65 and 13.76,respectively.展开更多
The increased accessibility of social networking services(SNSs)has facilitated communication and information sharing among users.However,it has also heightened concerns about digital safety,particularly for children a...The increased accessibility of social networking services(SNSs)has facilitated communication and information sharing among users.However,it has also heightened concerns about digital safety,particularly for children and adolescents who are increasingly exposed to online grooming crimes.Early and accurate identification of grooming conversations is crucial in preventing long-term harm to victims.However,research on grooming detection in South Korea remains limited,as existing models trained primarily on English text and fail to reflect the unique linguistic features of SNS conversations,leading to inaccurate classifications.To address these issues,this study proposes a novel framework that integrates optical character recognition(OCR)technology with KcELECTRA,a deep learning-based natural language processing(NLP)model that shows excellent performance in processing the colloquial Korean language.In the proposed framework,the KcELECTRA model is fine-tuned by an extensive dataset,including Korean social media conversations,Korean ethical verification data from AI-Hub,and Korean hate speech data from Hug-gingFace,to enable more accurate classification of text extracted from social media conversation images.Experimental results show that the proposed framework achieves an accuracy of 0.953,outperforming existing transformer-based models.Furthermore,OCR technology shows high accuracy in extracting text from images,demonstrating that the proposed framework is effective for online grooming detection.The proposed framework is expected to contribute to the more accurate detection of grooming text and the prevention of grooming-related crimes.展开更多
Software-defined satellite networks(SDSNs)play an essential role in future networks.Due to the diverse service scenarios,SDSN faces the demand of packet processing for heterogeneous protocols.Existing packet switching...Software-defined satellite networks(SDSNs)play an essential role in future networks.Due to the diverse service scenarios,SDSN faces the demand of packet processing for heterogeneous protocols.Existing packet switching typically works on one single protocol.For protocol-heterogeneous users,existing packet switch architectures have to construct multiple protocol-specific switching instances,resulting in severe resource waste.In this article,we propose the heterogeneous protocol-independent packet switch architecture(HISA).HISA employs a fast parsing structure to achieve efficient heterogeneous packet parsing and a novel match-action pipeline to achieve shared packet processing among heterogeneous users.HISA can also support the online configuration of switching behaviors.Use cases illustrate the effectiveness of applying HISA in SDSN.Numerical results show that compared to existing packet switching,HISA can significantly improve the resource utilization of SDSN.展开更多
Under the paradigm of Industry 5.0,intelligent manufacturing transcends mere efficiency enhancement by emphasizing human-machine collaboration,where human expertise plays a central role in assembly processes.Despite a...Under the paradigm of Industry 5.0,intelligent manufacturing transcends mere efficiency enhancement by emphasizing human-machine collaboration,where human expertise plays a central role in assembly processes.Despite advancements in intelligent and digital technologies,assembly process design still heavily relies on manual knowledge reuse,and inefficiencies and inconsistent quality in process documentation are caused.To address the aforementioned issues,this paper proposes a knowledge push method of complex product assembly process design based on distillation model-based dynamically enhanced graph and Bayesian network.First,an initial knowledge graph is constructed using a BERT-BiLSTM-CRF model trained with integrated human expertise and a fine-tuned large language model.Then,a confidence-based dynamic weighted fusion strategy is employed to achieve dynamic incremental construction of the knowledge graph with low resource consumption.Subsequently,a Bayesian network model is constructed based on the relationships between assembly components,assembly features,and operations.Bayesian network reasoning is used to push assembly process knowledge under different design requirements.Finally,the feasibility of the Bayesian network construction method and the effectiveness of Bayesian network reasoning are verified through a specific example,significantly improving the utilization of assembly process knowledge and the efficiency of assembly process design.展开更多
Accurate state of health(SOH)estimation is essential for the safe and reliable operation of lithium-ion batteries.However,existing methods face significant challenges,primarily because they rely on complete charge–di...Accurate state of health(SOH)estimation is essential for the safe and reliable operation of lithium-ion batteries.However,existing methods face significant challenges,primarily because they rely on complete charge–discharge cycles and fixed-form physical constraints,which limit adaptability to different chemistries and real-world conditions.To address these issues,this study proposes an approach that extracts features from segmented state of charge(SOC)intervals and integrates them into an enhanced physics-informed neural network(PINN).Specifically,voltage data within the 25%–75%SOC range during charging are used to derive statistical,time–frequency,and mechanism-based features that capture degradation trends.A hybrid PINN-Lasso-Transformer-BiLSTM architecture is developed,where Lasso regression enables sparse feature selection,and a nonlinear empirical degradation model is embedded as a learnable physical term within a dynamically scaled composite loss.This design adaptively balances data-driven accuracy with physical consistency,thereby enhancing estimation precision,robustness,and generalization.The results show that the proposed method outperforms conventional neural networks across four battery chemistries,achieving root mean square error and mean absolute error below 1%.Notably,features from partial charging segments exhibit higher robustness than those from full cycles.Furthermore,the model maintains strong performance under high temperatures and demonstrates excellent generalization capacity in transfer learning across chemistries,temperatures,and C-rates.This work establishes a scalable and interpretable solution for accurate SOH estimation under diverse practical operating conditions.展开更多
Accurate,up to date,and quick information related to any disaster supports disaster management team/authorities to perform quick,easy,and cost-effective response to enhance rescue operations to alleviate the possible ...Accurate,up to date,and quick information related to any disaster supports disaster management team/authorities to perform quick,easy,and cost-effective response to enhance rescue operations to alleviate the possible loss of lives,financial risks,and properties.Due to damaged infrastructure in disaster-affected areas,social media is the only way to share/exchange real time information.Therefore,‘X’(formerly Twitter)has become a major platform for disseminating real-time information during disaster events or emergencies,i.e.,floods and earthquake.Rapid identification of actionable content is critical for effective humanitarian response;however,the brief and noisy nature of tweets makes automated classification challenging.To tackle this problem,this study proposes a hybrid classification framework that integrates term frequency–inverse document frequency(TF-IDF)features with graph convolutional networks(GCNs)to enhance disaster-related tweet analysis.The proposed model performs three classification tasks:identifying disaster-related tweets(achieving 94.47%accuracy),categorizing disaster types(earthquake,flood,and non-disaster)with 91.78%accuracy,and detecting aid requests such as food,donations,and medical assistance(94.64%accuracy).By combining the statistical strengths of TF-IDF with the relational learning capabilities of GCNs,the model attains high accuracy while maintaining computational efficiency and interpretability.The results demonstrate the framework’s strong potential for real-time disaster response,offering valuable insights to support emergency management systems and humanitarian decision-making.展开更多
Elevation patterns and assembly processes of soil microbial community structures are essential for understanding biogeo-chemical processes in mountain systems.Differences in soil properties caused by elevation gradien...Elevation patterns and assembly processes of soil microbial community structures are essential for understanding biogeo-chemical processes in mountain systems.Differences in soil properties caused by elevation gradients can regulate the spatial distribu-tion and network complexity of the community structure.To explore the variations in soil microbial community structures and their as-sembly mechanisms across different elevations of the Changbai Mountains,as well as their responses to environmental factors,we col-lected microbial samples along an elevational gradient(seven elevations containing four vegetation zones)on the western slope of the Changbai Mountains using the method of metagenomic sequencing.The results showed a significant difference(P<0.05)for the Chao1 index across different elevations,but no significant difference was observed for the Shannon and Simpson indices.With increasing elev-ation,the number of nodes and links in the microbial network gradually decreased.Acidobacteria were highly connected to many nodes.The microbial communities indicated a significant distance-decay relationship(P<0.001)and were affected more by stochastic pro-cesses along the elevation gradient.The results of the Structural Equation Model(SEM)showed that elevation had direct significant ef-fect on carbon(C,P<0.01),nitrogen(N,P<0.01),and phosphorus(P,P<0.05)and weak negative effect on their ecological stoi-chiometry.Elevation was one of the major variables contributing to microbial network topology.The contribution of C and N to micro-bial network complexity was higher than that of P.Our study provides valuable insights into the responses of soil microbial communit-ies to elevation variations.展开更多
Temporal alignment of multisensor time series(MTS)is a critical prerequisite for accurate modeling and optimal control in subsequent data-driven applications.Nevertheless,many approaches frequently neglect to consider...Temporal alignment of multisensor time series(MTS)is a critical prerequisite for accurate modeling and optimal control in subsequent data-driven applications.Nevertheless,many approaches frequently neglect to consider the complex interdependencies between different sensors in MTS,and temporal alignment in many methods is typically treated as an isolated task disconnected from the downstream objectives,leading to unsatisfactory performances in follow-up applications.To address these challenges,this paper proposes a novel knowledge graph(KG)-guided iterative-updating graph neural network(GNN)for time-delay estimation(TDE)in MTS.Initially,a domain-specific KG is constructed from domain mechanism knowledge,providing a foundation for GNN's initialization.Next,capitalizing on the inherent structure of the graph topology,a GNN-based TDE method is developed.Then,a customized loss function is constructed,which synthesizes both the performances of downstream tasks and graph-based constraints.Moreover,an innovative algorithm for GNN structure learning and iterative-updating is proposed to renovate the graph structure further.Finally,experimental results across various regression and classification tasks on numerical simulation,public datasets,and the real blast furnace ironmaking dataset demonstrate that the proposed method can achieve accurate temporal alignment of MTS.展开更多
Based on the experimental data of Ti40 alloy obtained from Gleeble-1500 thermal simulator,an artificial neural network model of high temperature flow stress as a function of strain,strain rate and temperature was esta...Based on the experimental data of Ti40 alloy obtained from Gleeble-1500 thermal simulator,an artificial neural network model of high temperature flow stress as a function of strain,strain rate and temperature was established.In the network model,the input parameters of the model are strain,logarithm strain rate and temperature while flow stress is the output parameter.Multilayer perceptron（MLP） architecture with back-propagation algorithm is utilized.The present study achieves a good performance of the artificial neural network（ANN） model,and the predicted results are in agreement with experimental values.A processing map of Ti40 alloy is obtained with the flow stress predicted by the trained neural network model.The processing map developed by ANN model can efficiently track dynamic recrystallization and flow localization regions of Ti40 alloy during deforming.Subsequently,the safe and instable domains of hot working of Ti40 alloy are identified and validated through microstructural investigations.展开更多
The characteristic impedances of L-type and T-type networks are first investigated for a distributed amplifier design.The analysis shows that the L-type network has better frequency characteristics than the T-type one...The characteristic impedances of L-type and T-type networks are first investigated for a distributed amplifier design.The analysis shows that the L-type network has better frequency characteristics than the T-type one.A distribution amplifier based on the L-type network is implemented with the 2-μm GaAs HBT（heterojunction-bipolar transistor） process of WIN semiconductors.The measurement result presents excellent bandwidth performance and gives a gain of 5.5 dB with a gain flatness of ±1dB over a frequency range from 3 to 18 GHz.The return losses S11 and S22 are below-10dB in the designed frequency range.The output 1-dB compression point at 5 GHz is 13.3 dBm.The chip area is 0.95 mm2 and the power dissipation is 95 mW under a 3.5 V supply.展开更多
Underwater multi-target tracking logic and decision （UMTLD） has difficulty resolving multi-target tracking problems for underwater vehicles. Present methods assume factors in UMTLD are uncorrelated, when these are a...Underwater multi-target tracking logic and decision （UMTLD） has difficulty resolving multi-target tracking problems for underwater vehicles. Present methods assume factors in UMTLD are uncorrelated, when these are actually in a complex, interdependent relationship. To provide this, an index set of multi-target tracking decision characteristics and an analytic network process （ANP） model of the UMTLD method was -established. This method brings the index set of multi-target tracking decision into the ANP model, and the optimization multitarket tracking decision is achieved via computation of the resulting supermatrix. The rationality and robustness of decision results increase in simulations by 13% and 47% respectively with analytic hierarchy process （AHP）. These results indicate that the ANP method should be the preferred method when UMTLD factors are interdependent.展开更多
A hybrid neural network model,in which RH process(theoretical)model is combined organically with neural network(NN)and case-base reasoning(CBR),was established.The CBR method was used to select the operation mode and ...A hybrid neural network model,in which RH process(theoretical)model is combined organically with neural network(NN)and case-base reasoning(CBR),was established.The CBR method was used to select the operation mode and the RH operational guide parameters for different steel grades according to the initial conditions of molten steel,and a three-layer BP neural network was adopted to deal with nonlinear factors for improving and compensating the limitations of technological model for RH process control and end-point prediction.It was verified that the hybrid neural network is effective for improving the precision and calculation efficiency of the model.展开更多
In the real world, the inputs of many complicated systems are time-varying functions or processes. In order to predict the outputs of these systems with high speed and accuracy, this paper proposes a time series predi...In the real world, the inputs of many complicated systems are time-varying functions or processes. In order to predict the outputs of these systems with high speed and accuracy, this paper proposes a time series prediction model based on the wavelet process neural network, and develops the corresponding learning algorithm based on the expansion of the orthogonal basis functions. The effectiveness of the proposed time series prediction model and its learning algorithm is proved by the Macke-Glass time series prediction, and the comparative prediction results indicate that the proposed time series prediction model based on the wavelet process neural network seems to perform well and appears suitable for using as a good tool to predict the highly complex nonlinear time series.展开更多
In this work,the nickel-based powder metallurgy superalloy FGH95 was selected as experimental material,and the experimental parameters in multiple overlap laser shock processing(LSP)treatment were selected based on or...In this work,the nickel-based powder metallurgy superalloy FGH95 was selected as experimental material,and the experimental parameters in multiple overlap laser shock processing(LSP)treatment were selected based on orthogonal experimental design.The experimental data of residual stress and microhardness were measured in the same depth.The residual stress and microhardness laws were investigated and analyzed.Artificial neural network(ANN)with four layers(4-N-(N-1)-2)was applied to predict the residual stress and microhardness of FGH95 subjected to multiple overlap LSP.The experimental data were divided as training-testing sets in pairs.Laser energy,overlap rate,shocked times and depth were set as inputs,while residual stress and microhardness were set as outputs.The prediction performances with different network configuration of developed ANN models were compared and analyzed.The developed ANN model with network configuration of 4-7-6-2 showed the best predict performance.The predicted values showed a good agreement with the experimental values.In addition,the correlation coefficients among all the parameters and the effect of LSP parameters on materials response were studied.It can be concluded that ANN is a useful method to predict residual stress and microhardness of material subjected to LSP when with limited experimental data.展开更多
A neural network model of key process parameters and forming quality is developed based on training samples which are obtained from the orthogonal experiment and the finite element numerical simulation. Optimization o...A neural network model of key process parameters and forming quality is developed based on training samples which are obtained from the orthogonal experiment and the finite element numerical simulation. Optimization of the process parameters is conducted using the genetic algorithm (GA). The experimental results have shown that a surface model of the neural network can describe the nonlinear implicit relationship between the parameters of the power spinning process:the wall margin and amount of expansion. It has been found that the process of determining spinning technological parameters can be accelerated using the optimization method developed based on the BP neural network and the genetic algorithm used for the process parameters of power spinning formation. It is undoubtedly beneficial towards engineering applications.展开更多
Intelligent process planning(PP)is one of the most important components in an intelligent manufacturing system and acts as a bridge between product designing and practical manufacturing.PP is a nondeterministic polyno...Intelligent process planning(PP)is one of the most important components in an intelligent manufacturing system and acts as a bridge between product designing and practical manufacturing.PP is a nondeterministic polynomial-time(NP)-hard problem and,as existing mathematical models are not formulated in linear forms,they cannot be solved well to achieve exact solutions for PP problems.This paper proposes a novel mixed-integer linear programming(MILP)mathematical model by considering the network topology structure and the OR nodes that represent a type of OR logic inside the network.Precedence relationships between operations are discussed by raising three types of precedence relationship matrices.Furthermore,the proposed model can be programmed in commonly-used mathematical programming solvers,such as CPLEX,Gurobi,and so forth,to search for optimal solutions for most open problems.To verify the effectiveness and generality of the proposed model,five groups of numerical experiments are conducted on well-known benchmarks.The results show that the proposed model can solve PP problems effectively and can obtain better solutions than those obtained by the state-ofthe-art algorithms.展开更多
基金supported in part by the Pioneer Research and Development Program of Zhejiang(2025C01021)Zhejiang Province Postdoctoral Research Project Selection Fund(ZJ2025061)+3 种基金the National Science and Technology Major Project-Intelligent Manufacturing Systems and Robotics of China(2025ZD1602000,2025ZD1601800)the National Natural Science Foundation of China(61933015,62273030,62573387)the Natural Science Foundation of Zhejiang province,China(LY24F030004)the Fundamental Research Funds of Zhejiang Sci-Tech University(25222139-Y)。
文摘Ironmaking process(IP)is indispensable to modern iron and steel industry,where real-time monitoring is crucial for achieving high molten iron quality(MIQ)with low energy consumption.While neural network-based models show some promising results,they are generally limited by non-negligible drawbacks such as interpretability issues of feature learning.To address these issues,we propose a novel concept based on the shallow-to-deep correlation network representation regression(Sh-to-De CNRR).Our approach,shallow correlation network representation regression(ShCNRR),combines neural network and canonical correlation analysis thoughts to generate explainable features via shallow correlation network representation(CNR).A twin inverse network is then derived to obtain the explicit model output,leveraging the shallow CNR.To capture deeper nonlinear information,we extend ShCNRR into a hierarchical deep correlation network representation regression(DeCNRR)model that features stacked neural networks,enabling us to learn deeper CNR from process data.The feasibility and advantages of our proposals are validated by theoretical derivations and practical IP cases,which contain one MIQ regression and three MIQ-related fault detection tasks.The results reveal that highly fused statistical and neural network models yield superior monitoring performance compared to current state-of-the-art models,while statistical tests verify the convincing feature mining.
基金funded by scientific research projects under Grant JY2024B011.
文摘With the increasing complexity of industrial automation,planetary gearboxes play a vital role in largescale equipment transmission systems,directly impacting operational efficiency and safety.Traditional maintenance strategies often struggle to accurately predict the degradation process of equipment,leading to excessive maintenance costs or potential failure risks.However,existing prediction methods based on statistical models are difficult to adapt to nonlinear degradation processes.To address these challenges,this study proposes a novel condition-based maintenance framework for planetary gearboxes.A comprehensive full-lifecycle degradation experiment was conducted to collect raw vibration signals,which were then processed using a temporal convolutional network autoencoder with multi-scale perception capability to extract deep temporal degradation features,enabling the collaborative extraction of longperiod meshing frequencies and short-term impact features from the vibration signals.Kernel principal component analysis was employed to fuse and normalize these features,enhancing the characterization of degradation progression.A nonlinear Wiener process was used to model the degradation trajectory,with a threshold decay function introduced to dynamically adjust maintenance strategies,and model parameters optimized through maximum likelihood estimation.Meanwhile,the maintenance strategy was optimized to minimize costs per unit time,determining the optimal maintenance timing and preventive maintenance threshold.The comprehensive indicator of degradation trends extracted by this method reaches 0.756,which is 41.2%higher than that of traditional time-domain features;the dynamic threshold strategy reduces the maintenance cost per unit time to 55.56,which is 8.9%better than that of the static threshold optimization.Experimental results demonstrate significant reductions in maintenance costs while enhancing system reliability and safety.This study realizes the organic integration of deep learning and reliability theory in the maintenance of planetary gearboxes,provides an interpretable solution for the predictive maintenance of complex mechanical systems,and promotes the development of condition-based maintenance strategies for planetary gearboxes.
基金support from the following foundations:the National Natural Science Foundation of China(62322309,62433004)Shanghai Science and Technology Innovation Action Plan(23S41900500)Shanghai Pilot Program for Basic Research(22TQ1400100-16).
文摘Fault diagnosis in industrial process is essential for ensuring production safety and efficiency.However,existing methods exhibit limited capability in recognizing hard samples and struggle to maintain consistency in feature distributions across domains,resulting in suboptimal performance and robustness.Therefore,this paper proposes a fault diagnosis neural network for hard sample mining and domain adaptive(SmdaNet).First,the method uses deep belief networks(DBN)to build a diagnostic model.Hard samples are mined based on the loss values,dividing the data set into hard and easy samples.Second,elastic weight consolidation(EWC)is used to train the model on hard samples,effectively preventing information forgetting.Finally,the feature space domain adaptation is introduced to optimize the feature space by minimizing the Kullback–Leibler divergence of the feature distributions.Experimental results show that the proposed SmdaNet method outperforms existing approaches in terms of classification accuracy,robustness and interpretability on the penicillin simulation and Tennessee Eastman process datasets.
基金supported in part by the National Key Research and Development Program of China(2022YFB3304900)the Science and Technology Innovation Program of Hunan Province(2022RC1089)+1 种基金the Central South University Innovation Driven Research Programme(2023CXQD040)the Fundamental Research Funds for the Central Universities of Central South University(2025ZZTS0213).
文摘Complex industrial processes present typical uncertainty due to fluctuations in the composition of raw materials and frequently changing operating conditions.This poses three challenges for precise fault diagnosis,including random noise interference,less distinguishability between multi-class faults,and the new fault emerging.To address these issues,this study formulates fault diagnosis in uncertain industrial processes as a multilevel refined fault diagnosis problem.A hierarchical stochastic network approach is proposed to refine fault diagnosis of multiclass faults.This method considers the augmentation of fault categories as naturally following a hierarchical structure.At each hierarchical stage,stochastic network methods are designed according to the sources of uncertainty.For fault feature extraction,a doubly stochastic attention-based variational graph autoencoder is introduced to suppress noise during the messagepassing process,ensuring the extraction of high-quality fault features and providing the provision of differentiated information.Subsequently,multiple stochastic configuration networks are deployed to realize multi-level fault diagnosis from coarse to fine granularity via a hierarchical structure rather than treating all faults equally.This approach effectively enhances the precision of multi-class fault diagnosis and ensures its robust generalization capability.Finally,the feasibility and effectiveness of the proposed method are validated using two industrial processes.The results demonstrate that the proposed method can effectively suppress the random noise interference and adapt to the emergence of small samples and imbalanced extreme fault-type data,achieving a satisfactory fault diagnosis performance.
基金supported by National Natural Science Foundation of China(52274323 and 524743495)the Postdoctoral Fellowship Program of CPSF under Grant Number GZC20240231.
文摘In the electroslag remelting(ESR)process,it mainly relies on thermal experiments or analysis via mechanistic models to realize the physical fields simulation of the electromagnetic field and temperature field coupled transfer,which has the limitations of high cost,a large amount of calculating data and high computing power requirements.A novel network based on physics-informed neural network(PINN)was designed to realize the fast and high-fidelity prediction of the distribution of electromagnetic field and temperature field in ESR process.The physical laws were combined with the deep learning network through PINN,and physical constraints were embedded to achieve effective solution of partial differential equations(PDEs).PINN was used to minimize the loss function consisting of data error,physical information error and boundary condition error.The physical laws and boundary condition constraints in the ESR process were considered to maintain high PDE solution accuracy under different spatial and temporal resolutions.Automatic differentiation(Autodiff)technique and gradient descent algorithm were used to optimize the network parameters.The experimental results show that compared with the mechanistic models,PINN can effectively replace thermal experiments to realize the physical field simulation of ESR process with only a few experimental data,which can avoid the disadvantages of pure data-driven network simulation that requires a large amount of training data.Moreover,the solution of PINN has good physical interpretability and reliability of simulation results.For simulating electromagnetic field and temperature field distribution,the training time of the network is only 140 and 203 s,and the regression indicators of root mean square error can reach 12.65 and 13.76,respectively.
基金supported by the IITP(Institute of Information&Communications Technology Planning&Evaluation)-ITRC(Information Technology Research Center)grant funded by the Korean government(Ministry of Science and ICT)(IITP-2025-RS-2024-00438056).
文摘The increased accessibility of social networking services(SNSs)has facilitated communication and information sharing among users.However,it has also heightened concerns about digital safety,particularly for children and adolescents who are increasingly exposed to online grooming crimes.Early and accurate identification of grooming conversations is crucial in preventing long-term harm to victims.However,research on grooming detection in South Korea remains limited,as existing models trained primarily on English text and fail to reflect the unique linguistic features of SNS conversations,leading to inaccurate classifications.To address these issues,this study proposes a novel framework that integrates optical character recognition(OCR)technology with KcELECTRA,a deep learning-based natural language processing(NLP)model that shows excellent performance in processing the colloquial Korean language.In the proposed framework,the KcELECTRA model is fine-tuned by an extensive dataset,including Korean social media conversations,Korean ethical verification data from AI-Hub,and Korean hate speech data from Hug-gingFace,to enable more accurate classification of text extracted from social media conversation images.Experimental results show that the proposed framework achieves an accuracy of 0.953,outperforming existing transformer-based models.Furthermore,OCR technology shows high accuracy in extracting text from images,demonstrating that the proposed framework is effective for online grooming detection.The proposed framework is expected to contribute to the more accurate detection of grooming text and the prevention of grooming-related crimes.
基金supported by the National Natural Science Foundation of China(62101300,62341130)the Youth Fund Program of the Beijing National Research Center for Information Science and Technology under Grant BNR2021RC01012the Open Research Fund Program of the Beijing National Research Center for Information Science and Technology under Grant BNR2021KF02001.
文摘Software-defined satellite networks(SDSNs)play an essential role in future networks.Due to the diverse service scenarios,SDSN faces the demand of packet processing for heterogeneous protocols.Existing packet switching typically works on one single protocol.For protocol-heterogeneous users,existing packet switch architectures have to construct multiple protocol-specific switching instances,resulting in severe resource waste.In this article,we propose the heterogeneous protocol-independent packet switch architecture(HISA).HISA employs a fast parsing structure to achieve efficient heterogeneous packet parsing and a novel match-action pipeline to achieve shared packet processing among heterogeneous users.HISA can also support the online configuration of switching behaviors.Use cases illustrate the effectiveness of applying HISA in SDSN.Numerical results show that compared to existing packet switching,HISA can significantly improve the resource utilization of SDSN.
基金Supported by National Key Research and Development Program(Grant No.2024YFB3312700)National Natural Science Foundation of China(Grant No.52405541)the Changzhou Municipal Sci&Tech Program(Grant No.CJ20241131)。
文摘Under the paradigm of Industry 5.0,intelligent manufacturing transcends mere efficiency enhancement by emphasizing human-machine collaboration,where human expertise plays a central role in assembly processes.Despite advancements in intelligent and digital technologies,assembly process design still heavily relies on manual knowledge reuse,and inefficiencies and inconsistent quality in process documentation are caused.To address the aforementioned issues,this paper proposes a knowledge push method of complex product assembly process design based on distillation model-based dynamically enhanced graph and Bayesian network.First,an initial knowledge graph is constructed using a BERT-BiLSTM-CRF model trained with integrated human expertise and a fine-tuned large language model.Then,a confidence-based dynamic weighted fusion strategy is employed to achieve dynamic incremental construction of the knowledge graph with low resource consumption.Subsequently,a Bayesian network model is constructed based on the relationships between assembly components,assembly features,and operations.Bayesian network reasoning is used to push assembly process knowledge under different design requirements.Finally,the feasibility of the Bayesian network construction method and the effectiveness of Bayesian network reasoning are verified through a specific example,significantly improving the utilization of assembly process knowledge and the efficiency of assembly process design.
基金supported by the Shanghai Pilot Program for Basic Research(22T01400100-18)the National Natural Science Foundation of China(22278127 and 12447149)+1 种基金the Fundamental Research Funds for the Central Universities(2022ZFJH004)the Postdoctoral Fellowship Program of CPSF(GZB20250159).
文摘Accurate state of health(SOH)estimation is essential for the safe and reliable operation of lithium-ion batteries.However,existing methods face significant challenges,primarily because they rely on complete charge–discharge cycles and fixed-form physical constraints,which limit adaptability to different chemistries and real-world conditions.To address these issues,this study proposes an approach that extracts features from segmented state of charge(SOC)intervals and integrates them into an enhanced physics-informed neural network(PINN).Specifically,voltage data within the 25%–75%SOC range during charging are used to derive statistical,time–frequency,and mechanism-based features that capture degradation trends.A hybrid PINN-Lasso-Transformer-BiLSTM architecture is developed,where Lasso regression enables sparse feature selection,and a nonlinear empirical degradation model is embedded as a learnable physical term within a dynamically scaled composite loss.This design adaptively balances data-driven accuracy with physical consistency,thereby enhancing estimation precision,robustness,and generalization.The results show that the proposed method outperforms conventional neural networks across four battery chemistries,achieving root mean square error and mean absolute error below 1%.Notably,features from partial charging segments exhibit higher robustness than those from full cycles.Furthermore,the model maintains strong performance under high temperatures and demonstrates excellent generalization capacity in transfer learning across chemistries,temperatures,and C-rates.This work establishes a scalable and interpretable solution for accurate SOH estimation under diverse practical operating conditions.
文摘Accurate,up to date,and quick information related to any disaster supports disaster management team/authorities to perform quick,easy,and cost-effective response to enhance rescue operations to alleviate the possible loss of lives,financial risks,and properties.Due to damaged infrastructure in disaster-affected areas,social media is the only way to share/exchange real time information.Therefore,‘X’(formerly Twitter)has become a major platform for disseminating real-time information during disaster events or emergencies,i.e.,floods and earthquake.Rapid identification of actionable content is critical for effective humanitarian response;however,the brief and noisy nature of tweets makes automated classification challenging.To tackle this problem,this study proposes a hybrid classification framework that integrates term frequency–inverse document frequency(TF-IDF)features with graph convolutional networks(GCNs)to enhance disaster-related tweet analysis.The proposed model performs three classification tasks:identifying disaster-related tweets(achieving 94.47%accuracy),categorizing disaster types(earthquake,flood,and non-disaster)with 91.78%accuracy,and detecting aid requests such as food,donations,and medical assistance(94.64%accuracy).By combining the statistical strengths of TF-IDF with the relational learning capabilities of GCNs,the model attains high accuracy while maintaining computational efficiency and interpretability.The results demonstrate the framework’s strong potential for real-time disaster response,offering valuable insights to support emergency management systems and humanitarian decision-making.
基金Under the auspices of the National Natural Science Foundation of China(No.42430511,U20A2083)the National Key Research and Development Program of China(No.2022YFF1300900)the Science and Technology Development Program of Jilin Province(No.20210509037RQ,20230101348JC)。
文摘Elevation patterns and assembly processes of soil microbial community structures are essential for understanding biogeo-chemical processes in mountain systems.Differences in soil properties caused by elevation gradients can regulate the spatial distribu-tion and network complexity of the community structure.To explore the variations in soil microbial community structures and their as-sembly mechanisms across different elevations of the Changbai Mountains,as well as their responses to environmental factors,we col-lected microbial samples along an elevational gradient(seven elevations containing four vegetation zones)on the western slope of the Changbai Mountains using the method of metagenomic sequencing.The results showed a significant difference(P<0.05)for the Chao1 index across different elevations,but no significant difference was observed for the Shannon and Simpson indices.With increasing elev-ation,the number of nodes and links in the microbial network gradually decreased.Acidobacteria were highly connected to many nodes.The microbial communities indicated a significant distance-decay relationship(P<0.001)and were affected more by stochastic pro-cesses along the elevation gradient.The results of the Structural Equation Model(SEM)showed that elevation had direct significant ef-fect on carbon(C,P<0.01),nitrogen(N,P<0.01),and phosphorus(P,P<0.05)and weak negative effect on their ecological stoi-chiometry.Elevation was one of the major variables contributing to microbial network topology.The contribution of C and N to micro-bial network complexity was higher than that of P.Our study provides valuable insights into the responses of soil microbial communit-ies to elevation variations.
基金supported by the Young Scientists Fund of the National Natural Science Foundation of China(62303491)the Major Program of Xiangjiang Laboratory(22XJ01005)+1 种基金the Science and Technology Innovation Program of Hunan Province(2024RC1007)the Natural Science Foundation of Hunan Province(2025JJ10007)。
文摘Temporal alignment of multisensor time series(MTS)is a critical prerequisite for accurate modeling and optimal control in subsequent data-driven applications.Nevertheless,many approaches frequently neglect to consider the complex interdependencies between different sensors in MTS,and temporal alignment in many methods is typically treated as an isolated task disconnected from the downstream objectives,leading to unsatisfactory performances in follow-up applications.To address these challenges,this paper proposes a novel knowledge graph(KG)-guided iterative-updating graph neural network(GNN)for time-delay estimation(TDE)in MTS.Initially,a domain-specific KG is constructed from domain mechanism knowledge,providing a foundation for GNN's initialization.Next,capitalizing on the inherent structure of the graph topology,a GNN-based TDE method is developed.Then,a customized loss function is constructed,which synthesizes both the performances of downstream tasks and graph-based constraints.Moreover,an innovative algorithm for GNN structure learning and iterative-updating is proposed to renovate the graph structure further.Finally,experimental results across various regression and classification tasks on numerical simulation,public datasets,and the real blast furnace ironmaking dataset demonstrate that the proposed method can achieve accurate temporal alignment of MTS.
基金Project(2007CB613807)supported by the National Basic Research Program of ChinaProject(NCET-07-0696)supported by the New Century Excellent Talents in University,ChinaProject(35-TP-2009)supported by the Fund of the State Key Laboratory of Solidification Processing in Northwestern Polytechnical University,China
文摘Based on the experimental data of Ti40 alloy obtained from Gleeble-1500 thermal simulator,an artificial neural network model of high temperature flow stress as a function of strain,strain rate and temperature was established.In the network model,the input parameters of the model are strain,logarithm strain rate and temperature while flow stress is the output parameter.Multilayer perceptron（MLP） architecture with back-propagation algorithm is utilized.The present study achieves a good performance of the artificial neural network（ANN） model,and the predicted results are in agreement with experimental values.A processing map of Ti40 alloy is obtained with the flow stress predicted by the trained neural network model.The processing map developed by ANN model can efficiently track dynamic recrystallization and flow localization regions of Ti40 alloy during deforming.Subsequently,the safe and instable domains of hot working of Ti40 alloy are identified and validated through microstructural investigations.
基金China Postdoctoral Science Foundation (No.20090461048)Postdoctoral Science Foundation of Jiangsu Province (No.0901022C)Postdoctoral Science Foundation of Southeast University
文摘The characteristic impedances of L-type and T-type networks are first investigated for a distributed amplifier design.The analysis shows that the L-type network has better frequency characteristics than the T-type one.A distribution amplifier based on the L-type network is implemented with the 2-μm GaAs HBT（heterojunction-bipolar transistor） process of WIN semiconductors.The measurement result presents excellent bandwidth performance and gives a gain of 5.5 dB with a gain flatness of ±1dB over a frequency range from 3 to 18 GHz.The return losses S11 and S22 are below-10dB in the designed frequency range.The output 1-dB compression point at 5 GHz is 13.3 dBm.The chip area is 0.95 mm2 and the power dissipation is 95 mW under a 3.5 V supply.
基金Supported by the State Key Laboratory Foundation under Grant No.9140C2304080607the Aviation Science Foundation under Grant No.05F53027
文摘Underwater multi-target tracking logic and decision （UMTLD） has difficulty resolving multi-target tracking problems for underwater vehicles. Present methods assume factors in UMTLD are uncorrelated, when these are actually in a complex, interdependent relationship. To provide this, an index set of multi-target tracking decision characteristics and an analytic network process （ANP） model of the UMTLD method was -established. This method brings the index set of multi-target tracking decision into the ANP model, and the optimization multitarket tracking decision is achieved via computation of the resulting supermatrix. The rationality and robustness of decision results increase in simulations by 13% and 47% respectively with analytic hierarchy process （AHP）. These results indicate that the ANP method should be the preferred method when UMTLD factors are interdependent.
基金Item Sponsored by National Natural Science Foundation of China(50074026)
文摘A hybrid neural network model,in which RH process(theoretical)model is combined organically with neural network(NN)and case-base reasoning(CBR),was established.The CBR method was used to select the operation mode and the RH operational guide parameters for different steel grades according to the initial conditions of molten steel,and a three-layer BP neural network was adopted to deal with nonlinear factors for improving and compensating the limitations of technological model for RH process control and end-point prediction.It was verified that the hybrid neural network is effective for improving the precision and calculation efficiency of the model.
基金Project supported by the National Natural Science Foundation of China (Grant No 60572174)the Doctoral Fund of Ministry of Education of China (Grant No 20070213072)+2 种基金the 111 Project (Grant No B07018)the China Postdoctoral Science Foundation (Grant No 20070410264)the Development Program for Outstanding Young Teachers in Harbin Institute of Technology (Grant No HITQNJS.2007.010)
文摘In the real world, the inputs of many complicated systems are time-varying functions or processes. In order to predict the outputs of these systems with high speed and accuracy, this paper proposes a time series prediction model based on the wavelet process neural network, and develops the corresponding learning algorithm based on the expansion of the orthogonal basis functions. The effectiveness of the proposed time series prediction model and its learning algorithm is proved by the Macke-Glass time series prediction, and the comparative prediction results indicate that the proposed time series prediction model based on the wavelet process neural network seems to perform well and appears suitable for using as a good tool to predict the highly complex nonlinear time series.
基金Projects(51875558,51471176)supported by the National Natural Science Foundation of ChinaProject(2017YFB1302802)supported by the National Key R&D Program of China。
文摘In this work,the nickel-based powder metallurgy superalloy FGH95 was selected as experimental material,and the experimental parameters in multiple overlap laser shock processing(LSP)treatment were selected based on orthogonal experimental design.The experimental data of residual stress and microhardness were measured in the same depth.The residual stress and microhardness laws were investigated and analyzed.Artificial neural network(ANN)with four layers(4-N-(N-1)-2)was applied to predict the residual stress and microhardness of FGH95 subjected to multiple overlap LSP.The experimental data were divided as training-testing sets in pairs.Laser energy,overlap rate,shocked times and depth were set as inputs,while residual stress and microhardness were set as outputs.The prediction performances with different network configuration of developed ANN models were compared and analyzed.The developed ANN model with network configuration of 4-7-6-2 showed the best predict performance.The predicted values showed a good agreement with the experimental values.In addition,the correlation coefficients among all the parameters and the effect of LSP parameters on materials response were studied.It can be concluded that ANN is a useful method to predict residual stress and microhardness of material subjected to LSP when with limited experimental data.
基金Supported by the Natural Science Foundation of Shanxi Province Project(2012011023-2)
文摘A neural network model of key process parameters and forming quality is developed based on training samples which are obtained from the orthogonal experiment and the finite element numerical simulation. Optimization of the process parameters is conducted using the genetic algorithm (GA). The experimental results have shown that a surface model of the neural network can describe the nonlinear implicit relationship between the parameters of the power spinning process:the wall margin and amount of expansion. It has been found that the process of determining spinning technological parameters can be accelerated using the optimization method developed based on the BP neural network and the genetic algorithm used for the process parameters of power spinning formation. It is undoubtedly beneficial towards engineering applications.
基金supported in part by the National Natural Science Foundation of China(51825502,51775216)in part by the Program for Huazhong University of Science and Technology(HUST)Academic Frontier Youth Team(2017QYTD04).
文摘Intelligent process planning(PP)is one of the most important components in an intelligent manufacturing system and acts as a bridge between product designing and practical manufacturing.PP is a nondeterministic polynomial-time(NP)-hard problem and,as existing mathematical models are not formulated in linear forms,they cannot be solved well to achieve exact solutions for PP problems.This paper proposes a novel mixed-integer linear programming(MILP)mathematical model by considering the network topology structure and the OR nodes that represent a type of OR logic inside the network.Precedence relationships between operations are discussed by raising three types of precedence relationship matrices.Furthermore,the proposed model can be programmed in commonly-used mathematical programming solvers,such as CPLEX,Gurobi,and so forth,to search for optimal solutions for most open problems.To verify the effectiveness and generality of the proposed model,five groups of numerical experiments are conducted on well-known benchmarks.The results show that the proposed model can solve PP problems effectively and can obtain better solutions than those obtained by the state-ofthe-art algorithms.
