Addressing the limitations of inadequate stochastic disturbance characterization during wind turbine degradation processes that result in constrained modeling accuracy,replacement-based maintenance practices that devi...Addressing the limitations of inadequate stochastic disturbance characterization during wind turbine degradation processes that result in constrained modeling accuracy,replacement-based maintenance practices that deviate from actual operational conditions,and static maintenance strategies that fail to adapt to accelerated deterioration trends leading to suboptimal remaining useful life utilization,this study proposes a Time-Based Incomplete Maintenance(TBIM)strategy incorporating reliability constraints through stochastic differential equations(SDE).By quantifying stochastic interference via Brownian motion terms and characterizing nonlinear degradation features through state influence rate functions,a high-precision SDE degradation model is constructed,achieving 16%residual reduction compared to conventional ordinary differential equation(ODE)methods.The introduction of age reduction factors and failure rate growth factors establishes an incomplete maintenance mechanism that transcends traditional“as-good-as-new”assumptions,with the TBIM model demonstrating an additional 8.5%residual reduction relative to baseline SDE approaches.A dynamic maintenance interval optimization model driven by dual parameters—preventive maintenance threshold R_(p) and replacement threshold R_(r)—is designed to achieve synergistic optimization of equipment reliability and maintenance economics.Experimental validation demonstrates that the optimized TBIM extends equipment lifespan by 4.4%and reducesmaintenance costs by 4.16%at R_(p)=0.80,while achieving 17.2%lifespan enhancement and 14.6%cost reduction at R_(p)=0.90.This methodology provides a solution for wind turbine preventive maintenance that integrates condition sensitivity with strategic foresight.展开更多
This paper presents an artificial intelligence(AI)-based observer that combines fuzzy logic and neural networks to detect abnormalities in sensors embedded in an electrolyser.Electrolysers are hydrogen production plan...This paper presents an artificial intelligence(AI)-based observer that combines fuzzy logic and neural networks to detect abnormalities in sensors embedded in an electrolyser.Electrolysers are hydrogen production plants that require effective maintenance to guarantee suitable operation,prevent degradation,and avoid loss of efficiency.In this sense,predictive maintenance arises as one of the most advisable techniques for maintenance in electrolysers by using sensor data to predict potential abnormalities.However,if the sensor fails,there will be an incorrect forecasting of abnormalities.Among the different types of operational faults that sensors can present are drift-related faults,which are probably the most difficult to detect due to a slow but progressive loss of accuracy in measurements.Another problem with predictive maintenance is that it often requires enormous training data,which is not available at the early stage of plant operation.The developed fuzzy system is responsible for detecting faulty readings arising from drift sensor signals,while the neural network complements the function of the fuzzy system by predicting sensor signals when enough training data are available.The AI-based observer and the fuzzy rules are validated in an experimental case study with a 1 Nm^(3)/h electrolyser.The selected variables are electrolyser temperature and efficiency.Experimental results show that the rules of the fuzzy component of the AI-based observer guarantee an accuracy of±0.25 within the range of 0 to 1,and the neural network component predicted correct sensor values with a root mean square error(RMSE)as low as 0.0016.The authors’approach helps to determine drift faults without additional sensors or components installed in the plant.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
Modern industrial environments require uninterrupted machinery operation to maintain productivity standards while ensuring safety and minimizing costs.Conventional maintenance methods,such as reactive maintenance(i.e....Modern industrial environments require uninterrupted machinery operation to maintain productivity standards while ensuring safety and minimizing costs.Conventional maintenance methods,such as reactive maintenance(i.e.,run to failure)or time-based preventive maintenance(i.e.,scheduled servicing),prove ineffective for complex systems with many Internet of Things(IoT)devices and sensors because they fall short in detecting faults at early stages when it is most crucial.This paper presents a predictive maintenance framework based on a hybrid deep learning model that integrates the capabilities of Long Short-Term Memory(LSTM)Networks and Convolutional Neural Networks(CNNs).The framework integrates spatial feature extraction and temporal sequence modeling to accurately classify the health state of industrial equipment into three categories,including Normal,Require Maintenance,and Failed.The framework uses a modular pipeline that includes IoT-enabled data collection along with secure transmission methods to manage cloud storage and provide real-time fault classification.The FD004 subset of the NASA C-MAPSS dataset,containing multivariate sensor readings from aircraft engines,serves as the training and evaluation data for the model.Experimental results show that the LSTM-CNN model outperforms baseline models such as LSTM-SVM and LSTM-RNN,achieving an overall average accuracy of 86.66%,precision of 86.00%,recall of 86.33%,and F1-score of 86.33%.Contrary to the previous LSTM-CNN-based predictive maintenance models that either provide a binary classification or rely on synthetically balanced data,our paper provides a three-class maintenance state(i.e.,Normal,Require Maintenance,and Failed)along with threshold-based labeling that retains the true nature of the degradation.In addition,our work also provides an IoT-to-cloud-based modular architecture for deployment.It offers Computerized Maintenance Management System(CMMS)integration,making our proposed solution not only technically sound but also practical and innovative.The solution achieves real-world industrial deployment readiness through its reliable performance alongside its scalable system design.展开更多
With the progress of the times and the development of information technology, the computer has become a life tool. Due to its intelligent convenience, it is widely used in daily life, but the inevitable failure proble...With the progress of the times and the development of information technology, the computer has become a life tool. Due to its intelligent convenience, it is widely used in daily life, but the inevitable failure problems of the machine, or the loss of the computer and data, and causes immeasurable trouble and loss.展开更多
Scientific and technological advancements are rapidly transforming underground engineering,shifting from labor-intensive,time-consuming methods to automated,real-time systems.This timely and comprehensive review cover...Scientific and technological advancements are rapidly transforming underground engineering,shifting from labor-intensive,time-consuming methods to automated,real-time systems.This timely and comprehensive review covers in-situ testing,intelligent monitoring,and geophysical testing methods,highlighting fundamental principles,testing apparatuses,data processing techniques,and engineering applications.The state-of-the-art summary emphasizes not only cutting-edge innovations for complex and harsh environments but also the transformative role of artificial intelligence and machine learning in data interpretations.The integration of big data and advanced algorithms is particularly impactful,enabling the identification,prediction,and mitigation of potential risks in underground projects.Key aspects of the discussion include detection capabilities,method integration,and data convergence of intelligent technologies to drive enhanced safety,operational efficiency,and predictive reliability.The review also examines future trends in intelligent technologies,emphasizing unified platforms that combine multiple methods,real-time data,and predictive analytics.These advancements are shaping the evolution of underground construction and maintenance,aiming for risk-free,high-efficiency underground engineering.展开更多
Preventive maintenance(PM)enhances the reliability of an ultra-high-voltage DC(UHVDC)line.However,it introduces new states and complicates the task of determining the reliability parameters of the component with PM.Th...Preventive maintenance(PM)enhances the reliability of an ultra-high-voltage DC(UHVDC)line.However,it introduces new states and complicates the task of determining the reliability parameters of the component with PM.The hierarchical connection(HC)of the inverter improves the flexibility of UHVDC,but the state spaces of the inverter and UHVDC-HC are determined by capacity distribution between the high terminal(HT)and low terminal(LT),which has not been studied yet.In this paper,the reliability of UHVDC-HC with the PM is studied.First,with the impact of PM described by a nonlinear coefficient,the state space of the components with PM is aggregated.It is the same size as without PM.The only difference is the transition rate.Second,considering power distribution between the HT/LT and its dependence on the state of the rectifier,new capacity levels of HT/LT,e.g.,75%/2,50%/3,are defined.Third,with capacity levels of the HT/LT differentiated,2 existing reliability indices are extended,and 2 are newly defined.Finally,a sensitivity model of the reliability of the UHVDC-HC to its parameters and the PM period is proposed.展开更多
This study proposes a novel visual maintenance method for photovoltaic(PV)modules based on a two-stage Wiener degradation model,addressing the limitations of traditional PV maintenance strategies that often result in ...This study proposes a novel visual maintenance method for photovoltaic(PV)modules based on a two-stage Wiener degradation model,addressing the limitations of traditional PV maintenance strategies that often result in insufficient or excessive maintenance.The approach begins by constructing a two-stage Wiener process performance degradation model and a remaining life prediction model under perfect maintenance conditions using historical degradation data of PV modules.This enables accurate determination of the optimal timing for postfailure corrective maintenance.To optimize the maintenance strategy,the study establishes a comprehensive cost model aimed at minimizing the long-term average cost rate.The model considers multiple cost factors,including inspection costs,preventive maintenance costs,restorative maintenance costs,and penalty costs associated with delayed fault detection.Through this optimization framework,the method determines both the optimal maintenance threshold and the ideal timing for predictive maintenance actions.Comparative analysis demonstrates that the twostage Wiener model provides superior fitting performance compared to conventional linear and nonlinear degradation models.When evaluated against traditional maintenance approaches,including Wiener process-based corrective maintenance strategies and static periodic maintenance strategies,the proposed method demonstrates significant advantages in reducing overall operational costs while extending the effective service life of PV components.The method achieves these improvements through effective coordination between reliability optimization and economic benefit maximization,leading to enhanced power generation performance.These results indicate that the proposed approach offers a more balanced and efficient solution for PV system maintenance.展开更多
As an essential part of the urban infrastructure,underground utility tunnels have a long service life,complex structural performance evolution and dynamic changes both inside and outside the tunnel.These combined fact...As an essential part of the urban infrastructure,underground utility tunnels have a long service life,complex structural performance evolution and dynamic changes both inside and outside the tunnel.These combined factors result in a wide variety of disaster risks during the operation and maintenance phase,which make risk management and control particularly challenging.This work first reviews three common representative disaster factors during the operation and maintenance period:settlement,earthquakes,and explosions.It summarizes the causes of disasters,key technologies,and research methods.Then,it delves into the research on the intelligent operation and maintenance architecture for utility tunnels.Additionally,it explores the data challenges,monitoring technologies,and management platform architectures faced during the operation and maintenance process.This work provides new research perspectives for the long-term,healthy,and sustainable development of utility tunnels,which serve as the underground arteries of cities.展开更多
BACKGROUND Research examining the relationships among anxiety,depression,self-perceived burden(SPB),and psychological resilience(PR),along with the determinants of PR,in patients with chronic renal failure(CRF)receivi...BACKGROUND Research examining the relationships among anxiety,depression,self-perceived burden(SPB),and psychological resilience(PR),along with the determinants of PR,in patients with chronic renal failure(CRF)receiving maintenance hemodia-lysis(MHD)is limited.AIM To investigate the correlation between anxiety,depression,SPB,and PR in pati-ents with CRF on MHD.METHODS This study included 225 patients with CRF on MHD who were admitted between June 2021 and June 2024.The anxiety level was evaluated using the Self-Rating Anxiety Scale(SAS);the depression status was assessed using the Self-Rating Depression Scale(SDS);the SPB was measured using the SPB Scale(SPBS);and the PR was determined using the Connor–Davidson Resilience Scale(CD-RISC).The correlations among the SAS,SDS,SPB,and CD-RISC were analyzed using Pearson’s correlation coefficients.Univariate and multivariate analyses were performed to identify the factors that influence the PR of patients with CRF on MHD.RESULTS The SAS,SDS,SPB,and CD-RISC scores of the 225 patients were 45.25±15.36,54.81±14.68,32.31±11.52,and 66.48±9.18,respectively.Significant negative correlations were observed between SAS,SDS,SPB,and CD-RISC.Furthermore,longer dialysis vintage(P=0.015),the absence of religious beliefs(P=0.020),lower monthly income(P=0.008),higher SAS score(P=0.013),and higher SDS score(P=0.006)were all independent factors that adversely affected the PR of patients with CRF on MHD.CONCLUSION Patients with CRF on MHD present with varying degrees of anxiety,depression,and SPB,all of which exhibit a significant negative correlation with their PR.Moreover,longer dialysis vintage,the absence of religious beliefs,lower monthly income,higher SAS score,and higher SDS score were factors that negatively affected the PR of patients with CRF on MHD.展开更多
While Artificial Intelligence (AI) is leading the way in terms of hardware advancements, such as GPUs, memory, and processing power, real-time applications are still catching up. It is inevitable that when one aspect ...While Artificial Intelligence (AI) is leading the way in terms of hardware advancements, such as GPUs, memory, and processing power, real-time applications are still catching up. It is inevitable that when one aspect leads and other trails behind, they coexist in life, as is often the case. The trailing aspect cannot remain far behind because, without application and use, there would be a dead end. Everything, whether an object, software, or tool, must have a practical use for humans. Without this, it will become obsolete. We can see this in many instances, such as blockchain technology, which is superior yet faces challenges in practical implementation, leading to a decline in adoption. This publication aims to bridge the gap between AI advancements and maintenance, specifically focusing on making predictive maintenance a practical application. There are multiple building blocks that make predictive maintenance a practical application. Each block performs a function leading to an output. This output forms an input to the receiving block. There are also foundational parts for all these building blocks to perform a function. Eventually, once the building blocks are connected, they form a loop and start to lead the path to predictive maintenance. Predictive maintenance is indeed practically achievable, but one must comprehend all the building blocks necessary for its implementation. Although detailed explanations will be provided in the upcoming sections, it is important to understand that simply purchasing software and plugging it in might be a far-fetched approach.展开更多
This paper presents a project aimed at developing a trilingual visual dictionary for aircraft maintenance professionals and students.The project addresses the growing demand for accurate communication and technical te...This paper presents a project aimed at developing a trilingual visual dictionary for aircraft maintenance professionals and students.The project addresses the growing demand for accurate communication and technical terminology in the aviation industry,particularly in Brazil and China.The study employs a corpus-driven approach,analyzing a large corpus of aircraft maintenance manuals to extract key technical terms and their collocates.Using specialized subcorpora and a comparative analysis,this paper demonstrates challenges and solutions into the identification of high-frequency keywords and explores their contextual use in aviation documentation,emphasizing the need for clear and accurate technical communication.By incorporating these findings into a trilingual visual dictionary,the project aims to enhance the understanding and usage of aviation terminology.展开更多
Tunnel facility management(FM)is crucial for ensuring safety,efficiency,and resilience of tunnel infrastructure.Current FM practices,such as reactive and preventive maintenance,have limitations-reactive maintenance ca...Tunnel facility management(FM)is crucial for ensuring safety,efficiency,and resilience of tunnel infrastructure.Current FM practices,such as reactive and preventive maintenance,have limitations-reactive maintenance can't prevent failures,and preventive maintenance can't predict asset maintenance needs,leading to costly and inefficient processes.In addition,existing computerized tunnel FM systems face various challenges,including the lack of integrated real-time monitoring information,visualization of assets in a three-dimensional(3D)environment,supporting predictive maintenance,and scaling into long infrastructure with complicated spatiotemporal relationships.This study addresses these limitations by proposing a data-driven digital twin(DT)-based framework that supports predictive maintenance to improve tunnel FM processes and enhance resilience.The proposed framework consists of six layers,allowing the integration of data from monitoring system,FM system,and building information modeling(BIM)models.The framework proposes a flexible tunnel data model and classification system that hierarchically divides the tunnel models,ensuring an efficient data connection from the physical twin to the DT.The system was implemented in a tunnel case study that generates maintenance plans and work orders using historical and current condition monitoring data,and the 3D visualization technology suggests maintenance and repair processes,making the FM decision process more effective.The proposed system detected and predicted the twin state based on a data-driven analysis,and the prediction accuracy of the machine learning models was sufficiently high for use in real scenarios to make FM plans in advance and prevent asset failures.The proposed framework is contributing to the infrastructure resilience by enhancing the tunnel system ability to predict the maintenance tasks and prevent failures using data-driven DT technology.展开更多
Predictive maintenance plays a crucial role in preventing equipment failures and minimizing operational downtime in modern industries.However,traditional predictive maintenance methods often face challenges in adaptin...Predictive maintenance plays a crucial role in preventing equipment failures and minimizing operational downtime in modern industries.However,traditional predictive maintenance methods often face challenges in adapting to diverse industrial environments and ensuring the transparency and fairness of their predictions.This paper presents a novel predictive maintenance framework that integrates deep learning and optimization techniques while addressing key ethical considerations,such as transparency,fairness,and explainability,in artificial intelligence driven decision-making.The framework employs an Autoencoder for feature reduction,a Convolutional Neural Network for pattern recognition,and a Long Short-Term Memory network for temporal analysis.To enhance transparency,the decision-making process of the framework is made interpretable,allowing stakeholders to understand and trust the model’s predictions.Additionally,Particle Swarm Optimization is used to refine hyperparameters for optimal performance and mitigate potential biases in the model.Experiments are conducted on multiple datasets from different industrial scenarios,with performance validated using accuracy,precision,recall,F1-score,and training time metrics.The results demonstrate an impressive accuracy of up to 99.92%and 99.45%across different datasets,highlighting the framework’s effectiveness in enhancing predictive maintenance strategies.Furthermore,the model’s explainability ensures that the decisions can be audited for fairness and accountability,aligning with ethical standards for critical systems.By addressing transparency and reducing potential biases,this framework contributes to the responsible and trustworthy deployment of artificial intelligence in industrial environments,particularly in safety-critical applications.The results underscore its potential for wide application across various industrial contexts,enhancing both performance and ethical decision-making.展开更多
Objective:To investigate the clinical efficacy and cost-effectiveness of combined hemodialysis(HD)and hemoperfusion(HP)therapy in managing secondary hyperparathyroidism(SHPT)in patients undergoing maintenance hemodial...Objective:To investigate the clinical efficacy and cost-effectiveness of combined hemodialysis(HD)and hemoperfusion(HP)therapy in managing secondary hyperparathyroidism(SHPT)in patients undergoing maintenance hemodialysis(MHD).Methods:A total of 195 patients with MHD and SHPT at Deyang People's Hospital from April 2024 to April 2025 were enrolled.Patients were randomly assigned to a control group receiving standard HD treatment and an experimental group receiving HD combined with HP therapy.The experimental group underwent 1 year of observation(97 cases in the experimental group,98 cases in the control group).During treatment,changes in parathyroid hormone(PTH),serum calcium,serum phosphorus,and inflammatory factors were monitored,along with analysis of treatment-related economic benefits and safety.Results:The experimental group demonstrated significantly better reductions in PTH,serum phosphorus,and inflammatory factors compared to the control group(P<0.05).Although the total treatment cost was slightly higher,the unit cost per therapeutic effect was lower,resulting in a superior cost-effectiveness ratio.Conclusion:Combined HD and HP therapy can significantly improve SHPT-related indicators in MHD patients,demonstrating safety,controllability,and high cost-effectiveness,making it a clinically applicable and recommended treatment option.展开更多
On-device Artificial Intelligence(AI)accelerators capable of not only inference but also training neural network models are in increasing demand in the industrial AI field,where frequent retraining is crucial due to f...On-device Artificial Intelligence(AI)accelerators capable of not only inference but also training neural network models are in increasing demand in the industrial AI field,where frequent retraining is crucial due to frequent production changes.Batch normalization(BN)is fundamental to training convolutional neural networks(CNNs),but its implementation in compact accelerator chips remains challenging due to computational complexity,particularly in calculating statistical parameters and gradients across mini-batches.Existing accelerator architectures either compromise the training accuracy of CNNs through approximations or require substantial computational resources,limiting their practical deployment.We present a hardware-optimized BN accelerator that maintains training accuracy while significantly reducing computational overhead through three novel techniques:(1)resourcesharing for efficient resource utilization across forward and backward passes,(2)interleaved buffering for reduced dynamic random-access memory(DRAM)access latencies,and(3)zero-skipping for minimal gradient computation.Implemented on a VCU118 Field Programmable Gate Array(FPGA)on 100 MHz and validated using You Only Look Once version 2-tiny(YOLOv2-tiny)on the PASCALVisualObjectClasses(VOC)dataset,our normalization accelerator achieves a 72%reduction in processing time and 83%lower power consumption compared to a 2.4 GHz Intel Central Processing Unit(CPU)software normalization implementation,while maintaining accuracy(0.51%mean Average Precision(mAP)drop at floating-point 32 bits(FP32),1.35%at brain floating-point 16 bits(bfloat16)).When integrated into a neural processing unit(NPU),the design demonstrates 63%and 97%performance improvements over AMD CPU and Reduced Instruction Set Computing-V(RISC-V)implementations,respectively.These results confirm that our proposed BN hardware design enables efficient,high-accuracy,and power-saving on-device training for modern CNNs.Our results demonstrate that efficient hardware implementation of standard batch normalization is achievable without sacrificing accuracy,enabling practical on-device CNN training with significantly reduced computational and power requirements.展开更多
Spiking neural networks(SNN)represent a paradigm shift toward discrete,event-driven neural computation that mirrors biological brain mechanisms.This survey systematically examines current SNN research,focusing on trai...Spiking neural networks(SNN)represent a paradigm shift toward discrete,event-driven neural computation that mirrors biological brain mechanisms.This survey systematically examines current SNN research,focusing on training methodologies,hardware implementations,and practical applications.We analyze four major training paradigms:ANN-to-SNN conversion,direct gradient-based training,spike-timing-dependent plasticity(STDP),and hybrid approaches.Our review encompasses major specialized hardware platforms:Intel Loihi,IBM TrueNorth,SpiNNaker,and BrainScaleS,analyzing their capabilities and constraints.We survey applications spanning computer vision,robotics,edge computing,and brain-computer interfaces,identifying where SNN provide compelling advantages.Our comparative analysis reveals SNN offer significant energy efficiency improvements(1000-10000×reduction)and natural temporal processing,while facing challenges in scalability and training complexity.We identify critical research directions including improved gradient estimation,standardized benchmarking protocols,and hardware-software co-design approaches.This survey provides researchers and practitioners with a comprehensive understanding of current SNN capabilities,limitations,and future prospects.展开更多
BACKGROUND No clear guidelines for long-term postoperative maintenance therapy have been established for patients with lung oligometastases from colorectal cancer(CRC)who achieve radiological no evidence of disease af...BACKGROUND No clear guidelines for long-term postoperative maintenance therapy have been established for patients with lung oligometastases from colorectal cancer(CRC)who achieve radiological no evidence of disease after radiofrequency ablation(RFA)treatment.We compared the outcomes of patients with lung oligometa-stases from CRC after RFA plus maintenance capecitabine with RFA alone.AIM To determine whether adding capecitabine to RFA improves prognosis compared with RFA alone.METHODS This multicenter retrospective study included consecutive patients from two tertiary cancer centers treated for pulmonary oligometastases from CRC between 2016 and 2023.Subjects were assigned to RFA plus capecitabine(combined)or RFA alone(only RFA)groups.Primary outcomes included overall survival(OS)and progression-free survival(PFS)survival and the secondary outcome was local tumor progression(LTP).The OS,PFS,and LTP rates were compared between the two groups.In addition,prognostic factors were identified using univariate and multivariate analyses.RESULTS Combination therapy(RFA+capecitabine,n=148)and RFA monotherapy(n=99)were compared in patients with CRC and lung metastases.The median OS was 37.8 months(22.4,50.3),the PFS was 18.7 months(13.0,36.5),and the LTP was 31.5 months(20.0,52.4)in the Only RFA group.The OS increased significantly(P=0.011)and the LTP decreased at all time points(P<0.001)in the combined group.The multivariate cox analysis revealed that combined chemotherapy significantly improved OS,with hazard ratios ranging from 0.29 to 0.35(all P<0.015)after adjusting for demographic,tumor,and treatment-related factors.The risk of death was consistently lower in the combination therapy group compared to RFA monotherapy.CONCLUSION RFA prolongs survival and local control in patients with CRC pulmonary oligometastases.Adjuvant capecitabine increases OS and reduces LTP compared to RFA alone,but PFS did not significantly change.展开更多
Objective:To evaluate the effectiveness of digital-intelligent health education for patients undergoing maintenance hemodialysis.Methods:From December 2023 to December 2024,82 patients undergoing maintenance hemodialy...Objective:To evaluate the effectiveness of digital-intelligent health education for patients undergoing maintenance hemodialysis.Methods:From December 2023 to December 2024,82 patients undergoing maintenance hemodialysis in our hospital were selected and randomly divided into an observation group(n=41,receiving routine health education)and a control group(n=41,receiving digital health education).The levels of knowledge,belief,and behavior related to dry weight control,as well as changes in dry weight and complications,were compared before and after intervention.Results:After intervention,the observation group had higher scores for knowledge(40.96±6.43),belief(39.11±6.39),behavior(39.66±5.78),and total score(119.04±13.11)compared to the control group(p<0.05).The observation group also showed better dry weight control than the control group(p<0.05).The total incidence of complications in the observation group(4.88%,2/41)was lower than that in the control group(21.95%,9/41)(p<0.05).Conclusion:The rational application of digital-intelligent health education can effectively maintain dry weight in patients undergoing maintenance hemodialysis,reduce complications,and improve patients’knowledge,belief,and behavior levels.This approach is worthy of promotion.展开更多
BACKGROUND Eosinophilic esophagitis(EoE)is a chronic inflammatory disorder presenting as symptoms of dysphagia,esophageal food impaction,chest pain,and heartburn.After an initial trial of proton pump inhibitor(PPI)the...BACKGROUND Eosinophilic esophagitis(EoE)is a chronic inflammatory disorder presenting as symptoms of dysphagia,esophageal food impaction,chest pain,and heartburn.After an initial trial of proton pump inhibitor(PPI)therapy,swallowed topical corticosteroids(STC)are effective as induction therapy for EoE.However,out-come data for STC as a maintenance strategy is limited.RESULTS Three randomized control trials and one observational study were included,involving 303 patients(189 in the STC group,114 in the placebo-controlled group).Analysis showed that histologic recurrence was significantly lower with STC(OR:0.04,95%CI:0.01-0.28,P<0.00001,I^(2)=78%).Overall symptom recurrence was similar between groups(OR:0.23,95%CI:0.02-3.54,P=0.29,I^(2)=92%).On sensitivity analysis,symptom recurrence was significantly lower in the STC group(OR:0.05,95%CI:0.02-0.17,P=0.00001,I^(2)=39%).Odds of repeat dilation were significantly lower in the STC group(OR:0.14,95%CI:0.02-0.91,P=0.04,I^(2)=0%).Candida infection rates were similar between groups(OR:6.13,95%CI:0.85-44.26,P=0.07,I^(2)=24%).Proportion of concomitant PPI use was similar between groups(OR:1.64,95%CI:0.83-3.21,P=0.15,I^(2)=0%).CONCLUSION For patients who successfully achieved remission of EoE with STC induction therapy,maintaining treatment is effective in sustaining histologic remission,while newer regimens may be effective in preventing symptom recurrence compared to placebo.We found no significant difference for oropharyngeal/esophageal candidiasis with STC maintenance therapy.Future studies with longer follow-up periods are needed.展开更多
基金supported in part by the National Natural Science Foundation of China(No.52467008)Gansu Provincial Depatment of Education Youth Doctoral Suppo Project(2024QB-051).
文摘Addressing the limitations of inadequate stochastic disturbance characterization during wind turbine degradation processes that result in constrained modeling accuracy,replacement-based maintenance practices that deviate from actual operational conditions,and static maintenance strategies that fail to adapt to accelerated deterioration trends leading to suboptimal remaining useful life utilization,this study proposes a Time-Based Incomplete Maintenance(TBIM)strategy incorporating reliability constraints through stochastic differential equations(SDE).By quantifying stochastic interference via Brownian motion terms and characterizing nonlinear degradation features through state influence rate functions,a high-precision SDE degradation model is constructed,achieving 16%residual reduction compared to conventional ordinary differential equation(ODE)methods.The introduction of age reduction factors and failure rate growth factors establishes an incomplete maintenance mechanism that transcends traditional“as-good-as-new”assumptions,with the TBIM model demonstrating an additional 8.5%residual reduction relative to baseline SDE approaches.A dynamic maintenance interval optimization model driven by dual parameters—preventive maintenance threshold R_(p) and replacement threshold R_(r)—is designed to achieve synergistic optimization of equipment reliability and maintenance economics.Experimental validation demonstrates that the optimized TBIM extends equipment lifespan by 4.4%and reducesmaintenance costs by 4.16%at R_(p)=0.80,while achieving 17.2%lifespan enhancement and 14.6%cost reduction at R_(p)=0.90.This methodology provides a solution for wind turbine preventive maintenance that integrates condition sensitivity with strategic foresight.
基金support of(1)Grant Ref.PID2023-148456OB-C41 and(2)Grant Ref.RED2022-134588-T found by MICIU/AEI/10.13039/501100011033。
文摘This paper presents an artificial intelligence(AI)-based observer that combines fuzzy logic and neural networks to detect abnormalities in sensors embedded in an electrolyser.Electrolysers are hydrogen production plants that require effective maintenance to guarantee suitable operation,prevent degradation,and avoid loss of efficiency.In this sense,predictive maintenance arises as one of the most advisable techniques for maintenance in electrolysers by using sensor data to predict potential abnormalities.However,if the sensor fails,there will be an incorrect forecasting of abnormalities.Among the different types of operational faults that sensors can present are drift-related faults,which are probably the most difficult to detect due to a slow but progressive loss of accuracy in measurements.Another problem with predictive maintenance is that it often requires enormous training data,which is not available at the early stage of plant operation.The developed fuzzy system is responsible for detecting faulty readings arising from drift sensor signals,while the neural network complements the function of the fuzzy system by predicting sensor signals when enough training data are available.The AI-based observer and the fuzzy rules are validated in an experimental case study with a 1 Nm^(3)/h electrolyser.The selected variables are electrolyser temperature and efficiency.Experimental results show that the rules of the fuzzy component of the AI-based observer guarantee an accuracy of±0.25 within the range of 0 to 1,and the neural network component predicted correct sensor values with a root mean square error(RMSE)as low as 0.0016.The authors’approach helps to determine drift faults without additional sensors or components installed in the plant.
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.
文摘Modern industrial environments require uninterrupted machinery operation to maintain productivity standards while ensuring safety and minimizing costs.Conventional maintenance methods,such as reactive maintenance(i.e.,run to failure)or time-based preventive maintenance(i.e.,scheduled servicing),prove ineffective for complex systems with many Internet of Things(IoT)devices and sensors because they fall short in detecting faults at early stages when it is most crucial.This paper presents a predictive maintenance framework based on a hybrid deep learning model that integrates the capabilities of Long Short-Term Memory(LSTM)Networks and Convolutional Neural Networks(CNNs).The framework integrates spatial feature extraction and temporal sequence modeling to accurately classify the health state of industrial equipment into three categories,including Normal,Require Maintenance,and Failed.The framework uses a modular pipeline that includes IoT-enabled data collection along with secure transmission methods to manage cloud storage and provide real-time fault classification.The FD004 subset of the NASA C-MAPSS dataset,containing multivariate sensor readings from aircraft engines,serves as the training and evaluation data for the model.Experimental results show that the LSTM-CNN model outperforms baseline models such as LSTM-SVM and LSTM-RNN,achieving an overall average accuracy of 86.66%,precision of 86.00%,recall of 86.33%,and F1-score of 86.33%.Contrary to the previous LSTM-CNN-based predictive maintenance models that either provide a binary classification or rely on synthetically balanced data,our paper provides a three-class maintenance state(i.e.,Normal,Require Maintenance,and Failed)along with threshold-based labeling that retains the true nature of the degradation.In addition,our work also provides an IoT-to-cloud-based modular architecture for deployment.It offers Computerized Maintenance Management System(CMMS)integration,making our proposed solution not only technically sound but also practical and innovative.The solution achieves real-world industrial deployment readiness through its reliable performance alongside its scalable system design.
文摘With the progress of the times and the development of information technology, the computer has become a life tool. Due to its intelligent convenience, it is widely used in daily life, but the inevitable failure problems of the machine, or the loss of the computer and data, and causes immeasurable trouble and loss.
基金supported by Ministry of Education of Singapore,under Academic Research Fund Tier 1(Grant Number RG143/23).
文摘Scientific and technological advancements are rapidly transforming underground engineering,shifting from labor-intensive,time-consuming methods to automated,real-time systems.This timely and comprehensive review covers in-situ testing,intelligent monitoring,and geophysical testing methods,highlighting fundamental principles,testing apparatuses,data processing techniques,and engineering applications.The state-of-the-art summary emphasizes not only cutting-edge innovations for complex and harsh environments but also the transformative role of artificial intelligence and machine learning in data interpretations.The integration of big data and advanced algorithms is particularly impactful,enabling the identification,prediction,and mitigation of potential risks in underground projects.Key aspects of the discussion include detection capabilities,method integration,and data convergence of intelligent technologies to drive enhanced safety,operational efficiency,and predictive reliability.The review also examines future trends in intelligent technologies,emphasizing unified platforms that combine multiple methods,real-time data,and predictive analytics.These advancements are shaping the evolution of underground construction and maintenance,aiming for risk-free,high-efficiency underground engineering.
基金supported by the National Natural Science Foundation of China(51877061).
文摘Preventive maintenance(PM)enhances the reliability of an ultra-high-voltage DC(UHVDC)line.However,it introduces new states and complicates the task of determining the reliability parameters of the component with PM.The hierarchical connection(HC)of the inverter improves the flexibility of UHVDC,but the state spaces of the inverter and UHVDC-HC are determined by capacity distribution between the high terminal(HT)and low terminal(LT),which has not been studied yet.In this paper,the reliability of UHVDC-HC with the PM is studied.First,with the impact of PM described by a nonlinear coefficient,the state space of the components with PM is aggregated.It is the same size as without PM.The only difference is the transition rate.Second,considering power distribution between the HT/LT and its dependence on the state of the rectifier,new capacity levels of HT/LT,e.g.,75%/2,50%/3,are defined.Third,with capacity levels of the HT/LT differentiated,2 existing reliability indices are extended,and 2 are newly defined.Finally,a sensitivity model of the reliability of the UHVDC-HC to its parameters and the PM period is proposed.
基金supported by the National Natural Science Foundation of China(51767017)the Basic Research Innovation Group Project of Gansu Province(18JR3RA133)the Industrial Support and Guidance Project of Universities in Gansu Province(2022CYZC-22).
文摘This study proposes a novel visual maintenance method for photovoltaic(PV)modules based on a two-stage Wiener degradation model,addressing the limitations of traditional PV maintenance strategies that often result in insufficient or excessive maintenance.The approach begins by constructing a two-stage Wiener process performance degradation model and a remaining life prediction model under perfect maintenance conditions using historical degradation data of PV modules.This enables accurate determination of the optimal timing for postfailure corrective maintenance.To optimize the maintenance strategy,the study establishes a comprehensive cost model aimed at minimizing the long-term average cost rate.The model considers multiple cost factors,including inspection costs,preventive maintenance costs,restorative maintenance costs,and penalty costs associated with delayed fault detection.Through this optimization framework,the method determines both the optimal maintenance threshold and the ideal timing for predictive maintenance actions.Comparative analysis demonstrates that the twostage Wiener model provides superior fitting performance compared to conventional linear and nonlinear degradation models.When evaluated against traditional maintenance approaches,including Wiener process-based corrective maintenance strategies and static periodic maintenance strategies,the proposed method demonstrates significant advantages in reducing overall operational costs while extending the effective service life of PV components.The method achieves these improvements through effective coordination between reliability optimization and economic benefit maximization,leading to enhanced power generation performance.These results indicate that the proposed approach offers a more balanced and efficient solution for PV system maintenance.
基金financially supported by the Scientific Research Projects of the Education Department of Zhejiang Province(Grant No.Y202454744)the Ningbo Public Welfare Science and Technology Project(Grant Nos.2023S007 and 2023S165)the Key Research and Development Program of Zhejiang(Grant No.2023C03183).
文摘As an essential part of the urban infrastructure,underground utility tunnels have a long service life,complex structural performance evolution and dynamic changes both inside and outside the tunnel.These combined factors result in a wide variety of disaster risks during the operation and maintenance phase,which make risk management and control particularly challenging.This work first reviews three common representative disaster factors during the operation and maintenance period:settlement,earthquakes,and explosions.It summarizes the causes of disasters,key technologies,and research methods.Then,it delves into the research on the intelligent operation and maintenance architecture for utility tunnels.Additionally,it explores the data challenges,monitoring technologies,and management platform architectures faced during the operation and maintenance process.This work provides new research perspectives for the long-term,healthy,and sustainable development of utility tunnels,which serve as the underground arteries of cities.
基金Supported by Key Research Fund of Wannan Medical College,No.WK2021ZF15Research Foundation for Advanced Talents of Wannan Medical College,No.YR202213+3 种基金Foundation of Anhui Educational Committee,No.2023AH051759Excellent Youth Research Project of Anhui UniversitiesNo.2023AH030107Horizontal Project of Wannan Medical College,No.622202504003 and No.662202404013.
文摘BACKGROUND Research examining the relationships among anxiety,depression,self-perceived burden(SPB),and psychological resilience(PR),along with the determinants of PR,in patients with chronic renal failure(CRF)receiving maintenance hemodia-lysis(MHD)is limited.AIM To investigate the correlation between anxiety,depression,SPB,and PR in pati-ents with CRF on MHD.METHODS This study included 225 patients with CRF on MHD who were admitted between June 2021 and June 2024.The anxiety level was evaluated using the Self-Rating Anxiety Scale(SAS);the depression status was assessed using the Self-Rating Depression Scale(SDS);the SPB was measured using the SPB Scale(SPBS);and the PR was determined using the Connor–Davidson Resilience Scale(CD-RISC).The correlations among the SAS,SDS,SPB,and CD-RISC were analyzed using Pearson’s correlation coefficients.Univariate and multivariate analyses were performed to identify the factors that influence the PR of patients with CRF on MHD.RESULTS The SAS,SDS,SPB,and CD-RISC scores of the 225 patients were 45.25±15.36,54.81±14.68,32.31±11.52,and 66.48±9.18,respectively.Significant negative correlations were observed between SAS,SDS,SPB,and CD-RISC.Furthermore,longer dialysis vintage(P=0.015),the absence of religious beliefs(P=0.020),lower monthly income(P=0.008),higher SAS score(P=0.013),and higher SDS score(P=0.006)were all independent factors that adversely affected the PR of patients with CRF on MHD.CONCLUSION Patients with CRF on MHD present with varying degrees of anxiety,depression,and SPB,all of which exhibit a significant negative correlation with their PR.Moreover,longer dialysis vintage,the absence of religious beliefs,lower monthly income,higher SAS score,and higher SDS score were factors that negatively affected the PR of patients with CRF on MHD.
文摘While Artificial Intelligence (AI) is leading the way in terms of hardware advancements, such as GPUs, memory, and processing power, real-time applications are still catching up. It is inevitable that when one aspect leads and other trails behind, they coexist in life, as is often the case. The trailing aspect cannot remain far behind because, without application and use, there would be a dead end. Everything, whether an object, software, or tool, must have a practical use for humans. Without this, it will become obsolete. We can see this in many instances, such as blockchain technology, which is superior yet faces challenges in practical implementation, leading to a decline in adoption. This publication aims to bridge the gap between AI advancements and maintenance, specifically focusing on making predictive maintenance a practical application. There are multiple building blocks that make predictive maintenance a practical application. Each block performs a function leading to an output. This output forms an input to the receiving block. There are also foundational parts for all these building blocks to perform a function. Eventually, once the building blocks are connected, they form a loop and start to lead the path to predictive maintenance. Predictive maintenance is indeed practically achievable, but one must comprehend all the building blocks necessary for its implementation. Although detailed explanations will be provided in the upcoming sections, it is important to understand that simply purchasing software and plugging it in might be a far-fetched approach.
文摘This paper presents a project aimed at developing a trilingual visual dictionary for aircraft maintenance professionals and students.The project addresses the growing demand for accurate communication and technical terminology in the aviation industry,particularly in Brazil and China.The study employs a corpus-driven approach,analyzing a large corpus of aircraft maintenance manuals to extract key technical terms and their collocates.Using specialized subcorpora and a comparative analysis,this paper demonstrates challenges and solutions into the identification of high-frequency keywords and explores their contextual use in aviation documentation,emphasizing the need for clear and accurate technical communication.By incorporating these findings into a trilingual visual dictionary,the project aims to enhance the understanding and usage of aviation terminology.
文摘Tunnel facility management(FM)is crucial for ensuring safety,efficiency,and resilience of tunnel infrastructure.Current FM practices,such as reactive and preventive maintenance,have limitations-reactive maintenance can't prevent failures,and preventive maintenance can't predict asset maintenance needs,leading to costly and inefficient processes.In addition,existing computerized tunnel FM systems face various challenges,including the lack of integrated real-time monitoring information,visualization of assets in a three-dimensional(3D)environment,supporting predictive maintenance,and scaling into long infrastructure with complicated spatiotemporal relationships.This study addresses these limitations by proposing a data-driven digital twin(DT)-based framework that supports predictive maintenance to improve tunnel FM processes and enhance resilience.The proposed framework consists of six layers,allowing the integration of data from monitoring system,FM system,and building information modeling(BIM)models.The framework proposes a flexible tunnel data model and classification system that hierarchically divides the tunnel models,ensuring an efficient data connection from the physical twin to the DT.The system was implemented in a tunnel case study that generates maintenance plans and work orders using historical and current condition monitoring data,and the 3D visualization technology suggests maintenance and repair processes,making the FM decision process more effective.The proposed system detected and predicted the twin state based on a data-driven analysis,and the prediction accuracy of the machine learning models was sufficiently high for use in real scenarios to make FM plans in advance and prevent asset failures.The proposed framework is contributing to the infrastructure resilience by enhancing the tunnel system ability to predict the maintenance tasks and prevent failures using data-driven DT technology.
文摘Predictive maintenance plays a crucial role in preventing equipment failures and minimizing operational downtime in modern industries.However,traditional predictive maintenance methods often face challenges in adapting to diverse industrial environments and ensuring the transparency and fairness of their predictions.This paper presents a novel predictive maintenance framework that integrates deep learning and optimization techniques while addressing key ethical considerations,such as transparency,fairness,and explainability,in artificial intelligence driven decision-making.The framework employs an Autoencoder for feature reduction,a Convolutional Neural Network for pattern recognition,and a Long Short-Term Memory network for temporal analysis.To enhance transparency,the decision-making process of the framework is made interpretable,allowing stakeholders to understand and trust the model’s predictions.Additionally,Particle Swarm Optimization is used to refine hyperparameters for optimal performance and mitigate potential biases in the model.Experiments are conducted on multiple datasets from different industrial scenarios,with performance validated using accuracy,precision,recall,F1-score,and training time metrics.The results demonstrate an impressive accuracy of up to 99.92%and 99.45%across different datasets,highlighting the framework’s effectiveness in enhancing predictive maintenance strategies.Furthermore,the model’s explainability ensures that the decisions can be audited for fairness and accountability,aligning with ethical standards for critical systems.By addressing transparency and reducing potential biases,this framework contributes to the responsible and trustworthy deployment of artificial intelligence in industrial environments,particularly in safety-critical applications.The results underscore its potential for wide application across various industrial contexts,enhancing both performance and ethical decision-making.
基金supported by the Deyang City Science and Technology Planning Project[Grant Number 2023SZZ010].
文摘Objective:To investigate the clinical efficacy and cost-effectiveness of combined hemodialysis(HD)and hemoperfusion(HP)therapy in managing secondary hyperparathyroidism(SHPT)in patients undergoing maintenance hemodialysis(MHD).Methods:A total of 195 patients with MHD and SHPT at Deyang People's Hospital from April 2024 to April 2025 were enrolled.Patients were randomly assigned to a control group receiving standard HD treatment and an experimental group receiving HD combined with HP therapy.The experimental group underwent 1 year of observation(97 cases in the experimental group,98 cases in the control group).During treatment,changes in parathyroid hormone(PTH),serum calcium,serum phosphorus,and inflammatory factors were monitored,along with analysis of treatment-related economic benefits and safety.Results:The experimental group demonstrated significantly better reductions in PTH,serum phosphorus,and inflammatory factors compared to the control group(P<0.05).Although the total treatment cost was slightly higher,the unit cost per therapeutic effect was lower,resulting in a superior cost-effectiveness ratio.Conclusion:Combined HD and HP therapy can significantly improve SHPT-related indicators in MHD patients,demonstrating safety,controllability,and high cost-effectiveness,making it a clinically applicable and recommended treatment option.
基金supported by the National Research Foundation of Korea(NRF)grant for RLRC funded by the Korea government(MSIT)(No.2022R1A5A8026986,RLRC)supported by Institute of Information&Communications Technology Planning&Evaluation(IITP)grant funded by the Korea government(MSIT)(No.2020-0-01304,Development of Self-Learnable Mobile Recursive Neural Network Processor Technology)+3 种基金supported by the MSIT(Ministry of Science and ICT),Republic of Korea,under the Grand Information Technology Research Center support program(IITP-2024-2020-0-01462,Grand-ICT)supervised by the IITP(Institute for Information&Communications Technology Planning&Evaluation)supported by the Korea Technology and Information Promotion Agency for SMEs(TIPA)supported by the Korean government(Ministry of SMEs and Startups)’s Smart Manufacturing Innovation R&D(RS-2024-00434259).
文摘On-device Artificial Intelligence(AI)accelerators capable of not only inference but also training neural network models are in increasing demand in the industrial AI field,where frequent retraining is crucial due to frequent production changes.Batch normalization(BN)is fundamental to training convolutional neural networks(CNNs),but its implementation in compact accelerator chips remains challenging due to computational complexity,particularly in calculating statistical parameters and gradients across mini-batches.Existing accelerator architectures either compromise the training accuracy of CNNs through approximations or require substantial computational resources,limiting their practical deployment.We present a hardware-optimized BN accelerator that maintains training accuracy while significantly reducing computational overhead through three novel techniques:(1)resourcesharing for efficient resource utilization across forward and backward passes,(2)interleaved buffering for reduced dynamic random-access memory(DRAM)access latencies,and(3)zero-skipping for minimal gradient computation.Implemented on a VCU118 Field Programmable Gate Array(FPGA)on 100 MHz and validated using You Only Look Once version 2-tiny(YOLOv2-tiny)on the PASCALVisualObjectClasses(VOC)dataset,our normalization accelerator achieves a 72%reduction in processing time and 83%lower power consumption compared to a 2.4 GHz Intel Central Processing Unit(CPU)software normalization implementation,while maintaining accuracy(0.51%mean Average Precision(mAP)drop at floating-point 32 bits(FP32),1.35%at brain floating-point 16 bits(bfloat16)).When integrated into a neural processing unit(NPU),the design demonstrates 63%and 97%performance improvements over AMD CPU and Reduced Instruction Set Computing-V(RISC-V)implementations,respectively.These results confirm that our proposed BN hardware design enables efficient,high-accuracy,and power-saving on-device training for modern CNNs.Our results demonstrate that efficient hardware implementation of standard batch normalization is achievable without sacrificing accuracy,enabling practical on-device CNN training with significantly reduced computational and power requirements.
文摘Spiking neural networks(SNN)represent a paradigm shift toward discrete,event-driven neural computation that mirrors biological brain mechanisms.This survey systematically examines current SNN research,focusing on training methodologies,hardware implementations,and practical applications.We analyze four major training paradigms:ANN-to-SNN conversion,direct gradient-based training,spike-timing-dependent plasticity(STDP),and hybrid approaches.Our review encompasses major specialized hardware platforms:Intel Loihi,IBM TrueNorth,SpiNNaker,and BrainScaleS,analyzing their capabilities and constraints.We survey applications spanning computer vision,robotics,edge computing,and brain-computer interfaces,identifying where SNN provide compelling advantages.Our comparative analysis reveals SNN offer significant energy efficiency improvements(1000-10000×reduction)and natural temporal processing,while facing challenges in scalability and training complexity.We identify critical research directions including improved gradient estimation,standardized benchmarking protocols,and hardware-software co-design approaches.This survey provides researchers and practitioners with a comprehensive understanding of current SNN capabilities,limitations,and future prospects.
基金Supported by the National Natural Science Foundation of China,No.82072034。
文摘BACKGROUND No clear guidelines for long-term postoperative maintenance therapy have been established for patients with lung oligometastases from colorectal cancer(CRC)who achieve radiological no evidence of disease after radiofrequency ablation(RFA)treatment.We compared the outcomes of patients with lung oligometa-stases from CRC after RFA plus maintenance capecitabine with RFA alone.AIM To determine whether adding capecitabine to RFA improves prognosis compared with RFA alone.METHODS This multicenter retrospective study included consecutive patients from two tertiary cancer centers treated for pulmonary oligometastases from CRC between 2016 and 2023.Subjects were assigned to RFA plus capecitabine(combined)or RFA alone(only RFA)groups.Primary outcomes included overall survival(OS)and progression-free survival(PFS)survival and the secondary outcome was local tumor progression(LTP).The OS,PFS,and LTP rates were compared between the two groups.In addition,prognostic factors were identified using univariate and multivariate analyses.RESULTS Combination therapy(RFA+capecitabine,n=148)and RFA monotherapy(n=99)were compared in patients with CRC and lung metastases.The median OS was 37.8 months(22.4,50.3),the PFS was 18.7 months(13.0,36.5),and the LTP was 31.5 months(20.0,52.4)in the Only RFA group.The OS increased significantly(P=0.011)and the LTP decreased at all time points(P<0.001)in the combined group.The multivariate cox analysis revealed that combined chemotherapy significantly improved OS,with hazard ratios ranging from 0.29 to 0.35(all P<0.015)after adjusting for demographic,tumor,and treatment-related factors.The risk of death was consistently lower in the combination therapy group compared to RFA monotherapy.CONCLUSION RFA prolongs survival and local control in patients with CRC pulmonary oligometastases.Adjuvant capecitabine increases OS and reduces LTP compared to RFA alone,but PFS did not significantly change.
文摘Objective:To evaluate the effectiveness of digital-intelligent health education for patients undergoing maintenance hemodialysis.Methods:From December 2023 to December 2024,82 patients undergoing maintenance hemodialysis in our hospital were selected and randomly divided into an observation group(n=41,receiving routine health education)and a control group(n=41,receiving digital health education).The levels of knowledge,belief,and behavior related to dry weight control,as well as changes in dry weight and complications,were compared before and after intervention.Results:After intervention,the observation group had higher scores for knowledge(40.96±6.43),belief(39.11±6.39),behavior(39.66±5.78),and total score(119.04±13.11)compared to the control group(p<0.05).The observation group also showed better dry weight control than the control group(p<0.05).The total incidence of complications in the observation group(4.88%,2/41)was lower than that in the control group(21.95%,9/41)(p<0.05).Conclusion:The rational application of digital-intelligent health education can effectively maintain dry weight in patients undergoing maintenance hemodialysis,reduce complications,and improve patients’knowledge,belief,and behavior levels.This approach is worthy of promotion.
文摘BACKGROUND Eosinophilic esophagitis(EoE)is a chronic inflammatory disorder presenting as symptoms of dysphagia,esophageal food impaction,chest pain,and heartburn.After an initial trial of proton pump inhibitor(PPI)therapy,swallowed topical corticosteroids(STC)are effective as induction therapy for EoE.However,out-come data for STC as a maintenance strategy is limited.RESULTS Three randomized control trials and one observational study were included,involving 303 patients(189 in the STC group,114 in the placebo-controlled group).Analysis showed that histologic recurrence was significantly lower with STC(OR:0.04,95%CI:0.01-0.28,P<0.00001,I^(2)=78%).Overall symptom recurrence was similar between groups(OR:0.23,95%CI:0.02-3.54,P=0.29,I^(2)=92%).On sensitivity analysis,symptom recurrence was significantly lower in the STC group(OR:0.05,95%CI:0.02-0.17,P=0.00001,I^(2)=39%).Odds of repeat dilation were significantly lower in the STC group(OR:0.14,95%CI:0.02-0.91,P=0.04,I^(2)=0%).Candida infection rates were similar between groups(OR:6.13,95%CI:0.85-44.26,P=0.07,I^(2)=24%).Proportion of concomitant PPI use was similar between groups(OR:1.64,95%CI:0.83-3.21,P=0.15,I^(2)=0%).CONCLUSION For patients who successfully achieved remission of EoE with STC induction therapy,maintaining treatment is effective in sustaining histologic remission,while newer regimens may be effective in preventing symptom recurrence compared to placebo.We found no significant difference for oropharyngeal/esophageal candidiasis with STC maintenance therapy.Future studies with longer follow-up periods are needed.
