NetWork
Fusion localization system and experimental design based on multi-sensor fusion
YIN Shu;SUN Hu;FU Minglei;ZHANG Wen'an;[Objective]With the rapid development of intelligent robotics and autonomous driving technologies， accurate and robust vehicle localization has become one of the fundamental prerequisites for environment perception， motion planning， and autonomous decision-making. However， in practical applications， localization systems operating in complex environments are frequently affected by abnormal measurement， multipath interference， and sensor uncertainties， which significantly degrade localization accuracy and system reliability. In particular， GNSS-based positioning methods are highly susceptible to environmental disturbances in urban canyons， indoor–outdoor transition regions， and occluded scenarios， resulting in severe localization drift or even temporary signal loss. Therefore， the integration of sensor information through multi-sensor fusion techniques has become an important research direction for improving localization accuracy， robustness， and environmental adaptability in intelligent unmanned systems. To enhance students understanding of vehicle localization technologies and the theoretical foundations of fusion-based estimation， a multi-sensor-based localization experimental platform was designed as part of the robotics-related courses， focusing on a collaborative localization system based on multi-sensor integration. [Methods]This experiment introduces a fusion localization framework for intelligent unmanned vehicle， integrating Ultra-Wideband (UWB)， Global Navigation Satellite System (GNSS)， and Inertial Measurement Unit (IMU) sensors. The experiment aims to bridge the gap between theoretical learning and engineering implementation by enabling students to understand the practical workflow of sensor fusion localization systems， including data acquisition， state estimation， information fusion， and robustness analysis under disturbed environments. Specifically， GNSS measurements provide global positioning information， UWB measurements compensate for localization degradation under signal occlusion and interference conditions， while the IMU enables continuous motion-state propagation and short-term state estimation. By exploiting the complementary characteristics of heterogeneous sensor measurements， the proposed multi-sensor fusion framework effectively integrates multi-dimensional information to improve localization accuracy， reliability， and environmental adaptability. Through the collaborative utilization of multi-dimensional sensor information， the system fully exploits the complementary properties among different measurements， thereby improving localization accuracy， estimation consistency， and robustness against abnormal observations and environmental disturbances. The experiment also enables students to gain a deeper understanding of the practical implementation of fusion filtering algorithms， including state prediction， measurement update， and uncertainty propagation in unmanned vehicle localization tasks. In addition， students can intuitively analyse the influence of sensor noise， measurement uncertainties， and environmental interference on localization performance through experimental observations and comparative analyses. [Results and Conclusions]Experimental results demonstrate that the proposed multi-sensor fusion localization scheme can effectively suppress the influence of abnormal measurements and compensate for localization degradation caused by environmental interference. Compared with single-sensor localization methods， the proposed fusion framework exhibits superior localization accuracy， stronger robustness， and improved stability under complex environments. The designed experimental platform not only provides an effective educational tool for robotics and intelligent vehicle courses， but also offers practical guidance for understanding the engineering applications of multi-sensor fusion localization technologies in autonomous systems.
Design of a miniaturized fiber-optic photoacoustic gas sensing system for high-performance detection
YANG Beilei;WANG Yifan;SONG Jie;GUO Min;[Objective] This work aims to overcome the limitations of traditional gas sensing methods in harsh and confined environments. Semiconductor and electrochemical sensors are constrained by significant cross-sensitivity and poor electromagnetic compatibility， and they often struggle to meet the stringent intrinsic safety standards required for specialized environments. Although photoacoustic spectroscopy offers exceptional selectivity and sensitivity， its reliance on resonant cavity dimensions and electromagnetic microphones restricts its deployment in space-constrained scenarios. To address these limitations， this study focuses on developing a gas sensing device that integrates high sensitivity， compactness， intrinsic safety， and electromagnetic immunity， thereby enabling reliable trace gas monitoring essential for critical safety and process control in industrial applications. [Methods] The proposed sensor integrates a dual-enhancement mechanism within a miniaturized non-resonant photoacoustic cell of only 1 mL. A Herriott-type multi-pass configuration， employing two coaxially aligned concave mirrors， extends the effective optical path length to approximately 440 mm. This represents an order-of-magnitude enhancement compared to a single-pass configuration， significantly amplifying the light-gas interaction and consequently improving the photoacoustic excitation efficiency without increasing the physical dimensions of the cell. To achieve high-sensitivity signal detection， a cantilever-enhanced fiber-optic Fabry-Perot acoustic sensor is incorporated into the photoacoustic cell. Vibration of the cantilever modulates the length of the Fabry-Perot cavity formed between its surface and the end face of the optical fiber ferrule， enabling all-optical detection of the photoacoustic signal. The complete measurement system employs a distributed feedback laser as the excitation source， wavelength-modulated at half the cantilever's resonant frequency to enable second-harmonic detection， which effectively suppresses the fundamental frequency noise. [Results] Comprehensive experimental investigations were conducted to evaluate the performance characteristics of the proposed sensor using methane as the target analyte. Frequency response measurements clearly demonstrated resonant enhancement， with the second-harmonic photoacoustic signal amplitude peaking sharply at the cantilever resonance frequency of 1970 Hz， thereby confirming the optimal operating conditions. Spectral analysis of the time-domain signal acquired at a methane concentration of 50 μL/L revealed a distinct frequency component matching the second harmonic of the modulation frequency， validating successful signal excitation and detection. The sensor exhibited excellent linear response across methane concentrations ranging from 10 to 50μL/L， with a calibration sensitivity of 36.73 pm/(μL/L)， demonstrating outstanding measurement repeatability and linearity. Based on the characteristic absorption coefficient of methane at 1650.9 nm， the normalized noise equivalent absorption coefficient was calculated to be 9.96×10-10 cm-1 W/Hz1/2， representing state-of-the-art performance for non-resonant photoacoustic systems. Allan-Werle deviation analysis further revealed that extending the integration time to 100 seconds could improve the minimum detection limit to 7.24nL/L， demonstrating excellent long-term stability. [Conclusions] This study successfully demonstrates a miniaturized all-optical photoacoustic gas sensor， which enhances optical absorption through a Herriott-type multi-pass cell configuration and achieves mechanical resonance amplification via cantilever-enhanced fiber-optic acoustic detection. The sensor exhibits excellent methane detection performance while maintaining a photoacoustic cell volume of only 1 mL. The all-optical architecture ensures immunity to electromagnetic interference， intrinsic safety for operation in explosive environments， low transmission loss for remote monitoring， and passive operation at the sensing end. Moreover， the requirement of only a 1 mL sample volume makes it particularly valuable for trace gas analysis where sample conservation is essential， as well as for online laboratory testing applications. From an educational perspective， this sensor system integrates several fundamental principles， including the photoacoustic effect， multi-pass optical path design， fiber-optic interferometry， and mechanical resonance. It provides an ideal experimental platform for undergraduate and graduate students in areas such as optical system design， signal processing， and interdisciplinary applications. The demonstrated dual-enhancement strategy offers a promising technological pathway for developing next-generation high-performance miniature gas sensors suitable for harsh environments.
Teaching practice: Digital design and performance optimization of the cam mechanism of an on-load tap changer
ZHOU Yicong;MAO Guoxin;WANG Yichen;DUAN Jinyan;[Objective] Cam mechanisms are extensively employed in complex precision mechanical systems， due to their ability to deliver precise switching actions and high motion reliability. As a fundamental teaching content in mechanical engineering curricula—including courses such as Theory of Machines and Fundamentals of Mechanical Design—the conventional teaching mode of cam mechanism design has long relied on graphical and analytical techniques. These traditional modes， however， often suffer from limited digital integration， insufficient depth in performance optimization， and a lack of engineering relevance in pedagogical case studies. To bridge these gaps， this work presents an integrated digital design and performance optimization methodology focused on the main contact cam mechanism of an on?load tap changer (OLTC)， a critical component deployed in ultrahigh voltage (UHV) transmission systems. [Methods] This teaching case is inherently rich in engineering significance and possesses substantial educational value. UHV transmission technology represents China’s flagship achievement in globally leading high?end manufacturing， while the OLTC is the sole core component within UHV converter transformers that undergoes frequent operational cycling. By anchoring the instruction in a real?world major engineering project， this teaching design effectively merges theoretical principles with practical application. The methodological framework adopted in this work integrates 3D modeling， multibody dynamic simulation， finite element analysis， and fatigue life calculation， thereby encompassing the complete iterative workflow of “design–modeling–analysis–optimization” for cam mechanisms. This comprehensive digital framework provides practical guidance for engaging students in modern engineering practices of cam mechanism development within the context of nationally important infrastructure projects. [Results] This study systematically addresses the design challenges through the following key steps and corresponding findings: 1) A double?layer integrated cam mechanism is designed and modeled. This configuration employs distributed cam profiles on two distinct layers， combined with a staggered arrangement of moving contact pins. Such a layout achieves reciprocating motion and controlled intermittent movement of the moving contacts， ensuring reliable switching performance under operational loads. 2) Multibody dynamics simulations are conducted to obtain the collision load between the cam and follower， as well as the complete load–time history throughout the operating cycle. The resulting dynamic load data are subsequently used as inputs for finite element analysis， which determines the detailed stress distribution during collision. Building upon the stress results， a fatigue life assessment is then performed to predict the durability of the mechanism under repeated loading. 3) The motion profile of the moving contact is optimized. By evaluating various cam profile designs， a comparative analysis is conducted focusing on two critical metrics: the peak collision force between the cam and moving contact， and the computed fatigue life of the cam mechanism. This optimization process identifies the cam profile that best balances dynamic performance with structural longevity， ultimately yielding a design that meets the prescribed service life requirements while maintaining smooth and reliable motion transfer. [Conclusions] This work demonstrates how a strong engineering case， drawn from a national project， can be leveraged to enhance the teaching of cam mechanism design. The presented teaching case effectively addresses the shortcomings of the conventional teaching mode for cam mechanism design， providing students with a practical reference for applying theoretical knowledge to real engineering problems.
Design and implementation of virtual simulation-based experimental teaching for quarantine forest diseases
LIN Sixi;DING Xiaolei;CHEN Wei;ZHU Lihua;YE Jianren;[Objective] Forest pathology is a fundamental and compulsory course widely offered in forestry colleges and universities in China. However， due to the epidemic characteristics of forest diseases， experimental teaching in forest pathology has long faced several challenges. Firstly， it is challenging to systematically execute the teaching within a limited time， as it is not feasible to comprehensively demonstrate the entire disease process and development patterns. Secondly， it is difficult to accurately depict the real situation， as the existing experimental conditions cannot fully simulate the wild environment. Thirdly， the teaching style is constrained due to the irreversible nature of inoculation experiments. Moreover， major forest diseases involve quarantine pathogens， making it inappropriate to conduct such experiments in open environments. The introduction of virtual simulation experiments provides a viable technical solution to address these challenges. [Methods] This project has been initiated to address a pressing national concern: the control of pine wilt disease， a highly significant quarantine disease in China. It focuses on the most critical prevention step (disease diagnosis) of this disease to carry out the experiment design. The project is supported by extensive teaching and research achievements accomplished by our research group over the past two decades. In addition， the project leveraged virtual simulation technology to introduce classic disease diagnosis experiments into a virtual laboratory， thereby enabling full reproduction of the entire process of pine wilt disease diagnosis and pathogenicity determination. The system recreates highly realistic virtual environments， including wild forests， laboratories， and greenhouses. The diagnostic process is meticulously structured into three primary modules: disease cognition， disease diagnosis， and pathogenicity determination， comprising 14 interactive steps corresponding to 24 assessment points. The multilevel assessment method enables comprehensive evaluation of the experimental outcomes and students’ abilities to analyze and solve problems in error-tolerant scenarios， while also helping optimize the curriculum system and experimental teaching content. [Results] The experiment successfully addressed several issues， such as the prolonged duration of pine wilt disease diagnosis experiments， the irreversibility of destructive experiments， the stringent requirements for experimental environments， and the lack of suitable conditions in non-infected areas. It has effectively compensated for the disadvantages of traditional forest pathology experiments. The experimental background data are derived from long-term scientific achievements obtained by top national research teams， providing substantial data support for the establishment of experimental projects， facilitating the transfer of recent advances into teaching， and ensuring the validity of contents and the accuracy of results while also reflecting innovation and cutting-edge techniques. The virtual experiment begins with a distressing depiction of pine wilt disease， which has led to a substantial mortality of pine trees and a consequent ecological crisis. The sequence of events commences with a simulation request from the National Forestry and Grassland Administration. This configuration will help users in dedicating themselves to experiments in disease diagnosis and pathogen identification. The program enables students to engage in experimental simulation， online interaction， and experimental assessment， thereby achieving closed-loop verification of Koch’s postulates within constrained teaching time， satisfying the systematic integrity of experimental teaching content. [Conclusions] This virtual simulation experiment underscores the importance of innovation in experimental teaching models and the advantages of virtual simulation technology. It expands the breadth and depth of experimental teaching in forest pathology， achieving a profound integration of modern information technology with experimental teaching methodologies. Additionally， it incorporates political elements to enhance students’ identification with their major. Through a dual-dimensional training model of “technical ability + political literacy，” it effectively cultivates students’ sense of social and professional mission， stimulates their patriotic sentiments of being a tree doctor， and provides strong support for cultivating high-quality agricultural and forestry talents in the new era.
Development and verification of a simulation test platform for aerospace pyrotechnic shock environment
ZHANG Jiaming;XIE Lu;QI Kai;WANG Wenrui;[Objective] In aerospace mechanical environments， pyroshock has garnered significant attention because of its high frequency， high amplitude， and transient broadband behavior. Conducting efficient ground-based pyroshock environment tolerance assessments for aerospace units and components is therefore imperative. To support such assessments， this paper proposes a tunable resonant fixture based on the principle of mechanical impact. This device utilizes a projectile to impact a specific resonant structure (e.g.， a beam， plate， or shell) at high speed， thereby converting the impact kinetic energy into a broadband vibration response. By adjusting the structural modal parameters， it achieves envelope control over the target Shock Response Spectrum (SRS)， enabling accurate simulation of aerospace pyroshock environments and analysis of product shock resistance. [Method] Built upon classical gas dynamics and incorporating corrections for local corner resistance， variable cross-section losses， and virtual mass effects， an improved internal ballistic equation suitable for a horizontal L-shaped air gun was established. This equation provides a theoretical foundation for the experimental platform. A pyroshock environment simulation test platform was constructed， primarily consisting of an L-shaped air gun excitation device， a resonant plate assembly， test specimen fixtures and supports， and a measurement and control system. The storage chamber pressure is regulated by an automatic pressure regulation device. A pneumatic valve instantaneously releases compressed gas， propelling the projectile to the target speed. This impact on the resonant plate generates a specific shock effect. The resonant plate， leveraging its intrinsic modal characteristics， amplifies or spectrally shapes the shock， thereby establishing a pyroshock simulation environment for the test specimen. A finite element model was established based on the device structure to simulate chamber pressure and projectile velocity. In the context of air gun impact tests， the projectile’s velocity was measured using a high-speed camera. This measurement was then validated against both simulation results and theoretical calculations. Five key variables， including chamber pressure， buffer gasket thickness， projectile length， impact position， and connection method， were systematically studied to ascertain their influence on the SRS. A comparative test case was designed for each parameter type using the controlled variable method. SRS curves at characteristic locations were obtained through testing and compared with finite element simulation results， using simulation data for cross-validation of experimental conclusions. [Results] The device is capable of releasing the projectile at a predetermined target speed. The maximum relative error observed between the measured velocity and the improved theoretical/simulation calculations was 4.57%. The maximum projectile velocity dispersion was observed in three tests， with an average result of 0.224 m/s at 0.6 MPa. In the validation tests of the pyroshock simulation platform， the maximum relative error for the peak value at the knee frequency in three repetitive shock tests was 2.11%. By increasing the chamber pressure， the overall energy level is elevated， while the characteristic frequencies remain largely unaffected. Increasing the thickness of the gasket has been shown to enhance high-frequency attenuation characteristics and mitigate local overloads on the plate system within a certain range. By increasing the projectile length and altering the impact position， it is evident that the participation of the plate system’s dominant modes is noticeably adjusted. This adjustment is accompanied by a high degree of sensitivity to the knee frequency and peak value. The alteration of the connection method and stiffness level results in a modification of the energy distribution among the structural paths， thereby exerting a substantial influence on the response amplitude distribution across the plate surface. [Conclusion] The theoretical calculation model， finite element model， and experimental results established in this paper demonstrate a high degree of agreement， thus achieving high-precision predictions of projectile velocity and the SRS of the resonant plate. The device under consideration enables the precise modulation of projectile impact speed. Through the implementation of parameter control， the configuration of the SRS can be modified to emulate a variety of pyroshock environments. The development of this device is of significant importance for the analysis of shock resistance performance in aerospace products.
Optimization of Critical Parameters for Cryo-Focused Ion Beam Milling
LI Xiaomin;LEI Jianlin;[Objective] Although cryo-electron microscopy (cryo-EM) is widely used for determining three-dimensional structures of isolated and purified biomacromolecules， high-resolution structural studies of cells and tissues in their native context still depend on the preparation of high-quality ultrathin sections. Focused ion beam (FIB) milling has emerged as a key technique for producing ultrathin cryo-lamellae from in situ samples， due to its minimal artifacts and precise targeting capability. This study systematically investigates critical factors influencing the quality of cryo-lamellae prepared by FIB milling， including sample vitrification quality， accurate temperature control of the cryo-system， ice deposition and contamination. [Methods] We studied the entire process from sample preparation to lamella assessment using an Aquilos 2 cryo-FIB/SEM microscope. Diverse biological samples—yeast cells， 293T cells and isolated muscle fibers—were prepared. Cells were vitrified by plunge freezing， while muscle fibers were pre-treated with glycerol before vitrification in an ethane/propane mixture for homogeneous vitrification. A systematic FIB milling protocol was established: initial coarse milling (1-3 nA beam current) created trenches and stress relief cuts， followed by sequential thinning with beam current reduced stepwise from 1 nA to 50 pA， and final polishing at beam currents as low as 10 pA to produce 100–200 nm lamellae. With this protocol， key operating parameters were meticulously optimized based on experimental results: (1) Sample vitrification quality was first assessed during FIB-SEM milling and then correlated with cryo-ET outcomes to identify failure signatures. (2) A temperature sensor was directly instrumented on the shuttle to measure the true thermal conditions at the sample， guiding the optimal waiting period before loading according to the cooling kinetics profile. (3) Ice deposition inside the microscope chamber was evaluated by imaging the lamella at regular intervals after coarse milling. Rapid ice accumulation led to a significant increase in lamella thickness-nearly doubling within 1.5 hours， for example-and caused edge curling， both of which degraded sample quality. (4) Based on a study of ice contamination mechanisms， a custom integrated loading device with an anti-contamination lid was specifically designed to reduce ice formation during sample package and transfer procedure and fully tested against standard methods. [Results] Primary challenge in cryo-sample preparation stems from inherent limitations of plunge-freezing. Effective vitrification requires a coordinated strategy of sample pre-treatment and freezing medium optimization， while for larger biological specimens， high-pressure freezing is essential to achieve uniform vitrification. A consistent measurement bias was observed， with the system temperature underreported relative to the actual grid temperature， necessitating direct temperature calibration. A major finding was the detrimental impact of rapid ice deposition on lamella integrity. Enhancing the chamber vacuum via system upgrades effectively mitigated this issue， maintaining stable lamella thickness throughout extended chamber sessions. An integrated loading device was developed to reduce ice contamination during package and transfer. Additionally， key technical parameters—platinum coating uniformity， ion beam milling settings， and machining precision—were also identified as critical factors for producing high-quality lamellae. [Conclusions] By integrating systematic analysis with experimental data， we demonstrate that complete vitrification is prerequisite for successful milling， reveal the critical discrepancy and its practical implications between the displayed system temperature and the actual sample temperature， and confirm that enhancing chamber vacuum is vital for controlling ice deposition， thereby providing an effective solution to reduce ice contamination. This work presents a reliable workflow and optimized strategies， offering concrete guidance to improve the robustness and reproducibility of FIB-based thinning for in situ structural biology.
Experimental study of the deformation characteristics and bearing mechanism of gravel compaction in goafs
JI Guangrui;MA Zhenqian;ZHOU Jinlian;[Objective] The stability of gob-side roadways， especially that of narrow coal pillars， is critically influenced by the compaction behavior， crushing expansion characteristics， and lateral stress of crushed gangue in the goaf. Understanding the deformation characteristics and bearing mechanism of this crushed material is essential for ensuring mining safety and optimizing support designs in complex geological conditions. This study investigates the effects of lithology and particle size on the compaction deformation and load-bearing behavior of crushed rocks from a typical coal mine goaf， providing a foundational basis for analyzing the cooperative bearing mechanism between coal pillars and crushed gangue. [Methods] A self-developed gravel compression test system was used， which consists of an electro-hydraulic servo pressure testing machine， an axial compression device， and resistance strain instruments. The axial compression device included a pressure piston， a cylindrical wall， and a base. Crushed rock samples were prepared from three lithologies—mudstone (immediate roof)， siltstone (direct roof)， and muddy siltstone (basic roof)—obtained from the No. 7 coal seam of a Panjiang mine. The samples were crushed using a jaw crusher and sieved into four particle size ranges: 0–5 mm， 5–10 mm， 10–15 mm， and 15–20 mm. A total of 12 sample groups were prepared， each with a mass of 100 0 g， representing single-size distributions for each lithology. The axial compression tests were conducted under a constant axial stress of 15 MPa to simulate the overburden pressure. During testing， strain gauges attached to the outer surface of the cylindrical vessel measured lateral strain， and lateral stress was calculated based on Lamé’s solution for thick-walled cylinders under internal pressure. Key parameters， such as the bulking factor， residual bulking factor， stress–strain relationships， and lateral pressure coefficient， were analyzed to evaluate the compaction and bearing characteristics. [Results] The physical and mechanical properties of the three lithologies were determined. Mudstone had a density of 1 823 kg/m3， a uniaxial compressive strength of 44.47 MPa， and a tensile strength of 0.85 MPa; siltstone had a density of 2 213 kg/m3， a compressive strength of 73.34 MPa， and a tensile strength of 3.47 MPa; muddy siltstone had a density of 2 175 kg/m3， a compressive strength of 52.90 MPa， and a tensile strength of 1.84 MPa. The bulking factor ranged from 1.19 to 1.63 and increased with lithological strength and particle size. The residual bulking factor， however， first increased and then decreased with particle size， ranging from 1.11 to 1.30. The maximum compressive strain of the gravel was directly proportional to particle size but showed no clear correlation with lithological strength. For example， mudstone samples with particle sizes of 0–5 mm， 5–10 mm， 10–15 mm， and 15–20 mm exhibited maximum strains of 0.361， 0.428， 0.484， and 0.561， respectively. Under the same conditions， siltstone and muddy siltstone showed lower and similar strain values， respectively. The compaction stress–strain curves were divided into three stages: rapid， stable， and full compaction. Lateral stress was proportional to axial stress， and the lateral pressure coefficient decreased with increasing lithological strength but showed no notable relationship with particle size. The lateral pressure coefficients for mudstone， siltstone， and muddy siltstone were 0.427 6， 0.334 3， and 0.344 8， respectively. [Conclusions] The compaction deformation and bearing mechanism of goaf gravel are significantly influenced by lithology and particle size. Stronger lithologies and larger particle sizes result in higher bulking factors， while the residual bulking factor initially increases and then decreases with particle size. The maximum compressive strain is more dependent on particle size than lithological strength. The lateral pressure coefficient is inversely related to lithological strength but independent of particle size. These findings provide important insights into the load-transfer mechanisms in goaf areas and contribute to the design of stable gob-side roadways. Future research should focus on the cooperative bearing behavior between coal pillars and crushed gangue to further enhance the understanding of overall ground control in mines.
Investigation of a multisensor-based breather valve fault diagnosis method
LIU Xiumei;MA Chaoxin;LI Beibei;HE Jie;ZHAO Qiao;HONG Conghua;To meet the requirements of new engineering while developing students' abilities and practical skills， this study develops a multisensor fusion-based fault diagnosis method for oil storage tank breather valves. The breather valve is a critical mechanical component for reducing the volatilization loss of oil and ensuring tank safety. Due to its long-term working time， some failure phenomena are common， such as leakage， rust， or seizure. Therefore， it is important to investigate the real-time monitoring and fault diagnosis method， reducing the probability of damage to the storage tank caused by breather valve failure and ensuring production safety. A monitoring system for the breather valve based on a single-chip microcomputer is established， the functions of which include real-time monitoring， data acquisition， data storage， data preprocessing， and threshold alarms. Four typical failure phenomena of the breather valve—nonfaulty， leakage， rust， and seizure—are assessed. The characteristic signal analysis methods for breather valve disc movement under failure conditions are also discussed. Based on the measured displacement signals， the vertical acceleration signals of the valve disc are further used for feature extraction in the time， frequency， and time–frequency domains. Five dimensional parameter indicators of the acceleration signal (maximum value， minimum value， variance value， peak-to-peak value， and root mean square value) in the time domain are discussed， and three dimensionless parameter indicators (kurtosis， impulse factor， and margin factor) are also investigated. The maximum values of the acceleration signals for nonfaulty， leaking， rusty， and seized valves are most obviously. The minimum values of the valve disc acceleration signals are smallest for the seized valves and largest for the rusty valve. The average values of the valve disc acceleration under the four typical failure phenomena are quite similar. The variance values of nonfaulty valves and seized valves are small. The peak-to-peak values of the rusty valves are highest， while those of the seized valves are smallest. The root mean square (RMS) values are highest for the rusty valves and lowest for the seized valves. The RMS of the leaky valves is slightly larger than that of the nonfaulty valves. Compared with the dimensional parameter indicators， the dimensionless parameters are less affected by the environment. All the dimensionless parameters of the rusty valves are large， while the dimensionless parameters of the leaky， rusty， and seized valves are slightly different. The fault feature extracted in the frequency domain is the frequency standard deviation. In the time–frequency domain， complementary empirical ensemble mode decomposition and wavelet packet transform are used to extract the fault signals. The wavelet packet transform has a better signal decomposition effect and faster decomposition efficiency. Therefore， the fault feature extracted in the time–frequency domain is the third-layer band energy value after wavelet packet decomposition. Finally， the multisensor-based breather valve fault diagnosis method is assessed. The feature-level and data-level fusion methods are used to combine the eigenvalues of the multisensor signals into eigenvectors. Multiple acceleration signals are processed through feature-level fusion， extracting parameter indicators and then consolidating them into a feature vector. Meanwhile， the displacement signals using the data-level fusion are used to calculate the maximum displacement values and the valve disc angles during breather valve movement， which are then incorporated into the feature vector. Taking this merged new feature vector as the input， a breather valve fault diagnosis model was built using the least-squares support vector machine optimized by the particle swarm algorithm， and the fault state of the breather valve was accurately identified. The classification of 60 breather valve samples with typical failure phenomena was conducted. The recognition accuracy of the fault diagnosis model is 96.66%. This engineering case could help students understand sensor technology applications in real projects， improve their grasp of sensor operation and signal processing， and strengthen their capabilities in practical projects.
Research on the architecture design and operational mechanism of an intelligent safety management system for university laboratories
LIU Daijun;LU Yi;[Objective] University laboratories， as critical hubs for scientific innovation， face escalating safety management challenges. Traditional paradigms， reliant on periodic manual inspections and static compliance checklists， are increasingly inadequate， suffering from inherent deficiencies: static risk perception， fragmented management elements， and a reactive response mode. The misalignment between safety management and the dynamic life cycle of experimental activities makes it difficult to adequately address the coupled and complex risks in modern research， transforming safety from an enabler of research into a heavy operational and administrative burden.Therefore， this study aims to transcend these deep-seated limitations by proposing a systematic， integrated framework to drive a fundamental paradigm shift in laboratory safety management from static， passive compliance to proactive， dynamic， and intelligent governance， thereby unifying safety assurance with scientific development.[Methods] This research constructs a comprehensive solution comprising a novel theoretical model， a supporting technological architecture， and a defined operational mechanism. First， the core innovation is the WSR-T dynamic safety management model. It integrates the Time/Process (T) dimension-operationalized into three sequential phases: Pre-experiment (T1: Prevention and Preparation)， During-experiment (T2: Monitoring and Execution)， and Post-experiment (T3: Restoration and Learning)-orthogonally with the Wuli (Physical)， Shili (Procedural)， and Renli (Human) dimensions of the WSR systems methodology. This integration forms a dynamic management matrix that reframes safety as continuous control over the entire experimental life-cycle trajectory. Second， to enable this model， a Multi-Agent collaborative Digital Twin-enabled Cyber-Physical System (MA-DT-CPS) architecture is designed. Its foundation is a high-fidelity digital twin， which integrates five core computational models: a Geometric model for spatial semantics， a Physical model for real-time monitoring and simulation， a Rule model that formalizes regulations and procedures， a Behavior model for quantifying human actions and states， and a Process model that creates intelligent digital threads for each experiment's life-cycle. Within this digital environment， a collaborative multi-agent system operates， featuring specialized agents for Situational Awareness， dynamic Risk Assessment， Compliance Execution， Emergency Decision-making， and system Learning & Optimization. This system is designed to operate in a three-tier hybrid-intelligence mode: full automation for routine tasks， suggestion mode for uncertain scenarios， and co-creation for novel situations. Finally， the study details the "Perception-Mapping-Analysis-Decision-Execution-Learning" event-driven closed-loop operational mechanism， specifying how the architecture implements the logic of the WSR-T model for dynamic， intelligent control. [Results] The study yields a holistic and actionable framework. Theoretically， the WSR-T model provides a novel， structured lens， making complex laboratory safety events analyzable as specific spatio-temporal couplings of W， S， and R elements， thereby moving beyond static checklist compliance. Technologically， the designed MA-DT-CPS architecture translates this theoretical model into a concrete， actionable implementation path. The key results include: 1) A systematic methodology that re-contextualizes safety as a process of dynamic control; 2) A sophisticated technological blueprint enabling high-fidelity digital representation， autonomous agent collaboration， and human-machine synergy; and 3) A well-defined operational mechanism that transforms management from a periodic， plan-driven activity to an event-driven， intelligent closed-loop process. This integrated "theoretical model-technical architecture-operational mechanism" framework systematically addresses the root causes of static， fragmented， and passive management， providing a clear pathway for the intelligent transformation of laboratory safety systems. The framework enables proactive risk inference， virtual strategy testing via the digital twin sandbox， and a system capable of self-learning and continuous improvement based on operational feedback. [Conclusions] This research presents a systematic， end-to-end solution designed to overcome the fundamental challenges in modern university laboratory safety management. By integrating the time dimension into the WSR methodology， the WSR-T model establishes a robust theoretical foundation for life-cycle-encompassing， dynamic safety governance. The MA-DT-CPS architecture provides the necessary technological enablers， fusing digital twin and multi-agent system concepts to create a platform for realizing the model's logic. Together， they form a coherent framework that paves the way for the development of intelligent management systems that seamlessly embed safety requirements into the entire scientific workflow. This paradigm aims to unify the goals of safety assurance and scientific innovation at a higher level. Future work should focus on developing prototypes and conducting long-term empirical studies in real laboratory environments to validate， optimize， and iteratively improve the proposed system， with particular attention to the quantitative modeling of complex human factors and ensuring cost-effectiveness for widespread adoption.
Application of the PDCA cycle in university laboratories safety management
KUANG Zhiqi;XIONG Lina;HE Jian;ZHENG Shiyong;[Objective] To address the inadequacies in management of safety hazards at Sun Yat-sen University， this study innovatively introduced and extensively applied the plan–do–check–act (PDCA) cycle theory. This approach was used to construct a precise， sustainable， and systematic closed-loop management system for laboratory safety hazards. The core value of this system lies in achieving full-cycle， dynamic governance of safety hazards and progressively enhancing the hazard management capabilities of university laboratories. [Methods] A four-dimensional dynamic management model was constructed to encompass strategy formulation， rectification implementation， inspection and supervision， and continuous improvement. This model established a hierarchically nested and collaboratively interconnected management network， ensuring comprehensive closed-loop management control of the four core elements of laboratory safety: personnel， materials， environment， and management. A target-oriented strategy was formulated in Stage P by identifying key hazards， conducting multidimensional root-cause analysis， establishing actionable rectification objectives， and systematically planning hazard identification and rectification strategies. This approach clarified the overall direction， defined key tasks， and outlined principles for resource allocation. In Stage D， precision in implementation was achieved through a graded early-warning mechanism that distinguished different risk levels and guided differentiated rectification measures. Thorough process tracing was mandated to identify root causes and assign explicit accountability， ensuring accurate corrective action and responsibility alignment. In Stage C， a dual-track mechanism integrating data-driven verification and supervision was established to monitor and validate the sustained effectiveness of hazard rectifications. In Stage A， effective practices that had been proven through implementation were systematically consolidated， and safety culture development was reinforced to elevate overall safety awareness and competence. Meanwhile， unresolved issues and newly emergent risks underwent in?depth root-cause analysis to support the optimization of rectification strategies， thereby driving the management system toward higher levels of refinement and adaptability. [Results] The implementation of a closed-loop management system for laboratory safety hazards based on the PDCA cycle theory yielded substantial outcomes: (1) The overall number of laboratory safety hazards markedly decreased， the occurrence rate of repeated hazards was substantially reduced， the upward trend of basic safety hazards was contained， and critical hazards were effectively controlled; (2) The capacity of individual laboratories to manage risk sources， identify and rectify safety hazards， and respond to emergencies was significantly enhanced， demonstrating a strengthened ability for autonomous safety prevention and control; (3) The closed-loop management system comprehensively encompassed the core elements of personnel， materials， environment， and management. It established 13 distinct and operable closed-loop pathways that ensured the substantive implementation， rigorous verification， and continuous tracking of corrective actions. [Conclusions] The implementation of the PDCA cycle management resulted in a substantial reduction in the overall incidence of safety hazards and the occurrence rate of repeated hazards. Additionally， it enhanced laboratories’ capacity for self-identification， prevention， and continual improvement. Consequently， the effectiveness of closed-loop safety hazard management was substantially improved. The PDCA cycle management system provides a replicable tiered management framework and a precise rectification pathway， offering a valuable reference for improving laboratory safety management systems in universities.