NetWork
Study and experimental validation of the magnetic and thermal performance of UHV transformers under DC bias
HAO Liang;DU Zhenbin;WANG Youhua;WANG Jianmin;ZHAO Zhigang;[Objective] As a key component of ultrahigh-voltage (UHV) transmission systems， the operational reliability of UHV transformers directly affects power grid security， stability， and power quality. With China’s rapid development of hybrid UHV AC/DC power grids and the increasing occurrence of geomagnetic disturbances， the issue of DC bias has become highly important. The resulting surge in core and structural component losses， along with increased local temperature rises， poses a serious threat to grid safety. Therefore， an in-depth investigation into the magneto–thermal behavior of UHV transformers under DC bias conditions is essential. [Methods] This study focuses on a 1000 kV UHV main transformer. A three-dimensional magneto–thermal coupling model is developed based on actual product parameters， including structural components such as the tank， belly plates， tie plates， support plates， footings， and magnetic shields， along with a segmented-frame core. First， the magnetic characteristic curves of the ferromagnetic core material under different DC bias currents are obtained through single-sheet measurements， then corrected and extrapolated for accuracy. Second， under high to medium operating conditions， different levels of DC excitation are applied to the high-voltage side to calculate excitation currents for various DC bias conditions， followed by a comparative analysis of magnetic flux density and loss characteristics of the core and structural components at maximum excitation. Third， the calculated losses under various DC bias levels are directly coupled to the thermal field—ignoring oil flow effects—to perform magneto–thermal simulations and assess temperature distribution and hotspot formation. Finally， to verify the accuracy and reliability of the proposed model and method， experiments are conducted on a 10 kV transformer model to examine excitation currents and loss behavior under different DC bias conditions. [Results] The findings indicate that (1) without DC bias， the excitation current exhibits a symmetric peaked waveform with mainly odd harmonic components; as the DC component increases， the waveform becomes distorted， the amplitude increases significantly， and harmonic components emerge. (2) Introducing a 4 A DC bias current causes a 49.1% increase in core loss compared to the unbiased state. At the same time， the longitudinal leakage flux in the main channel between the high- and medium-voltage windings increases by 27.4%， significantly intensifying eddy current losses in structural parts， especially near winding ends. (3) Magneto–thermal coupling simulations show that with a 4 A DC bias， the overall core temperature rises notably， with the middle core limb experiencing the most significant increase， reaching 66.28 K. Distinct local hotspots appear in structural components， particularly near winding ends. (4) Experimental validation confirms consistent trends in excitation current and a gradual rise in no-load loss， with the growth rate decreasing over time. Although some numerical deviations occur， they remain within acceptable limits， validating the accuracy of the proposed approach. [Conclusions] Simulation and experimental results demonstrate that DC bias causes significant increases in losses and temperature， worsens local overheating， and accelerates insulation aging risks. Therefore， the negative effects of DC bias must be considered in the design， evaluation， and long-term operation of UHV transformers.
Teaching experimental design of magnetic anomaly detection based on a rotor UAV platform
WANG Chao;CHENG Linhan;WANG Yanzhang;WAN Yunxia;[Objective] In the talent training system for geophysical exploration and instrumentation majors， the integration of professional theory and engineering practice is a core requirement. The magnetic anomaly detection experiment serves as an important link between theory and practice， laying a solid foundation for subsequent professional learning and engineering applications. This paper presents the design of a full-process teaching experiment scheme for magnetic anomaly detection， covering “instrument system integration— instrument performance testing—target information acquisition—data preprocessing—result demonstration.” [Methods] Using a uniformly magnetized sphere as the research object， this study simulates the spatial distribution of magnetic anomalies. An experimental teaching system capable of synchronously acquiring magnetic field and spatial position data was designed and integrated， comprising an optically pumped magnetometer， a positioning module， a multichannel data acquisition instrument， and a UAV. Field experiments， including sensor performance testing， sensor placement testing， dynamic flight consistency testing， and regional magnetic anomaly surveys， were conducted to collect magnetic anomaly data. [Results] Due to differences between the field experimental environment and ideal working conditions， and because manually placed small vehicles cannot fully replicate standard uniformly magnetized spheres， certain numerical deviations exist between the measured data and the simulation results. However， both exhibit overall variation trends in the spatial distribution of the magnetic field that are highly consistent， thereby verifying the reliability and trend accuracy of the detected data. [Conclusions] The teaching demonstration and practical training mode designed in this paper， which combines “modular” instrument system integration， “process-based” instrument performance testing， and “scenario-based” field magnetic anomaly detection， effectively improves the cognitive level of students majoring in instrumentation on the design concept of advanced specialized instrument systems. It also deepens the understanding of students majoring in geophysical exploration-related disciplines regarding the fundamental principles of magnetic anomaly detection， and comprehensively tempers students’ comprehensive application capabilities in real-world engineering scenarios.
Construction of a virtual simulation experiment for meteorological observation lidar
WANG Meng;WANG Yufeng;YAN Qing;WANG Li;LIU Jingjing;[Objective] In response to the strategic demands of national ecological civilization， meteorological observation lidar， as a key technology for atmospheric environmental monitoring， has become a core support for promoting the high-quality development of the meteorological industry. To address the challenges of high cost， high risk， and difficulty in reproducing experimental phenomena in traditional practical teaching， this study aims to develop a virtual simulation experimental system for meteorological observation lidar， thereby compensating for the shortcomings of existing virtual simulation systems in meteorological remote sensing. The system is developed based on a national first-class course and the Lidar Remote Sensing Research Center at the Xi’an University of Technology. It is designed to enhance students’ comprehensive design and innovation capabilities and to provide core teaching support for cultivating high-level interdisciplinary meteorological talents in the new era. [Methods] This study adopts an experimental approach based on “modular design and progressive training” and constructs a three-tiered experimental framework comprising “cognition， design， and exploration，” corresponding to three core modules: lidar system cognition， multispectral spectroscopic system design， and detection and data processing. (1) At the cognitive level， students master the structure and principles of lidar systems via immersive navigation and interactive model demonstrations. (2) At the design level， a task-driven mode is adopted to enable students to independently select detection targets and complete the design of a multispectral spectroscopic system. (3) At the exploration level， students assemble a functional lidar system and undergo a comprehensive training in data acquisition， inversion， and systematic error analysis. Throughout the experiment， scientific methods—including observation， modeling， comparison， and induction—are seamlessly integrated， with theoretical knowledge embedded into interactive tasks， thereby effectively enhancing the students’ capabilities from basic cognition to comprehensive innovation. [Results] This study has achieved substantial teaching effectiveness through an innovative design that integrates science and education， virtual reality， and learning with assessment. The system employs high-fidelity modeling to authentically reproduce lidar structures and detection processes， effectively addressing the challenges of high cost， safety risks， and limited repeatability in traditional experimental teaching. Within the simulated environment， students engage in a complete workflow， including system cognition， optical design， data acquisition， and inversion analysis. Their operational behaviors are recorded in real time and automatically evaluated， generating comprehensive multidimensional assessment reports. The experiment thus achieves a deep integration of theoretical knowledge acquisition， practical skill training， and process-oriented evaluation. [Conclusion] The implementation of this study has not only advanced the systematization and practical application of virtual simulation experiments for meteorological observation lidar， but has also explored an innovative talent development pathway characterized by “virtual augmentation of real practice and the integration of science and education.” This experimental framework significantly enhances students’ capabilities in system design， data processing， and scientific inquiry in complex meteorological detection scenarios. It provides robust support for cultivating high-quality talents in emerging engineering majors such as instrumentation， optoelectronics， and meteorology， thereby offering an important teaching and practical platform for the independent development of meteorological detection equipment and technologies in China.
Experimental platform for enhancing resilience of distribution network under flood disasters based on power-transportation coupling
LIAN Xianglong;CHEN Jie;FU Weifeng;LIU Lijun;[Objective] With the increasing frequency and intensity of flood disasters， distribution networks are exposed to complex cascading failures caused by the strong coupling between power and transportation systems. However， conventional experimental pedagogy in electrical engineering focuses primarily on deterministic operation and single-system analysis， which limits students’ understanding of disaster-induced risks， recovery processes， and resilience-oriented decision-making. This study aims to design an experimental teaching platform for enhancing the resilience of distribution networks in flood-disaster scenarios by explicitly considering power–transportation coordination. The goal is to support interdisciplinary learning and improve students’ ability to analyze and manage complex coupled systems under extreme conditions. [Methods] An integrated experimental platform is developed by coupling distribution and transportation networks with emergency resource systems within a unified modeling framework. The evolution of flood disasters is first described using a grid-based hydrological model， which captures the spatial and temporal accumulation of surface water. The simulated flooding depth is then mapped to the probability of failure of distribution nodes and degradation of road traffic， enabling the construction of a coupled power–transportation failure model. Emergency resources， including repair crews and mobile energy storage systems (MESS)， are incorporated to represent both structural repair and temporary power supply capabilities. To reflect uncertainty in the impact of disasters， Monte Carlo simulation is employed to generate multiple failure scenarios. By implementing scenario clustering， the computational burden is reduced while preserving representative and high-risk characteristics. Based on these scenarios， a two-stage emergency scheduling framework is established， in which pre-disaster resource allocation and post-disaster dynamic dispatch decisions are jointly optimized. A risk-aware strategy based on conditional value-at-risk (CVaR) is introduced to emphasize low-probability but high-impact scenarios. The platform enables comparative experimental analysis by varying key dimensions， including the consideration of road flooding effects and the number of MESS units and repair crews deployed. [Results] The simulation results under different flood disaster scenarios demonstrate that the proposed platform effectively captures the influence of cross-system coupling on the resilience of the distribution network. When road flooding and traffic degradation are considered， the arrival of emergency resources is delayed， and the power restoration process is significantly slowed， leading to larger resilience loss. Increasing the number of MESS units improves early-stage power supply by providing temporary support to critical loads， which helps mitigate initial service interruptions. In contrast， increasing the number of repair crews mainly accelerates mid- and late-stage structural recovery， enabling the system to reach full restoration earlier. The best overall performance is achieved through the coordinated deployment of repair crews and MESS， combining early power support with faster recovery. Moreover， the CVaR-based scheduling strategy provides enhanced robustness by prioritizing high-impact disaster scenarios， resulting in more stable recovery trajectories across different scenarios. These results clearly illustrate the complementary roles of transportation conditions， emergency resources， and risk-aware decision-making in resilience enhancement. [Conclusions] The proposed experimental teaching platform integrates flood-disaster modeling， coupled power–transportation failure analysis， and risk-aware emergency scheduling into a coherent framework. It transforms abstract concepts of system resilience into observable experimental phenomena， enabling students to intuitively understand how disaster evolution， transportation accessibility， and resource coordination jointly affect the recovery of distribution networks. The platform effectively supports comparative experiments and decision analysis in uncertainty， fostering interdisciplinary thinking and resilience-oriented engineering skills. This study provides a practical reference for advancing experimental teaching reform in electrical engineering and cultivating students’ ability to analyze and manage complex coupled energy systems in the event of extreme events.
Simulation and optimization practice of electromagnetic force stroke characteristics of the pilot solenoid valve in an electronic controlled air suspension gas distribution valve
LIU Yuechao;ZHAO Leilei;YU Yuewei;DING Fan;JIN Mingqing;SHAN Xiyu;[Objective] The pilot solenoid valve is a critical component of the electronic controlled air suspension (ECAS) air distribution valve， governing the inflation and deflation processes and， consequently， the inflation and deflation rates. Its electromagnetic force–stroke profile directly determines the accuracy and stability of spool-position control， which in turn affects the rapid and precise regulation of ECAS ride height and air-spring stiffness. Owing to the limited load capacity of commercial vehicle electrical systems， the available drive current is constrained， resulting in insufficient electromagnetic force at the beginning of the armature stroke and excessive force near the final pull-in. During the initial motion， an inadequate electromagnetic driving force cannot effectively overcome resistance， thereby deteriorating the armature start-up behavior， manifested as a delayed response or even start-up failure. During final pull-in， the electromagnetic force increases rapidly as the air gap decreases， causing the armature to accumulate excessive kinetic energy and collide with the stationary core at high speed. This induces stop-pin impact and energy loss， thereby reducing the operational stability and reliability of the solenoid valve. Therefore， without increasing the drive current， this study aims to improve the electromagnetic force–stroke distribution and to conduct structural optimization of the pilot solenoid valve. [Methods] Based on equivalent magnetic-circuit theory， the magnetic-circuit characteristics of the pilot solenoid valve were systematically analyzed. A finite-element electromagnetic model was established， considering magnetic nonlinearity and variations in the working air gap， and its accuracy was validated through comparison with static electromagnetic force experiments. On this basis， evaluation indices for characterizing the electromagnetic force–stroke distribution were defined. Pearson correlation coefficients were employed to perform sensitivity analyses of key structural parameters， and parameters with high sensitivity were selected as optimization variables to clarify the relationships between structural parameters and electromagnetic force distribution. Considering the coupling effects among multiple structural parameters， a response-surface prediction model for the average electromagnetic force over typical working air-gap intervals was developed using a Box–Behnken design. To maximize the average electromagnetic force in the initial and mid-stroke air-gap intervals and minimize it in the final pull-in interval， the NSGA-II multiobjective optimization algorithm was adopted to search for Pareto-optimal combinations of design variables. Subsequently， electromagnetic force simulations were conducted for the optimal parameter combination， and the simulation results were compared with the corresponding response-surface predictions to verify the effectiveness of the proposed optimization method. [Results] The optimization results indicate that， under an unchanged drive current， the average electromagnetic force in the initial working air-gap segment increased by 31.83% compared with the pre-optimization design， while it increased by 25.72% in the mid-stroke working air-gap segment. In contrast， the average electromagnetic force in the final pull-in air-gap segment was reduced by 20.98%. The improved electromagnetic force–stroke distribution effectively alleviates the imbalance between insufficient driving force at the initial stage and excessive pull-in force at the final stage. Consequently， the start-up capability of the pilot solenoid valve under current-limited conditions is significantly enhanced， the responsiveness and stability of armature motion in the intermediate stroke are improved， and the end-stage impact and associated energy loss are effectively reduced. [Conclusions] This study addressed the suboptimal electromagnetic force–stroke distribution of a small-orifice pilot solenoid valve in an ECAS air distribution valve under current-limited conditions through modeling， experimental validation， simulation， and optimization. The proposed method significantly improved the force distribution across the full stroke， providing a valuable reference for the design， simulation， testing， and application of pilot solenoid valves. Additionally， it serves as a typical case for solenoid valve simulation experiment teaching， enhancing students’ engineering analysis skills and innovative thinking.
Observation of confined CO2 jets in an annular structure and plume identification method
BAI Xiang;JIANG Zhiming;LI Jie;FAN Jianchun;CHEN Sen;LI Min;ZHANG Liwei;[Objective] In carbon capture， utilization， and storage operations， the continuous injection of high-pressure CO2 into subsurface reservoirs exposes wellbore tubing to high-pressure， low-temperature， and corrosive CO2-rich environments. This increases the risk of local damage or perforation. If a tubing leak occurs， high-pressure CO2 is instantly released into the annular space between the tubing and casing， forming a highly transient confined jet. Unlike free jets， the confined jet in an annulus is strongly affected by geometric constraints， especially during the early leakage stage before wall impingement. This early-stage flow is characterized by rapid evolution， strong unsteadiness， and complex mixing， which are challenging to measure directly in field conditions. Therefore， it is necessary to investigate the early development characteristics of confined CO2 jets induced by tubing leakage under controlled experimental conditions. [Methods] A laboratory-scale experimental system utilizing a Z-type schlieren optical configuration was established to visualize the early-stage confined jet formed by CO2 leakage in a coaxial annular geometry. The actual wellbore structure was simplified into a concentric tubing-casing model， focusing on jet evolution prior to interaction with the casing wall. High-purity CO2 served as the working fluid. Experiments were conducted under representative operating conditions: a supply pressure of 4 MPa， a leakage orifice diameter of 1 mm， and an ambient temperature of approximately 18 ℃. High-speed schlieren image sequences were acquired at 24 390 frames/s to capture jet initiation and early development. Based on these schlieren images， a plume identification and jet-front displacement extraction method suitable for high-frame-rate image sequences was developed. This method combines background subtraction and adaptive threshold segmentation to extract candidate plume regions. A nozzle connectivity constraint ensures physical consistency between the identified plume and the leakage origin， whereas a temporal consistency veto mechanism suppresses abnormal segmentation results caused by wall-related schlieren interference and transient noise. A calibrated pixel-to-length conversion factor is used to convert the axial displacement of the jet front into physical distance， and the cumulative average propagation velocity is calculated. [Results] The experimental results show that the confined CO2 jet exhibits pronounced， rapid axial development immediately after leakage initiation. During the early stage， the jet front advances quickly along the annular axis， driven primarily by high initial momentum at the leakage orifice. As the jet develops， geometric confinement becomes increasingly significant， resulting in a gradual reduction in axial propagation rate and the emergence of lateral spreading and recirculation. Quantitative analysis indicates that the cumulative average jet-front velocity increases rapidly during the early stage and then approaches a quasi-stable level of approximately 60 m/s before wall impingement occurs. The calculated velocity’s evolution trend is consistent with the schlieren-observed plume development， confirming the reliability of the proposed identification and measurement method. [Conclusions] A schlieren-based experimental approach， combined with a robust plume identification and jet-front tracking method， is developed to investigate early-stage confined CO2 jets induced by tubing leakage. Without relying on complex physical modeling assumptions， this method effectively suppresses wall-induced schlieren artifacts and transient disturbances， enabling the stable extraction of plume morphology and jet-front displacement from high-speed image sequences. The experimental results provide valuable insights into the transient evolution characteristics of confined CO2 jets in annular geometries， thereby offering a reliable experimental and analytical basis for further studies on leakage behavior and flow characterization in CO2 injection wells.
Failure mechanism and mesoscopic response of limestone with filled joints under triaxial stress
QIN Zelong;ZHANG Chao;CUI Yinxiang;GUO Yongcheng;LI Qiyang;HU Jin;[Objective] This study addresses the common structural defect of filled joints in limestone and investigates how the inward extension length of such joints affects the mechanical properties of the rock. Specifically， the intrinsic relationship among filled joint length， stress concentration， and crack propagation paths is analyzed to reveal the damage evolution process at the joint tips. The patterns of energy accumulation， release， and dissipation during loading are also examined to clarify how filled-joint length influences the energy storage limit of the rock mass and the risk of sudden failure. [Methods] A combined methodology of laboratory conventional triaxial compression tests and PFC3D particle-flow numerical simulations is employed to systematically investigate the influence of filled-joint length and inclination angle on the mechanical properties of limestone. Limestone specimens containing filled joints of varying lengths and inclination angles are prepared， and a series of triaxial compression tests are conducted to obtain their macroscopic mechanical responses， failure modes， and crack evolution patterns. Based on the experimental results， mesoscopic parameters are calibrated to construct the corresponding PFC3D numerical models， and simulations under identical conditions are performed to reveal， from a mesoscopic perspective， the mechanisms of force chain evolution， crack initiation and propagation， and energy transformation. Mutual verification and complementary analysis of the physical test and numerical simulation results elucidate the controlling effect of defect geometric characteristics on the strength， deformation， and failure of limestone. [Results] Laboratory experiments and numerical simulations indicate that the failure mode of limestone is jointly governed by joint length and inclination angle. With increasing joint length， the peak strength and elastic modulus of the specimens generally decrease， indicating a substantial degradation in the load-bearing capacity and deformation resistance of the rock mass as the defect scale expands. During deformation， pronounced displacement concentration zones tend to form around the joints， accompanied by nonuniform stress distribution. Microcracks preferentially initiate at the joint tips and propagate along the joint orientation， causing rock deformation to shift gradually from globally coordinated behavior to localized concentration. The resulting damage evolution paths are strongly influenced by the spatial morphology of the joints. As joint extension increases， the influence range of the DFN III displacement field gradually diminishes， indicating a continuous reduction in the integrity and continuity of the rock mass structure， while the controlling effect of structural planes on rock deformation and failure becomes considerably more pronounced. Concurrently， energy dissipation in the initial stage is enhanced， whereas the capacity for elastic energy storage is reduced， and the balance of energy accumulation， release， and dissipation within the rock mass system is progressively disrupted. [Conclusions] A comprehensive analysis of the influence of filled joints on limestone is conducted from both macroscopic and mesoscopic perspectives， revealing the coupled controlling effects of joint geometric parameters on the mechanical properties of limestone. On this basis， a quantitative response expression describing the deterioration of limestone strength and deformation modulus with variations in joint geometric parameters is established， providing a reliable quantitative basis for evaluating the mechanical behavior of limestone with filled joints. Furthermore， from the mesoscopic perspectives of acoustic emission and crack evolution， the failure mechanism and damage evolution process of limestone with filled joints are systematically analyzed， clarifying the internal correlation between microcrack initiation and propagation and the macroscopic mechanical response. Overall， the findings deepen the understanding of the interaction mechanism between filled joints and limestone and provide theoretical support and practical reference for engineering stability analysis.
Fire risk assessment and management practices for university laboratories involving fire based on the FAHP-FCE model
LIU Caiwei;LIANG Wenhao;LIU Yanchun;WANG Junfu;MIAO Jijun;[Objective] University laboratories involving fire are indispensable infrastructures for engineering education and scientific research. These laboratories are typically characterized by high-temperature operations， open-flame processes， combustible materials， high-power electrical equipment， and complex experimental procedures， resulting in significantly higher fire risks compared to conventional laboratory settings. Existing fire risk assessment methods for university laboratories often rely heavily on historical accident data， fail to adequately consider emerging and compound risk scenarios， and have limited adaptability to diverse laboratory functions and operational modes. To address these challenges， this study develops a systematic， generalizable， and practical fire risk assessment approach specifically designed for university laboratories facing fire hazards. By integrating the fuzzy analytic hierarchy process (FAHP) with the fuzzy comprehensive evaluation (FCE) method， key fire risk factors are systematically identified， and their relative importance is quantitatively determined， thereby providing a scientific basis for fire risk classification， graded control， and safety management decision-making in university laboratories. [Methods] This study establishes a comprehensive fire risk assessment framework for university laboratories by synthesizing relevant national policies， safety regulations， technical standards， representative accident analyses， and expert insights to construct a fire risk evaluation indicator system. The proposed system consists of four primary indicators， namely direct disaster-causing factors， vulnerability of the exposed body， prevention and management capability， and special scenario risks， and is further detailed into 26 secondary indicators. These indicators encompass essential risk elements， including ignition source control， equipment condition， experimental environment characteristics， human factors， management mechanisms， and scenario-specific hazards. The FAHP is employed to determine the weights of each indicator. Expert judgment data are collected through structured questionnaires administered to specialists with extensive experience in laboratory safety management， and fuzzy complementary judgment matrices are constructed to address uncertainty and subjectivity in expert assessments. The FCE method is then applied to quantitatively evaluate laboratory fire risks. Based on expert scoring， membership functions and fuzzy relation matrices for each indicator are established. The study then conducts a weighted fuzzy synthesis operation to obtain comprehensive membership vectors and performs defuzzification using the weighted average method to determine the final fire risk level and overall risk score. The Disaster Prevention and Mitigation Laboratory of Qingdao University of Technology is selected as a case study to validate the feasibility and applicability of the proposed model. [Results] The weighting results indicate that direct disaster-causing factors and prevention and management capabilities are critical in determining overall fire risk， highlighting the importance of ignition source control， electrical safety， equipment condition， and management enforcement on laboratory fire safety. The evaluation results reveal that the case laboratory achieves an overall score of 91.38， corresponding to a fire risk level classified as “no risk.” This assessment aligns closely with the laboratory’s actual safety performance， reflecting a high level of safety in fire compartmentation design， firefighting facility configuration， personnel access control， and implementation of safety responsibility systems. However， some indicators， such as fire risk awareness among faculty and students and the closed-loop management of hazard identification and rectification， received relatively lower scores， suggesting a need for improvement in personnel awareness and enhanced safety management practices. [Conclusions] The FAHP-FCE model effectively differentiates among various fire risk levels and accurately identifies critical risk factors. It demonstrates strong scientific validity， practical operability， and the potential for wide application， providing robust technical support for graded fire risk control and safety management in university laboratories. The findings of this study not only contribute to improving fire risk prevention and control in university laboratories but also serve as valuable references for advancing safety governance in research facilities at higher education institutions.
Experiment assistance system based on large language models and voice interaction
WANG Xiaonian;XU Zhiyu;SHA Ruizhi;[Objective] This paper presents a novel intelligent laboratory assistant system designed to revolutionize experimental teaching through a synergistic combination of Large Language Models (LLMs)， voice interaction， and state machine control. The proposed system addresses the perennial challenges in hands-on education， including the high demand on instructor time for repetitive guidance， the difficulty in providing immediate and personalized student feedback， and the critical need for real-time safety monitoring. By developing a voice-enabled companion， we aim to transition from static， paper-based manuals to a dynamic， interactive， and intelligent tutoring experience that guides students through experiments step-by-step， answers contextual questions， and proactively mitigates risks. [Methods] The core innovation of this system resides in its tri-domain adapter architecture， a hierarchical design that harmonizes the broad reasoning capabilities of Large Language Models (LLMs) with the stringent precision required for experimental instruction. First， the General Knowledge Adapter? establishes the foundational layer， utilizing a locally deployed and fine-tuned DeepSeek model. This adapter is systematically infused with comprehensive electrical engineering expertise， encompassing circuit analysis， analog and digital electronics， and power systems. This injection of domain-specific knowledge ensures the accurate interpretation of complex terminology—such as distinguishing between "series-aiding" and "series-opposing" configurations in mutual inductance experiments—providing a reliable theoretical bedrock. Second， the Experimental Context Adapter? dynamically bridges theory and practice. By interfacing with the system’s state machine， it loads standardized procedural workflows corresponding to the student's selected experiment. This resolves critical contextual ambiguities where the same component (e.g.， "resistor R1") requires different measurement protocols depending on the experimental setup. It facilitates a "co-piloted" guidance process， replacing static manuals with interactive， step-by-step voice instructions and performing real-time validation of student measurements to prevent error propagation. Third， the Real-Time Safety Filter? functions as a dedicated risk mitigation module. It continuously monitors the audio stream for lexical indicators of hazards—such as "unusual smell，" "overheating，" or "sparks"—triggering immediate voice alerts and simultaneous instructor notifications to preempt accidents. Orchestrating these domains is an Intelligent Information Router. By analyzing the student's query alongside the current state machine status， the router dynamically arbitrates between the general knowledge base and the specific experimental database. This ensures that every response is both contextually aware and grounded in verified instructional materials， effectively aligning generic AI capabilities with specific educational objectives. [Results] The proposed system has been implemented and evaluated in ”Fundamentals of Electronic Technology” laboratory. Experimental results demonstrate that the system performed stably across 120 controlled tests: student operation accuracy increased to 97.7% (baseline: 83.3%)， equipment damage rate dropped to 0.8% (baseline: 4.3%)， and the experimental report quality score improved by 29.5 points (on a 100-point scale). [Conclusions] Even with limited initial datasets that were subsequently expanded and enhanced through data augmentation techniques， demonstrate a marked improvement in student experiment completion efficiency， a reduction in repetitive queries to instructors， and a notable decrease in simulated safety incidents. This work provides a reusable technological paradigm for the future of intelligent， personalized， and safer laboratory education across various scientific and engineering disciplines.
Experiences and implications of the operation and management of nuclear laboratories in the United States
ZHOU Chen;WANG Luo;ZHANG Jingrui;SU Yongning;LI Bin;[Objective] Nuclear science and technology is a strategic field that underpins national security， drives energy transition， and enables cutting-edge technological breakthroughs. Accelerating the development of world-class nuclear research institutions with clear positioning， flexible governance mechanisms， and efficient collaboration structures is crucial for strengthening national security， advancing energy technology transformation， and securing strategic technological leadership. Currently， China’s major nuclear research platforms face several practical challenges， including rigid institutional mechanisms， underutilized innovation system efficiency， and insufficient collaborative innovation. This study provides theoretical support and practical guidance for China in establishing a modern nuclear research system tailored to national conditions and in achieving high-level self-reliance in nuclear science and technology. [Methods] This paper employs a comprehensive methodological approach integrating literature review， comparative analysis， case studies， and inductive reasoning. It systematically traces the developmental trajectory of the nuclear-related national laboratory system in the United States. By examining three core dimensions—governance architecture， contract management， and operational mechanisms—the study distills key practices in operational management. Drawing on the actual conditions of China’s nuclear innovation platforms， targeted insights and recommendations are proposed across four dimensions: constructing a modern governance framework， optimizing resource allocation models， cultivating talent development ecosystems， and upgrading mechanisms for transforming scientific and technological achievements. [Results] The findings indicate that the development of nuclear laboratories in the United States has consistently aligned with national strategic needs， evolving from emergency-driven establishment to institutionalized development， and from single-function entities to comprehensive innovation enablers. This evolution can be divided into three phases: embryonic inception (World War II era， 1943–1946)， institutionalization (Atomic Energy Commission era， 1947–1974)， and integration and expansion (Department of Energy era， 1977–present). At present， the United States has formed an operational management system tailored to the nuclear sector’s highly sensitive， heavily regulated， long-cycle， and security-critical characteristics. This system includes a government-owned， contractor-operated governance model and a full-cycle contract management framework adapted for nuclear regulation; a “dual-track” resource allocation model supporting nuclear capability development; and an open collaborative innovation mechanism that balances nuclear safety with operational efficiency. These systematic practices provide a valuable reference for similar institutions worldwide. China’s nuclear laboratory development has entered a critical stage， requiring accelerated progress in institutionalization and the establishment of legal entities to build a modern governance structure with clearly defined responsibilities and accountability. Innovative investment and funding management models are needed to establish a mission-driven resource allocation system. Talent recruitment， cultivation， and incentive mechanisms should be further optimized to foster a globally competitive nuclear research ecosystem. In addition， an open and collaborative innovation environment should be developed， supported by a refined innovation system that effectively integrates research outputs with practical applications. [Conclusions] The success of nuclear laboratories in the United States lies in sustained institutional innovation， which has produced a hybrid organizational model that effectively fulfills national missions， safeguards academic freedom and scientific exploration， and ensures efficient operational capabilities. This model addresses three fundamental tensions: balancing national oversight with professional autonomy， reconciling long-term strategic objectives with short-term performance demands， and harmonizing security and confidentiality requirements with innovation and openness. In response to emerging challenges and strategic demands， China must develop a dynamic and modern nuclear science and technology innovation system that promotes frontier breakthroughs. Such a system should be firmly grounded in China’s national conditions and the unique characteristics of the nuclear sector， while selectively and critically drawing on relevant United States experiences and practices.