NetWork
Development and Application of a Crushing Specific Work Test Device for Phase Change Aggregate Concrete
Liu Chengbin;Chen Ju;Zhang Ruiqiang;Yang Ziming;[Objective] To address the limitation that existing chiseling-specific test equipment can only operate at room temperature and thus cannot support quantitative studies of phase-change-aggregate concrete breakability under complex thermal conditions， this study presents a newly developed test apparatus specifically designed to assess the specific crushing work of phase-change aggregate concrete. [Methods] The apparatus is composed of three critical elements: the specimen area， the crushing device area， and the support structure. To achieve accurate temperature control of the specimen， a cast-aluminum electric heating plate is utilized in conjunction with a highly efficient thermal insulation system. This setup ensures precise maintenance of the desired temperature throughout the testing process. The apparatus features a unique synchronous belt lifting-drop hammer impact system that enables uniform crushing operations in multiple directions. This innovative system works in tandem with an adaptable rotatable central shaft structure， which further enhances the device's ability to conduct crushing actions effectively. The synchronous belt lifting-drop hammer impact system guarantees consistent and controlled crushing actions， while the rotatable central shaft structure allows for flexible positioning of the crushing device to optimize its performance. [Results] To verify the comprehensive performance of the device， a series of performance verification tests were conducted. Phase-change aggregate concrete specimens with different mix proportions were selected for the specific energy of fragmentation test under different temperature conditions， including room temperature， 40℃， 80℃ ，120℃， 160℃ and 200℃. Three groups of parallel specimens were set for each temperature gradient to ensure the representativeness of the test data. The test results show that the testing device operates stably， is convenient to operate， and can accurately measure the specific energy of fragmentation data of phase-change aggregate concrete at different temperatures. The test error is controlled within 5%， indicating good reliability and repeatability. As the test temperature increases， the specific energy of fragmentation of the phase-change aggregate concrete shows a significant downward trend， and the crushability of the material gradually improves. When the temperature is below 120℃， the decrease of the specific energy of fragmentation is relatively gentle， with a reduction range of 15% - 28%. When the temperature is above 120℃， the decline rate of the specific energy of fragmentation accelerates significantly， and the crushability of the material is significantly enhanced. At 200℃， the specific energy of fragmentation decreases by 57% compared to that at room temperature， indicating that the mechanical properties of the phase-change aggregate concrete are significantly weakened in high-temperature environments， showing excellent dismantlability. This pattern is consistent with the phase-change characteristics of phase-change aggregates at high temperatures and the change pattern of the internal structure of the concrete. [Conclusions]The testing apparatus can effectively simulate a variety of complex thermal conditions and accurately perform quantitative measurements and data analyses of the specific crushing energy of phase-change aggregate concrete specimens under different temperature environments， substantially enhancing the stability and accuracy of the experimental results. It not only addresses the technical gap in quantifying the fracture characteristics of phase-change aggregate concrete under coupled thermal effects—thereby establishing a critical technical foundation for systematic investigations， mechanism analyses， and performance evaluations of the material’s crushability—but also provides a feasible approach and an important engineering reference for optimizing mechanical tests， developing dedicated testing equipment， and standardizing test methods for various specialty concretes such as low-temperature and phase-change-modified materials.
Research and practice on the operation and management of the deep Earth laboratory
LI Yasong;KANG Kai;LI Zaiqiang;[Objective] Deep earth laboratories use thick rock formations to shield cosmic rays， providing an experimental environment with an extremely low radiation background for frontier fields such as particle physics， nuclear physics， and life sciences. They are excellent venues for conducting cutting-edge basic scientific research. These laboratories exhibit unique characteristics in operational management， such as extreme environmental conditions， interdisciplinary nature， and complex risks. This paper takes the China Jinping Underground Laboratory as the research object. As a major national science and technology infrastructure， it boasts the deepest rock cover， the largest underground space， and the lowest cosmic ray flux in the world. Located 2400 meters underground， it has a total construction area of 40，000 square meters and a space volume of 300，000 cubic meters. It has attracted research teams from more than ten top scientific research institutions in China， including Tsinghua University， Shanghai Jiao Tong University， Beijing Normal University， and Sichuan University. [Methods] To address the core issues in laboratory operation and management， this research focuses on analyzing the key operational management challenges based on the deep-earth characteristics of the Jinping Laboratory， such as pre-management and settlement guarantee for project access， long-term maintenance of the extremely low-radiation background experimental environment， safety risk management and control during laboratory operation， and the construction and management of the operation support team. This paper proposes a series of efficient and advanced management measures in terms of project access， background environment maintenance， safety risk management and control， and support team construction. These measures include: ① a scientific approval and resource scheduling system with hierarchical coordination， covering project approval， access， exit and equipment installation; ② low-background safeguard measures including radon suppression， dedicated ventilation and radiation monitoring; ③ a safety system tailored for deep underground scenarios， supporting experimental operation， earthquake monitoring and fire prevention; ④ a professional operation management and support team for integrated management and implementation guarantee. [Results] Endowed with the unique advantage of 2400-meter vertical rock coverage， the Jinping Laboratory faces the challenge of reconstructing the management paradigm in the extreme deep-earth environment while creating an "ultra-low background" scientific research environment. By systematically analyzing and addressing prominent issues in core dimensions such as scientific collaboration， background maintenance， safety control， and team building， this paper develops an operational management solution suitable for deep underground scenarios. This solution ensures the safe， efficient， and reliable operation of the laboratory， supports the achievement of more high-level scientific experimental results， and improves the efficiency of laboratory operation and management. [Conclusions] This research breaks through the management thinking of traditional ground-based scientific research facilities. Guided by scientific research needs， based on safety and stability， and supported by technological innovation， it realizes the in-depth integration of scientific management and engineering operation and maintenance. The operational management practice of the Jinping Laboratory will not only lay a solid foundation for its own cutting-edge scientific research but also contribute to the professional and refined operation and management of major science and technology infrastructure， and provide a practical paradigm with Chinese characteristics for the operation and management of deep earth laboratories worldwide.
Development and experimental research of a vacuum infiltration sintering apparatus
WANG De;YANG Puyuan;ZHEN Zhen;XIAO Xiaofeng;LIU Huan;WANG Wenqin;[Objective] Wear-resistant seal coating on the tip of single-crystal turbine blades plays a crucial role in ensuring the airtightness and operational efficiency of aeroengines. With the continuous increase in turbine inlet temperature， the harsh service environment imposes increasingly stringent requirements on the coating， such as superior high-temperature wear resistance， oxidation resistance and adhesion. However， existing preparation technologies face prominent challenges: thermal spraying and laser cladding produce coatings with flat surfaces where abrasive particles are uniformly distributed inside， failing to meet the protruding morphological requirement; electrodeposited yields coatings that suffer from insufficient adhesion and increased brittleness with increasing thickness; brazing， while improving interface bonding， damages the base metal because of the diffusion of melting-point-lowering elements. Additionally， Al?O? ceramic particles， as ideal reinforcing phases， exhibit poor wettability with metal melts， and their particle size and content significantly affect coating quality， yet relevant systematic research is scarce. To address these issues， this study aims to develop a high-performance preparation technology for NiCoCrAlYTa–Al?O? blade-tip wear-resistant coatings， synergistically integrating the high-temperature wear resistance of particles and protective performance of the coating while avoiding damage to the single-crystal base metal. [Methods] A visualized high-vacuum infiltration sintering apparatus was developed based on a traditional vacuum tube furnace. Key improvements included equipping a 10× visual window， a high-speed camera (maximum shooting rate of 3980 frames/s)， and a synchronous light source for real-time recording of the entire experimental process， as well as integrating a high-flux diffusion pump， vacuum gauge， and vacuum meter to achieve a high-vacuum environment of 1 × 10?3 Pa with real-time monitoring. Al?O? particles with three sizes (1， 10， and 30 μm) and five weight percentages (0%， 3%， 5%， 8%， and 10%) were selected as reinforcing phases， NiCoCrAlYTa as the coating matrix， and NiCrSi as the infiltration alloy. Electroless Ni–P alloy plating was applied to Al?O? particles to improve their wettability with the metal melt. The coating preparation followed a specific thermal cycle: heating to 420 °C at 10 °C/min for 30 min， subsequent heating to 1200 °C at 10 °C/min for 2 h， cooling to 600°C at a rate not exceeding 5°C/min， and final natural cooling to room temperature. The microstructure and surface morphology of the coatings were characterized using precision image measuring instruments， scanning electron microscopy (SEM) equipped with energy dispersive spectroscopy (EDS)， and 3D profilometers. Friction and wear tests were conducted on a self-designed rig with a normal load of 1.5 N， sliding speed of 1 m/s， and total sliding distance of 1000 m; the wear resistance was evaluated by measuring the weight loss of coatings and mating graphite disks. [Results] The electroless Ni–P plating effectively improved the wettability of Al?O? particles with the metal melt， reducing the contact angle from 93.425° to 80.371°. Particle size had a considerable impact on coating formation: coatings reinforced with 1 and 10 μm Al?O? particles contained numerous pores due to particle agglomeration， while those with 30 μm Al?O? particles did not exhibit pore defects， resulting in dense coatings with protruding Al?O? particles exhibiting an exposure height of 220–240 μm， which met the morphological requirements. Regarding mass percentage， when the Al?O? content was ≤5%， the total coating porosity remained stable at approximately 0.6% with gas pores as the main defects; beyond 5%， particle agglomeration intensified， clogging seepage channels and leading to a sharp increase in porosity (0.56% for the 3% sample and 2.22% for the 10% sample). Friction and wear test results showed that all coatings containing Al?O? particles exhibited considerably lower wear rates than the coating without Al?O?. The sample with 8wt% 30 μm Al?O? particles achieved the lowest wear rate of 0.00228 mg·N?1·m?1， and the mating graphite disks formed effective wear marks with a depth of 60–90 μm， indicating excellent wear resistance and cutting–sealing performance. The main wear mechanism of the coatings was the spallation of protruding Al?O? particles， and the wear rate increased at 10% Al?O? content because of excessive particle agglomeration.[Conclusions] This study successfully developed a visualized high-vacuum infiltration sintering apparatus that enables real-time monitoring of the sintering process. The optimal preparation parameters were determined as follows: 30 μm Al?O? particles with a mass percentage of 5–8%， sintering temperature of 1 200°C， and holding time of 2 h. The NiCoCrAlYTa–Al?O? coating prepared under these parameters exhibits excellent comprehensive performance， including low porosity， high wear resistance， and strong adhesion without damaging the single-crystal base metal. This technology solves the key technical problems of existing preparation methods and provides important theoretical and experimental support for the engineering application of high-performance wear-resistant seal coatings on single-crystal turbine blade tips.
Influent shock forecasting in wastewater treatment plants: A case study of coupling of shock-preserving preprocessing with a SARIMAX model
WU Hanjiang;LU Donghui;SHAN Ning;XIE Min;ZHU Aifen;XIA Xiuyun;LU Lichao;[Objective] Wastewater treatment plants (WWTPs) are essential for urban water environmental protection， yet their influent is frequently subjected to multi-source disturbances such as rainfall-induced infiltration and atypical discharges， triggering short-term quality shocks that impose significant operational risks. These shock events can rapidly alter organic and nutrient loading， causing mismatches in dissolved oxygen， sludge recirculation， and chemical dosing controls， thereby elevating the risk of effluent non-compliance. Accurate forecasting of influent shocks is therefore critical for stable WWTP operation. Data-driven methods have shown promise in influent prediction， but their deployment is often constrained by high data requirements and reliance on external driving information unavailable at most plants. Under the realistic condition of using only in-plant online monitoring data， ARIMA-family models retain unique advantages in interpretability， low computational cost， and suitability for online deployment. However， two bottlenecks hinder their application to shock prediction: conventional preprocessing uniformly removes outliers through smoothing， inadvertently clipping genuine sustained shock peaks and causing systematic underestimation during high-risk intervals; and fixed-parameter models exhibit response lag and error amplification during shock-induced structural breaks. [Methods] This study proposes a shock forecasting framework coupling three optimization strategies with a SARIMAX model， using 4，901 hourly records of COD， NH?-N， TN， and TP from a WWTP online monitoring system (80/20 chronological split). The first strategy is shock-preserving preprocessing: a duration-based discrimination logic classifies outlier segments persisting ≤2 hours as transient instrumental spikes for local repair， while those exceeding 2 hours are recognized as genuine shocks and entirely preserved. Outlier detection follows the Pauta criterion applied to a 24-point sliding window， and two metrics—shock retention rate and peak preservation ratio—quantify preprocessing fidelity. The second strategy extends ARIMA to SARIMAX with a 24-hour seasonal period for diurnal-cycle modeling to capture daily periodicity driven by urban water-use rhythms. The third strategy is shock-triggered refitting: upon shock detection during the prediction phase， the model automatically refits parameters on recent data to mitigate parameter mismatch caused by structural breaks. Four model configurations are systematically compared: M0 (conventional preprocessing + ARIMA)， M1 (shock-preserving preprocessing + ARIMA)， M2 (conventional preprocessing + SARIMAX with strategies 2～3)， and M3 (all three strategies combined). A three-factor factorial ablation experiment decomposes single-factor， pairwise， and three-way interaction effects to quantify each mechanism’s contribution. [Results] Shock-preserving preprocessing raised the COD shock retention rate from 34.29% to 97.14%， with NH?-N， TN， and TP all increasing from 0% to 100%; peak preservation ratios rose to 1.000 across all indicators. For shock-period prediction， M3 achieved MAE reductions over M0 of 54.8% for COD (from 57.869 to 26.151)， 61.2% for NH?-N (from 11.999 to 4.659)， 51.1% for TN (from 5.330 to 2.608)， and 65.2% for TP (from 0.470 to 0.164). Factorial analysis revealed that shock-preserving preprocessing was the dominant contributor to shock-period improvement， with the largest single-factor MAE reductions for COD (25.430)， NH?-N (5.952)， and TP (0.231). A clear positive synergy was observed between preprocessing and shock-triggered refitting， particularly for COD (interaction effect 8.196) and NH?-N (1.389)， indicating that refitting better tracks structural changes when shock morphology is preserved in the training data. Diurnal-cycle modeling exhibited mostly negative synergy during shock periods but contributed to overall prediction mainly through combined interactions with other strategies. [Conclusions] The proposed method， relying solely on in-plant monitoring data， raises shock retention rates to 97%～100% and reduces shock-period MAE by 54.8%～65.2% compared with the conventional baseline. Shock-preserving preprocessing is the primary contributor to shock-period performance gains， with notable positive synergy with shock-triggered refitting， while diurnal-cycle modeling enhances overall prediction through synergistic interactions.
Research and development of internal inspection devices for large cylinder-type vessels and steel cylinders of long-tube trailers
SHI Kun;ZHONG Maohua;ZHOU Yunyi;HE Yu;CAI Kangjian;[Objective] With the increasing demand for gaseous energy， the demand for equipment for storing and transporting gaseous substances is also growing. Cylinder-type vessels and tube trailer cylinders are the main equipment for storing and transporting gaseous substances， and both have highly similar structural types and damage patterns. Although these two types of equipment have different safety technical requirements， an external inspection mode is always adopted due to factors such as structural type， traditional inspection mode， and operability. However， using an external inspection mode for detecting high-risk internal surface defects does not yield satisfactory results， as it is prone to missing the detection of defects， with certain drawbacks. If an internal inspection is to be conducted， specific inspection equipment must be used， and a series of problems， such as the contracting and expanding of inspection device components， stable support， and multi-directional driving， must be solved. [Methods] To address the deficiencies existing in external inspection and develop an inspection mode that can replace the external inspection mode， we focus on the characteristics of “small mouth and large belly” of cylinder-type vessels and tube trailer cylinders and develop a dedicated internal inspection device based on the required functions of an internal inspection platform. This device is constructed using a modular combination of a detection sensor module， an internal support module unit， a cylinder mouth support module， a motor module， and a drive rod module. The drive rod comprises multiple short rods connected together， and its length can be continuously increased as the detection progresses. Considering its structural strength， stiffness， and lightweight， the device’s main material is aluminum alloy. The analysis involves technologies such as mechanics， electronics， control， simulation， and testing by adopting a method that combines theoretical calculation， simulation modeling， and experimental analysis. [Results] The internal inspection device is easy to install， and the support frame can be flexibly contracted and expanded to achieve “in and out” capabilities. It can also be equipped with various detection sensors. The device can adopt two detection operation modes: axial stepwise circumferential detection and circumferential stepwise axial detection. Although the device has a long rod structure， the maximum deflection generated by the drive rod is acceptable， and the positioning error caused by the drive rod’s torsional deformation can be automatically compensated through the program. We installed multichannel eddy current detection sensors， conducted experiments using comparison specimens， and performed tests at the equipment site. After repeated experiments and tests， the device was proven to be practical， reliable， and stable， capable of performing automatic detection and meeting the internal inspection and detection requirements of cylinder-type vessels and tube trailer cylinders. [Conclusions] In summary， the internal inspection device provides a new platform for inspecting large-volume cylinder-type vessels and long-drum trailer cylinders， solving the long-standing problem of the inability of internal inspection for these systems. The internal inspection mode not only complements external inspection but can also replace the traditional inspection mode， providing an effective means for inspecting new composite material cylinder-type vessels.
Research on SSDPT-based acoustic detection and adaptive thresholding methods for internal leakage of hydropower auxiliary valves
HAN Changlin;MENG Yifei;ZHANG Weijun;LI Yifan;YAN Yanan;[Objective] Early-stage internal leakage in valves used within hydropower-unit auxiliary systems (e.g.， compressed-air subsystems) is typically small， concealed， and strongly affected by operating-condition variability， which makes it difficult to detect using fixed empirical thresholds and difficult to model using supervised learning due to the scarcity of on-site fault samples. This study aims to establish a practical acoustic anomaly-detection route that (i) learns “healthy” valve acoustic signatures mainly from normal data， (ii) remains usable under multi-pressure operating conditions， and (iii) supports robust alarm decision-making through condition-aware and online-adaptive thresholding. [Methods] A controllable experimental platform was constructed by replicating the layout of a hydropower-station auxiliary compressed-air system and integrating a real in-service valve (a DN200 hemispherical valve) as the test object. Valve sounds were recorded using two microphones placed symmetrically around the valve body at approximately 1.0–1.1 m， with synchronized timestamps to ensure consistent multi-channel acquisition. Data were collected under multiple pressure levels from 0.2 to 0.7 MPa. Two representative states were considered: a fully closed valve as the normal condition and a 10% opening to emulate internal leakage. To improve the stability of acoustic inputs， multi-channel waveforms were aligned via cross-correlation， DC offsets were removed， and a band-pass filter was applied to retain diagnostic frequency content; recordings were then resampled to a unified sampling rate and segments with clipping or long saturation were discarded. The processed signals were segmented using a fixed-length sliding window (about one second per segment) with different strides for training and evaluation so that normal data could provide sufficient training diversity while test-time analysis retained high temporal coverage. Log-Mel spectrograms were extracted as compact time–frequency representations using short-time Fourier transform， Mel filter-bank projection， logarithmic compression， and per-channel standardization based only on normal data statistics. On the modeling side， a self-supervised dual-path Transformer (SSDPT) was employed to alternately capture dependencies along time and along frequency， enabling fine-grained characterization of leakage-induced spectral structures and their temporal evolution. Training combined a discriminative identification objective (learning to recognize normal operating signatures across groups/conditions) with a reconstruction objective under random patch masking， which encourages robust representation learning without requiring extensive labeled fault samples. For inference， anomaly scores were computed primarily from the classification-based confidence decay (i.e.， lower confidence in the learned “healthy ID” implies higher abnormality)， and the scoring strategy was analyzed against reconstruction-inclusive alternatives. [Results] The classification-based score provided the most reliable separation between normal and leakage segments. On the complete dataset， the overall area under the ROC curve reached 0.707， while the partial AUC in the low-false-alarm region (false positive rate ≤ 0.10) reached 0.417， indicating meaningful discrimination capability under practical low-false-alarm constraints. Performance， however， was strongly pressure-dependent: medium and high pressures exhibited clearer separability and more stable high-score tails associated with leakage， whereas certain low and mid pressures showed substantial score overlap between normal and leakage segments， limiting the effectiveness of a single global threshold. Pressure-wise analysis highlighted near-complete separability at the highest pressure level and useful separability at some medium pressures， while other lower-pressure settings approached chance-level ordering except for a small subset of strongly abnormal segments detectable under very low false-positive rates. To translate scores into actionable alarms， two complementary thresholding mechanisms were developed. First， pressure-level-specific thresholds were designed to compensate for systematic distribution shifts across pressures and to reduce mismatches caused by mixed-condition score scaling. Second， an online adaptive threshold scheme was formulated to update alarm boundaries during long-term operation by tracking a rolling high quantile of recent scores and calibrating robustness via median absolute deviation， thereby improving stability against gradual background changes and intermittent disturbances. [Conclusions] The study demonstrates that SSDPT-based self-supervised acoustic anomaly detection can serve as a feasible and engineering-oriented approach for internal leakage monitoring in hydropower auxiliary valves when fault labels are limited. Multi-pressure experiments confirm that operating conditions significantly affect score distributions and detection separability， making condition-aware thresholding essential for reliable deployment. The proposed pressure-level and online adaptive threshold strategies improve decision robustness across operating regimes and over time. Remaining challenges are concentrated in low-pressure scenarios where leakage signatures may be weak or masked by background noise; future work can target these regimes through richer sensing configurations and acoustic–vibration multimodal fusion， as well as validation in longer-term field monitoring pipelines.
Design of an Aerodynamic Experimental Teaching Platform for Wind Energy Drag Reduction Based on the Magnus Effect
SUN Huawei;ZHAO Xingyu;HAN Yang;ZHAO Dagang;ZHOU Guangli;Under the background of the “New Engineering Education” initiative， traditional experimental teaching in fluid mechanics and aerodynamics has been facing challenges such as insufficient integration with engineering practice， an overemphasis on verification-based experiments， and limited comprehensiveness and innovation. To address these issues， this study designs and develops an aerodynamic experimental teaching platform for wind-assisted drag reduction devices based on the Magnus effect. Centered on rotor wind tunnel experiments， the platform integrates fluid mechanics theory， aerodynamic measurement techniques， and ship drag reduction engineering applications. Through a modular experimental framework， the aerodynamic characteristics and flow interference phenomena of rotating cylinders under varying inflow velocities， spin ratios， and rotor arrangements are systematically investigated， providing intuitive and efficient experimental support for teaching complex aerodynamic mechanisms. In terms of system design， the platform integrates a variable-speed rotating cylinder device， a multi-component force balance， rotational speed and wind speed measurement units， and a data acquisition and processing system. It enables synchronous measurement of lift， drag， and aerodynamic torque with good stability and repeatability. Systematic experiments on single-rotor and dual-rotor configurations were conducted to obtain aerodynamic response characteristics under different parameter combinations. The results demonstrate that the platform effectively reveals the physical mechanism of rotation-induced lift in the Magnus effect， as well as the influence of inter-rotor flow interference on aerodynamic performance， providing reliable experimental evidence for wind-assisted propulsion and ship drag reduction applications. In teaching practice， the experimental platform has been incorporated into fluid mechanics and ship engineering–related courses through a three-level experimental framework comprising fundamental verification， parametric analysis， and engineering extension. This approach guides students progressively from theoretical understanding to engineering application. During the experiments， students systematically acquire skills in wind tunnel testing， rotating system control， multi-component force measurement， data processing， and uncertainty analysis， significantly enhancing their experimental design capability， engineering thinking， and teamwork skills. Moreover， the close integration of experimental content with research problems enables students to gain initial exposure to complex nonlinear flows and engineering optimization issues， stimulating their interest in cutting-edge technologies and innovative research. Overall， the proposed experimental teaching platform achieves an effective integration of “research-driven teaching and teaching-supported research.” It not only enhances the depth， challenge， and engineering orientation of experimental education， but also provides strong support for cultivating innovative and well-rounded engineering talents in ship and ocean engineering and related disciplines. The study indicates that the platform has strong demonstrative significance and broad applicability for experimental curriculum development and engineering education reform under the New Engineering Education framework.
Design and experimental evaluation of uniform magnets based on finite element analysis
YUAN Bo;REN Xiuyan;WU Dan;TIAN Yaqi;WANG Guobao;[Objective] The generation of a uniform magnetic field plays a pivotal role in various engineering applications and experimental endeavors， contributing to scientific progress and technological innovation. In the context of plasma experiments， a uniform magnetic field significantly enhances the intensity of gas discharge processes， effectively minimizes the diffusion losses of plasma particles to chamber walls， and markedly improves the overall efficiency of plasma generation. Plasma experiments have stringent requirements in terms of the three-dimensional distribution of the magnetic field and uniformity of magnetic induction intensity. To meet these demands， a dedicated magnet system must be designed to produce a magnetic induction intensity of 1 000 Gs with a uniformity of ±1% within a cylindrical spatial volume of Φ45 mm×150 mm. [Methods] The Helmholtz coil is an effective device for generating uniform magnetic fields over small， localized areas. Finite element modeling was utilized to systematically compute the magnetic induction intensity distributions for coils of varying diameters. The analysis revealed that increasing the coil radius improves magnetic field uniformity while correspondingly decreasing the overall magnetic induction intensity in a nonlinear correlation. This insight allows for informed trade-offs in optimizing the design of devices to balance cost and performance. By integrating the specified design requirements alongside considerations of cost- effectiveness and manufacturing feasibility， a viable engineering scheme was proposed. The three-dimensional distribution of the magnetic field was discussed in detail for a configuration featuring four pancakes per coil and a coil radius of 250 mm. Through iterative simulations， the final magnet design achieved a uniformity of 1 000 Gs±0.52%， demonstrating superior precision exceeding initial expectations. The key structural components of the comprehensive system are the magnet coils， water cooling mechanisms， power supply units， and adjustable supporting brackets. Detailed specifications were provided for the power supply， ensuring stable and efficient operation， as well as for the water cooling system， which maintains thermal stability to prevent overheating and ensure long-term reliability. [Results] To validate whether the magnetic field distribution meets the established design criteria， experiments were performed using a magnet system. Measurements were conducted using a high-precision three-axis Gauss meter， which provided accurate readings across the targeted volume. The results conclusively demonstrated that within the Φ45 mm×150 mm three-dimensional cylindrical space， a magnetic field uniformity of 1 000 Gs±0.66% was attained， fully complying with and even surpassing the required ±1% tolerance in practical implementation. The designed uniformity (±0.52%) was slightly superior to the experimental value (±0.66%)， where three primary contributing factors were identified and analyzed in depth. Overall， the experimental data underscore the high performance of the system and confirm the successful integration of all system components. [Conclusions] This paper details the process for optimizing the design of a uniform magnet system based on finite element calculations. Detailed analyses of three-dimensional magnetic field distributions for various structural configurations were presented， illustrating the nuanced interplay between design parameters and performance outcomes. Experimental analyses following the completion of machining and assembly yielded results that align closely with the designed parameters. Specifically， the measurements confirm that within the prescribed three-dimensional spatial domain (Φ45 mm×150 mm)， the magnetic field uniformity reaches 1 000 Gs±0.66%， where the distribution pattern and induction intensity fully satisfy design specifications. The magnet system has been successfully deployed in real-world applications. Future work can build upon this foundation to explore scalable designs， thereby creating a positive feedback cycle for advancing magnet design.
Experiments on multi-UAV path planning Using DRL integrating graph neural networks and curriculum learning
Fu Mingjian;Chen Wentao;Zhuo Xiaoxin;Chen Hengsheng;Chen Fei;[Objective] With the widespread application of Unmanned Aerial Vehicles (UAVs) in disaster rescue， industrial inspection， and other scenarios， multi-UAV path planning in limited airspace faces dual challenges of avoiding dense static obstacles and improving operational efficiency. Traditional path planning methods based on environmental priors struggle to adapt to dynamically generated scenarios with randomly distributed obstacles. Existing reinforcement learning algorithms predominantly rely on simplified 2D planar assumptions， neglecting 3D spatial constraints for obstacle avoidance. This study proposes a collaborative decision-making framework integrating Graph Neural Network (GNN) architecture optimization and Progressive Curriculum Learning for multi-UAV path planning in 3D static dense obstacle environments. The key research contributions and innovations are summarized as follows: [Methods] Firstly， a 3D path planning model based on the Markov Decision Process (MDP) is constructed by incorporating altitude dimensions into state representations and designing node-type identification mechanisms. This enables UAVs to distinguish heterogeneous characteristics between themselves and obstacles. Addressing limitations of conventional GNNs in spatial relationship modeling， this work couples edge features (including relative velocity， position， and distance) with neighbor node features (containing relative centroid position， velocity， and type identifiers) and employs multilayer perceptrons to generate joint representations. This approach replaces the linear superposition of independently encoded features used in existing algorithms， thereby enhancing the network's capability to analyze complex obstacle spatial distributions. Secondly， a safety-efficiency-balanced reward function is formulated by integrating multidimensional metrics such as target proximity， first-arrival time， dwell duration， velocity alignment， collision risks， and proximity penalties. This design guides UAVs to achieve optimized trade-offs between obstacle avoidance and navigation objectives， improving trajectory rationality and policy convergence speed. Thirdly， a three-stage progressive training framework is developed， transitioning from sparse to dense obstacle scenarios. UAVs initially learn basic obstacle avoidance strategies in simplified environments， gradually progressing to moderate-difficulty environments， and ultimately generating cooperative paths balancing safety and efficiency in complex obstacle configurations. This methodology addresses suboptimal policy issues caused by excessive exploration space in high-dimensional environments. Finally， a 3D multi-UAV path planning test environment is established using the PyBullet high-fidelity physics simulation platform， featuring randomly distributed static obstacles with varying density levels. [Results] Experimental results demonstrate that the proposed Edge-Enhanced Informative Multi-Agent Proximal Policy Optimization (EC-InforMAPPO) framework outperforms baseline algorithms across all scenario difficulty levels. Its edge feature encoding mechanism， coupling relative motion parameters and spatial relationships， enhances trajectory safety in dense obstacle environments， offering a novel technical pathway for environmental perception modeling in multi-agent systems. Additionally， the progressive curriculum learning framework enhances policy stability in challenging scenarios. The Edge-Enhanced Informative Multi-Agent Proximal Policy Optimization with Curriculum Learning (EC-InforMAPPO-CL) achieves higher obstacle avoidance success rates and faster convergence compared to direct training using equivalent computational resources. This establishes a reusable training paradigm for reinforcement learning in high-dimensional state spaces. [Conclusions] This paper proposes a collaborative decision-making framework that combines edge feature coupling based on Graph Neural Networks with progressive curriculum learning， addressing the challenges in path planning for multiple UAVs in three-dimensional dense obstacle environments. This research provides new insights and technical support for intelligent collaborative navigation of multiple UAVs in complex environments， holding significant application potential and promotional value.
Research on the pedagogical path of AI programming assistants in EDA software development education
HE Zhonghai;XU Zhifu;YU Wenjie;FAN Zehui;CAO Sheng;LIU Leyuan;ZHANG Xiaosong;[Objective] The teaching of large-scale Electronic Design Automation (EDA) software development faces substantial challenges， including massive codebases， complex architectures， and steep algorithm learning curves， which severely constrain students' engineering practice capabilities. While AI programming assistants such as GitHub Copilot or Cursor offer unprecedented opportunities to transform software engineering education， critical questions remain: How can these tools be rationally integrated into complex software development pedagogy? What are the optimal usage patterns that enhance learning without undermining independent problem-solving capabilities? This study systematically investigates the pedagogical pathways and application boundaries of AI programming assistants in EDA software development education. [Methods] Grounded in progressive capability construction and intelligent augmentation enhancement principles， this research designed a three-tiered task system: source code analysis (tracing Yosys synthesis flow)， algorithm comprehension (reverse engineering ABC’s And-Inverter Graph rewriting algorithms)， and functional extension (developing Yosys statistical commands). A controlled experiment with 30 undergraduate students stratified by programming proficiency involved three groups: GitHub Copilot–assisted， Cursor-assisted (with Claude 3.5)， and non-AI-assisted control. The four-week experiment assessed pre- and posttest task completion quality， time efficiency， comprehension depth， and subjective experience via Likert-scale questionnaires. AI interaction frequencies were logged to analyze usage patterns. [Results] AI-assisted groups demonstrated substantial improvements: task quality increased from 16.9% (Copilot) to 21.2% (Cursor)， with Cursor showing particular advantages in architecture analysis (20.3%) and algorithm comprehension (20.7%). Time efficiency gains were remarkable—Cursor reduced completion time by 71.0% and Copilot by 47.3%. Post-test scores measuring deep comprehension increased by 8.7 points (Copilot) and 10.8 points (Cursor)， representing 1.9× and 2.3× improvements over the control group's 4.6-point gain. Subjective metrics showed enhanced self-efficacy and reduced frustration. Critically， correlation analysis identified an optimal usage range of 20–22 interactions; students exceeding this threshold exhibited declining performance， suggesting that excessive reliance undermines independent development capabilities. [Conclusions] This study establishes an empirically validated pathway integrating AI programming assistants into EDA education through progressive task design， moderate AI assistance， and process-based monitoring. The findings reposition AI tools as cognitive scaffolds rather than knowledge providers， offering actionable insights for complex software engineering pedagogy and providing empirical evidence for AI integration in computing education.