This PDF file contains the front matter associated with SPIE Proceedings Volume 13275, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Different from optical images, detection performance in synthetic aperture radar (SAR) images is poor. To solve issues above, an improved ship target detection method is proposed in this paper. Firstly, wavelet decomposition is introduced, the advantages of using wavelet pooling in the down-sampling process of image feature maps were analyzed, replacing the normal pooling with the wavelet pooling in the standard Faster R-CNN and improving the detection performance of large ship targets. Besides, the wavelet decomposition is iterated to obtain the wavelet hierarchical decomposition, forming a wavelet convolutional neural network (WCNN). The obtained feature maps of different scales are fused with the feature maps of the backbone to improve the accuracy of small object detection. Finally, combining the above two methods, the algorithm proposed in this paper is obtained. Through ablation experiments, the target detection performance of different methods was compared, and the results showed that compared to the standard Faster R-CNN algorithm, the algorithm proposed in this paper can improve the mean average precision by 2.2%.

To guarantee the stable functioning of extensive comprehensive collaborative power communication network equipment linked to the power grid, this paper suggests a defect correlation analysis approach for comprehensive collaborative power communication network equipment, enhanced by the FP-Growth algorithm. Firstly, the fundamental framework and underlying principles of the comprehensive collaborative power communication network equipment are examined, and the malfunction data of the comprehensive collaborative power communication network equipment are gathered. Then, the original data is preprocessed, quantized and encoded into FP-Growth algorithm, and its support and confidence are calculated by constructing FP-Tree. Finally, the effectiveness of the method is verified by the simulation example of the actual full-level collaborative power communication network equipment defect data. This method can analyze and trace the information related to the core equipment defects of the fullness collaborative power communication network in an efficient and intelligent way, thus offering theoretical backing for operational and maintenance staff to devise maintenance strategies.

This study first describes the application and development level of software defined radio; Secondly, it analyzes the problems of traditional radio and the advantages of software defined radio. Again, several software defined radio platforms based on different signal processors are discussed and compared in terms of performance and application. Finally, the text transmission system of software radio platform based on FPGA (Field Programmable Gate Array) is designed in detail, and text data sending and receiving are realized.

Unmanned aerial vehicle (UAV) refers to an unmanned aircraft operated by radio remote control equipment and self-provided program control devices. The development of unmanned aerial vehicle, which originated from the verification application, has a history of 100 years. In the 1930s, the United Kingdom first used it as a remote control target aircraft. With the rapid development of integrated circuits, information communication, material science and other technologies, UAV technology has shown the characteristics of networking, clustering and intelligence. Uav has low cost, low risk of casualties, strong adaptability to bad weather, and remarkable asymmetric nonlinear countermeasures. Countries continue to develop a variety of UAVs to adapt to a variety of task environments. In view of such characteristics, combined with different distances and deployment, point-to-point confrontation, genetic algorithms are used to optimize the countermeasure strategy, which significantly reduces the risk of countering UAVs.

In the design of automatic weeding machine, in order to realize its autonomous operation function, it is necessary to ensure that the weeding machine can independently distinguish weeds from crops. Considering the problem of small sample identification, the study addresses this point based on the meta-learning algorithm, through which the machine can accurately identify weed targets and thus automatically lock the job objects. And considering the green and efficient weed removal, the study used laser weeding. After the target weeds are detected and identified, the laser actuator is started to burn out the weeds and the laser used is a point laser to ensure that the weeds will not harm the crops, so as to achieve the weeding method without pesticide residues and improve the planting environment of crops.

In view of the problems such as the huge amount of data generated by traditional Nyquist sampling and the complex structure of the existing array compression sampling system, this paper proposes an L-array compression sampling system based on improved Modulated Broadband Converter (MWC), and uses the combination method to realize the carrier frequency estimation of multiple sources with the same carrier frequency. At the same time, the matrix enhancement and matrix pencil two-dimensional(2-D) direction of arrival (DOA) estimation algorithm based on the system is proposed, which can directly use the combined compressed sampling data to achieve 2-D DOA estimation. The complexity of the proposed system is low, which is beneficial to the realization of hardware engineering, and the simulation experiment proves that the parameter estimation algorithm based on the system has good estimation performance.

In the daily inspection process of urban cable pipelines, due to the high concealment of urban underground cable networks, the use of traditional underground cable network GIS information not only has certain map recognition thresholds, but also cannot determine the specific actual location of underground cable networks in GIS information due to environmental changes, geological changes, vegetation changes, and other environmental reference changes in real-life scenarios. At the same time, due to the lack of spatial sense in viewing pipeline GIS information, the quality of line patrol inspection and the accuracy of pipeline excavation cannot be guaranteed. This article relies on GIS high-precision positioning technology, intelligent north pointing calibration, augmented reality technology based on ORB-SLAM3, data analysis and calculation, and multi-spatial data conversion visualization display technology to create cable network data platform, augmented reality visualization mobile terminal and other systems. This article mainly demonstrates that by combining GIS data of cable pipelines with high-precision positioning of mobile terminals, and transforming GIS two-dimensional spatial data of electricity pipeline into intuitive and visualized augmented reality SAR data, through ORBSLAM3, it provides a theoretical method for more stable and reliable real-space binding for SAR, laying the foundation for reducing the maintenance difficulty of urban cable pipeline data, improving the efficiency of cross-department business collaboration within power enterprises, and providing intensive data application capability support for urban management.

Aiming at the problem of high efficiency and energy saving in WSN, a multi-hop cooperative transmission strategy is proposed based on CMIMO technology. The multi-hop cooperative transmission model which is more widely used in practice is adopted. The simulation results show that the multi-hop cooperative MIMO transmission strategy greatly improves the network transmission efficiency and saves energy consumption.

As the key equipment in the power system, the safe operation of gas-insulated metal-enclosed switchgear (GIS) is important. Timely and correct fault recognition of insulation defects in GIS is of great importance for early fault warning. In this paper, partial discharge of 5 typical insulation faults in GIS equipment is experimentally simulated. By calculating statistical feature characteristics, an efficient classification method based on agglomerative hierarchical clustering is proposed. The results show that the calculation of the characteristics of discharge samples is simple, the physical meaning is intuitively clear, and the clustering method is able to distinguish different samples based on the differences in characteristics without prior knowledge. The statistical feature characteristics can accurately characterize the features of different typical defects in GIS and can distinguish different defects through agglomerative hierarchical clustering analysis. The method can provide a certain reference for the diagnosis and early warning of insulation faults in GIS equipment.

Coverage performance is an important index to evaluate the reliability and stability of the service provided by the megaconstellation. Different from traditional constellations, mega-constellation has a large number of satellites distributed in several different orbital planes, which leads to a sharp increase in the calculation amount of traditional coverage performance analysis methods. In this paper, the position of the satellite is deduced by TLE (Two-Line Elements) data and SGP4 (Simplified General Perturbations Satellite Orbit Model 4) model. Then, a multi-precision method for dividing grid points is proposed, which is different from the traditional method of dividing grid points at equal intervals. The longitude interval is set according to the latitude in the method, so as to avoid the excessive division of longitude in the high latitude region. Meanwhile, the selection of grid points near the 180° longitude line is optimized in view of the characteristics of the coincidence of the 180° east-west longitude line. On this basis, an improved visibility function model is constructed which comprehensively considers the beamwidth of satellite antenna and the minimum elevation angle of grid point to analyze the coverage performance of mega-constellation, overcoming the problem that the traditional visibility function model cannot consider the restriction of the beamwidth of satellite antenna to the coverage performance. Finally, this paper takes OneWeb constellation as an example to analyze the proposed method, and compares it with the traditional method. The simulation results show that the proposed method can greatly reduce the computational complexity while ensuring the accuracy of the calculation results.

Real-time monitoring of circuit breaker mechanical characteristics enables early diagnosis of gradual mechanical failure, ensuring stable and reliable operation. This involves studying multi-physical intelligent sensing and wireless sensing technologies, integrating intelligent monitoring methods with multi-state variables, and developing compact online monitoring devices. A dynamic simulation model and deep learning-based calculation method for circuit breaker mechanical characteristics are proposed. Finally, a real-time diagnosis and evaluation method incorporating circuit breaker parameters is developed, along with a new management system for accurate analysis of operation state.

The large transformer is a critical component of substations and has historically suffered significant damage during earthquakes. This study develops an enhanced finite element model of a 220kV transformer to accurately assess its seismic performance. The study examines the seismic response characteristics and equipment resilience by utilizing four sets of typical seismic wave records. This enables the identification of critical failure locations and responses. The study results indicate that the high-voltage bushing, a critical component of the 220kV transformer, is susceptible to equipment failure due to excessive stress at the root and displacement at the top. Under the seismic action of a 9-degree design-level earthquake, the seismic resistance of the transformer in the X direction is insufficient.

When the distribution network line fails, the island nodes generated by the line disconnection will lose load due to the inability to get the superior power supply, which further leads to the failure of the corresponding information nodes to effectively monitor and control, seriously affecting the safe and stable operation of the power system. On the basis of this problem, this paper proposes a method to improve the resilience of urban power grid by taking into account the state of information nodes. In this method, the objective function is established by considering the information node state and load reduction comprehensively, and the influence of line disconnection caused by extreme disasters on the system is reduced by the cooperative regulation of building load, energy storage and distributed generator. Therefore, the source-load-storage combined optimization model is established and the revised IEEE33-bus system is used to validate the model. The results show that the model proposed in this paper can effectively reduce the impact of typhoon disasters, and is conducive to the application of comprehensive consideration of budget and system disaster resistance in practical projects.

For a narrow-beam radar working in a down-looking view, the antenna pattern modulates the ground clutter apparently in the delay/range dimension, causing amplitude/intensity distortion of radar echoes and performance loss of target detection. This paper proposes a range-dimensional clutter distortion correction method based on clutter pre-screening and beam pattern demodulation. This method realizes the correction of clutter distortion, to improve the target detection performance of Narrow-beam radar. The effectiveness of the method is verified through simulations and experimental measurements. The proposed method demonstrates practical significance in improving the detection probability of radar targets under clutter distortion in narrow-beam broadband radar systems.

With the continuous evolution of supercomputers, traditional public-key encryption becomes insecure. Quantum key distribution (QKD) and physical layer security (PLS) technologies, however, are expected to innovate encryption system for they both protect confidential information without relying on computational security: the unobservability of quantum states ensures QKD providing secure keys for partners Alice and Bob, while PLS generates keys by channel's characters. The noise in optical fiber and wireless channel makes the initial key between partners inconsistent, and the same key can be obtained only after key reconciliation process. Traditional reconciliation methods have disadvantages either in reconciliation speed or data utilization, and new algorithms based on error correction coding are prone to error diffusion when the initial key disagreement rate (KDR) is high. To solve the above problems, this paper proposes a novel secret-key reconciliation algorithm based on two-dimensional error correction (2DEC), in this approach, data is arranged in two dimensions so that a faster reconciliation is realized. We analyze its theoretical bound of error correction capacity, and then design a scheme for accurate error correction. Thereafter, we combine 2DEC with the optimized Cascade, the joint algorithm achieves even better performance. It not only has a good robustness to the initial KDR, but also can effectively improve the reconciliation speed and data retention rate. Simulations show that the speed and data preservation is increased by 100% and 30% respectively when SNR is higher than 8dB. Hence, the proposed method can help QKD and PLS system improve secret-key generation rate, which is of great significance to realize a safer confidential communication in the future.

Addressing the issue of inconsistent energy storage requirements and dishonest transactions in microgrids, a multimicrogrid optimization considering shared energy storage and cooperative fraud is proposed. Firstly, the framework of multi-microgrid system considering shared energy storage is established. Secondly, the optimization model of multimicrogrid system considering shared energy storage and false information is established. Based on Nash negotiation theory, the problem split into two sub-problems: minimizing system costs and cooperative fraud distributing revenue. Finally, the model is solved using the Alternating Direction Method of Multipliers. The simulation results show that the proposed method can maximize the system benefits and ensure the interests of all subjects.

The paper focuses on exploring the design of an ultrasonic liquid level measurement system, introducing its universal design concepts, hardware design, software design, system composition, and detection principle. This study can provide necessary guidelines for the design of ultrasonic liquid level measurement systems.

To address the problem that the existing Linear Frequency Modulated Continuous Wave (LFMCW) signal detection method needs to search in the multidimensional parameter domain leading to large computational complexity, this paper improves the delay structure of the discrete polynomial phase transform (DPT) and proposes the LFMCW signal detection method based on the periodic DPT (PDPT). The proposed detection method retains the advantage of low computational complexity and eliminates the interference term, which improves the problem of poor anti-noise performance of LFMCW signals due to the inability of optimal energy aggregation. In this paper, we first analyze the computational complexity and detection performance of the proposed detection method, and then verify the effectiveness of the proposed method by using simulation results and measured data.

With the rapid construction and development of China's energy interconnection system, more than 600 million end-side equipment are currently connected to the power system. It is expected that by 2030, the number of equipment will reach 2 billion. However, due to the large scale, complex types, and limited resources of the power end-side devices, the traditional identity authentication based on PKI digital certificates has poor scalability, high management and maintenance costs, and CA security problems, and is not applicable to the power end-side business scenario. Meanwhile, the end-side devices with similar service and electrical characteristics are effectively classified and grouped. It has become an urgent need to carry out efficient group authentication to realize group control.

In view of the above problems, this scheme studies and proposes a lightweight access authentication management method for complex power end-side devices. Firstly, on the basis of analyzing the specific performance of power end-side devices, the authentication requirements of devices are analyzed, including identity authentication and authority control requirements, as well as batch authentication requirements in the face of group tuning and group control scenarios. Then, a lightweight access authentication management method for power end-side devices is proposed, which realizes the initial key distribution of devices based on lightweight cryptographic resource pool, and a lightweight access authentication management method based on dynamic group is designed to realize batch authentication of end-side devices and data upload, and group device authentication. Finally, the performance analysis and security analysis show that compared with the existing typical group authentication schemes, the proposed scheme can ensure the security and reduce the calculation cost.

This study aims to improve the accuracy and energy efficiency of body temperature measurement and localization algorithms in wearable devices for health monitoring and user experience needs. The current algorithm has the problems of limited accuracy and vulnerability to interference. Therefore, a comprehensive algorithm optimization scheme is proposed. Through data preprocessing, temperature compensation and model establishment, the temperature measurement algorithm is optimized. At the same time, the Kalman filter algorithm and multi-point positioning are used to improve the positioning accuracy. Experimental verification shows that the optimized algorithm has a significant improvement in temperature measurement and positioning, and the research results provide important technical progress for the practicality of wearable devices.

The HUST Field-Reversed Configuration (HFRC) is able to form a Field-Reversed Configuration (FRC) plasma by the theta-pinch method, and improve the FRC parameters formed by the collision fusion process. The pre-ionization phase is a pivotal stage in FRC formation, and the characteristics of FRC plasma depend largely on the specific conditions of gas breakdown in the pre-ionization stage. The breakdown process of plasma undergoes a short time, the parameter change amplitude is very large, and the breakdown process data obtained by experimental measurement is very limited, so it is important to use numerical simulation method to adjust the controllable parameters of the discharge process to study the parameter evolution of plasma under different discharge conditions. Based on the structural characteristics and operating parameters of the HFRC platform, particle models and Monte Carlo will be used to calculate the gas breakdown process in the pre-ionization stage, showing the evolution of parameters during the breakdown process, and comparing and analyzing with the experimental data measured in the HFRC formation region to study the initial bias magnetic field, pre-ionized magnetic field, It is hoped that the influence of induced electric field and gas-charged air pressure on the plasma behavior during breakdown can improve the captured magnetic flux under the condition of sufficient ionization of the gas, and improve the characteristics of FRC plasma formed in subsequent experiments to a certain extent.

In power line communication system, ensuring support for multi-user concurrent communication entails the precise channel state detection to determine whether the channel is available. However, power line communication system is greatly affected by the impulse noise, which is a challenge not fully addressed by existing channel state detection schemes. To improve the accuracy of discerning the state of power line channels, unlike the traditional channel detection scheme that divides the channel state into two states: with signal or without signal, this paper categorizes channel states into three classifications: Signal, Impulse noise, and Gaussian noise. This classification is achieved through the application of deep learning. Our method leverages both time domain and frequency domain information of the signal to refine accuracy. Simulation results demonstrate a notable improvement in the accuracy of our proposed channel state detection method compared to conventional methods. Moreover, the proposed method can adapt to different signal modulation types, demonstrating strong robustness.

Power line communication (PLC) is essential for smart grid applications but struggles with noise due to its use of electrical lines for data transmission. PLC systems are particularly affected by bursty impulse noise. In order to estimate the impulse noise, this paper proposes a variational inference generalized approximate message passing (VIGAMP) based algorithm that simplifies complexity and improves performance for PLC-OFDM system. Compared to existing Sparse Bayesian Learning algorithms, Simulations reveal that the proposed method offers better Bit Error Rates performance and faster convergence.

Compared with the traditional modulation and demodulation methods, Gaussian minimum frequency shift keying has the advantages of narrow bandwidth, fast lobe attenuation, easy FPGA implementation and strong anti-interference ability, so it is widely used in frequency hopping communication. At the same time, the reason why GMSK is called efficient digital technology is mainly because it is developed from MSK modulation and has a modulation index of 0.5.

As one of the major air pollutants, PM2.5 poses serious threats to human health and daily life. Improving the accuracy of PM2.5 predictions is of significant importance for pollution control, as it can effectively reduce its harmful impact. Long Short-Term Memory (LSTM) networks are capable of handling long-term dependencies and are commonly used for modeling and remembering long sequence data, making them a key tool for PM2.5 prediction. Existing LSTM models often use a single parameter for PM2.5 prediction, without considering the coupling effects of multiple meteorological factors with different pollutants and their concentrations. Additionally, using a single optimization algorithm may lead to optimization parameters getting stuck in local optimal solutions, disrupting the optimization results. In this study, multiple meteorological factors highly correlated with PM2.5 data using Pearson correlation coefficients (PCCs) and pollutants with concentrations were selected to construct a coupled correlation model. Furthermore, to improve the reliability of optimization results, the Sparrow Search Algorithm (SSA) with strong local search capability and the Particle Swarm Optimization algorithm (PSO) with strong global search capability were employed to simultaneously optimize LSTM hyperparameters and obtain the optimal solution. Using hourly meteorological data from multiple stations in Beijing from March 1, 2013, to February 28, 2017, as the data source, the SSA-PSO algorithm was used to obtain the optimal solution for network hyperparameters. The coupled correlation model of multiple meteorological parameters with multiple pollutants and concentrations was established. The research results indicate that considering the coupled correlation model of meteorological parameters with pollutants and concentrations can effectively improve the prediction accuracy of PM2.5, and the more types of input parameters, the higher the model prediction accuracy. The R² of the coupled correlation model with seven parameters increased to 0.9556, and the RMSE decreased by 7.68% compared to the single-parameter model.

This article proposes an indoor 5G and UWB fusion positioning algorithm to address the issues of a single positioning source being easily affected by environmental factors, low positioning accuracy, and poor robustness in indoor environments. This algorithm filters and aligns the data of 5G and UWB positioning units, and sends them to a BP neural network optimized by LM for training, outputting the positioning results. And conduct static and dynamic experiments at a driving school in Zengcheng District, Guangzhou. The experimental results show that under static testing, the LM-BP algorithm has a significant effect on accuracy improvement, with a 32.84% increase compared to the 5G ratio and a 24.24% increase compared to the UWB ratio. Under dynamic testing, the LM-BP algorithm has an 30.81% increase compared to the 5G ratio and an 18.89% increase compared to the UWB ratio.

To study the time-domain characteristics of spectrum-spread interference after matched filtering, this paper uses the stationary phase method to estimate its time-domain output. Then, the interference signal is simulated by matched filtering operation to obtain the corresponding pulse compression results, which are similar to the estimated results of the stationary phase method. The results show that the stationary phase method is a simple and effective way to analyze the pulse compression properties of spectrum-spread interference signals.

To ensure the safety, efficiency, and stability of batteries, accurate estimation of battery health status is crucial. To address this, this paper proposes a method for estimating battery state of health by integrating interval local information. Firstly, the full-cycle discharge data of the battery are divided into voltage intervals, and the time differences within each voltage interval segment are extracted as features to characterize the battery's health status. Then, the features within each voltage interval segment are fused. Finally, a Long Short-Term Memory neural network algorithm is employed to construct a battery health status prediction model, achieving accurate estimation of the battery health status. Experimental results demonstrate that the proposed method can accurately estimate the health status of individual cells and battery modules, with R-squared values (coefficient of determination) exceeding 0.99 and root mean square errors within 1%.

A refined Whale Optimization Algorithm (MWOA), employing a hybrid strategy, is introduced to address challenges such as slow convergence, weak global search capability, low accuracy in problem-solving, and susceptibility to local optima. Leveraging the unique properties of the Chebyshev chaotic sequence, the algorithm strategically optimizes the initial population's placement. This ensures enhanced diversity among individuals, leading to a more rational distribution of locations, thereby boosting the overall optimization efficiency. Furthermore, the application of the reverse learning strategy from pinhole imaging enhances the leader's capacity to escape local optimal regions, subsequently elevating the algorithm's accuracy in problem-solving. During the whale spiral stage, an adaptive weight parameter is introduced to augment the algorithm's local search capabilities and refine convergence accuracy. The experimental results of 7 benchmark functions indicate that the MWOA algorithm outperforms any single-strategy improved algorithms mentioned above, thus confirming the effectiveness and superiority of the improvement method.

Diversified recommendations can alleviate issues such as filter bubbles and redundant recommendations in recommendation systems. However, the enhancement of diversity is often accompanied by a decrease in relevance. One commonly used method for diversified recommendations is re-ranking, with most re-ranking algorithms aiming to balance the diversity of the list and its relevance to the user. However, these algorithms often overlook the user's acceptance of diversity and ignore hidden information in the ranking scores generated during the candidate list generation stage. To address these limitations, we propose the UPCC (User Preference Cluster Classification) framework to optimize existing re-ranking models for diversity recommendations. UPCC utilizes long short-term neural networks to predict the user's relevance demand for items in the next moment. It filters candidate clusters based on this demand and re-ranks the filtered clusters to obtain a set of recommended clusters. Finally, the highest-ranked items from each recommended cluster are selected to form the recommendation list. Through experiments conducted on ML-1M and ML- 100k datasets, the results indicate that the re-ranking model using UPCC improves the speed and accuracy of recommendations. Additionally, it generates recommendations that better align with users' diverse preferences.

This study employs machine learning methods with the aim of detecting daily-life sleepiness in the daily study life of college students, facilitating individuals to adjust their states sensibly and enhance work efficiency. Initially, we collected electrocardiogram (ECG) and tri-axial accelerometer data from 83 college students during their daily activities. We extracted 21 features from the ECG and 3 features representing the level of physical activity from the tri-axial accelerometer data. Subsequently, we trained five classifiers using these features and employed a backward selection algorithm through Leave-One-Subject-Out (LOSO) cross-validation to choose the most crucial feature subset. Ultimately, utilizing the best-performing classifier and the selected key feature subset, we constructed a model for detecting daily-life sleepiness, achieving the F1 score of 77.61% on the independent test set. The results suggest that employing machine learning to detect the sleepiness of individuals is feasible.

The cyclotron is an important device in the field of high-energy physics experiments and radiation medicine, and its performance is directly affected by the uniformity and stability of the magnetic field. In this paper, a Hall sensor-based magnetic field measurement system is introduced to achieve high-precision measurement and real-time monitoring of the magnetic field of cyclotrons. The system realizes the accurate measurement and dynamic analysis of the magnetic field distribution in the cyclotron through the combination of sophisticated mechanical and advanced data acquisition technology and magnetic field reconstruction algorithm. On the basis of full consideration of environmental factors and systematic errors, the measurement results show excellent accuracy and repeatability, which effectively guides the commissioning and optimization of the accelerator.

In order to improve the power supply reliability of the distribution network and build a digital distribution network, power supply companies in most provinces have carried out the construction of distribution automation system. The feeder terminal unit (FTU) has relay protection function. The incorrect operation of breaker sometimes caused by the wrong configuration of FTU parameters. The paper analyzes a distribution network failure involving the incorrect operation of the FTU relay protection. Firstly, the process of distribution network fault is introduced. Then, the paper analyzes the results of the accuracy test, protection value analysis, SOE analysis, and device disassembly analysis of the FTU. Finally, the paper provides a detailed analysis of the fault result and proposes rectification suggestions. Through the fault analysis, the paper aims to optimize the parameter configuration of FTU relay protection and improves the reliability of distribution network.

This paper proposes a method for EMI source localization on DCDC converter circuit boards based on machine vision. By analyzing the operating principle of DCDC converters and the mechanism of EMI generation, the application of vision technology in EMI source localization is discussed. A complete system including hardware platform construction, image processing, and EMI source localization algorithms is designed. Simulations and experiments show that this method can effectively identify and locate the EMI source with high precision and real-time performance. There is a good correspondence between the magnetic field distribution images and the EMI source locations. This paper provides new ideas and methods for the diagnosis and resolution of EMI issues.

To meet the requirements of multiple-input multiple-output (MIMO) radar target detection and parameter estimation in a multi-target scenario, this study establishes a multi-objective optimization model with the weighted sum of dual mutual information and signal-to-clutter-plus-noise ratio (SCNR) as the criterion function, and proposes a waveform optimization method to meet the given constraints. Firstly, based on the multi-objective optimization model, the maximum marginal allocation (MMA) algorithm is used to obtain the optimal waveform in the frequency domain. Then the sequential quadratic programming (SQP) algorithm is used to generate the continuous phase coded signal to approximate the optimal waveform by considering the constant mode constraint and orthogonality constraint with the minimum mean-square error as the criterion function. Finally, the simulation shows that the proposed method can get the time-domain waveforms for different target characteristics, as well as achieve well performance both in multi-target detection and parameter estimation performance.

Research on dialogue mode focuses on understanding the ways people engage in conversation and their learning processes, with exploratory dialogue serving as an external indicator of learning occurrences. With the continuous expansion of dialogue data in different learning scenarios, automatic recognition of dialogue modes becomes increasingly crucial. Building upon previous research on dialogue mode encoding schemes, this paper identified six categories of dialogue modes in meeting scenes. This paper collected and cleaned 11K utterances, annotated them with DM labels, and constructed a dataset called MRDM (Meeting Record Dialogue Mode). Additionally, this paper proposed the CAHDMR (Context-Attention Hierarchical Dialogue Mode Recognition) model, which integrates different levels of context into utterance representations using encoders and contextual attention layers, to generate more robust representations of dialogue. The experimental results indicate that the CAHDMR model demonstrates excellent performance on the constructed MRDM dataset.

Due to the interaction between AC and DC systems, the transient characteristics of power systems have become increasingly intricate, leading to higher safety risks. Transient simulation that yields high-dimensional and sparse data, serves as the foundational approach for investigating system dynamics. However, there is still a lack of effective methods for mining knowledge of simulation data. To address the above limitations, this paper proposes a knowledge extraction and application method for massive simulation data. Initially, a general method is proposed for applying the event-driven architecture in transient simulation. On this basis, an AC line auto-tripping mechanism is designed to simulate the development of faults. Subsequently, characteristic events (CEs) are predefined and continuously recorded throughout the simulation so that information can be quickly extracted from simulation results. After obtaining extensive sequences, the paper utilizes natural language generation techniques to process the textual sequences. An LSTM-based model has been built to learn the logic of CEs and predict the CE at the next time point. Finally, the validity of the proposed method is verified using the case of the IEEE 39-bus system.

In response to the significant errors caused by altitude measurement jumps in low altitude radar and the limited filtering accuracy caused by the lack of latest measurement information in traditional interactive multi model (IMM) particle filtering algorithms, the paper improves the interactive multi model particle filtering algorithm according to the mixed Kalman particle filtering and proposes an interactive multi model mixed Kalman particle filtering algorithm. Firstly, this algorithm uses Unscented Kalman Filter (UKF) to generate system state estimation, and then uses EKF to obtain the importance recommendation distribution of particles, fully utilizing the latest measurement information to update the particle state. The experimental results prove that the improved algorithm significantly reduces the height error of low altitude radar compared to traditional data filtering methods. This method can improve the height measurement accuracy of low altitude radar and has high engineering application value in the field of low altitude measurement.

This paper presents the principles of Space Vector Pulse Width Modulation (SVPWM), conducts a comprehensive analysis of overmodulation control strategies, and examines computational methods for various stages of overmodulation. The proposed SVPWM overmodulation algorithm is simulated and analyzed using MATLAB/Simulink, validating its effectiveness and practicality. The simulation results demonstrate that the introduction of overmodulation techniques effectively increases the inverter's maximum output voltage.

Data Link network is the foundation and key to support the efficient aggregation and sharing of various resources on the battlefield. Unmanned cluster intelligent data chain under strong confrontation environment will become the key to support unmanned system to form combat capability. At present, the existing unmanned system communication is susceptible to the Doppler effect due to the orthogonal frequency division multiplexing (OFDM) modulation technique, while the orthogonal time frequency space (OTFS) has better performance than OFDM in high mobility wireless communication scenarios. In this paper, we propose a deep neural network-based symbol detection method for OTFS system to address the problems of high complexity and poor detection performance of the symbol reception detection algorithm of OTFS for unmanned system intelligent data chain. The method uses a training set to train the constructed deep neural network, detects the received symbols in the time-delayed Doppler domain, and obtains the estimated value of the transmitted symbols. The method is able to reduce the overall symbol detection complexity of the OTFS system, increase the speed of detecting received symbols, and significantly improve the reception performance of OTFS.

Vegetables, characterized by short shelf life and easily perishable features, pose operational challenges for fresh food supermarkets in replenishing and pricing vegetable products within limited sales space. This study proposes an automated decision-making framework based on machine learning, data mining, and intelligent optimization algorithms, using a random forest model for regression prediction and an LSTM-based Seq2Seq model for time series prediction to address these challenges. The random forest model is trained using historical wholesale price data for regression prediction, with an average root mean square error (RMSE) of 0.7046 on the test set. In the time series prediction based on historical sales data, an LSTM-based Seq2Seq model is used, resulting in an average RMSE of 0.9631 on the test set. Subsequently, the random forest model is trained to predict actual sales volume based on the relationship between sales volume, wholesale price, and sales price, combining it with a multi-objective optimization model. The NSGA-II intelligent optimization algorithm is employed to solve the model. Ultimately, we determined various vegetables' sales prices and replenishment quantities by maximizing profits and minimizing supply surplus losses. The method proposed in this study helps improve supermarkets' operational efficiency and profits, providing new insights and methods for fresh food management decision-making

This article proposes a solution to the cooling problem of high-power air-cooled server racks by designing a three-layer parallel Z-type ventilated cabinet and optimizing the heat dissipation path and structure. Through thermal simulation and experimental tests, it was found that using Z-type air ducts, static pressure chambers, and optimized heat dissipation structures in a single sub-rack can reduce the maximum temperature and the maximum temperature difference of chips by about 7°C and 2°C, respectively. After the cabinet is equipped with parallel air ducts, the maximum temperature and maximum temperature difference of chips are reduced by 7-22°C and 7-10°C, respectively, and the maximum heat dissipation capacity of the cabinet is increased from 4.6kW to 6.9kW. Additionally, the noise level of the entire cabinet is also improved due to the reduction in the number of fans.

The problem of ambiguity or incompleteness often exists in customer addresses, which affects the accuracy and efficiency of address matching. This method provides a systematic solution through the steps of data collection, preprocessing, feature extraction, model training and fuzzy address verification. First, the address data of electricity customers from power companies and preprocess them to standardize and correct errors. Second, features such as street names and door numbers are extracted from the address data. Then, the known accurate addresses are trained using suitable data mining algorithms to build models to identify and speculate the possible accurate locations of the fuzzy addresses. Finally, through fuzzy address verification, a list of most likely accurate addresses or candidate addresses is given based on the similarity of address features and the analytical ability of the model. The experimental results show that the method can improve the accuracy and efficiency of address verification for electricity customers and provide more reliable address matching service for power companies.

There are complex image noises in LDCT images. Among the existing deep learning models, although there are many excellent models that can restrain the noise of LDCT images, the denoised images of these models will produce a smooth effect and some will be lost. Image details. To improve the image quality and strengthen the correlation between feature channels, inspired by the REDCNN network and CBAM module, this paper proposes an improved REDCNN network (CASA-REDCNN) based on the attention mechanism. By adding channel attention and spatial attention to the residual connection part of REDCNN, the attention module can distinguish the differences between feature channels, focus more attention on channels containing noise and artifacts, and reduce attention on other channels. force. Experiments show that this network can effectively reduce noise and artifacts in LDCT images, and can retain more image details in key areas. It also has good denoising capabilities for LDCT images with different noise levels, proving that this network has good denoising ability and improves the performance of the original model.

With the continuous reduction of traditional energy sources, coupled with our country's goal of achieving carbon peak and carbon neutrality in recent years, wind power has been increasingly emphasized in recent years, especially offshore wind power, due to the sufficient and stable wind energy, the installed capacity has risen sharply. However, the working environment of offshore wind turbines is more severe, the structure is more complex, the failure rate is higher, and their operation and maintenance costs are also higher. Therefore, this paper is aimed at the method of offshore wind power fault early warning based on vibration monitoring, the research content is as follows:

First of all, data processing is carried out to convert the collected vibration monitoring data into RMS Using the average method of adding the window to denoise the generated raw data, and then adding the window Fourier transform of the noise reduction data, the time domain signal is converted into a frequency domain signal to obtain the spectrogram. Finally, Spectrum analysis is carried out to determine the fault location according to the spectral characteristics of the spectrum and combined with the actual verification. According to the experimental results and then combined with the example analysis, a certain early warning strategy, used for the future parts of the turbine may fail to early warning, reducing the cost of operation and maintenance.

The offshore wind turbine fault diagnosis method proposed in this paper can effectively diagnose a variety of faults in various parts of offshore wind turbines, and make full use of big data to improve diagnostic accuracy.

The paper develops a platform for calculating the distribution of underwater corrosion electrostatic fields in submarines to address the issue of time-consuming large-scale and structurally complex modeling in submarine underwater corrosion electrostatic field distribution calculations. The development requirements of the simulation calculation platform are analyzed by analyzing the submarine underwater electrostatic field modelling process and characteristics. The complex physical modeling, mathematical modeling and grid dissection process is encapsulated, and the influence of different sea areas and different working conditions on the submarine underwater corrosion electric field is distinguished through the setting of polarization curves and electric field calculation module, so as to calculate the submarine corrosion electric field under different working condition.

The study of matching power consumption attributes of station users plays an important role in smart grids, providing an important basis for tasks such as power dispatch and energy optimisation. Traditional classification tasks usually use hand-designed feature extractors to classify the users in the station area. To solve the problem that its performance is limited by the feature expression, the daily power data of the users are transformed into the form of two-dimensional images by introducing the Gramian Angular Field (GAF) in order to capture their time-varying characteristics, and the image features are learnt by combining with the generalized self-attention mechanism, and the network model based on Transformer is used for the analysis of the generated feature images to achieve a more accurate classification. The generated feature images are analysed to achieve more accurate user category classification. The simulation experiments show that the provided model and method have a good effect of station classification.

Typhoon is one of the major threats to the safe and stable operation of power systems among all kinds of natural disasters. This paper proposes a prediction model for the damage of transmission towers under typhoon disasters based on the Random Forest (RF) algorithm with 1km*1km refined spatial grids. The model is applied to predict the data related to the damage of transmission towers caused by Typhoon "Meihua" hitting Zhoushan City, Zhejiang Province, China. The results show that compared with models based on four other commonly used algorithms, the model based on the random forest algorithm is the most scientific, effective, and accurate.

This paper first selects some features with classification effect according to the flight paths of drones and birds. In terms of echo intensity and movement, it selects features such as zero crossing number and dispersion coefficient, and then uses each feature for separate classification to compare which feature has better classification effect. Then, these features are combined for processing, and CNN network is used for classification, and it is verified that these features have a good classification effect. Finally, SVM, KNN and CNN-T are also used to identify the UAV, and the results of the four classification methods are compared. When CNN-T model is used, the recognition rate of the UAV is the best, reaching 0.91.

To address the high rate of missed detections and low accuracy in traditional shipboard fire alarm systems, this paper proposes a ship compartment fire alarm system based on the fusion of multi-sensor data. The hardware components of this system include environmental data modules, wireless communication modules, display modules, and alarm system application software, enabling the acquisition of compartment environmental data, real-time display of parameters, and backdraft warning functions. In terms of algorithm, a combination of BP neural network and fuzzy reasoning is adopted to enhance system performance and address the issue of high false alarm rates in traditional ship fire alarm systems. Experimental results demonstrate that this algorithm can accurately identify the actual status of fires, predict fire scales and development trends, and exhibit high reliability, fully meeting the expected requirements of the system.

Embedded systems and web application development are two crucial directions in the rapidly developing computing field. The Python language, known for its simplicity, readability, and flexibility, has a wide range of applications in both fields. This thesis compares Django and Flask, two major frameworks in Python web development, and evaluates their applicability in different scenarios through systematic experiments on Shannon Orange Pie. The study analyzes the differences in response time, development speed, and flexibility. The purpose of this text is to assist developers in comprehending the appropriate scenarios for various frameworks. It aims to provide practical references for developers when selecting the most suitable framework for their projects.

Medical image segmentation, as an integral component of medical image processing, plays a pivotal role in clinical diagnosis, surgical planning, and treatment program development. Its accuracy directly impacts diagnostic outcomes, enabling precise identification and extraction of anatomical structures or lesion regions from existing medical images. This process assists physicians in quantitative analysis, facilitating enhanced understanding of patient anatomy and pathology for accurate diagnosis and treatment planning. PCNN (Pulse-coupled neural networks), leveraging neuronal activity characteristics, hold significant promise for application in this domain. Hence, this paper proposes an image segmentation method based on VPFC-MSPCNN (Vision-Perceptual FCMSPCNN), derived from the FC-MSPCNN (Fire-controlled MSPCNN) model. This approach simplifies relevant parameters, reduces computational complexity, and effectively controls neuron activity, resulting in higher-quality detection images. Experimental results demonstrate its superior image processing capabilities, establishing it as an effective algorithm for medical image segmentation.

This paper introduces a novel approach for predicting momentum shifts in time-series data by integrating an enhanced Cumulative Sum (CUSUM) algorithm with a Light Gradient Boosting Machine (LightGBM) model, augmented with SHAP values. Our methodology refines CUSUM through a composite sliding window to improve detection accuracy and utilizes LightGBM for efficient prediction, with SHAP values providing interpretability of model predictions. Tested on a dataset of 1701 tennis matches from the 2023 Wimbledon Championships, our method demonstrated superior performance, achieving near-perfect prediction accuracy with an R2 score of 0.999, and significantly lower prediction errors (MAPE and RMSE) than traditional models. These results underline the method's effectiveness in not only accurately predicting momentum shifts but also offering detailed insights into the driving factors behind these changes. This advancement opens new avenues for time-series analysis across various domains, extending well beyond the initial sports analytics application.

For the monolithic architecture in the power information system, the amount of information processed is more limited, which leads to the inability to comprehensively monitor the situation of grid link information. Therefore, this paper proposes a dynamic collection method of microservice invocation link data based on high-performance agent technology. First, the agent device is deployed on the key nodes of the microservice invocation links that need to be monitored. Then, the agent network is configured to establish communication between the devices. At each key node, the agent device captures information about the microservice invocation, and at a central node of the network (e.g., a gateway), all the packets are aggregated and transmitted to a server. On the server, the collected data can be processed and analyzed. Based on the analyzed results, the strategy of microservice invocation can be dynamically adjusted. The experimental results show that the method can accurately monitor the number of abnormal packets exactly matching the number of abnormal packets injected manually, with an accuracy rate of 99.8%, which realizes the real-time monitoring of the virtual network in cloud environment On an end-to-end path.

As a heterogeneous computing architecture, FPGA accelerators are widely used for AI and other fields. Compared with traditional CPU clusters, the architecture of the virtual FPGA cluster is relatively simple, and thus lacks the design of relevant mechanisms for data backup. In our work, a data backup optimization mechanism for the virtual FPGA cluster is proposed to select local sub-kernels and remote sub-kernels of other FPGA accelerators as data backups. During the selection of remote sub-kernel backups, the data transmission delay between the sub-kernel under evaluation and the subkernel to be backed up, the statistical probability of the historical errors of the sub-kernel under evaluation, and the internal environment influence factors in its FPGA accelerator are fully considered. The simulation results demonstrate that the proposed data backup optimization mechanism has shown good data disaster recovery performance.

Diffusion models have emerged as a leading method for generating diverse and realistic samples, revolutionizing tasks like text-to-image generation, super-resolution, and inpainting. However, their slow sampling process hampers implementation. This work focuses on post-training quantization (PTQ) optimization for diffusion models. We propose an error-contribution-aware calibration dataset sampling scheme, effectively reducing calibration dataset size while preserving model precision. Experimental results demonstrate the effectiveness of our approach, achieving performance comparable to uniform sampling methods with a significant reduction in calibration dataset size.

Due to the heavy haul rail's large single train capacity and strong impact on the track, the rails of heavy haul rail are prone to injuries, and the accurate detection and identification of injuries is the basis of intelligent operation and maintenance of the rails of heavy haul rail. However, the existing rail surface injury recognition methods mainly based on machine vision have the disadvantages of strong noise and limited field of view. Therefore, this paper proposes a detection method based on the fusion of profile data and vibration signals. The profile data has high accuracy, but it has the disadvantages of low detection rate and low accuracy of a single profile. The vibration signal is easy to measure, but it is easy to be interfered, and the vibration signals of some injuries will reflect the same characteristics. The respective defects are overcome by fusion of the two. The experimental results show that the detection accuracy is effectively improved after the fusion of the two. The vibration signals are extracted and dimensionality reduced using stacked self-coding network, and then input into the classifier to obtain the classification results based on the vibration data. At the same time, the profile data and standard rails are matched, and then the wear sequence is generated by the DTW algorithm, and then the features are extracted by the PCA algorithm and classified to obtain the results based on the profile data. Finally, the two methods are fused at the decision level to obtain a new rail surface injury recognition and detection method based on the fusion of the two.

The stochastic and fluctuating nature of wind energy limits its large-scale grid connection, and the precise prediction of wind speed is crucial for advancing the wind energy sector. A novel wind speed combination prediction model utilizing quadratic decomposition is introduced. Firstly, the wind speed sequence is decomposed into various frequency components through adaptive noise-complete ensemble empirical modal decomposition (CEEMDAN) and then reconstructed according to the sample entropy. Subsequently, the high-frequency elements obtained from the reconstruction is decomposed secondly using the variational modal decomposition (VMD), and then, the predictions are carried out for the obtained components. Finally, predictions are made for each of these components, and a final forecast is compiled by combining these individual predictions. Comparative experimental analysis shows that the proposed model outperforms others in accuracy, with lower root mean square error (0.368), mean absolute error (0.238), and mean absolute percentage error (5.899%), and the value of the coefficient of determination is 0.984, which is close to 1, which indicates a strong model fit.

In order to solve problems of output chattering, difficulty in tuning and poor noise suppression ability of tracking differentiators (TDs), a TD based on the Gaussian error function (ERFTD) is designed. The ERFTD has a simple form and applies the Gaussian error function to construct the acceleration function, and introduces a terminal attractor function. The application of Gaussian error function reduces chattering caused by high-frequency signals of TD, and the terminal attractor function enhances noise suppression capability of the TD. The stability of ERFTD is proved by theoretical derivation. The parameter tuning rules were summarized through frequency sweep testing. Numerical simulations show that ERFTD has good tracking capability for signals contaminated by noise, with fast response speed and high tracking accuracy; The differential ability of ERFTD is strong, which effectively suppresses the noise. The tracking differentiator based on the Gaussian error function has excellent performance.

The development of modern heavy-duty gas turbine technology has led to a gradual increase in combustion temperature. This increase has resulted in a higher incidence of gas turbine ablation failures. Ablation produces a large number of high-temperature particles that flow downstream with the gas and presents a safety hazard to gas turbine components and the stable operation of the gas turbine. This paper investigates the deposition behaviour of ablative particles of three different sizes on gas turbine blades through numerical simulation. The results indicate that the deposition rate is highest for 10μm particles on the stator blade pressure surface. The overall deposition rate on the dynamic blade pressure surface decreases, and the highest deposition rate is observed for 2.5μm particles.

Suicide has become one of the severe global social issues. This study focuses on users' social media text data and adopts natural language processing techniques to conduct research on suicide risk identification, proposing a suicide risk recognition model based on dynamic segmentation and Transformer architecture. Users' social media posts over a period can often reach chapter length, but traditional Transformer-based models have input limitations concerning text length. To address this, the paper introduces a dynamic segmentation algorithm that minimizes cost to split all user posts within a month, sequentially inputs them into the RoBERTa model for feature extraction, and then utilizes a bidirectional Long Short-Term Memory network to extract features between segments that include contextual associations. The model also embeds its own attention mechanism to capture important local feature information. Eventually, it achieves an overall representation of the user's posting content, thus enabling effective identification of users' suicide risks. Compared to baseline models, there is a significant improvement in both F1 score and accuracy, reaching 73.1% and 74.67%, respectively.

Currently, image detection based on deep learning has a good application prospect for railroad fault diagnosis. However, the existing methods of applying deep learning to fastener fault detection more or less have some drawbacks, such as computational complexity, slow detection speed, low detection accuracy, poor robustness, etc. To address these issues, this paper introduces a fastener fault identification method for railroad lines based on CenterNet. By introducing the Convolutional Block Attention Module and the jump connection mechanism, and conducting comparative experiments on different networks, it is proved that the network proposed in this paper has a better speed and accuracy for fastener detection.

Aiming at the problem of low target detection accuracy in unmanned aerial vehicle (UAV) inspection, this paper proposes a target detection algorithm based on improved YOLOv5s for transmission line routes and trees. Firstly, a squeeze-excitation (SE) module is introduced into the backbone network of YOLOv5s, which selectively emphasizes feature information using global information, suppresses non critical features, and recalibrates features. Secondly, the traditional pyramid feature fusion structure of YOLOv5s is replaced with a bifrustum feature fusion (BFF) structure to enhance the extraction ability of validity information and reduce the influence of interference information on recognition accuracy. By constructing an inspection dataset for experimental verification, the results showed that the improved YOLOv5s achieved a MAP of 91.26%, and the detection accuracy of lines and trees reached 94.5% and 89.37%, respectively. The improved model improves detection accuracy while balancing speed, demonstrating the effectiveness of the algorithm.

The prompt and accurate detection of bone fractures is vital in patient care, especially in facilities with few experienced diagnosticians. Addressing this need and utilizing Depthwise convolution (DWConv), this study presents DW-YOLO, a novel adaptation of the YOLO object detection model, fortified with depthwise convolutional layers to enhance fracture identification in medical imaging. Leveraging a diverse dataset of 4,148 cranial images categorized into seven fracture types, DW-YOLO underwent comprehensive training, validation, and testing. The model's architecture enables rapid image processing while maintaining high accuracy, as evidenced by achieving a mAP of 0.889 in the test dataset. In addition, the model size is only 7.5MB. The study confirms DW-YOLO's proficiency in identifying fractures across diverse bone structures, emphasizing its potential to enhance clinical diagnostics.

Increase Accuracy of Transformer Ground Temperature Prediction, this paper proposes a new approach to forecast transformer earth temperature using Variable Mode Decay (VMD), two-way gated recurrence (BGRU) and multi-attention mechanism (MHA). In the VMD-BiGRU-MHA model, firstly, an environmental monitoring device is built to obtain the weather data near the transformer and eliminate the weakly correlated parameters, and then the raw data are decomposed into several sub-sequences using VMD to get the more stable individual components, which are then inputted into BiGRU for training, and then finally, MHA is introduced to mine the features of the time-series long-range data of the transformer surface temperature. Thus, prediction precision is increased. Comparison between BiGRU and VMD-BiGRU indicates that the predicted error is smaller than that of VMD-BiGRU.

The new power system is characterized by a substantial volume of measurement data and a sophisticated storage and interaction ecosystem. There are major security risks in the process of cross-system data flow. Given the above problems, this paper constructs a safety risk evaluation model and proposes a fuzzy reasoning algorithm based on an improved fuzzy Petri network (FPN) to realize comprehensive risk monitoring on the storage side and circulation link of measurement data. The experimental findings demonstrate that the risk evaluation approach based on FPN gives more accurate and scientific results than the traditional comprehensive risk evaluation method. Therefore, this method is more suitable for practical measurement data security risk evaluation.

The data for load forecasting is characterized by randomness and burstiness, and it is often required to have an accurate and fast load forecasting response in integrated energy and power market trading. Therefore, a solution is proposed for this scenario, which is based on an edge computing approach using an algorithmic model combining variable modal decomposition (VMD) and long and short-term memory network (LSTM). Because of the use of VMD, it is necessary to train multiple LSTM networks and complete the forward inference as simultaneously as possible, and the parallelization feature of Field Programmable Gate Array (FPGA) is considered for hardware acceleration design. First, edge computing has low latency, fast response, etc., and the data is decomposed into multiple intrinsic modal functions using variable modal decomposition so that the LSTM models can be trained separately to better handle nonlinear and nonstationary data. This is used to achieve computational acceleration and hardware resource consumption optimization. The results of the comparative analysis show that the edge-based VMD+LSTM model hardware architecture exhibits excellent performance in terms of computational speed for load prediction.

Electricity load data is typically analyzed on a temporal scale, which can vary with seasons or different times of the day. When examined on a spatial scale, electricity load data display distinct regional characteristics. Therefore, ensuring accurate data correction under time-series conditions is paramount. In this paper, we present a sequential correction method for load resource data based on minimum enclosing ball clustering. This method is structured into four phases. In the initial phase, data feature extraction is conducted using load similarity based on a dynamic time planning algorithm to identify the data for processing. Subsequently, the Singular Value Decomposition (SVD) dimensionality reduction algorithm is applied in the second phase to decrease the dimensionality of the data for the subsequent steps. The third phase involves utilizing the minimum enclosing ball clustering technique to cluster and analyze the data obtained from the previous step. Finally, in the fourth phase, the uncorrected values are adjusted.

Current approaches to document-level event extraction face challenges such as scattered argument parameters across sentences and multiple event records within a single document. In this study, we propose a solution that combines context paragraph information modeling with argument extraction to tackle these issues. Firstly, we employ advanced techniques for segmenting lengthy documents into smaller parts. Additionally, we utilize pre-trained models to obtain paragraph vector matrices for each segment. By integrating these models, annotated sentence vector matrices and paragraph matrices are combined through a dot product fusion model and concatenated together to generate improved sentence vector embeddings using pooling operations. Next, the entities extracted through named entity recognition are used to select pseudo-trigger words and construct adjacency similarity matrices and composite graphs. Finally, the task of populating event table elements is transformed into a multi-label classification task to address the issue of overlapping roles. Experiments on the chFinAnn dataset demonstrate an overall F1 improvement of 1.5 percentage points, with a significant increase in the model's inference speed.

In order to ensure that the dry quick coupling device in the LNG pipeline transportation system can quickly establish a transportation channel during LNG transportation, this study conducted an in-depth analysis of its structural strength. We established ten different working conditions and determined the three most dangerous working conditions with the maximum load force and the lowest allowable stress through linear superposition. Subsequently, we conducted a detailed analysis of whether the dry quick coupling device can meet the structural safety requirements under the most dangerous working conditions. The results of finite element analysis show that under the most dangerous working conditions, the load force and allowable stress of the device are within the design range, confirming the rationality of the structural design. However, we observed that the maximum stress in the entire device occurs at the contact positions between the guiding wheel and the guiding groove, as well as between the valve core and the spiral groove of shell 2. Therefore, we recommend strengthening the strength of these components to improve the overall structural safety.

To address the limitations of traditional Droop Control (DC) methods stemming from differences in transmission line impedance within parallel inverter systems, which often result in inaccurate reactive power distribution, we propose an enhanced DC strategy incorporating adaptive virtual impedance. Firstly, the mathematical model of inverter parallel system is established to analyses the reason why the reactive power of traditional DC cannot be equally divided; Secondly, adaptive virtual impedance is integrated into the control loop to streamline the complex decoupling control algorithm. The value of the virtual impedance is adjusted based on the feedback of active power, thereby enhancing the accuracy of power allocation and achieving equitable power distribution. Finally, a simulation model of the inverter parallel system is constructed using MATLAB/Simulink. The simulation results validate the effectiveness of the proposed method.

This paper introduces the pulsating high-frequency signal injection (HFI) method, which does not require a position sensor system. By carefully selecting the HFI, the estimation error of the rotor position can be improved, leading to enhanced accuracy in rotor position detection. The principle of the pulsating HFI method is explained, followed by a theoretical analysis of how the amplitude and frequency of the HFI affect the position estimation error. Finally, a simulation model is constructed to verify the findings.

In electronic information systems, sine waves are widely used as test signals, reference signals, and carrier signals. The most direct method for generating sine signals is to use an operational amplifier with a small number of external components to form a sine wave oscillation circuit. This article uses Multisim virtual simulation software to design and simulate the RC sine wave oscillation circuit. Dynamically reproduce the working process of the oscillation circuit, achieving the process of circuit oscillation and balance. Then, the circuit is improved to further achieve the output of square and triangular waves. Not only can it visually observe the output waveform, but it can also appreciate the influence of circuit parameters on the waveform, further verifying theory through practice, and better achieving the combination of theory and practice.

The prediction of wind power is influenced by various factors, such as climate conditions, geographical location, and the status of adjacent wind turbine generators, which exhibit complex dependencies. In addition, the data distribution of wind power generation may also shift over time. Prediction models trained on previous data may not have good prediction performance on new data. Therefore, a Graph Convolutional Neural Network (GCN) Feature Fusion and Incremental Self-Attention based Transformer (GCN-IncreFormer) is proposed in this paper for wind power continual prediction. The GCN based feature fusion module effectively integrates and refines key features from different channels while retaining temporal information, thereby enhancing prediction accuracy. Additionally, we introduce incremental learning into the self-attention mechanism of the Transformer and propose IncreFormer. Consequently, model can adapt to continuously changing data distributions by continuously updating with newly collected data to accommodate dynamic environmental changes. Experimental results demonstrate that GCN-IncreFormer can effectively explore the deep correlations of different features in wind power data, thereby improving power prediction accuracy. The introduction of continual learning mechanism enables the model to adapt to data distributions shifts over different time periods. Compared to baselines, GCN-Increformer achieves a reduction in Mean Absolute Error (MAE) of over 2.40. Comprehensive comparative experiments and ablation studies validate the reliability and practicality of the proposed method.

In a variable frequency drive system without electrolytic capacitor, the bus voltage fluctuates periodically due to the decrease of the bus capacitance. The current beat frequency phenomenon generally occurs when the frequency of the output voltage of the inverter is similar to the fluctuation frequency of the bus voltage. A multi-rotation PI harmonic suppression method is proposed to suppress the current beat frequency of non-electrolytic capacitor variable frequency driver. This method takes dq coordinate system as reference, uses the characteristics of fundamental current in dq coordinate system, builds a current closed loop under rotating coordinate system respectively, extracts the main harmonic current, and represses harmonic current through multi-rotating PI method, so that the harmonic in winding current can be suppressed, thus suppressing the motor beat frequency phenomenon. Through simulation and theoretical deduction, this method can obtain better beat suppression effect in non-electrolytic capacitor variable frequency driver, and the control structure is simpler, the parameter tuning and engineering implementation are easier.

Adaptive dynamic programming (ADP) technique is utilized in this paper to obtain optimal coordinated control for low-voltage distribution system with distributed photovoltaic and battery energy storage. The control algorithm has been packaged into an embedded APP and inserted into the smart distribution transformer combine terminal unit (SCU). Moreover, compared to the former studies, the algorithm is firstly applied in a low power distribution area, the effectiveness of the algorithm is verified. Through the training of the algorithm, the optimal energy allocation strategy can be obtained in power distribution area, which provides a new idea for the intelligent application of power system.

In actual industrial environments, noise interference poses a challenge for fault diagnosis models to effectively extract precise fault features from vibration signals. Accordingly, this paper proposes a fault diagnosis method based on Multiscale Residual Shrinkage Network (MsRSN). Firstly, a wide convolution kernel is used in the first layer to extract the short-time features in the signal and initially suppress the high-frequency noise, and then three branches of Residual Shrinkage Network are designed to extract the multi-scale features, and adaptive feature fusion is carried out on the extracted multi-scale features by Efficient Channel Attention. The soft thresholding function is introduced into residual module and the attention mechanism is constructed for adaptive noise reduction. The standard convolution with 256 channels in the residual module is replaced with Depthwise separable convolution, which reduces the number of parameters and speeds up training. Next, Meta-ACON activation function is added to the adaptively fused features to adaptively activate the model output of the neurons to enhance the generalisation performance of the model, finally the classification is performed by softmax. Experiments show that the multiscale residual contraction network can still accurately identify fault categories under strong noise interference and has good generalisation ability.

In today's emerging new energy electric vehicles, silicon-based MOSFETs are an indispensable presence. Due to its extremely short switching time and very fast operating frequency, it generates large oscillations in the parasitic inductance and capacitance, which results in large transient voltages at the moment of turn-off. Voltage spikes can have a number of adverse effects on the system. It is important to suppress overvoltage or switching stress on power semiconductors connected in series and parallel. In this study, a spike voltage absorption circuit based on active embedding is proposed to reduce the spike voltage generated at the moment of mos turn-off in a step-down isolated automotive DC-DC power circuit with large spike voltage at the low voltage side. In this study, a classical doublepulse test topology, a circuit topology based on the active embedded absorption circuit, and the main power topology of an isolated automotive DC-DC are constructed based on LTSPICE, and the theoretical feasibility is analyzed, and the practical application of the circuit is also verified by simulation, and the simulation results show that the absorption circuit can inhibit the spike voltage of the mos very well.

Reactive voltage optimization is often based on optimal power flow calculation to achieve safe and economical operation of the power system. When the system parameters change, the optimal power flow calculation involves more parameters and is time-consuming. In the distribution network, low-voltage photovoltaic power stations with distributed access can only rely on changing the reactive power compensation capacity for voltage control. The reactive power compensation amount obtained from the voltage regulation coefficient acts quickly, but it can cause the problem of multiple actions of the reactive power compensation device. This article is based on automatic identification of system reactance parameters. When the system parameters change, the system reactance can be automatically identified to reduce the impact of system operation on reactive power compensation. An improved adjustment method for reactive power compensation is constructed, which can ensure voltage compliance and reduce the number of actions of reactive power compensation devices.

VPP is an important form of energy organization in the new power system. This paper presents a medium/long-term virtual power plant aggregation method, which forms a VPP through capacity leasing. A VPP risk decision-making model is provided to select suitable distributed energy resources to form the VPP. The calculation based on actual data shows that: 1) Although VPPs with medium to long-term alliances can operate normally under existing conditions, there are significant risks involved; 2) The main tools for controlling risks are: selecting appropriate distributed resource combinations, utilizing forward contracts, and participating in frequency regulation; 3) The risk decision-making tool for distributed energy resource alliance provided in this paper can provide effective decision support.

This article discusses the optimization evaluation of dual active bridge (DAB) converter under current stress optimization and minimum backflow power scheme. A large-signal model with the transferred power as the control quantity is built, and a closed-loop control method based on direct power control is proposed to realize the precise modulation of the output voltage. Based on this, the current stress optimization scheme and minimum backflow power scheme are designed by modulating the inner phase shift ratio and outer phase shift ratio. The computational results show that current stress minimization performs better compared to the minimum backflow power scheme.

As the field of power electronics continues to evolve, new energy generation technologies are gaining increasing attention. This has made the study of inverter control strategies more crucial than ever. Traditional control strategies for inverters operating in parallel have encountered limitations due to the need for output power calculation, which hampers dynamic performance. This paper introduces an inverter control strategy based on improved virtual oscillator control that enables autonomous parallel operation of inverters. Compared to conventional inverter control methods, this strategy offers rapid dynamic response and strong disturbance rejection capabilities. Through parallel simulation of improved virtual oscillator control inverter, the proposed control strategy has been demonstrated to provide excellent dynamic performance and reliability under parallel operating conditions.

In response to the problem of abnormal detection data and inaccurate detection data in temperature and pressure changing environments using traditional electrochemical principle oxygen concentration detection sensors, an oxygen concentration demodulation technology based on multiple information perception of atmospheric pressure, temperature, and oxygen concentration is designed, and a long-life mining oxygen concentration sensor with a wide temperature and pressure range is developed. This sensor adopts a distributed microprocessor data acquisition architecture and atmospheric pressure mutation suppression technology based on pressure change trend prediction, achieving real-time collection, compensation, and data fusion demodulation of measurement data, achieving higher environmental adaptability and high-precision measurement of oxygen concentration. The atmospheric pressure range of the sensor is more than twice that of ordinary electrochemical sensors, and the working life is doubled.

Electromagnetic testing is the most widely used technology for wire rope inspection. However, with the existing technology, there are still some problems in practical applications; one of the urgent needs to solve the problem is how to suppress the relative fluctuations of the wire rope and detector. To solve this problem, a wire rope damage detection device (WRDDD) based on hybrid maglev is developed. First, the control force and key structural parameters of the WRDDD based on hybrid maglev are analyzed using theoretical formulas. Then, the effect of current on the excitation magnetic field of the wire rope is analyzed using the finite element method. Finally, the correctness of the WRDDD is verified by comparing the difference between finite elements and theoretical calculations in the current-force relationship, which shows extraordinary advantages for the inspection of wire rope.

Fusion diagnosis of medium and low voltage distribution switchgear (MLVDS) is an important technology for identification the power equipment operating status by infrared thermal image analysis, combined the partial discharge method for complex faults. The objective of this research is to develop fusion diagnosis and techniques for detect the state of 10kV and 400V distribution equipment in the switch-gear and its bus, identifying its abnormal conditions, by the infrared camera method and UHF partial discharge detection. The condition data of the equipment from these sensors is collected and analyzed to from the operating conditions by intelligent monitoring data technology for power distribution equipment. Under the background of actual failure case, overheat defects and discharge defects were captured and deeply analysis by infrared method. The fusion recognition by comprehensive detect for a MLVDS contactors power line and terminal board were verified in power distribution site. This paper provides a fault monitoring and diagnosis instance for MLVDS to detect overheat detects and partial discharge defects.

A low-temperature drifting bandgap reference circuit with high-order temperature compensation was designed to solve the problem of poor temperature characteristics of traditional Brokaw bandgap reference. In this design, a high-order temperature compensation was formed by adding a base current of the transistor, which reduces the temperature coefficient of the bandgap reference. The loop negative feedback method was used to replace the clamping of the op amp to the circuit, which simplifies the circuit and reduces the impact of operation amplifier misalignment. The design of the low-temperature drifting bandgap reference was completed based on the Dongbu 0.18 um BCD process, and the circuit was simulated by Candence spectre, and the simulation results showed that the temperature coefficient of the bandgap reference was 1.918×10-6/°C in the temperature range of -55°C~125°C. In the low frequency range of 0~10kHz, the power supply rejection ratio reaches 87.3dB. At 1Mhz, the power supply rejection ratio reaches 62.2dB.

In response to the challenges of large data volumes and insufficient accuracy of traditional prediction methods in defect size detection of ferromagnetic materials, this study proposes an optimized residual network model for precise prediction of defect sizes in ferromagnetic materials. The model enhances the ResNet18 network by integrating a channel attention mechanism, utilizing adaptive average pooling and fully connected layers to determine channel weights, and applying the Sigmoid activation function to process the input feature maps. Furthermore, by multiplying the obtained weights with the input feature maps, the model dynamically weights the importance of features along the channel dimension. In the experiments, one-dimensional signals were converted into two-dimensional grayscale images, and the features of these images were extracted to eliminate biases in manual feature selection. The results demonstrate that the optimized network achieved prediction accuracies of 95.65%, 100.00%, 97.01%, 78.07%, and 93.22% for defect sizes with a length of 50mm, depth of 2mm, and varying widths of 1mm to 5mm, based on 12,500 grayscale images. Similarly, for defects measuring 50mm in length, 2mm in width, and varying depths of 1mm to 5mm, the network attained prediction accuracies of 98.50%, 99.22%, 98.31%, 100.00%, and 99.22%, respectively, on a set of 12,500 grayscale images. Compared to the original ResNet18 model, the optimized model in this study exhibits significant improvements in prediction accuracy.

This paper proposes a generalized state-space average model (GSSAM) of the VRM power stage based on a multi-phase interleaved BUCK converter (IBC). By analyzing the operating modes of the two-phase IBC, and considering the parasitic parameters of the components, the GSSSAM of the N-phase IBC, which is not affected by the duty cycle of the control signal, is constructed and proposed. Meanwhile, ageneral expression for the power stage transfer function of N-phase IBC is derived by small signal analysis. Finally, the correctness of the theoretical model is verified through the comparison between the proposed model and equivalent inductance model.

To achieve a more precise assessment of the comprehensive advantages of electric energy substitution technology, the article begins by conducting a thorough analysis of the various factors that influence it, ranging from a macro-level perspective to a micro-level perspective. It then proceeds to comprehensively summarize and analyze the interactions among these influencing factors. Subsequently, the article identifies the key indicators that play a decisive role in evaluating the benefits of electric energy substitution and develops a three-level evaluation network architecture model. This model combines both subjective and objective weights and employs the cloud model and entropy weight method to assign weights to the three-level indicators. This approach effectively addresses the issue of traditional assignment methods disregarding the subjective aspect. Finally, the article employs the fuzzy comprehensive evaluation method to assess the environmental benefits of electric energy substitution technology. This methodology can serve as a valuable reference for devising well-founded plans for electric energy substitution.

Gas turbine blades work in the harsh conditions of high temperature, high pressure and rapid rotation. The impact of the surrounding environment will bring a lot of interference to the process of radiation thermometry, which will produce reflection on the surface of the measured blades. Aiming at the problem of temperature assessment within the operational setting of turbine blades, it is analyzed that the radiation of the surrounding environment will change with the position of the target blade, which brings challenges to the calculation of reflected radiation. In this paper, we examine the reflected radiation errors at different locations and at different temperatures by building a three-dimensional blade model and calculating the angular coefficients between the blades.

The coreless axial flux motor has the advantages of small torque ripple and high power density. It has wide application value in the fields of high torque, high overload, limited installation size and so on. The axial flux motor with double rotor and single stator is widely used at present, and its power density is higher than that of other axial motors with topology structure, so it has high research and application value. The fixing of stator coil is a difficult problem of single-stator dual-rotor axial flux motor without iron core. The traditional methods have the disadvantages of using epoxy resin to bond the stator winding to the shell or using support plate to support the winding, which occupy the axial air gap of the motor and are easy to produce vibration. This paper presents a new stator fixed structure, which has the advantages of no air gap, high strength and low fabrication difficulty.

Variable reluctance resolver is an important device for obtaining rotor position information in a brushless motor system. And high measurement accuracy is its optimization goal. In order to avoid the disadvantages of traditional optimization methods of variable reluctance resolver, this paper uses Taguchi method for multivariable simultaneous optimization, shortening the design cycle for resolver optimization while ensuring a high level of accuracy. Providing a reference for the design and manufacture of variable reluctance resolver.

Thermal resistance is critical to the life cycle and performance of LED devices. The case thermal resistance of LEDs is commonly used to assess heat dissipation capability, therefore, accurate measurement of the case thermal resistance of LED devices is important to assess heat dissipation performance. The methods used to measure the thermal resistance parameters of semiconductor devices include the transient double interface method and the thermocouple method. When the chip bonding layer of the LED to be measured is a high thermal conductivity material, the results of the transient dual interface method are very close to the standard case thermal resistance of the device itself. However, when the bonding layer of the LED chip is made of materials with low thermal conductivity, the thermal impedance curves obtained from two measurements by the transient dual interface method cannot separate the obvious differences, and the separation point is not easy to be recognized, and the measurement results deviate from the real value. In view of the above problems, this chapter analyzes the integral structure function of the transient thermal impedance curve on the basis of the transient dual-interface method, and thus researches a kind of improved crust thermal resistance measurement method based on the integral difference method, and proves that the method improves the measurement accuracy by using the experimental results.

During high-speed flight, the friction between the aircraft surface and space particles can accumulate static charges, resulting in electrostatic discharge, which poses a great threat to the safe flight of the aircraft. In this paper, the friction between the aircraft surface and the ice crystal particles is equivalent to the friction between the metal and the medium, and the charge distribution on the aircraft surface is regarded as the exponential distribution. Based on the solid-state electrification theory, this paper creates an advanced friction charging current model and infers the charging current on the aircraft surface during electrostatic balance. The frictional electrification current between ice crystals and aircraft surface metals is related to factors such as aircraft speed, altitude, and ice crystal concentration. This paper mainly studies the characteristics of aircraft friction charging current under the condition that the flight height of the aircraft is located in the range of 15 km ~ 30 km in the near space and the flight speed range is 1 Mach ~ 7 Mach.

Smart power station are poised to become the mainstream of future power grids. High-voltage equipment is the basic unit of substations, and its intelligent transformation is a crucial component of substation intelligence. This paper begins by analyzing the characteristics and composition of smart power station of high-voltage equipments, and then explores the architecture of smart power station networks and methods for high-voltage equipment intelligence. A design scheme is proposed for high-voltage equipment intelligence and establishes intelligent implementation methods for transformers, circuit breakers, lightning arresters, and current transformers. The research effectively promotes the development of high-voltage equipment intelligence technology and provides a theoretical basis for substation operation and maintenance personnel in their actual work.

The AC to DC transmission project uses the original AC transmission tower to carry out DC transmission, and realizes the significant improvement of transmission power without large-scale increase in investment, which has become an important solution to realize energy efficient interconnection. The use of AC tower for DC transmission puts forward different requirements for the combination of overvoltage and insulation. Based on Yangzhou-Zhenjiang ±200kV AC to DC transmission project, the generation mechanism of DC line overvoltage is analyzed, and the level and distribution characteristics of DC line overvoltage are calculated based on PSCAD electromagnetic transient simulation platform. This paper provides a reference for the research on the overvoltage of DC transmission lines in AC to DC transmission project, and the research results have provided a basis for the design and construction of the first AC to DC Yangzhen HVDC transmission project in China.

In response to the problems of cable slippage, unsmooth winding, and uneven tension caused by the weak self-made force of the current cable drum self-tracking reversible bin belt conveyor, this paper studies a tension following based permanent magnet variable frequency cable reel control method, namely a fuzzy closed-loop and PID selection control system. It conducts Simulink simulation and industrial cable winding and unwinding tests. The test results show that it can solve the above problems and significantly improve its information and intelligence level.

In order to improve the accuracy and reliability of electricity meter measurement, ensure the stable operation of the power system and the economic interests of users, a smart electricity meter measurement error correction method based on adaptive weight FAHP is proposed. Construct a measurement error correction index system for smart energy meters, use comparative analysis methods to quantify and process fuzzy comprehensive evaluation indicators, determine the weight of evaluation indicators based on AHP algorithm, apply membership function method to calculate the membership degree of evaluation indicators, obtain a set of fuzzy comprehensive evaluation indicator membership degrees, and calculate the measurement error correction results of smart energy meters through fuzzy comprehensive method. The experimental results show that this method maintains high accuracy while reducing calibration time, providing strong support for the stable operation of the power system and the economic benefits of users.

The problem of estimation accuracy of lithium-ion battery state of charge (SOC) for electric vehicles is addressed. An equivalent network of the second-order RC fractional-order model is established and parameter identification is performed using the adaptive genetic algorithm identification parameter (AGA) method. Based on this model, adaptive Schmidt orthogonalization of untraceable particle filtering (ASOUPKF) is applied to estimate the SOC. The accuracy of the estimation of SOC is improved by four aspects, such as improving the model, automatic matching of noise information covariance, reducing the problems arising from the effects of battery aging, and reducing the number of sampling points, thus realizing the accurate joint estimation of SOC and SOH. Under the New European Driving Cycle (NEDC) conditions and the Urban Driveway Cycle (UDDS) conditions, compared with the EKF estimation algorithm, the joint estimation algorithm based on the fractional-order model ASOUPKF reduces the average error of the SOC by about 1.5%, and the maximal errors of the final SOCs of the fractional-order model ASOUPKF algorithm are all less than 2.2%, and the average errors are all less than 0.8%, providing an excellent estimation method for the state estimation of the battery.

During the submarine sailing, it is difficult to get help from the outside world. If there is a cable problem, it cannot be checked in time, it will seriously affect the internal communication in the cabin, as well as the combat training tasks such as weapon system testing. In order to solve this problem, aiming at the possible problems of submarine cable, a submarine cable test equipment is designed and developed, and the usage scenario of the test equipment is studied and analyzed. It can improve the troubleshooting efficiency of the cable and accurately locate the fault, which has great military significance.

A new energy system that is safe, low-carbon and clean is the goal of energy transformation. In recent years, Shandong Province has continued to promote the green transformation of energy, and has taken a series of measures in the development of new and renewable energy and the improvement of quality and efficiency in key industrial industries. This article conducts research on the evaluation of the energy and power transformation effects of Shandong Province. First, it constructs a fully quantified relative index system, and then proposes a weighting method based on the correlation between indicators and a comprehensive index comprehensive evaluation method. Finally, it evaluates the energy transformation situation of Shandong Province from 2015. The results show that the comprehensive score of energy transformation in Shandong Province increased by as much as 38%, which is remarkable. The scores of each secondary indicator show an increasing trend. Among them, the improvement of energy structure has the most obvious effect. The growth rate of energy consumption intensity, energy efficiency and carbon emission intensity is limited, and the level of energy security still needs to be improved.

This research endeavors to devise and engineer a sophisticated power meter gateway utilizing the FreeRTOS operating system coupled with Bluetooth Low Energy (BLE) technology, with the objective of augmenting the proficiency and dependability of electric power data governance and oversight. The constructed gateway facilitates the acquisition, assimilation, and dissemination of data from electricity meters, and is proficient in engaging in data communications with an array of power meters. Harnessing the multitasking prowess inherent in FreeRTOS, the system guarantees the promptness and consistency of data processing operations. Incorporating BLE technology enables the gateway to conduct wireless data transmission with minimal energy expenditure. Furthermore, the integration of an LCD permits the contemporaneous display of data, thereby enhancing the data management and surveillance framework.

Due to the complexity and nonlinearity of hydroelectric generators, timely and effective design of controller parameters is crucial for maintaining system stability. This paper adopts the theory of quadratic optimal systems and uses the Python Control library as a numerical calculation tool to optimize the controller. The numerical calculation results of the optimization algorithm for the water turbine governor show that the method is effective. As a result, the performance indicators of the water turbine regulation system have been improved. Nowadays, Python can run directly on many opensource microprocessor platforms. The research results of this article contribute to the practical application of optimal control in turbine speed control.

The Transformer architecture is widely used in the field of computer vision due to its ability to relate context and global modeling, and Transformer-based object detection methods have achieved very bright results. However, the existing methods have the problems of underutilizing background images and the model structure is too complex and redundant. In order to solve such problems, an improved high-performance object detection method based on Transformer is proposed. The method first extracts the depth features of the object by means of feature pyramid extraction, then extracts the large, medium and small object regions present in the object image by means of RPN and ROI Pooling, and finally seeks the extrinsic connection between different regions by means of the attention mechanism in the Transformer method, and predicts the object location and category using the proposed loss function. Compared with state-of-the-art object detection methods on MS-COCO datasets, The effectiveness and superiority of the proposed method were demonstrated.

In the practical application of electromagnetic transient real-time simulation computing platform, due to the large amount of calculation and long calculation time of electromagnetic transient simulation calculation, in order to improve the calculation speed and accuracy, it is usually necessary to use multiple CPU cores for parallel calculation at the same time, and ensure that each task synchronization and coordination among the computing cores. So, it is of great significance to select an appropriate task synchronization method for the realization of the electromagnetic transient real-time simulation computing platform. The paper puts forward a kind of based on TSC (Time Stamp Counter) Time Stamp Counter polling Time slot task synchronization method, this method supports high precision Time synchronization and coordination between different tasks, and can adapt to the transmission and efficient processing of large amounts of data. The test results prove that this method can avoid the overhead and delay of a large number of interrupts and thread switching, improve the efficiency, stability and predictability of task execution, and meet the performance requirements of the task synchronization method based on the centralized tightly coupled electromagnetic transient real-time simulation prototype.

This paper proposes the utilization of a dual closed-loop control approach, combining speed and current feedback, to address the issue of poor dynamic performance in the switched reluctance motor (SRM) and enhance the response characteristics and stability of its speed control system (SRD). The focus is on achieving effective control of a 6/4 pole SRM. The outer loop introduces fractional order integral on the basis of integer order PI(IOPI), and proposes a speed loop controller based on fractional order PI(FOPI). The introduced fractional order integral can improve the steady-state accuracy of the system and weaken the large overshoot caused by integral saturation. The results show that on the basis of the same K_p and K_i, the designed FOPI speed outer loop controller has good dynamic performance. Compared with the IOPI, the system overshoot is reduced by about 78.13 %.