Methods: Therefore, pilot protection of a negative sequence and additional network considering PV is proposed. The scheme is based on the feature that the PV model only outputs positive sequence components after a fault. For asymmetrical faults, the negative sequence impedance detected at both ends of the protection is utilized to construct a comparative negative sequence impedance protection criterion. For symmetrical faults, the voltage characteristics of the faulty additional network are utilized to construct a protection criterion.

Discussion: The protection method requires less data information, low dependence on communication, and can quickly identify asymmetric faults occurring in the area. The operation results have high reliability and simple calculation; additional criteria can effectively avoid the impact of load current changes on the protection, can effectively withstand different transition resistance access conditions, and different penetration rates of photovoltaic power access. This study has two limitations: (1) The model considers only PV; (2) The proposed protection scheme applies only to local circuits.

Results: Finally, an actual distribution network model with PV is constructed in PSCAD to verify the effectiveness and reliability of the protection method.

Conclusion: The protection method is selective and reliable and is not affected by high penetration rate and PV fault characteristics.

Background: The calibration of low-voltage electromagnetic current transformers for metering under energized conditions is affected by primary current harmonics and noise, which makes it difficult to accurately carry out full-range calibration.

Methods: In order to address this problem, this paper first introduces the equivalent circuit of a current transformer and carries out the error analysis at different frequencies. Then the current transformer, based on the background current offset full-range charged conditions of the frequency calibration method and the heterodyne current calibration method, isproposed, and the advantages and disadvantages of the two different methods are compared and analyzed.

Results: A calibration method based on primary current cancellation and heterodyne current injection is proposed by combining the advantages of these two calibration methods and experimentally verified.

Conclusion: The results show that the IF method based on background current offset can be applied to some conditions, but the heterodyne method without primary current offset is not suitable for calibration, while the calibration method based on primary current offset and het-ero-dyne current injection can be applied to a wide range of conditions.

Methods: The super twisting algorithm (STA) is introduced to improve the traditional SMO, and the super twisting sliding mode observer (STA-SMO) is constructed to solve the chattering problem of the traditional SMO. According to different speeds, the sliding mode variable gain coefficient is designed, and a continuous function L(x) is introduced as a switching function to make the switching of the sliding mode surface smoother. Considering the problem of stator current distortion caused by dead zone, the harmonic suppression strategy of adaptive notch filter (ANF) based on the least mean square (LMS) algorithm is studied and combined with the STA-SMO method to construct a position sensorless control system considering current harmonic compensation. Comparative verification under different speed conditions is carried out to verify the control performance of the method studied in this study under a wide speed range.

Results: Firstly, the speed information is introduced as a variable into the gain coefficient of the traditional STA-SMO, and the parameters are adjusted with speed, which solves the parameter matching problem in different speed domains of STA-SMO and effectively improves the stability of the observer. On this basis, the current harmonic compensation strategy based on LMS-ANF is introduced. According to the characteristics of the adaptive filter, the harmonic current of a specific wave can be extracted, and the acquisition current is compensated to suppress the influence of current harmonics on the estimation results of the observer, which further improves the accuracy of the observer.

Discussion: The proposed STA-SMO with LMS-ANF harmonic compensation demonstrates superior performance over traditional SMO, effectively reducing chattering and improving stability across wide speed ranges. Experimental results confirm its robustness under dynamic loads and adaptability to speed transitions, with chattering reduced by 1.1%. The LMS-ANF strategy mitigates current harmonics, enhancing low-speed accuracy. While the method balances simplicity and reliability, future work could address near-zero-speed performance and computational efficiency for broader industrial applications.

Conclusion: The STA-SMO + LMS-ANF proposed in this study can effectively improve the anti-interference ability of the observer, adapt to the application of a wide speed range, and have strong robustness and higher observer accuracy.

Methods: Innovative technologies like image processing, clever algorithms for computer and embedded vision (i.e., MobileNet, Canny Edge Detection, FaceMesh Model (Mediapipe), and Raspberry Pi 4 model B 8 GB are all included in the suggested solution. The system is made up of modules for online data visualization interfaces, driver monitoring systems, vehicle health and speed monitoring, and image processing for safety.

Results: Real-time interaction, hardware implementation, model training, and web app integration are among the project's benchmarks.

Conclusion: Deliverables include creating a reliable IoT device, installing sensors for vital metrics, setting up a centralized interface for monitoring in real-time, and creating a clever algorithm that will produce alerts promptly. The project entails extensive testing and validation to guarantee dependability, accuracy, and compliance with safety and privacy requirements by providing valuable information to law enforcement authorities, improving the road safety and effectiveness of high-speed bus operations on highways.

Methods: Firstly, in this work, the frequency-dependent feature of underground cables was analysed. Secondly, the correction algorithm for TW attenuation was obtained. Thirdly, a detailed partial discharge location algorithm was derived.

Results and Discussion: Using PSCAD/EMTDC, a ±400kV MMC-based power grid simulation model has been constructed, followed by performing a typical case study to verify the robustness of the proposed algorithm. To overcome these shortcomings, a novel partial discharge location principle for transmission line integrated with MMC-based station has been illustrated. However, it should be noted that the proposed method has only achieved frequency domain corrections, rather than the time-frequency domain, which still requires further research. Furthermore, the calculation has contained too many iterations, causing significant computational pressure.

Conclusion: The comprehensive frequency correction algorithm has exhibited the ability to recover the initial frequency spectrum information from highly attenuated TW signal, and the proposed fault location principle has been found suitable for transmission line integrated with MMC-based station.

Methods: The study initially developed an electromagnetic wave propagation simulation model for partial casing discharge to analyze the dynamic electromagnetic wave propagation process comprehensively. Subsequently, the Time Difference of Arrival (TDOA) localization algorithm underwent optimization through integration with the neural network, simulated annealing algorithm, and Bayesian algorithm.

Results: Simulation results demonstrate that electromagnetic wave signals propagate into outer space as spherical waves through the oil gap between the flange end screen and the upper casing section. The maximum electric field intensity direction exhibits substantial variations between the casing surface and the far end. The enhanced algorithms demonstrate improved localization accuracy. The neural network-based TDOA achieves a reduced Mean Absolute Percentage Error (MAPE) of 5%, with over 80% of errors contained within 0.5 units of the actual position in each coordinate direction. The Bayesian-based improvement demonstrates a MAPE of 8%, with 70% of errors within 0.8 units. The simulated annealing-based enhancement achieves a MAPE of 6%, with 85% of errors within 0.6 units.

Discussion: Based on the characteristics of the electromagnetic wave signal propagation process in the internal and external space of the oil-paper insulation sleeve, this article further improves the of the TDOA positioning algorithm based on the neural network.

Conclusion: The enhanced TDOA localization algorithm, incorporating neural network, simulated annealing, and Bayesian algorithms, successfully improves the accuracy of partial discharge localization in bushings.

Methods: The various typical hydraulic quadruped robots and their characteristics that have been published are summarized in this study. Additionally, the whole-body stable control methods of hydraulic quadruped robots are summarized, and the characteristics of various control methods are analyzed, especially the widely used model predictive control method and whole-body control method. Moreover, a few research results on hybrid control methods are introduced.

Results: By summarizing the research results of hydraulic quadruped robots, it is evident that different control methods have different characteristics. The single control method is suitable for simple control tasks of hydraulic quadruped robots on flat road surfaces.

Discussion: Due to the nonlinearity and time-varying parameters of hydraulic drive, hydraulic drive errors are inevitably present. There are still shortcomings in the development of hydraulic quadruped robots, such as energy utilization efficiency, lightweight design of structures, joint servo drive, and intelligent control.

Conclusion: In order to achieve stable control and industrial application of hydraulic quadruped robots, the control methods of whole body are summarized and analyzed. And it is pointed out that the future research work of hydraulic quadruped robots mainly focuses on the lightweight design of the system and the study of intelligent control algorithms with stronger adaptability.

Methods: The study employs a three-step approach: (1) The pearson correlation analysis was utilized to evaluate multi-dimensional influencing factors and reduce dataset dimensionality; (2) implementation of a BiLSTM neural network for temporal feature extraction and preliminary prediction; and (3) application of the Adaboost algorithm to construct a weighted combination of strong classifiers. The model’s effectiveness was validated through comprehensive simulation tests using real-world charging station data.

Results: The proposed Pearson feature selection-based BiLSTM-Adaboost model outperforms traditional benchmark models (LSTM and SVM), effectively reduces data redundancy through feature selection, achieves better performance in key indicators (MSE, RMSE, and MAPE), and demonstrates strong generalization ability and robustness while maintaining high accuracy.

Discussion: Experimental results demonstrate that the proposed method effectively extracts key features of charging loads, achieving superior prediction accuracy and generalization ability compared to traditional methods. This provides a reliable decision-making tool for power grid operation, effectively supporting the resilience planning needs of urban power grids under continuously increasing EV penetration rates. But further research is needed to address robustness under extreme weather conditions.

Conclusion: This study provides an effective load forecasting methodology for power systems to address the challenges of large-scale electric vehicle integration. Future research will explore more robust feature engineering methods and deep learning architectures, such as combining other more advanced time series prediction models and improving optimization algorithms to enhance model adaptability and generalization capability for complex data patterns.

Methods: This provides an effective means to solve the problem of different ownership entities and regulation objectives between distribution networks and DER. Existing researches mainly focus on the active power operating envelopes while ignoring the reactive power. This paper studies the calculating method for the active-reactive power operating envelope of distribution networks. First, the rectangle model of active-reactive power operating envelopes is constructed to calculate the operating envelopes that provide more flexibility for DER control, while satisfying the voltage constraints robustly and meeting the actual demands in the operation of DER control. Second, the area that the active-reactive operating envelope covers is enlarged by finding a new vertex on the right side of the p-q plane to generate a pentagon operating envelope which is called the expanded operating envelope.

Results: This leads to a higher active power output limit for DER control.

Conclusion: Finally, the effectiveness and safety of the proposed calculating method for distribution network operating envelopes is verified in distribution networks with different sizes

Methods: Dielectric breakdown tests revealed a significant inverse relationship between porosity and breakdown strength: the sample with the highest porosity (29.89 vol%) exhibited a breakdown strength of 0.83 kV/mm, while the sample with the lowest porosity (23.18 vol%) achieved 1.85 kV/mm.

Results: The analysis using the Weibull distribution showed a strong correlation between porosity and breakdown strength, with consistent fitting across the data. Electric field simulations demonstrated that pore size, distribution, and alignment with the electric field direction significantly enhance local electric field intensities, leading to reduced breakdown thresholds.

Conclusion: These results underscore the critical role of porosity in determining the dielectric performance of MgO and guide for optimizing its applications in insulation materials.

Objective: This study aims to explore how quantum computing can enhance AI and ML models, particularly in the domain of predictive analytics. It proposes a comprehensive Quantum-AI framework that integrates quantum technologies into existing predictive systems to overcome the challenges posed by classical approaches.

Methods: The study employed hybrid quantum-classical models using frameworks like Qiskit, TensorFlow Quantum, and Pennylane. Datasets were sourced from real-world applications in finance, healthcare, and climate science. Experiments were designed to compare classical and quantum- enhanced models on metrics such as accuracy, scalability, and computational time. Specific algorithms used include Quantum Support Vector Machines (QSVM) and Quantum Neural Networks (QNN).

Results: The proposed quantum-enhanced models showed significant improvements: • Accuracy increased from 85% (classical) to 92% (quantum) in finance-related predictions. • Diagnostic models in healthcare achieved a 94% accuracy rate and reduced training time by 50%. • Climate modeling tasks achieved 90% improved accuracy and 20% faster simulation times. The results validate the effectiveness of quantum models in processing large, complex datasets more efficiently.

Conclusion: Quantum-enhanced AI presents a transformative approach to predictive analytics by addressing the core limitations of classical ML systems. This study demonstrates the practical feasibility and industrial relevance of integrating quantum computing into AI workflows. It opens avenues for further research in developing scalable, domain-specific quantum algorithms and hybrid AI models for real-world applications.

Methods: Phasor Measurement Units (PMUs) are utilized to obtain active power-frequency data at generator nodes. The active output power of the generators serves as the input to the identification model, while frequency disturbances act as the output, thereby constructing a controlled autoregressive identification model for the generators. The model order is determined using the Akaike Information Criterion (AIC). Finally, a diminishing memory recursive least squares algorithm is employed to calculate the equivalent inertia of the system.

Results: The proposed method can determine the order of the identification model and estimate the equivalent inertia of synchronous machines and wind farms.

Discussion: The simulation results demonstrate that the proposed method in this paper is suitable for inertia assessment in power systems with renewable energy integration, providing significant reference value for the accurate evaluation of equivalent inertia in practical power grids. Compared to existing inertia assessment methods, the proposed approach effectively reduces the error impact caused by model order and achieves a higher fitting degree of the identified model to the system's frequency response characteristics, thereby improving the accuracy of equivalent inertia estimation.

Conclusion: The proposed identification model and algorithm demonstrate rapid tracking capability and strong convergence, facilitating precise estimation of the quickly time-varying equivalent virtual inertia of wind farms and continuous monitoring of the system's time-varying equivalent inertia during disturbances.

Methods: This research work introduces a novel design that incorporates IoT technology to improve the capabilities of portable ventilators. The proposed system integrates IoT technology in a novel manner, with major features including real-time remote monitoring and closed-loop control of critical parameters, including blood pressure (BPM) and oxygen saturation (SpO2), which are not currently available in many portable ventilators. The ventilator is characterized by extensive automation capabilities and facilities for remote control, proving to be useful in both clinical and home usage.

Results: This prototype represents a great improvement in respiratory care and provides a practical solution to the problem of working under resource constraints.

Discussion: The design of the system is supposed to bring benefits to both the patient and the healthcare providers by allowing continuous monitoring and quick interventions in dangerous situations. Further, employing inexpensive technologies ensures feasibility, especially for vulnerable groups of society. The limitation of the study includes its operation for sustained SpO2 levels above 90% highlighting the future research directions in more demanding clinical situations and when SpO2 is likely to be less than 90%.

Conclusion: Due to the combination of features and their implications for healthcare, it may be appropriate to patent this innovation as a novel concept for portable respiratory care.

Methods: This study presents a strategy to calculate the spoofed position and attitude of a drone with a GPS-spoofed error model. To mitigate GPS spoofing in networked drones, a two-stage control strategy that includes attitude and position control is suggested. This control strategy implements particle swarm optimization for tuning of proportional and derivative gain with inputs as a deviation in pseudorange at different waypoints. This control strategy generates appropriate thrusts of the drone’s rotor and then responds with a new rectified path to spoofed position detection. The suggested design technique finds the deviation in the flying path at different altitudes that cyber threats could cause. Drone route design based on multi-logit regression (MLR) with cross-entropy loss function is suggested to consider the spoofing errors between the waypoints of the expected and spoofed (hijacked) path to predict elevations and angles. By generating the appropriate thrusts of the drone’s rotor and then responding with a new rectified path to a spoofed position detection, a proportional and derivative (PD) control is developed for the attitude and position control of drones. For the swarm of drones, blockchain-based delegated proof of location (DPoL) as a consensus mechanism with GPS spoofing mitigation capability and navigation integrity at various waypoints existing on different locations and intervals has been proposed. Using a decentralized network of validators, Delegated Proof of Location (DPoL) for drone applications works on spoofing mitigation capability at different 3D locations of drones. For the detection and prevention of UAV GPS spoofing, the suggested solution combines machine learning, a two-stage control strategy, and blockchain-based cryptography techniques.

Results: The suggested method uses machine learning for desired path prediction to identify the spoofing attack and reroute the UAV to the intended location. In order to achieve a new course, the control techniques generate the required thrust. The data of newly created waypoints is encrypted using the blockchain approach. The effectiveness of the proposed work has been tested with simulation work using the UAV Toolbox of MATLAB.

Discussion: The synergy between blockchain and machine learning addresses both the accuracy of spoofing detection and the resilience of mitigation strategies. Blockchain ensures tamper-proof logging and trust consensus, while machine learning provides real-time pattern recognition and decision- making.

Conclusion: Before and after a GPS spoofing attack, a two-stage control technique and an integrated machine learning are used to produce the desired latitudes, longitudes, elevations, and altitude angles for drone trajectories. In conclusion, the proposed adaptive GPS spoofing detection and mitigation strategy using blockchain-enabled machine learning offers a promising direction for resilient UAV operations in contested environments.

Methods: In response to this, this study proposes an active distribution network flow optimization method based on energy storage regulation. This method takes the total system operation cost, which includes elements such as the cost of photovoltaic power generation, network loss cost, energy storage regulation cost, and curtailment cost, as the optimization objective. It uses the bus voltage, line flow, and transformer load of the active distribution network as the constraint for accommodating space. Additionally, it considers the charging and discharging regulation capacity and sensitivity index of distributed energy storage. Based on this, a flow optimization model for the active distribution network is established, which is driven by the joint cost and sensitivity of the minimum total adjustment amount and the minimum number of control nodes. The model is solved using the path-tracking method.

Results: In this paper, the improved IEEE39-bus system is used as an example for analysis, verifying the correctness and effectiveness of the proposed method.

Discussion: The findings demonstrate that coordinated control of distributed energy storage systems can optimize power distribution network flow, which not only fulfills photovoltaic generation accommodation requirements but also ensures secure, reliable, and economically efficient system operation. However, current research limitations reside in the failure to systematically account for energy storage operation dynamics across multiple temporal nodes. Subsequent research will extend the investigation into time-series operational characteristics of DES systems under varying temporal constraints.

Conclusion: Research shows that this method can fully consider the technical and cost factors of highproportion distributed photovoltaic grid integration. By optimizing the power flow of the distribution network through coordinated control of distributed energy storage, the proposed method achieves the minimum control adjustment amount while minimizing the number of nodes involved in the adjustment. This ensures the safe, reliable, and economical operation of the system.

Methods: The study proposes a methodology for integrating renewable energy generation through distributed pumped-storage hydropower stations. A day-ahead photovoltaic output forecasting model, based on the LSTM-Transformer architecture and designed for multi-seasonal scenarios, is also developed. Additionally, the operational characteristics of fixed-speed and variable-speed pumped-storage units within distributed pumped-storage hydropower stations are analyzed. A dynamic coordination strategy for different types of units is proposed to optimize the coordination of various pumped-storage units, aiming to maximize the operational economy of the distribution network.

Results: For the photovoltaic forecasting model, day-ahead output predictions of distributed photovoltaic systems are made under multi-seasonal scenarios, and these predictions are compared with actual data. The proposed scheduling model and strategy are validated using the IEEE33-bus distribution network system, and the performance is analyzed through various indicators, including operational costs and curtailment of distributed photovoltaic generation.

Conclusion: The proposed photovoltaic forecasting model demonstrates strong accuracy and high generalization ability under multi-seasonal scenarios. Under the proposed scheduling strategy, distributed pumped-storage hydropower significantly enhances the system's capacity for renewable energy integration and improves its economic efficiency.

Methods: In this study, the coupling relationships between energy sources are quantified using the Maximal Information Coefficient (MIC), and the energy load influencing factors are identified through Grey Relation Analysis (GRA). Furthermore, to enhance frequency separation and reduce data noise, the time-varying filter approach for empirical mode decomposition (TVF-EMD) is applied. Additionally, an Attention mechanism is integrated with Temporal Convolutional Networks (TCN) to develop a multi-task hybrid model, which effectively extracts temporal features and assigns varying degrees of attention to shared features. The proposed TVF-EMD-ATCN method is validated using publicly available IES datasets from Arizona State University.

Results: Compared to traditional single-task and data decomposition methods, the proposed model achieved Mean Absolute Percentage Errors (MAPE) of 0.0181, 0.0209, and 0.0221 for cooling, heating, and electricity load forecasting, respectively. Other evaluation metrics also demonstrated superior performance.

Discussion: The improvement of the prediction accuracy of multiple energy sources in IES significantly reduces energy waste and enhances the efficiency of multi-energy complementarity.

Conclusion: The experimental results indicate that the proposed method offers excellent forecasting accuracy and strong reliability for short-term IES load prediction.

Methods: An optimization framework is developed for both the power grid and user sides. The power grid aims to minimize voltage deviation and economic costs, while the user side seeks to maximize revenue through energy storage management. A leader-follower game model is proposed, where the power grid serves as the leader, setting the grid feed-in price for energy storage users. The users, acting as followers, optimize their response strategies based on the pricing structure provided. To overcome the challenge of local optima, an adaptive particle swarm optimization algorithm is employed to enhance the solution process.

Results: The simulation results demonstrate the effectiveness and practicality of the proposed collaborative optimization approach. The method successfully mitigates voltage deviations while optimizing both the grid’s economic costs and user-side revenues.

Discussion: Results confirm the proposed collaborative optimization's efficacy in simultaneously mitigating voltage deviations and optimizing grid/user economics. Leveraging price-based demand response and game theory, this approach offers a novel, market-compatible paradigm for coordinating distributed energy resources, differentiating it from traditional methods. Limitations pertain to real-world implementation complexity and comprehensive user behavioral modeling.

Conclusion: The proposed strategy not only mitigates voltage violations in power systems with high penetration of distributed energy but also enhances the economic and operational efficiency of both the power grid and user-side energy storage systems. By leveraging price-based demand response and an adaptive game-theoretic framework, the method effectively coordinates energy exchanges and optimizes overall system performance. The results demonstrate the practical applicability and superiority of this approach for future energy systems.

Methods: The method effectively extracts multi-source information from the inverter modules to enable precise fault localization and identification. First, parallel one-dimensional convolutional neural networks are constructed to extract key fault features from multi-source raw data. Next, an attention mechanism is employed to perform a weighted fusion of the extracted features. Finally, the fused features are classified to complete fault identification.

Results: Semi-physical simulation experiments for inverter module faults demonstrate that the proposed method can accurately extract high-discriminative fault features from multi-source sensor signals, achieving a diagnosis accuracy of 99.25%, significantly outperforming traditional methods.

Conclusion: Compared with other diagnostic techniques, the proposed method exhibits superior diagnostic accuracy and demonstrates enhanced fault diagnosis capability, providing an effective solution for inverter module fault diagnosis in complex operating environments.

Methods: According to the construction process of the overhead line, the accidental process of the broken line is analysed, and for the assumed defects of the existing simulation and the simulation difficulties, the sequence solution is introduced, and through the implicit-explicit sequence solution method, combined with the theory of nonlinear dynamic analysis, the simulation flow of the conductor-seal network collision is established.

Results: Based on the relevant parameters presented in this paper, the equivalent load and the overhead line state equation were used to calculate the bending stress and cable tension under accidental conditions. The results showed that the bending stress in brace section 1 was -119.4 MPa, and the bearing cable tension was 32,329 N. These values differed from the simulation results by less than 5%, which is within the permissible error range, thereby demonstrating the feasibility and validity of the simulation calculations.

Discussion: This study used numerical simulation to analyze the impact of a falling wire on the sealing network structure, highlighting displacement and stress changes. The implicit-explicit sequential solution method successfully simulated the entire process and addressed multi-step loading issues that traditional methods couldn't handle. The results provide key insights for sealing network design and advance transmission line construction safety research. However, the use of specific Dyneema ropes and epoxy resin sealing networks may limit the generalizability of the simulation results. Future studies should compare multiple materials and consider environmental factors affecting material performance.

Conclusion: This study effectively addresses the multi-step simulation challenges related to structural prestressing and wire-network collision. The complete collision process between the wire and the sealing network has been accurately modeled and simulated.

Methods: In this research, K-Means clustering algorithm is used for performing the segmentation. The predictive performance of machine learning models (Random Forest, Logistic Regression and XG Boost) is determined. The dataset belongs to the Commcreative company that is a marketing and advertising company. RFM (Recency, Frequency, Monetary Value) marketing analysis technique is used to identify and segment customers based on their purchasing behavior. The purpose of segmentation is to enable a marketer to design and implement targeted, effective marketing strategies. The process of customer segmentation allows an audience to be related into groups of people who share characteristics associated with the advertisement, including gender, age, interests, and spending patterns. This allows businesses to target particular groups with offers, goods, or services that are specifically designed to appeal to them. The program learns using a dataset that originates from the retail trade and includes characteristics such as an average amount of monthly customers and purchasers. Furthermore, the temporal patterns and the adaptability of the applied algorithms to various dimensionalities of the data sets are analyzed. The methodology used in this research involves a four-phase process, and it comprises information (data) gathering, client representation, customer analysis through segmentation, and customer targeting based on the collection of this study. A thorough rundown of the techniques is offered and that is applied in the context of client visualization and examination via segmentation process. Future enhancements may include the integration of new data sources, such as IoT devices and social media interactions, that will provide a richer, more comprehensive view of customers. In this paper, we also discuss the Marketing Campaign Process. Once the market is segmented, the next step is to create targeted Marketing Campaigns. A well-planned marketing campaign typically involves goal setting, audience research, strategy development, execution and monitoring. An effective campaign aligns with the specific needs and preferences of the target segments, ensuring that the message resonates and prompts desired actions. In this research, we have used machine learning for the Marketing Campaign Process.

Results: The predictive performance of machine learning models (Random Forest, Logistic Regression and XG Boost) is determined on a dataset of 2206 customers for marketing factors prediction from Tarai food manufacturing company. XGB Classifier (Extreme Gradient Boosting) shows a balanced performance with a high ROC AUC (0.906777) where ROC stands for Receiver Operating Characteristics and ROC AUC stands for Area Under the ROC Curve, which slightly surpasses Logistic Regression in distinguishing between the classes. However, according to the precision metric, Random Forest Classifier has the highest precision (0.728477), meaning it has the lowest false positive rate among the models tested. Logistic Regression offers a better balance between precision (0.714714) and recall (0.485584), resulting in a higher F1 score (0.577060). This indicates a better trade-off between false positives and false negatives.

Discussion: The findings from this study underscore the substantial benefits of applying machine learning (ML) techniques to customer segmentation and personalized targeting in e-commerce platforms. By implementing models such as k-means clustering, decision trees, and collaborative filtering algorithms, this research achieved a measurable improvement in key performance metrics including customer engagement. Compared to traditional rule-based segmentation, ML-driven approaches allowed for more dynamic, data-driven, and scalable segmentation. These techniques captured complex customer behaviors and preferences, resulting in more accurate groupings. Personalization strategies based on these segments led to significantly better alignment between customer needs and marketing efforts, validating the hypothesis that machine learning can outperform conventional methods in customer targeting.

The findings align with and extend prior research that suggests ML enhances predictive accuracy in consumer behaviour modelling. Further, this study contributes to the literature by: • Integrating multiple ML techniques into a unified e-commerce framework. • Demonstrating the real-world impact of these models using live platform data or high-fidelity simulations. • Evaluating not only segmentation quality but also downstream marketing outcomes.

Conclusion: Graphical data visualization is done that helps in feature selection of both categorical and non-categorical data. Some features show a clear relationship with the response. This research examines the current and potential applications of AI and ML in marketing, emphasizing how these technologies can automate tasks, reduce costs, and improve workflows. It also explores the role of ML in micro-segmentation and personalized marketing campaigns.

Methods: Firstly, an Ampere-hour Integral-based model is developed for state of charge estimation, and a capacity decay model is constructed for state of health estimation. Secondly, the LSSVM and the UKF fusion algorithm are applied to estimate the SOC of the battery. Simultaneously, the CNN and LSTM fusion algorithms are used to estimate the current SOH. Finally, the maximum available capacity of the battery, as estimated by the SOH module, is integrated into the physical model of SOC.

Results: This integration leads to a refined SOC estimate, enhancing the estimation accuracy. This paper uses a 150mAh button-type sodium-ion battery as an example for experimental verification. Compared with the measured values, the SOC estimation error for sodium-ion batteries at chargedischarge rates of 0.1C, 0.5C, and 1C is less than 1.1%, and the error accuracy estimate for the SOH estimation model is below 1.3%. Overall, this method has good applicability and accuracy.

Discussion: This study achieves joint estimation of SOC and SOH for energy storage batteries through a deep learning heterogeneous algorithm, demonstrating promising accuracy and stability. However, the SOH estimation exhibits declining fitting performance in later cycles, indicating insufficient relevance of the current health indicators. Future work should integrate multi-dimensional features to enhance model generalization capabilities.

Conclusion: This study proposes a high-precision joint modeling framework for battery state estimation and validates the collaborative effectiveness of heterogeneous deep learning algorithms. This achievement not only significantly improves the estimation accuracy of SOC and SOH but also provides an effective solution for enhancing battery usage efficiency and extending battery lifespan, thereby strongly advancing the development of energy storage technology and the electric vehicle industry.

Methods: This paper proposes a voltage support scheme for low-voltage distribution networks, taking into account the resistive characteristics of distribution lines. Based on the distribution network impedance ratio (R/X), the optimal active and reactive current reference values are estimated. A voltage support strategy is planned according to the severity of the voltage drop, providing coordinated voltage support by simultaneously delivering active and reactive power during voltage sags and fault conditions.

Results: A digital-physical hybrid experimental platform and a simulation model of the low-voltage distribution network were established to validate the effectiveness of the proposed scheme. The results indicate that the voltage support capability of the proposed scheme is superior to that of traditional reactive power voltage support methods.

Discussion: The simulation results demonstrate that the proposed method can effectively provide voltage support at the point of common coupling in low-voltage distribution networks during voltage faults, thereby preventing system instability. However, under severe fault conditions, the voltage support capability remains constrained by the converter's rated capacity limitations, preventing perfect voltage restoration.

Conclusion: The low-voltage distribution network voltage support scheme proposed in this paper considers the distribution network impedance ratio to calculate the required injected active and reactive power, thereby providing maximum support to the voltage. In addition, the active and reactive currents are adjusted according to the severity of the voltage drop using a droop curve. The effectiveness of the proposed method on the operation of the distribution network is validated through simulations. The scheme can effectively provide voltage support during voltage dips in the distribution network, preventing voltage instability.

Methods: Firstly, the dynamic model of the SC subsystem is established. Secondly, a mathematical model including all the unstable states of reference input, step, and double line frequency disturbance signals is built. Then, the extended state space model is constructed based on these two models. Furthermore, the optimal state feedback controller is designed by using the optimal tracking strategy.

Results: In the case of the DC-bus voltage control test, when the PV output power fluctuates, the SC makes a rapid response and compensates for the power deficiency. After such a dynamic adjustment process, the DC-bus voltage stabilizes at the reference value of 400 V, which achieves the accurate regulation of the DC-bus voltage. The maximum absolute error is 1 V, the maximum relative error is 0.25%, the maximum overshoot is 0.75%, and the settling time is less than 10 ms. In the case of the gird-connected operation test, the dynamic responses of the DC-bus voltage and its error also demonstrate the excellent performance of accurate tracking and suppression of multiple fluctuations. Accordingly, the MPPT control functions well. For instance, taking 1.5 s, the PV output power drops suddenly, and the active power demand increases at the same time, which makes the power difference reach the maximum. Then, the SC promptly increases its output current to provide power support. At this time point, the maximum overshoot of the voltage tracking is 4.5%, and the maximum absolute error is less than 1 V.

Discussion: This study provides a novel idea and approach to address the problem of suppressing the multiple fluctuations coupled to the DC-bus voltage in the two-stage grid-connected PV system via the single-phase DC/AC inverter. However, the results are limited to simulations. Hence, based on the work of this article, further studies should be conducted in applying the method in real operation systems and adapting it to deal with the multiple fluctuations suppression of the grid-- connected PV systems under asymmetrical grid conditions.

Conclusion: The proposed control scheme can not only effectively restrain the fluctuations in the PV output power and load demand but also suppress the double line frequency ripple so that the DC-bus voltage can always be accurately regulated to its reference value, which also guarantees the reliable realization of the PV’s MPPT control. Additionally, the dynamic and static performances of the system, such as overshoot, rapidity, and tracking error, are satisfactory.

Methods: We built 3D models of NTCs and examined the stresses in the chip inside with varying structures, on varying load conditions; structural bending load, temperature cycling load, separately, and both. In the resistor ceramic body around termination regions which were connected with solders, the stresses had relatively significant values than the other regions and it was considered as a main factor of the failure of the NTC body.

Results: When the number of the inner electrodes increased, the maximum principal stress that occurred on the inside of NTC was decreased gradually and when the number of electrodes was equal to 24, it had the lowest value of 524 MPa. Also, the stress changes inside at varying times were investigated when the periodic temperature of the chip body was changed from -60°C to 120°C. In this case, the maximum principal stress had the lowest value as 538 MPa when the number of inner-electrode was also equal to 24. In addition, stress distribution patterns of bending load were compared with temperature cycling load.

Conclusion: Through this study, optimized geometry and failure condition of NTC were obtained.

Methods: An in-depth analysis of the operational characteristics of individual microgrids (MG) must be conducted and a demand response model that integrates both electrical and thermal loads should be developed. Ahierarchical master-slave game optimization framework with the system operator as the upper-level leader and microgrid load aggregators, energy storage stations, and wind farms as lower-level followers should be established. Power interaction channels among entities to enable coordinated energy exchange should be designed. At the upper level, an adaptive particle swarm optimization algorithm to determine optimal energy pricing and demand response compensation has to applied. At the lower level, overall operational costs using energy scheduling strategies should be minimized.

Results: To validate the proposed method, four scenarios were simulated using the control variable method. Taking Scenario 3 as an example, where energy exchanges between parks were eliminated and each park operated independently, the total profit decreased by 14387 CNY (Chinese Yuan) compared to the proposed master-slave game optimization strategy. This demonstrates the critical role of coordinated energy exchanges in improving system efficiency and highlights the effectiveness of the proposed method.

Discussion: The bi-level game framework in this study demonstrates that coordinated energy exchange and integrated demand response could significantly enhance system-wide economic efficiency. The approach effectively balances collective benefits with individual agent interests. However, the results rely on a deterministic model, highlighting the need for future research to address system uncertainties and more complex network constraints.

Conclusion: The proposed strategy not only addresses the inherent conflicts in multi-agent integrated energy systems but also enhances system-wide economic and operational efficiency through a robust optimization framework.

Methods: This paper presents a novel approach to transform the product of the winding function and air-gap flux density into a cosine series. Then, the integral can be calculated with the aid of the periodicity of trigonometric functions.

Results and Discussion: The analytical results of the winding inductance are validated by the finite element method (FEM) simulations and experimental measurements. The relative error for the three-phase peak self-inductances is 4.8% between analytical and FEM results and does not exceed 4.9% between analytical and experimental results.

Conclusion: With this approach, the Fourier series of inductance can be calculated analytically rather than numerically. The relationship between the harmonic amplitude/order of inductance and machine parameters is revealed. This analytical model can provide the preliminary basis for the inductance design of VFRMs.

Methods: The research methodology for investigating cybersecurity risks and mitigation strategies for 5G-connected electric vehicles adopts a qualitative approach, utilizing in-depth expert interviews, content analysis, framework development, and validation processes. Overall, this methodology aims to provide valuable insights and practical recommendations for addressing cybersecurity challenges associated with the integration of 5G technology into electric vehicles.

Results: The outcomes of this study propose practical and robust mitigation strategies to address the cybersecurity risks associated with 5G-connected electric vehicles. The resulting framework is intended to guide stakeholders, including automakers, network providers, cybersecurity experts, and regulatory bodies, in implementing effective security protocols and best practices.

Discussion: The integration of 5G connectivity in EVs introduces new attack vectors and vulnerabilities, posing significant challenges for automakers, network providers, and cybersecurity experts. As EVs become more interconnected and reliant on high-speed, low-latency communication networks, they are increasingly susceptible to unauthorized access, data breaches, and cyberattacks. Therefore, it is essential for stakeholders to remain vigilant and proactive in identifying and mitigating these risks to ensure the safety, privacy, and trustworthiness of connected vehicles.

Conclusion: The findings of this study provide valuable insights and guidance for stakeholders, including automakers, network providers, cybersecurity experts, and regulatory bodies, in implementing effective security protocols and best practices for 5G-connected EVs.

Methods: In this regard, an optimized Radix-4 Booth Multiplier (RBM) is designed using a modified booth encoder and selector blocks along with the proposed improved version of Square Root Carry Select Adders (SRCSLA). The design metrics of the multiplier, which is used to multiply the inputs and coefficients of the filter also improved by using the proposed SRCSLA. For the proposed SRCSLA, The Carry Look Ahead (CLA) adder and Carry Skip Adder (CSA) are combined and modified for the proposed SRCSLA. Different bit size-based SRCSLA adders are implemented for the proposed multiplier architecture and to accumulate the final filter output. The adder structure speed is improved by modified carry-producing and propagating blocks. The retiming- based FIR filter architecture with order N = 32 is coded by HDL and synthesis is carried out using Genus synthesis tools in 45nm CMOS technology provided by CADENCE.

Results: Area Complexity (AC), Delay Complexity (DC), and PC are estimated by the reports generated by the software tool. The trade-off design metrics Area-Delay-Product (ADP) and Power- Delay-Product (PDP) are also estimated and correlated with the conventional filter structures and existing filter architectures.

Discussion: The proposed higher-order FIR filter architecture demonstrates significant improvements in latency and power efficiency for ECG signal processing by leveraging retiming techniques and optimized arithmetic blocks. The integration of an enhanced RBM with a novel SRCSLA, which combines features of CLA and CSA, results in faster and more power-efficient computation. Synthesized in 45nm CMOS using Cadence Genus, the design shows reduced Area, Delay, and Power. However, the complexity of SRCSLA logic and increased design overhead for retiming may present scalability and implementation challenges for ultra-high-order filters.

Conclusion: The proposed work is better comparatively existing works using the retiming concept.

Methods: In this work, a microstrip patch antenna with wide coverage, wide bandwidth, and high efficiency was designed to operate at a 2.4 GHz frequency. While antennas operating at higher frequencies have a narrower coverage area, the proposed 2.4 GHz frequency antenna has a wide coverage and penetrates better through obstacles such as walls, solid objects, and floor heights. While most standard 2.4 GHz patch antennas have a 20 MHz bandwidth and 30% efficiency, the recommended design has a 57.5 MHz bandwidth and 50% efficiency. The microstrip patch antenna developed for S-band wireless communication uses FR4 substrate, an inset feed line, and a lumped port. The antenna is rectangular in shape; further design parameters are explained in detail in the study. The proposed antenna design is made using the HFSS program.

Results: A detailed analysis was performed by simulations for the designed 2.4 GHz microstrip patch antenna. The S11 reflection coefficient simulation, bandwidth simulation of the antenna, total radiated electric field simulation, antenna gain simulation, radiation pattern simulations for E and H planes, co- and cross-polarization radiation patterns for E-plane and H-plane, and antenna efficiency simulation are performed. The proposed antenna has been proven to perform well, being 20% more efficient and approximately 40 MHz wider band than standard 2.4 GHz antennas.

Discussion: The simulation results show that the designed 2.4 GHz microstrip patch antenna works successfully. While the efficiency of most 2.4 GHz microstrip patch antennas in the existing literature ranges from 25% to 32%, this study achieves an efficiency of 50%. While most of the 2.4 GHz antennas in the existing literature have a bandwidth of 20 MHz, the designed antenna has a bandwidth of 57.5 MHz. Although the designed 2.4 GHz antenna has higher efficiency and a wider band than standard 2.4 GHz antennas, this research is limited by the simulation results.

Conclusion: The proposed microstrip antenna design has demonstrated 20% higher efficiency compared to standard 2.4 GHz antennas. It also achieved 57.5 MHz bandwidth, which is approximately 3 times the bandwidth of standard 2.4 GHz antennas. The designed antenna can be used in S-band communication applications such as cordless phones, Bluetooth headsets, garage door openers, keyless car locks, baby monitors, medical diathermy machines, microwave ovens, Wi-Fi applications, airport surveillance radar, weather radar, ship radar, and communication satellites. The production of the antenna is planned for future studies.

Methods: Firstly, the XGBoost algorithm is used to calculate the correlation between features and quantitatively analyze the specific impact of meteorological data on the load sequence. This helps screen out features with higher correlation, thereby improving prediction accuracy. Secondly, the filtered feature data is provided as input into the ETSformer model for training and prediction. By introducing the Exponential Smoothing Attention (ESA) mechanism and the Frequency Attention (FA) mechanism, efficient and accurate short-term load prediction is achieved. Finally, the COA algorithm is employed to optimize the hyperparameters of the ETSformer model, further enhancing its prediction accuracy.

Results: Comparative analysis with baseline models, such as ARIMA and BiLSTM, demonstrates that the model proposed in this article exhibits superior performance in short-term load forecasting tasks. Especially compared with the ARIMA model, this model reduced key indicators such as MAE, RMSE and MAPE by 44.6, 58.85, and 10.91%, respectively.

Discussion: The proposed method demonstrates strong performance in the field of short-term load forecasting. It addresses the limitations of traditional point forecasting methods, which are often difficult to apply effectively in real-world scenarios. Moreover, it overcomes the challenges of manual parameter tuning and the shortcomings of conventional optimization algorithms when dealing with high-dimensional data, thereby avoiding local optima. As a result, this method offers an efficient and practical solution for short-term load forecasting.

Conclusion: Our study demonstrates that the ETSformer model exhibits excellent performance in short-term load forecasting tasks. Integrating XGBoost for feature selection effectively mitigates the interference caused by features with low relevance to the prediction process. Furthermore, employing the COA algorithm to optimize model hyperparameters significantly enhances prediction accuracy. Compared to other baseline models, the proposed model achieves superior performance in short-term load forecasting tasks.

Methods: In this research work, a hybrid deep learning network is recommended for segmenting and classifying brain tumors. The main novelty of the research is the proposal of a hybrid network with a residual connection for classifying brain tumors, which can be used to provide proper therapies to affected patients. The recommended model provides accurate classification results, enabling the appropriate clinical decision to enhance the patient's life expectancy and quality of life. From the public data links, the brain MRI is garnered. Furthermore, the Spatial Attention-based Bilateral Network (SABNet) is proposed for segmenting the tumor region in the input images. Consequently, three sets of features, including morphological, shape, and texture, are extracted, and then feature fusion takes place with the help of optimal weights. Here, the weight is optimally determined by employing the Adaptive Fitness-aided Red Panda Optimization Algorithm (AF-RPOA) for acquiring the weighted feature fusion. At last, the fused features are passed to a Hybrid Residual Network (HRNet), which is built using the Long Short-Term Memory (LSTM) and Capsule Network (CapsNet) architectures. Henceforth, validation of the developed framework is conducted to reveal the simulation outcomes and analysis.

Results: According to the evaluation outcome, the proposed AF-RPOA-HRNet achieves an accuracy of 97%.

Discussion: The comparative validation of the proposed hybrid deep learning network with conventional works has verified its effectiveness in segmentation and classification. Moreover, the adoption of a weighted feature selection approach has yielded effective performance in classifying brain tumors.

Conclusion: On the contrary, the recommended system provides better results that can prove the system's supremacy in detecting different types of brain tumours.

Methods: First, a mathematical model of the multi-VSG parallel control system is established, considering the influence mechanism of different parameters of multi-VSG. Next, the inertia center model is introduced to calculate the inertia center values of each VSG parameter and control the error with the original values, dynamically adjusting the parameters to ensure consistency and minimize the impact of parameter differences. Then, high-pass filter feedback is introduced to optimize the parameters in the low-frequency band, improving the response capability and reducing communication delay. This approach can effectively reduce the interaction between parameters, thereby improving disturbance rejection and dynamic response capabilities.

Results: Simulation experiments under various operating conditions demonstrate that the proposed method can enhance the dynamic response speed of the system's power, frequency, and voltage, suppress overshoot, and make the parameters more consistent, thereby reducing their mutual influence and outperforming traditional methods.

Discussion: Through comparisons under different operating conditions, it is evident that the proposed control strategy can be effectively applied to various scenarios and solve the power-frequency fluctuation problems in the system under multi-VSG parallel control.

Conclusion: Compared with traditional VSG control, the proposed method demonstrates higher dynamic response capability and improves the disturbance rejection ability of the multi-VSG parallel system, providing theoretical and simulation basis for practical engineering applications.

Methods: To address these challenges, we propose a deep learning-based convolutional reconstruction network incorporating Spatial Pyramid Pooling (SPP). This architecture utilizes a fully connected block to model the complex nonlinear relationship between boundary voltage measurements (from the EIT system) and the pixel distribution within the imaging domain. Convolutional blocks and an SPP block are integrated to enhance feature dimensionality and diversity, thereby mitigating the ill-posedness of the inverse problem.

Results: Simulation and experimental results demonstrate that the proposed method significantly improves reconstruction accuracy compared to established techniques, including Tikhonov Regularization (TR), Total Variation (TV), and V-Net.

Discussion: The nonlinear relationship between the measurement information obtained from the electrical impedance tomography system and the pixel information of the measured space is implemented by a fully connected block. The convolution block and the spatial pyramid pooling block increase the dimension and diversity of features, thereby improving the ill-posedness of image reconstruction.

Conclusion: The proposed convolutional network with spatial pyramid pooling provides a robust and accurate solution for EIT image reconstruction in multiphase flow visualization, outperforming conventional and deep learning-based benchmarks. This approach holds promise for enhancing EIT applications requiring high-fidelity imaging.

Methodology: This research proposes a robust approach for detecting R peaks in QRS complexes using a recurrent neural network. Our proposed methodology was applied to the well-known MITBIH Arrhythmia Database (MIT-DB) dataset and the China Physiological Signal Challenge (2020) database, which contains over a million beats. The hybrid linearization technique used an adaptive filter and Discrete Wavelet Transform (DWT) to remove noise from the ECG signal. The next step was to use principal component analysis (PCA) to extract characteristics from the ECG data. Lastly, the R peak signals were classified using Long Short-Term Memory (LSTM) to improve accuracy through optimization techniques like Grey Wolf Optimization (GWO). The algorithm's performance was also evaluated using the MIT-BIH Arrhythmia database and the China Physiological Signal Challenge (2020).

Results: The suggested formal technique yielded the best results for R-peak detection on CPSC- DB, with an F1-score of 95.3%, recall of 96.8%, accuracy of 99.5%, and precision of 95.3%.

Discussion: The results achieved using these methodologies are really excellent and competitive with other technologies. In some areas, they are even better than the existing methodologies.

Conclusion: The F1-score, recall, and precision of the algorithms on MIT-DB are all equivalent to, or better than, those of the competing methods.

Methods: First, the frequency response characteristics of various frequency-regulating resources are analyzed in the new energy system under typical control modes. Secondly, a system frequency response model is established that takes into account the differences in the frequency response characteristics of synchronous units and wind power, revealing the influence mechanism of different frequency-regulating resources on the system frequency characteristics. Finally, considering the spatio-temporal distribution characteristics of inertia, and by using the post-disturbance network and the sudden-variable power allocation analysis method based on the superposition principle, with the rate of change of frequency as the key calculation index, a node inertia calculation method is proposed applicable to the new energy system.

Results: The simulation results indicate that the aggregated system frequency response model and the sudden-change power allocation analysis method presented in the paper are feasible.

Discussion: The proposed method represents a remarkable advancement. It can accurately model resources and simplify system analysis. Moreover, simulation comparisons under different scenarios have been carried out to verify their accuracy. However, this method currently overlooks emerging resources such as energy storage, and future research should address these issues.

Conclusion: The proposed approach allows for a comprehensive quantification of inertia demands from various power sources while ensuring the dynamic stability of the system frequency, providing a more accurate and reliable way to manage new energy system operations.

Methods: The research investigates and compares the performance of three prominent deep learning models—ResNet-50, VGG16, and AlexNet—on MRI image datasets. Based on the high accuracy of VGG16 and ResNet-50, a novel hybrid model combining both architectures, termed VGG16- ResNet-50, was developed. The model was trained and tested using large, publicly available clinical datasets. Performance metrics, such as F1-score, sensitivity, specificity, and overall accuracy, were used for model evaluation. Additionally, the hybrid model’s performance was benchmarked against other classical models, including DenseNet121, ResNet-101, and GoogleNet.

Results: The proposed hybrid VGG16-ResNet-50 model demonstrated exceptional performance, achieving an F1-score of 99.98%, sensitivity of 99.98%, specificity of 99.99%, and an overall accuracy of 99.99%. These results indicate a significant improvement over traditional models, validating the effectiveness of the hybrid deep learning approach for early and precise brain tumor classification.

Discussion: The proposed hybrid VGG16-ResNet-50 model demonstrated exceptional performance, achieving an F1-score of 99.98%, sensitivity of 99.98%, specificity of 99.99%, and an overall accuracy of 99.99%. These results indicate a significant improvement over traditional models, validating the effectiveness of the hybrid deep learning approach for early and precise brain tumor classification.

Conclusion: The hybrid VGG16-ResNet-50 model proves to be a powerful tool for the detection and classification of brain tumors using MRI images. Its superior accuracy and robustness, validated through comparison with other state-of-the-art models, suggest its potential as a reliable solution in clinical settings to support radiologists in early diagnosis and treatment planning.

Methods: This article examines the present applications of AI and ML in radiology, ranging from image identification to predictive analytics, along with issues such as data protection, legal restrictions, and the need for transparency in algorithmic decision-making.

Results: This study also examines emerging developments, including how AI can improve access to healthcare in remote locations and its potential applications in telemedicine. The AI and ML results represent exciting new possibilities, and resolving technical, moral, and legal concerns is necessary for their effective application.

Discussion: The integration of Artificial Intelligence (AI) and Machine Learning (ML) into radiology is significantly reshaping diagnostic practices in healthcare. AI, particularly deep learning techniques such as Convolutional Neural Networks (CNNs), is proving highly effective in analyzing complex imaging data, enhancing both speed and accuracy of diagnostic techniques.

Conclusion: Radiology could undergo a revolution due to AI and ML, which could improve diagnostic skills and healthcare administration in general. AI and ML hold transformative potential for the field of radiology, promising improved diagnostic accuracy, operational efficiency, and expanded access to care, particularly in underserved areas.

Objective: The scientific aim of the work is to apply the new model to a case study photovoltaic (PV) system with two input scenarios: accurate historical meteorological information and projected meteorological data. The research aimed to develop a model designed for short-term solar energy output projections, using the Long Short-Term Memory (LSTM) network algorithm, while accounting for fluctuations in meteorological conditions. This study extends the existing research by evaluating the performance of the LSTM-based model through an analysis of prediction accuracy and error rates and comparing them with various LSTM configurations and conventional timeseries forecasting techniques within the test dataset.

Methods: The architecture of the model, grounded in the LSTM algorithm, has been customized to align with the distinct meteorological trends and operational parameters of a region situated in Southwest China. The performance of the LSTM-based model is assessed by analyzing the prediction accuracy and error rates against those of various LSTM configurations and conventional time-series forecasting techniques within the test dataset. The most suitable hyperparameters of the LSTM model are finally selected. Finally, the proposed model is applied to a case study photovoltaic (PV) system with two input scenarios: accurate historical meteorological information and projected meteorological data.

Results: The optimal parameters for the LSTM model in forecasting PV power are four hidden layers with 100 nodes. The comparative analysis of the outcomes reveals the average absolute percentage error is significantly higher in the second scenario compared to the first (10.314% versus 3.316%). To address the inaccuracies introduced by the weather forecast data, new features are proposed to enhance the accuracy of the LSTM model. Applying this prediction model to the case system reduces the PV power error ranging from 10.314% to 9.32%.

Discussion: The optimal parameters for the LSTM model in forecasting PV power are four hidden layers with 100 nodes. Integrating forecast meteorological data into the training process enhances the model's ability to predict meteorological conditions.

Conclusion: During the establishment of the prediction model, the consideration of meteorological fluctuations and the selection of appropriate models are of beneficial to the improvement of the predicting accuracy of the photovoltaic power. Moreover, this study not only provides a direction for selecting suitable characteristics and structure of LSTM model, offering an effective method for short-term output prediction of photovoltaic power generation systems.

Methods: A finite element model of a coil-solenoid system was developed. Single-factor and orthogonal experiments analyzed the effects of winding height, thickness, radial turns, and axial layers. Linear regression and multi-objective optimization were used to minimize force/current peaks and maximize displacement.

Results: The optimal parameters were: solenoid winding height = 6 mm, thickness = 2.5 mm, radial turns = 6, axial layers = 11. This configuration achieved the best driving characteristics.

Discussion: The results demonstrate that increasing radial turns improves displacement, while other parameters significantly reduce repulsive force and current, thus balancing breaking capacity and mechanical longevity.

Conclusion: The orthogonal experiment-based approach effectively optimizes solenoid parameters, ensuring long-stroke rapid breaking capacity and extended mechanical service life.

Methods: This study presents a comprehensive optimization strategy that coordinates PSS with additional power oscillation damping controllers for both Static Var Compensator (SVC-POD) and Thyristor Controlled Series Capacitor (TCSC-POD). The proposed multi-objective function incorporates critical stability indicators, including generator rotor speed deviation, eigenvalue real components, damping ratios associated with low-frequency oscillation modes, and system bus voltage deviations. To optimize the controller parameters, an Enhanced Whale Optimization Algorithm (EWOA) is implemented.

Results: The effectiveness of this optimization strategy is validated through eigenvalue analysis and time-domain simulations conducted on the IEEE 4-machine 2-area test system under various operating conditions.

Discussion: The proposed EWOA-based coordinated control effectively suppresses low-frequency oscillations while improving voltage stability in renewable-integrated power systems. Although validated on a standard test system, the approach shows promising potential for real-world applications. This work advances damping control strategies for modern power grids with high levels of renewable energy penetration.

Conclusion: Simulation outcomes verify that the EWOA-optimized controller parameters substantially improve system stability, effectively attenuate low-frequency oscillations, and improve voltage regulation capabilities.

Methods: The system uses a Random Forest classifier with SHAP (SHapley Additive exPlanations) to forecast the top-3 courses with the greatest expected scores for appropriateness. The system uses SHAP (SHapley Additive exPlanations) values to integrate Explainable AI (XAI) to guarantee openness and trust. The proposed model solves the cold start problem and provides data security.

Results: In terms of precision (0.90), recall (0.92), and F1-score (0.92), Random Forest fared better than any other classifier. By combining predictions from several decision trees, its ensemble learning technique increases stability, decreases overfitting, and improves generalization across a broad range of input data patterns.

Discussion: A major drawback of conventional black-box machine learning models is their lack of transparency, which is addressed by the incorporation of SHAP into the course recommendation system. In this model, students get suggestions in addition to information on which topic scores had the biggest influence on their choice.

Conclusion: This study presents a practical and interpretable course recommendation system designed for 12th-grade students transitioning into higher education. Through the use of Explainable AI-enhanced Random Forest, the model provides precise, transparent, and customized recommendations. It addresses the main drawbacks of conventional systems, such as their lack of explainability and their cold start problems. The model's excellent precision, recall, and F1-score indicate its superior performance, which makes it a useful tool for student guidance and academic advice. For even more thorough recommendations, future research may consider combining long-term academic objectives, aptitude test results, and student interests.

Methods: This study analyzes the electromagnetic performance of the novel HE-RSVLFPMSM under three operating conditions: rated condition, 1.72 times overload condition with field winding (90s duration) and flux-weakening peak-speed condition with field winding. Based on heat transfer theory, a solution domain and fluid-solid coupling mathematical model are established to investigate temperature distributions under different conditions. Five channel structures are designed and modeled, and their impacts on temperature and fluid fields are simulated and compared.

Results: The results indicate that under auxiliary excitation overload, the maximum temperature in windings reaches 124.77°C, exceeding the class E insulation limit. Under extreme flux weakening with I^f = -20A, the maximum winding temperature is 118.72°C, approaching the class E insulation threshold. Comparative analysis shows that the axial Z-shaped channel is the optimal cooling structure, in which the maximum temperature in windings is reduced by 14°C under the most severe conditions, and decreases the overall maximum temperature by 26°C. This significantly lowers winding insulation costs.

Discussion: This study explores optimal thermal management designs for HE-RSVLFPMSM (medium-small power motors), with results demonstrating significant practical implications. Simulations confirm that the axial Z-shaped channel reduces peak winding temperature by 26°C under extreme operating conditions, enabling safe operation with lower-cost Class A insulation --enhancing reliability while reducing manufacturing costs. Compared to prior research, this work comprehensively analyzes five structural configurations and introduces the power-normalized metric η to quantify thermal performance across comparable designs. However, conclusions remain primarily simulation-based, with constrained physical interpretation of quantitative metrics. Final performance validation awaits prototype thermal experiments in the future.

Conclusion: The proposed axial Z-shaped channel design can effectively address the abovementioned thermal challenges. The proposed design not only enhances cooling efficiency but also reduces insulation costs. Finally, a comprehensive comparison is made between the recent researches related to the water channel structures and this study. The results demonstrate that this study is more comprehensive and reliable, and the axial Z-shaped channel exhibits superior thermal management characteristics, particularly for medium-to-small power-level motors.

Methods: The methodology involved collecting data, preparing the data, creating both single and combined models, assessing with ADNI data, and confirming with additional datasets. The approach was chosen by comparing various scenarios. The top six individual ConvNet-based classifiers were combined to form the ensemble model. The evaluation showed high accuracy rates across various classification groups. Validation of additional data showed impressive accuracy, exceeding results from numerous previous studies and aligning with others.

Results: Although ensemble methods outperformed individual models, there were no notable distinctions among different ensemble approaches. The ensemble model was constructed using the top six individual ConvNet-based classifiers in deep learning (DL), achieving high accuracy rates across various classification categories: 98.66% for Normal control | AD, 96.56% for Normal control | Early MCI, 94.41%

Conclusion: Early MCI/Late MCI, 99.96% for Late MCI | AD, 94.19% for four-way classification, and 94.93% for three-way classification. Validation results underscored the limited effectiveness of individual models in practical settings, contrasting with the promising outcomes of the ensemble method.

Methods: In this research, samples were randomly selected and grouped in pairs. The proposed method extracted features through two symmetrical back propagation neural networks to get multiple pairs of feature vectors, then predicted the state of furnace hearth activity by measuring the Manhattan distance and cosine similarity between each pair of feature vectors.

Results: The experimental results showed that the proposed method can extract sample features stably, it still has a good analysis effect in scenarios with few samples. The optimal accuracy rate was 99.18%, which can meet the monitoring of hearth activity in actual production.

Discussion: The measurement based on distance and angle can reflect the characteristics of each category more accurately. It also achieved good performance in scenarios with few samples and provided a theoretical basis for the reasonable regulation of the hearth activity.

Conclusion: Compared with the related methods, the proposed method based on distance-angle metric learning in this research has better performance. It can meet the application requirements in actual production scenarios and has great research value.

Methods: In this paper, a two-layer optimization model is developed. The upper-layer model considers the time-shifting characteristics of demand response to optimize the total electricity costs of users during all time periods after load response. The lower-layer model considers the distance and importance of load reduction to minimize economic losses caused by power shortages. Combined with the two-layer load optimization model and the Sequential Monte Carlo (SMC) algorithm, a reliability assessment process is proposed.

Results: An improved reliability assessment index system incorporating economic factors was adopted and validated through simulation in the IEEE RBTS BUS6 F4 system. The simulation results showed that the proposed two-layer load optimization model reduces the system reliability index by 14.38% to 48.9% and the economic index by 40.6%.

Discussion: This model overcomes the limitations of traditional methods, which only consider a single load control strategy, achieving coordinated optimization of load shifting and load reduction to enhance system reliability.

Conclusion: Compared to existing methods, the proposed two-layer load optimization model demonstrates significant advantages in improving reliability while also considering economic efficiency, thereby achieving a win-win situation for both the power grid and users.

Methods: To achieve faster data rates, increased throughput, and reduced latency, 6G networks may transition from orthogonal multiple access (OMA) to non-orthogonal multiple access (NOMA), which enhances spectral efficiency. Through a combination of analytical framework and intensive simulations, this paper examines the effects of admitting low-data- rate IoT devices into NOMA-based clusters of high-data-rate cellular users (CUs).

Results: The simulation results indicate that admitting IoT devices with favorable locations can increase the total achievable rate of the cluster by about 30%, while maintaining each CU’s target rate.

Discussion: Spatial-based recommendations for establishing an effective device admission strategy are provided. The results pave the way for device scheduling and association with the objective of maximizing the network utilization.

Conclusion: Admitting IoT devices into a NOMA cluster of cellular users is achievable without compromising the CUs target rates, a result that render NOMA techniques to play an important role in the 6G future deployments.

Objective: Renewable energy generation power experiences great fluctuation, which has caused serious harm to grid-connected operations. It is very important to establish a reliable and accurate prediction model for the safe operation of the power grid.

Methods: Since the weather changes are not correlated and the data required for long-term prediction is too large, this paper forecasts the power of renewable energy generation based on neural networks and short-term weather. The wind power station selected the actual situation of a power station in Guizhou. The solar power station selected the actual situation of a photovoltaic power station in Guizhou.

Results: The results show that the trend of wind power generation and solar power generation can be well predicted. The absolute error between the predicted value and the power of the wind power station did not exceed 3.0, and the relative error value did not exceed 8%. The absolute error value between the predicted value and the power of the wind power station did not exceed 1.25, and the relative error was not more than 2.6%.

Discussion: The results show that the trend of wind power generation and solar power generation can be well predicted based on the neural network and short-term meteorological analysis. However, the model does not take into account sudden changes in weather, and only takes into account the southwest region, and further investigation is needed in other regions. In general, this model can effectively predict the renewable energy power generation in the process of steady change of weather in southwest China, which is rich in renewable energy.

Conclusion: The simulation results show that the model can be prepared to predict the output power of renewable energy generation. On the other hand, the application of the model to the power grid can provide meaningful guidance for the regulation and operation of the power grid.

Methods: Feature extraction is conducted using Capsule Networks augmented with attention mechanisms to emphasize crucial spatial regions within the iris. Extracted features are subsequently fed into a Support Vector Machine (SVM) classifier for iris classification. Additionally, data augmentation and transfer learning techniques are employed to enhance model generalization. This will encourage additional uses of this technology in banking, security, and immigration control, among other areas. The next generation of iris recognition systems should be highly accurate and robust as research progresses due to the synergy between deep learning and practical deployment techniques.

Results: The efficacy of the proposed approach is evaluated through cross-validation on benchmark datasets, demonstrating promising results in terms of accuracy and robustness. While comparing the performance of different iris recognition techniques, including IrisParseNet, MADNet, ACNN-PCA, and the proposed CapsNet-AM-SVM approach.

Discussion: We find that CapsNet-AM-SVM consistently outperforms other methods, achieving the highest accuracy of 97.5% with 100 test images.CapsNet-AM-SVM yields an impressive precision of 99.0% in visible spectrum iris recognition. CapsNet-AM-SVM consistently exhibits elevated recall, achieving a notable 97.0% with 100 test images. CapsNet-AM-SVM consistently demonstrates elevated F1 scores, achieving an impressive 98.0% with 100 test images.

Conclusion: Thus, the results show that the proposed method accurately captures discriminative spatial features and mitigates the effects of illumination variations and occlusions.

Methods: The study went through steps such as system construction, user experience testing, cognitive load assessment, task completion, and system stability verification.

Results: The results showed that the system received high evaluations regarding user experience, task completion, and system stability, but cognitive load still needed to be optimized. The experimental comparison also found that the VR cognitive test system had a cognitive recovery degree of 30% to 60% for the elderly, while the traditional system was 10% to 25%.

Conclusion: From these data, it can be seen that the VR elderly cognitive test system does contribute to the cognitive health of the elderly.

Methods: The proposed method combines the Gabor Transform (GT) with local extrema, effectively capturing the directional extrema relationships between a reference pixel and its surrounding neighbors.

Results: The results demonstrate that the LEGP outperforms state-of-the-art techniques like LBP, LDP, LTP, and their Gabor-enhanced variants (GLBP, GLDP, GLTP), achieving a significantly higher recognition rate (RR).

Discussion: Existing approaches primarily use Gabor Transform (GT) features from the transform domain and Local Binary Patterns (LBP) from the spatial domain. In contrast, our proposed method combines both domains by introducing LEGP, a new spatial feature. While LBP captures pixel relationships, LEGP extracts directional extrema from Gabor transform responses.

Conclusion: The proposed Local Extrema Gabor Patterns (LEGP) descriptor enhances palmprint recognition by integrating the Gabor Transform with local extrema analysis to capture directional features. Evaluated on the HKPU Palmprint database, LEGP outperforms existing pattern-based methods in terms of recognition rate, achieving 99.53% accuracy with a processing time of 76.2 ms.

Methods: Based on the multivariate differential method, a dynamic analysis model of MMC- HVDC transmission considering the influence of outer loop control and limiting link is established. Based on the phase plane method and space vector diagram method, the influence of multiple control links and parameters of MMC-HVDC transmission on the transient power angle stability of receiving end power network are analyzed.

Results: A more accurate transient model of hybrid system was established, and the influence law of MMC-HVDC parameters on SG power angle was obtained through analysis. Based on this, a new control strategy was proposed to enable SG to restore stability more quickly.

Discussion: The significance of the research results is reflected in three aspects: first, a high-precision transient model including the power interaction term between SG and MMC is established, which lays a foundation for the study of stability mechanisms ; second, it reveals laws such as that MMC reactive current can reduce the acceleration area of SG and excessively high active current can narrow the stability boundary of SG, clarifying the direction of regulation; third, a coordinated control strategy based on SG power angle estimation and a current loop amplitude limiting strategy are proposed, and simulation verifies that it makes the system recover stability more easily than traditional control, providing a reliable scheme for engineering applications and contributing to the reliable operation of power systems.

Conclusion: The limitations of this paper are as follows: it ignores some control links, such as the current inner loop of MMC, and assumes an ideal scenario where the phase-locked loop (PLL) of MMC achieves real-time phase locking. Additionally, this paper only considers the coupling relationship between a single generator and is not applicable to the interaction analysis among multiple generators.