The immediate charging mode of electric vehicle battery swap stations can easily exacerbate the peak-valley difference in grid load and the imbalance in the utilization of battery energy storage resources,thereby restricting the economic operation of the swap stations and the security of the grid.To address this issue,this paper constructs a mixed-integer linear programming model with the objective of minimizing the cost of charging and purchasing electricity,while considering the constraints of battery state coordination and resource balance.It also proposes a battery charging optimization strategy based on dynamic time windows and battery collaborative management.Through the dynamic time window mechanism,the charging tasks are temporally and spatially shifted,and combined with battery state collaborative management,the overall charging load distribution is optimized while meeting the users' battery swap demands,reducing the cost of purchasing electricity and enhancing the security of the grid.Compared with the rapid response of the immediate charging mode,this strategy focuses more on system-level resource balance, providing a feasible solution for large-scale battery swap stations to participate in power demand response.Taking an actual battery swap station in a certain area as the verification object,the load regulation effect and comprehensive benefits of different time window lengths were evaluated.The results show that this strategy significantly improves load balance and reduces the cost of charging and purchasing electricity.In addition,through sensitivity analysis,the intluence laws of key parameters on system performance were revealed,providing a practical optimization method and decision support for large-scale battery swap stations to participate in power demand response,which is of great significance for promoting the coordinated development of the electric vehicle industry and the grid.

Aiming at the simulation problem of electric vehicle charging load in time and space distribution,this paper proposes a simulation method of electric vehicle charging load based on geographic information and multi-level trip chain.By establishing a regional road network model and dividing functional blocks,the vehicle movement trajectory is combined with road information to construct a complete mobile chain model of travel starting point-multi-destination stay-return path'.The path planning algorithm is used to simulate the user's daily travel mode,and the space-time simulation model of charging demand is constructed.The results show that this method can effectively simulate the charging load change trend of different regional types and urban functional areas,and accurately identify the electricity consumption characteristics of residential areas,work areas and other scenes.Compared with the traditional single trip chain model,the multi trip chain model constructed in this paper can more accurately simulate the spatial and temporal distribution characteristics of electric vehicle charging load.

As an environment-friendly vehicle,electric vehicle has achieved large-scale development,at the same time,the access of a large number of charging loads has caused a great impact on the traffic network and distribution network.Due to the lack of charging infrastructure and the disorderly charging of car owners,car owners may also face the problem of waiting in line for a long time during the charging peak.For this reason,an electric vehicle booking charging guidance strategy considering dynamic traffic information and available charging resources of charging stations is proposed in this paper.Firstly.the user travel characteristics.dynamic traffic network.charging station and distribution network are modeled；secondly,a global optimal path planning algorithm considering cross-period road resistance update is proposed；then,based on the maximum acceptable load of charging stations in the region,a hierarchical adjustment model of dynamic service charge is established according to the real-time load rate of each charging station.Finally,combined with the charging needs of car owners,a reservation charging guidance strategy aiming at the lowest charging cost,total travel time and slow charging cost is established,and the optimal charging scheme is planned for car owners.Ihe simulation results show that the proposed reservation charging guidance strategy can effectively improve the charging satisfaction of car owners,reduce the load variance between charging stations,stabilize the load fluctuation of distribution network,and maintain the safe and stable operation of distribution network.

In response to the issue of large-scale failures in power grids with a high proportion of new energy caused by severe sandstorms and the rapid recovery after a significant decline in the output of new energy units, this paper proposes a collaborative optimization method for line fault repair and new energy unit cleaning aimed at enhancing the resilience of transmission networks. Firstly, a high-proportion new energy grid model incorporating concentrated solar power plants and energy storage systems is constructed, which can effectively characterize the dynamic changes in the output power of concentrated solar power plants and energy storage after load variations. Then, the impact of severe sandstorm weather on new energy output and transmission line faults is quantitatively characterized. Further considering factors such as system grid strength and source-load imbalance, a post-disaster transmission line repair and unit cleaning sequence model is proposed with the objective function of maximizing system load recovery. Finally, the effectiveness of the proposed method is verified through an improved IEEE-30 node system, demonstrating that the proposed resilience enhancement strategy can accelerate system recovery, reduce system load loss, and achieve resilience enhancement in power grids with a high proportion of new energy.

Driven by the "dual-carbon"strategic goals,the penetration of renewable energy in power grids continues to increase,placing unprecedented pressure on the regulation capability of power systems.Most existing flexibility assessment and planning methods for power systems focus primarily on supply-demand balance,while paying insufficient attention to the role of transmission capacity in ensuring flexibility.To address this issue,this paper proposes a joint storage-transmission planning method that considers both resource and transmission flexibility margins.First,starting from the system operation mechanism,the influence of transmission network capability on meeting flexibility requirements is analyzed,and the node-level flexibility demand is mapped onto transmission lines.On this basis,two quantitative indices-resource flexibility margin and transmission flexibility margin-are constructed to enable multidimensional flexibility assessment.Subsequently,a joint storage-transmission planning model incorporating both resource and transmission flexibility margins is developed,in which energy storage deployment and transmission expansion are treated as coordinated decision variables to achieve a balance between flexibility enhancement and economic optimization.Finally,simulations based on an improved IEEE RTS-24 bus system are conducted.The results demonstrate that wc piopuscu icuou call mpiuve ucAwuy iaigis wic coiuvig wolal sysicm cust,cuccuvcly aucviaung upciauvnal stress caused by renewable energy fluctuations,and enhancing the accommodation capability of new energy sources.

In the high-proportion renewable energy (High Proportion Renewable Energy, HPRE) power system, thermal power units are largely replaced, and the contradiction between supply and demand of system daily regulation capacity is becoming increasingly prominent. Configuring multi-time scale regulation resources on demand can effectively alleviate the above contradiction. However, the strong uncertainty and volatility brought by large-scale wind power and photovoltaic access make it difficult to evaluate the daily regulation capacity requirements of the system. In this paper, a multi-time scale assessment method for daily regulation capacity demand is proposed for HPRE power system. Firstly, based on the perspective of 'adjustable' matching 'non-adjustable' intra-day power balance, a scenario set generation method is proposed to describe the uncertainty of non-adjustable power through Gaussian mixture model (Gaussian Mixture Model, GMM) and K-means clustering. Secondly, according to the hourly fluctuation characteristics of daily non-adjustable power, a multi-time scale evaluation system of daily regulation capacity demand is established. Then, a multi-time scale decoupling evaluation model of daily regulation capacity demand based on Haar wavelet coefficient is constructed by quantifying demand intensity with regulation power and regulation electricity as indicators. Finally, the example results show that the proposed method can provide some reference for the planning and allocation of multi-time scale adjustment resources.

With the increasing number of electric vehicles,the substantial charging demand has imposed certain impacts on the power grid while introducing challenges for grid dispatching.Consequently,predicting the charging load of electric vehicles has gradually become a research hotspot.To address the randomness of electric vehicle charging loads and the accuracy issues in deterministic predictions,an interval prediction method based on FMD-LSTM-Bootstrap is proposed.Using real historical electric vehicle charging load data collected,the dataset is preprocessed to eliminate erroneous data caused by equipment factors.The processed data is then divided into training and testing sets according to a certain ratio.Aiter tuning with the training set,the testing set is used to validate the prediction pertormance of the FMD-LSTM-Bootstrap method.The performance of the prediction intervals is further verified by combining the average bandwidth and interval coverage rate.Simulation results demonstrate the superiority of the proposed algorithm.

Wind, solar resource-rich areas usually appear a variety of new energy (wind, solar) co-generation, in-depth study of wind-solar joint power fluctuation characteristics of power grid safety operation is of great significance. In this paper, based on the measured data of large-scale wind power base in the northern part of Shanxi Province, the time-frequency characteristics of wind and light co-generation power are empirically analyzed by using db6 discrete wavelet transform. First, the wavelet transform is used to decompose the output power of different power generation clusters, and at the same time, the fluctuation rate, the extreme value difference and the fluctuation confidence interval index are calculated for each frequency band. Then, the integrated weight coefficient method is used to analyze the impact of the integrated fluctuation characteristic indexes of each frequency band on the power grid. Finally, based on the results of the time-frequency characterization of different wind farms in the northern part of Shanxi, the integrated fluctuation characteristic indexes of each frequency band of the wind-solar co-generation cluster are compared with those of the single wind/photovoltaic cluster, and the optimal capacity allocation ratio of the wind-solar co-generation cluster is obtained. The results show that compared with single wind/photovoltaic power generation cluster, the overall volatility of wind/photovoltaic co-generation cluster is smaller, which is more conducive to the safe operation of the power grid, and there exists an optimal ratio of wind/photovoltaic installed capacity.

Aiming at the problems of large demand side load fluctuations, high carbon emission base, and conflicting interests of multiple parties in the integrated energy system of multi microgrids with electricity and hydrogen energy interaction, this paper proposes an optimization of the integrated energy system of multi microgrids with zero carbon target considering self elasticity coefficient and peak shaving fluctuation incentives. Firstly, the benefit objectives and decision variables of each interaction party are analyzed, and an upper level cooperative game framework is constructed with the energy system operator of multi microgrids as the leader, and a lower level one master multi slave game model with user aggregator and shared energy storage operator as followers. Secondly, considering the dispersion coefficient of demand side loads, Pearson correlation coefficient with regional distribution networks, and load self elasticity coefficient, an incentive based and price based comprehensive demand response model is constructed. Finally, genetic algorithm was used to solve the problem, and simulation analysis was conducted. The comparison of the four schemes showed that the proposed strategy not only considers the benefits of multiple agents under the constraint of zero carbon emissions, but also improves the ability of demand side load fluctuations and grid peak shaving and valley filling through the reduction of system dispersion coefficient and Pearson correlation coefficient.

Aiming at the current mismatch between new energy output and base load,a set of multi-level refined dynamic peaking compensation mechanism is proposed；an additional dynamic incentive mechanism is proposed between aggregators and grid trading centers to set the peaking compensation price according to the peaking shortfall, which guides the aggregators to integrate the dispatchable resources to provide peaking auxiliary services for the grid； and a response compensation apportionment strategy is proposed between the aggregators and the electric vehicle users.It also proposes a response compensation sharing strategy between aggregators and EV users,which defines the concept of sharing coefficient from the dimensions of dispatchable capacity and load transfer demand during the access time to explore the dispatch potential of the user group.At the optimization level of charging and discharging scheduling scheme,Adjusted Grey Wolf Optimization （AGWO）is used to solve the control model with the objectives of maximizing the revenue of EV aggregator and minimizing the load fluctuation of the system；combined with the user dispatch response model,it is proved that the proposed strategy can tap the potential of EV dispatching according to the demand of grid,which makes the management more flexible.Combined with the user dispatch response model,the proposed strategy proves that it can exploit the potential of EV dispatch according to the grid demand,manage vehicle charging and discharging more flexibly,reduce the charging cost of EV users,increase the revenue of aggregators, reduce the net load fluctuation,and increase the rate of new energy consumption.

A high percentage of new energy is accessed, which decreases the system's inertia and speeds up response times. This can lead to several issues, such as poor unit-tripping regulation strategy and an erroneous evaluation of the system stability boundary. The traditional unit-tripping regulation strategy is no longer sufficient for the current power system. To ensure quick stable operation after optimal unit-tripping regulation, provide a boundary energy-based power system transient stability unit-tripping strategy. First, define the individual machine energy function to determine the system's boundary energy. through calculating and analyzing the boundary energy, assess the system's stability using the stability index; Second, create the generator sensitivity index and determine how the unit's active power and stability index relate to one another, and through the calculation of the sensitivity index, screen out the units that need to be regulated as a priority. Use the optimization model to calculate the minimum required unit-tripping amount to restore system stability, and take corresponding unit-tripping operation based on the judgment results; Finally, it has been verified that this method can effectively regulate unit-tripping quantity to achieve the goal of restoring stable system operation in actual power system.

To suppress inter-area low-frequency oscillations in power systems containing doubly-fed induction generator (DFIG)-based wind turbines, this paper proposes a wide-area supplementary damping control strategy based on reinforcement learning. A multi-machine system model incorporating DFIG dynamic characteristics is established, and a wide-area supplementary damping control loop is designed using inter-area tie-line power as the feedback signal. The deep deterministic policy gradient (DDPG) algorithm from reinforcement learning is employed to construct an offline parameter training mechanism for the supplementary damping controller. The trained intelligent agent enables dynamic parameter tuning of the controller, addressing the limitation of fixed parameters in conventional supplementary damping controllers. Simulation results demonstrate that compared with traditional methods, the proposed strategy can online update parameters in real-time according to system states under both three-phase short-circuit fault and synchronous generator step disturbance scenarios. This approach effectively enhances system damping characteristics and successfully suppresses low-frequency oscillations.

As China’s electricity market advances toward the full settlement operation of the spot electricity market and the frequency regulation (FR) auxiliary service market, this study investigates how different clearing mechanisms influence power resource allocation, renewable energy absorption, and overall economic performance. To this end, a sequential clearing strategy and a joint clearing strategy for the spot-FR market framework with energy storage participation are proposed, and conducts an analytical assessment of these aspects through case-based simulations. First, a spot-FR market trading system with multiple market participants is constructed, and the corresponding market organization processes are proposed. Second, separate clearing models for the spot electricity market and the FR market are formulated, upon which the joint clearing mechanism for the two markets is designed together with the price-settlement methodology. Finally, simulation cases are established to obtain the energy and FR capacity clearing results, market clearing prices, and renewable absorption performance under both clearing modes. The results demonstrate that, compared with sequential clearing, the unified optimization in the joint clearing mechanism enables more efficient allocation of generation resources across the two markets, significantly enhances the utilization of energy storage, maintains a high level of renewable energy absorption, reduces overall system procurement costs, and markedly improves the economic benefits of energy storage.

To address the issues of muddy sites and construction organization difficulties in summer,as well as low-temperature curing challenges in winter for the foundation construction of water field transmission towers in Northeast China,a prefabricated modular concrete foundation was designed.Field tests were conducted to measure the uplift bearing capacity of cast-in-place foundations and prefabricated modular foundations under combined vertical and horizontal loads.The study found that to reduce the cost of modular foundations and enhance lifting safety,it is necessary to first design a cast-in-place flexible foundation,which can then be divided into modular foundations.Under vertical and horizontal loads,the uplift bearing capacity of the modular foundation is 181 kN vertically and 12.5 kN and 12.3 kN horizontally,both lower than the 190 kN vertically and 12.8 kN and 12.6 kN horizontally required for cast-in-place foundations,but higher than the required vertical 158.07 kN and horizontal 10.57 kN and 8.78 kN.The relevant research findings can serve as references for the design and construction of prefabricated concrete step foundations,and can also be used for numerical simulation verification purposes.

To address the issues of the Supervisory Control and Data Acquisition (SCADA) system in wind farms, such as a large number of data acquisition parameters, strong coupling, and response time delay, this paper proposes an early warning model for wind turbine gearbox oil temperature based on delay time estimation. This model integrates the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Bayesian Optimization Algorithm (BOA), and Long Short-Term Memory (LSTM) network, establishing a multi-sensor collaborative method for temperature prediction and early warning of wind turbine gearboxes. Firstly, based on maximal mutual information, the delay time between input variables and output is estimated, and the time-series data is reconstructed to improve prediction accuracy. Secondly, the improved LSTM network model is used to realize online prediction of oil temperature, and this prediction model is extended to other temperature measurement points in the wind turbine gearbox. Finally, by analyzing the statistical characteristics of prediction residuals of the gearbox’s drive-end bearing temperature, oil temperature, and non-drive-end bearing temperature, a multi-sensor collaborative dynamic threshold early warning mechanism is constructed. Experimental results show that compared with traditional prediction models, this method can more effectively realize online monitoring and early warning of the gearbox’s operating status, providing effective support for the health monitoring and safe operation of wind turbine gearboxes.

The accuracy of predicting ice phase changes on transmission lines is influenced by the interaction of various meteorological factors.However,existing models,which ignore the interactions between multiple physical fields, result in insufficient prediction accuracy and other issues.To address this,a multi-physical field coupling ice phase change prediction model has been developed.First,by setting up multi-physical field coupling boundary conditions,the model simulates various meteorological conditions to create a three-dimensional geometric model of ice phase changes on transmission lines.Then,different kernel functions are selected to verify the reliability of the support vector machine（SVM）prediction algorithm.Finally,the multi-physical field coupling model is used for ice phase change prediction experiments,and its effectiveness is verified through comparison with actual measurement data and SVM prediction data.The experimental results show that the model has an average prediction accuracy of 98.12%and an average precision of 98.54%，significantly improving upon traditional methods.This model provides high-precision decision support for the early warning and protection of ice disasters on transmission lines,making it highly valuable for engineering applications.

To address the influence of deterioration state difference of wind turbines and output uncertainty on the overall maintenance schedules, an optimization method for maintenance scheduling of wind farms considering both power output and failure uncertainties is proposed. Initially, a large amount of power output scenarios are generated using Monte Carlo sampling based on the Weibull distribution of wind speed, and the Kantorovich distance is applied for scenario reduction, resulting in a set of typical output scenarios that capture the uncertainty in wind power generation. Secondly, a state enumeration method is employed to generate a set of wind farm failure scenarios, which incorporate the impact of maintenance strategies on the failure probability of wind turbines. Finally, an optimization model for wind farm maintenance scheduling is developed, aiming to minimize the overall costs consisting of preventive maintenance cost, post-failure repair cost, and the outage power loss cost due to maintenance or failure of wind turbines. Case studies demonstrates that the proposed maintenance schedule accounting for output and failure uncertainties results in a 1.6% and 5.4% reduction in wind farm operation and maintenance costs compared to that only considering failure uncertainty or output uncertainty, respectively.

In the research of traditional distribution network,reactive power optimization and network reconfiguration are usually discussed separately,and there is a lack of effective coordination of optimization technology.Based on this situation,a bi-level optimization mathematical model framework for reactive power compensation and network reconfiguration of distribution network is constructed.The construction goal of the upper reconstruction model is to minimize the active power loss and voltage offset of the distribution network.The model also takes into account the coordination and control between the new power supply access and the reactive power compensation device.The establishment of the lower-level optimization model takes the minimization of the total network loss and the voltage offset amplitude as the core objective,and the model fully integrates the related technologies of reactive power optimization.Then,the moth-flame optimization algorithm is improved by integrating multiple strategies to promote the rapid solution of the two-layer mathematical model.Finally,the IEEE 33-node distribution system is selected as an example.The analysis results show that the bi-level optimization model can significantly optimize the voltage distribution of the distribution network and effectively reduce the active power network loss.At the same time,the improved algorithm also shows its superior performance in terms of convergence rate and global optimal solution search ability.

Energy system resilience is a crucial guarantee for the energy transition and the achievement of the "dual carbon"goals.This study focuses on the period from 2012 to 2023 and employs the entropy-weighted TOPSIS method to establish an evaluation model from four dimensions:energy supply,energy consumption,energy technology,and energy environment,conducting a comprehensive assessment of the energy system resilience in Jilin Province.The results indicate that the overall resilience of the energy system exhibited a trend of "initial decline followed by an increase，"surpassing the initial level by 2023.Specifically,supply resilience,constrained by the scarcity of traditional energy resources,has not fully recovered.Consumption resilience underwent a three-stage evolution:stability,decline, and recovery.Technological resilience saw significant improvement due to the large-scale application of new energy and smart grid technologies.Environmental resilience showed a fluctuating rise before stabilizing into a dynamic equilib-rium.The study concludes that the improvement in Jilin Province's energy system resilience is largely attributable to the optimization of the energy structure and technological breakthroughs brought about by major projects such as the "onshore wind and solar power base."Jilin Province must continue to promote clean energy transition and technolo-gical innovation to consolidate and enhance the comprehensive resilience of its energy system.

When unloaded cable is put into operation,medium frequency oscillation will likely occur in direct drive wind farm with Static Var Generator （SVG），which can seriously threaten the stability of wind farm.In order to analyze this problem from the perspective of nonlinearity of grid-connected converter,firstly,considering the control of the inner and outer loops of grid-side converter the wind turbine,the SVG,and the impedance of the unloaded line,a state-space model of the wind farm system is established.Based on the characteristic root distribution of the state-space matrix,it is confirmed that the interaction between the wind turbine,SVG,and unloaded line leads to medium frequency oscillation；Then,using the participation factor method combined with the medium frequency oscillation mode,it is confirmed that the SVG current inner loop cannot be ignored in system modeling；To determine the key factors affecting medium frequency oscillation in the system,the typical state variables is selected which can characterize medium frequency oscillation,and phase space trajectories is also drawn,and the impact of wind turbine power output,system strength,and parameter changes in SVG inner and outer loop control is analyzed on medium frequency oscillation.Research shows that medium frequency oscillations are prone to occur in wind farm containing SVG under unloaded line operation,strong wind,and weak power grid conditions.At the same time,the proportional coefficient of the inner and outer loops of SVG has a significant impact on medium frequency oscillations.Finally, simulation verifies the correctness of theoretical modeling and analysis.

With the continuous growth of China’s new energy installed capacity, the power system is always facing the problem of new energy consumption and coordination. In order to further explore the advantages and potentials of the coupling and integration of new energy and thermal power, and to improve the level of new energy consumption, the paper proposes the integrated coupling of new energy and thermal power as a unified whole model to participate in the monthly market centralized bidding from the perspective of the monthly market. Based on the quotation method of new energy and thermal power independently participating in the market transaction, the unit generation cost model of the whole life cycle of the coupled system is established. On this basis, the cost components of the coupled system and single thermal power unit are compared, and the impact of baseline changes under the carbon market on the generation cost of the coupled system as well as the order of market transaction clearing is analyzed. Finally, considering the impact of the enterprise’s unit generation cost and the user’s inverse demand characteristics on the centralized market offer, the coupled system monthly centralized bidding market high and low matching clearing model is established, and the market transaction clearing is simulated. The clearing results show that the coupled system and thermal power units together participate in the monthly centralized market, has the competitive advantage of power, the volume of power traded accounted for the largest market share, and the coupled system to participate in the market transaction of the average rate of return to meet the requirements of the power generation enterprise profitability.

Aiming at the problems of increasing volatility of generation output and expanding peak-to-valley ratio of load curve in the new power system,a method of designing tariff package based on complete rationality of users under the new power system is proposed to guide the users to participate in shaving and filling in the peaks and valleys of the system load.Firstly,the characteristic values of electricity load of power users are extracted,and the two-stage clustering method is used to generate typical user curves.Secondly,Latin hypercube sampling technology is used to generate new energy output under three typical days of spring,fall,summer and winter,calculate the net load of the system in combination with the load demand,and use fuzzy theory to divide the net load curves of the three typical days into peaks and valleys and equal segments.Then,the demand response model of fully rational users is established and the user utility is calculated,on the basis of which the Logit discrete choice model is used to calculate the proportion of tariff packages chosen by users in a fully rational situation.Finally,with the goal of realizing a win-win situation for both the electricity sales company and the users,a model of the tariff package that takes into account the fully rational users in the context of a new type of electric power system is established.The results of the example show the effectiveness of the proposed electricity price package formulation method.

While the gas-charged heat pipe has high thermal conductivity,it can also achieve the function of variable thermal conductivity by adjusting the effective working length according to the temperature of the evaporation section.In the past decade,it has been gradually applied in systems such as mines,boilers,and aerospace thermal control systems.Among them,the starting temperature and allowable operating temperature of the potassium heat pipe perfectly match the inlet and outlet temperatures of the gas-steam combined cycle unit,and it can be used for the thermal protection of downstream equipment under the accident conditions of the turbine.However,due to the relatively late start of the research on variable thermal conductivity heat pipes and the complex manufacturing process,the research achievements are not sufficient to establish a complete theoretical system.Currently,most of the studies are simplified based on the front model,but this will cause relatively large errors when analyzing gas-charged heat pipes with a medium length-diameter ratio.In order to correct and improve the theoretical model of the gas-charged heat pipe, potassium-argon gas-charged heat pipe samples with absolute pressures of 500 Pa and 1250 Pa were fabricated,and a conventional potassium heat pipe was additionally set as a control group.Steady-state characteristic experiments and cold start experiments were carried out in the atmospheric environment.The experimental results show that the working part of the gas-charged heat pipe has good temperature uniformity,and the non-condensable gas does not affect its normal operation.Moreover,compared with the conventional potassium heat pipe,it has better cold start ability.Finally, a mathematical model of the gas-charged heat pipe was established based on the experimental data,and the average content of potassium vapor in the working area at different temperatures was obtained.

High-Frequency Transformer (HFT) is a core component of power electronic transformers and dual-active-bridge DC-DC converters. The distribution characteristics of its temperature and stress fields are crucial for system reliability and environmental friendliness. In this study, a 5kHz/10kVA HFT was analyzed using finite element simulation software. A full-chain multiphysics coupling simulation covering electromagnetic, fluid, thermal, structural, and acoustic domains was conducted to obtain the distribution patterns of electromagnetic, temperature, stress, and other multiphysics fields under nonsinusoidal excitation. The results reveal that the core loss of the high-frequency transformer is 52.7W, with high-voltage and low-voltage winding losses of 13.23 and 3.01W, respectively. The maximum magnetic flux density under sinusoidal excitation reached 0.746T, while under square wave excitation it reached 0.923T. The hotspot temperature was located approximately two-thirds along the core column at 115℃, with the lowest temperature recorded at 92.3℃ beneath the high-voltage winding. Forced air cooling demonstrates significant cooling efficiency below 4m/s wind speed, with cooling rates slowing thereafter. At 4m/s wind speed, core stress amplitudes under sinusoidal and square wave excitation were 2.81×106 and 3.65×106N/m2, respectively, with sound pressure level amplitudes of 78.1 and 94.9dB. When further considering the effect of thermal stress, under natural air cooling and square wave excitation, the core stress rises to 3.92×106N/m2, the vibration acceleration increases to 64.3m/s2, and the maximum sound pressure level reaches 97.3dB, representing increases of 7.4%, 6.8%, and 2.4dB, respectively, compared with the forced air cooling condition. Owing to its rich harmonic content, square wave excitation leads to significantly greater deformation, stress and noise than sinusoidal excitation, while thermal stress, as a secondary factor, further exacerbates vibration and noise levels. Through multiphysics simulation, the influence mechanism of waveform excitation and thermal stress on the comprehensive performance of transformers is revealed, providing a theoretical basis for multi-objective optimization of high-frequency transformers.