With the development of new energy and power electronics, new power systems exhibit typical "dual high" characteristics, but at the same time, they also bring problems such as low inertia and weak stability. Grid forming control technology （GFM） can enable the inverter to independently establish frequency and control voltage, improve dependence on external voltage sources, and improve system stability. The current situation and development trend of grid type converters are summarized. Firstly, a comparative analysis was conducted between grid following control technology （GFL） and grid building control technology, and it was pointed out that grid building control technology is an effective way to solve the stability of new power systems. Then, the existing grid control technology was introduced, and the application of grid control technology in various fields were analyzed. Finally, the challenges faced when applying grid control technology to converters were pointed out.

In view of the decline of frequency stability in the power grid caused by the shortage of frequency regulation resource, this paper takes electrolytic water hydrogen production system as the research object, through establishing a refined dynamic model and the targeted control strategy, so that electrolytic water hydrogen production system can provide fast frequency response, which fully exploits the frequency regulation potential of electrical, hydrogen and heat energy within the electrolyzers system. In addition, the mechanism and ability of the fast frequency response provided by the electrolyzers are analyzed. The results show that the established model can reveal the electric-hydrogen-thermal coupling mechanism and characterize the frequency response characteristics of the electrolyzers system during the fast frequency response. The fast frequency response provided by the electrolyzers system can improve the frequency stability of the power grid under multiple disturbance scenarios.

In response to the significant randomness and uncertainty issues in photovoltaic power, this paper proposes an improved LSTM based photovoltaic power prediction method to improve the accuracy of photovoltaic power prediction. Firstly, meteorological features with strong correlation with photovoltaic output were analyzed, and the t-distribution nearest neighbor embedding dimensionality reduction technique was used to reduce the selected feature data to 2D to reduce data complexity. Then, the reduced dimensionality data is automatically clustered into three categories through density peak clustering to help train the LSTM prediction model. Compared with the traditional Recurrent neural network and Long short-term memory neural network models, the model proposed in this paper shows higher prediction accuracy in photovoltaic power prediction. MSE decreases by 49.00% and 31.77%, RMSE decreases by 28.59% and 17.41%, and MAE decreases by 62.35% and 53.52%. The research results indicate that the model has good applicability in photovoltaic power prediction, providing valuable reference for the optimization and scheduling of integrated energy systems.

A photovoltaic array composed of YL250P-29b type polycrystalline silicon solar cells packaged in series at our school is taken as the research object in order to improve the efficiency of photovoltaic power generation. The rationality of the numerical model was verified by comparing the experimental data and simulation results of artificial ash accumulation. Based on this model, the effects of wind speed, irradiance, ambient temperature and ash density on the temperature of photovoltaic modules are simulated and analyzed. According to the empirical formula of photoelectric conversion efficiency and output power, the influence of the above environmental factors on the output characteristics of photovoltaic modules was explored. The results show that there is a negative correlation between PV module temperature, photoelectric conversion efficiency with wind speed, and positive correlation with irradiance, ambient temperature and ash density. There is a negative correlation between the output power of photovoltaic modules with ambient temperature and ash density, and a positive correlation with wind speed and irradiance. Within the scope of this study, the correlation order between each environmental factor and PV module temperature and photoelectric conversion efficiency is： ambient temperature> irradiance>wind speed>ash density, and the correlation order with output power is： irradiance>ash density>ambient temperature>wind speed; For every 1 ℃ increase in ambient temperature, the temperature of photovoltaic modules also increases by about 1 ℃, and the photoelectric conversion efficiency and output power decrease by about 0.06% and 0.4%, respectively. Ash accumulation reduces the temperature of photovoltaic modules and increases the photoelectric conversion efficiency, but it will greatly reduce the output power of photovoltaic modules.

Relying on the existing engineering experience data or the short-term heat transfer test results of middle and deep borehole heat exchanger, an extensive method is typically used to design the middle and deep borehole geothermal heat pump heating system. It leads to the unreasonable configuration of the borehole heat exchange system and the reduction of the reliability and economy of heat pump heating system. In order to improve the design status of the middle and deep borehole heat exchange system, a dynamic simulation method is adopted in this paper, considering the design load, cumulative load and hourly load distribution law, to analyze the unsteady characteristics of deep-buried geothermal resources in the long-term dimension and the temperature-self-recovery performance of rock and soil, and determine the designed water supply and return temperature at the source side based on the "annual periodic cooling and heat balance of final state" of underground rock and soil. And optimization method of source-side design flowrate based on the optimization of the average annual life-cycle cost on the ground source side is proposed. Taking an engineering project as an example, the optimization design of source-side parameter is given.

Addressing the impact of nonlinear characteristics such as dead-zones on the power quality of photovoltaic grid-connected inverters, this paper combines data-driven compensation methods with traditional control to investigate a dynamic and static characteristic optimization approach for grid-connected inverters. Firstly, a repetitive controller is utilized as the basis for online data training, elucidating the mechanism and validity of the data source. Secondly, an approximate linear regression method is employed to obtain a data model, reducing the dependence on storage space for data-driven methods, ensuring necessary compensation bandwidth, and solving the feasibility of data application. This model is then applied to the compensation loop of a traditional low-order controller, enabling the system to achieve precise control with sufficient stability margin. Data correlation analysis and experimental results demonstrate the feasibility and effectiveness of this compensation method.

In order to change the traditional operation mode of “source with load” and increase energy storage, and to realize the coordination and interaction of power grid, load, energy storage and other links, this paper establishes an electric-thermal-hydrogen coupled integrated energy system （ETHC-IES） for optimal scheduling. The use of hydrogen energy storage to realize the safe and stable operation of a new type of integrated energy system “source network storage” has become a current research hotspot. The aim of this paper is to reduce the operation cost of IES and the waste of wind and light. The ETHC-IES optimal scheduling problem is transformed into a Markov decision process （MDP）, and an optimal scheduling method based on continuous action proximal policy optimization （PPO） is proposed for the integrated energy system. Firstly, a mathematical model of each part of the electric hydrogen storage system is established. Then, the model is solved using a deep learning proximal policy optimization algorithm with the optimization objectives of economy and reduction of wind and light waste. The action space, state space, and reward function of the deep reinforcement learning model are set up. Intelligent agents are trained and learned to achieve dynamic scheduling optimization decisions for ETHC-IES. Finally, the effectiveness and applicability of the proposed model and method are verified by simulation.

A method is proposed for the packaging aging of IGBT modules to be monitored, based on the combination of EMR and V_{eE_peak} electrical parameters, with the aim of the health status of IGBT modules being monitored when multiple aging events occur simultaneously. Firstly, the generation mechanism of V_{eE_peak} and radiated interference signal （EMR） and the influence of parasitic parameters inside the module on V_eE and EMR were analyzed. Secondly, the impact of bond wire lift-off and solder layer voids on the internal parasitic parameters of IGBT modules is analyzed. Finally, the feasibility of this method was verified through experiments. This is a non-invasive monitoring method that does not require complex monitoring circuits, which can effectively reduce the errors caused by various aging coupling on the health status monitoring results of IGBT modules.

Aiming at the problems that the fault features of the main bearings are interfered by the harmonics of other transmission components in wind turbines' condition monitoring and the raw vibration signal is non-stationary, making it difficult to accurately extract the characteristic frequency of fault impacts, a combined method of harmonic removal and spectral kurtosis analysis （HR-Fast-Kurtogram） is proposed. Offline angle domain resampling is used to convert non-stationary time domain signals into stationary angle-order signals. Comb notch filters are introduced in the order domain to remove gear meshing frequencies, and Fast-Kurtogram based on short-time Fourier transfer （STFT） is applied to extract the fault sub-bands where the kurtosis peak is located in order to extract bearing fault features. The combined method is able to avoid the interference of non-periodic impacts and distinguish between non-bearing amplitude modulation components with lower impact and bearing fault impact components. The comparative analysis results in the application of real main bearing fault data shows that the proposed method can improve the accuracy of fault features extraction and fault type identification in a short time scale of seconds.

Phase change heat storage technology is one of the efficient means of heat storage in engineering applications. However, phase change materials have the problem of low thermal conductivity, which seriously limits their work efficiency. Therefore, it is necessary to optimize the structure of phase change heat storage devices to improve heat storage efficiency. Na₂S₂O₃·5H₂O and CH₃COONa·3H₂O were selected as phase change materials. Six phase change heat storage models were designed based on bionic structure. The melting and solidification processes of the two phase change materials in the six heat storage models were simulated and analyzed. Finally, based on the numerical simulation results and taking into account the heat storage and release processes, a small biomimetic cross seasonal phase change heat storage experimental platform was built. This article presents temperature changes of two phase change materials at each temperature measuring point in two models under the operating conditions of 75 ℃ and 0.1 m³/h. The results indicate that one heat storage model is more suitable for using Na₂S₂O₃·5H₂O as phase change material, while the other model is more suitable for using CH₃COONa·3H₂O as phase change material.

A comprehensive review was conducted on the technology of hydrogen production from biomass thermochemical conversion. The review included an analysis of the influencing factors of hydrogen production through pyrolysis and reforming, gasification, and supercritical water gasification. These factors comprised raw materials, temperature, catalyst, etc., as well as the challenges faced, such as tar formation, catalyst deactivation, and input cost. It is suggested that the selection of hydrogen production methods should be based on the characteristics of the raw materials and the hydrogen production conditions. Subsequently, pretreatment and regulation of reaction conditions can be employed to enhance the efficiency of hydrogen production. A comparison of the three technologies indicates that gasification exhibits relatively higher efficiency. Additionally, gasification and pyrolysis are more suitable for biomass with low moisture content, while supercritical water gasification demonstrates advantages in the treatment of toxic wastewater. Moreover, future research directions encompass addressing tar formation, developing catalysts, and investigating equipment preparation.

Accurate assessment of urban rooftop PV development potential is one of the key problems to be solved for urban rooftop PV construction. Most of the existing studies are limited to the assessment of urban rooftop PV potential by a single data source, such as satellite remote sensing images or geographic information, which is difficult to accurately respond to the urban rooftop PV potential under the influence of multiple factors. Considering the shadow effect, geographic location and meteorological conditions, this paper proposes an urban rooftop PV enhancement method based on the fusion of multi-source information, which mainly includes three links： roof information extraction, regional building shadow calculation and PV power generation potential calculation, and a large amount of satellite remote sensing imagery, geographic information data, meteorological data and other data are introduced into the analyses of the links, which greatly enhance the accuracy of the assessment of the urban photovoltaic potential. The accuracy of urban PV potential assessment is greatly improved. Finally, the method is applied to Changsha City as an example, and the evaluation results show that the power generation potential of urban rooftop photovoltaic in Changsha City is 14480.4 GWh, which is a large potential for development.

Primarily, the current status of development for the hydrogen storage and transportation technology are reviewed in this paper, including the storage and transportation manners of gaseous, liquid, solid, and hybrid, respectively. Subsequently, based on the index requirements of on-board hydrogen storage by US Department of Energy, the comprehensive performance of varied hydrogen storage technologies is compared, and the advantages and disadvantages of the existing hydrogen transportation ways are summarized. Finally, five suggestions are put forward for the future research direction of hydrogen storage and transportation technology. Overall, new hydrogen storage cylinders with superior comprehensive performance and hybrid hydrogen storage technologies should be the main focus of current research in the field of hydrogen storage. Liquid hydrogen transportation and natural gas mixed with hydrogen transportation are significant research hotspots for the future and present, respectively.

Taking a large-scale photovoltaic power station in alpine and high-altitude area with flat uniaxial photovoltaic racking system as the background, relying on ABAQUS finite element simulation software, we constructed an integrated three-dimensional finite element model of large-span flat uniaxial photovoltaic bracketing-foundation-panel, explored the loading characteristics of the key components of the flat uniaxial photovoltaic tracking system and the mechanism of instability deformation in an extreme environment, and provided the targeted optimization design suggestions. The results show that, under extreme wind load, the maximum stress of purlins in this large-span PV system reaches 544.4 MPa and the maximum displacement of columns reaches 76.3 mm at the maximum tracking angle, resulting in the potential risk of yield damage of local purlins and H-type steel piles; the deformations of superstructure members are related to their locations, and the overall performance shows that they are distributed in a wavy pattern along the long axis of the main girder, the mid-span position of the two columns also has a large displacement. The maximum deformation under wind load is even 153.2 mm, which is most obvious at the cantilever end of the PV module, and it is proposed to enhance the strength of the local components or change the local structural characteristics to enhance the overall stability of the large-span flat uniaxial tracking PV system.

The principle and state-of-art of PV monocrystalline silicon slicing processing are reviewed herein. The core wire diameter of the diamond wire saw has been reduced to 37 μm. The as-cut half G12 wafer thickness of PV monocrystalline silicon has been reduced to 110 μm. It is elaborated that the main technical approaches for high wafer yield slicing are to reduce the as-cut wafer thickness and the diameter of diamond wire saw. The effects of surface crack damage and fracture strength of sliced silicon wafer, liquid bridge interaction at diamond wire saws and as-cut silicon wafers, and machine vision detection on the high wafer yield slicing processing of PV monocrystalline silicon are discussed. The key technologies faced in the high wafer yield slicing processing of PV monocrystalline silicon are proposed： 1） to develop low-cost tungsten core wire diamond wire saw; 2） to develop new coolants and lubricating technologies; 3） to develop new slicing processing technology; 4） to establish a quantitative relationship between the distribution of abrasives and the machining performance of electroplated diamond wire saw.

To effectively analyze and utilize the prediction errors distributed in a specific pattern in the photovoltaic power prediction model, a photovoltaic power prediction model based on LSTM-Attention and CNN-BiGRU error correction is proposed. Firstly, the LSTM-Attention mechanism is introduced to compensate for the shortcomings of the long short-term memory （LSTM） network, which is difficult to retain key information in the input sequence. Secondly, the advantages of convolutional neural network （CNN） in non-linear feature extraction are combined with the advantages of Bidirectional gated recurrent unit （BiGRU） in preventing multiple features from interfering with each other to build a CNN-BiGRU error prediction model is used to predict the possible errors and to correct the initial prediction results. The experimental results indicates that the CNN-BiGRU error prediction model can effectively improve the prediction accuracy in different weather types when compared with the prediction results without error correction.

Hydrogen fuel cell rail vehicle is a new type of rail vehicle powered by hydrogen as fuel. Its characteristics of zero emission, low noise, and high efficiency have significant social and economic benefits, and it is also one of the important ways for rail transportation to save energy and reduce emissions. This paper expounds the working principle and system composition of hydrogen fuel cells, as well as the research and development progress and trends of hydrogen fuel cell systems. Systematically combs the research and application status of hydrogen fuel cell rail vehicles at home and abroad, mainly summarizes the differences in terms of vehicle type, fuel cell power, battery life, maximum operating speed, power mode, hydrogen storage capacity and pressure, and analyzes the output characteristics of hydrogen fuel cell hybrid power mode and power matching in various environments , confirming that the hybrid power system formed by hydrogen fuel cells and power batteries/super capacitors can better adapt to various working conditions such as start-stop, idling, continuous high speed and climbing of rail vehicles. Finally, the research difficulties of hydrogen fuel cells, the storage and transportation challenges of hydrogen, and the development potential of hydrogen fuel cell hybrid power in rail vehicles are discussed.

Aiming at the problems of low accuracy and poor generalization of wind speed forecast due to randomness and volatility, a combined prediction model based on variational mode decomposition （VMD）, whale optimization algorithm （WOA）, long short-term memory neural network （LSTM） and sparrow search algorithm （SSA） was proposed. Firstly, WOA is used to automatically optimize the core parameters of VMD （K value and penalty coefficient α）. After decomposing the wind speed time series, SSA is introduced to optimize the core learning parameters of LSTM, and finally, the predicted wind speed data of each subcomponent is integrated to obtain the final predicted wind speed, which is verified by a number of model evaluation indicators. The RMSE, MAE, MAPE and R² of the model are 0.0758 m/s, 0.0578 m/s, 1.492% and 0.979, respectively. Compared with other single optimization prediction models WOA-VMD-LSTM and VMD-SSA-LSTM, the relevant evaluation indicators have significantly improved.

The research focused on reducing energy consumption in greenhouse buildings by designing a solar-ground source heat pump phase-change heat storage heating system （SGSHPP-CHSH） based on an existing solar-ground source heat pump heating system （SGSHPH） in a greenhouse located in Hebei Province. Simulation models of both systems were built using TRNSYS software, and the simulation results and influencing factors were compared and analyzed. The findings indicate that the solar collector heat collection of the SGSHPP-CHSH is 10.75% higher than that of the SGSHPH throughout the entire heating season. Additionally, the phase change storage tank's temperature decreased from 49.4 ℃ to 34.4 ℃, resulting in reduced energy consumption for the heat pump unit （from 10145 kWh to 7843 kWh） and increased COP of the heat pump system （from 2.68 to 3.36）. Similarly, for the SGSHPH, there was a decrease in water supply temperature from 45 ℃ to 30 ℃, reduced energy consumption for the heat pump unit （from 12837 kWh to 8739 kWh）, and an increased COP of the unit （from 2.39 to 3.43）. Furthermore, When the solar collection area of the SGSHPH is 48 m², and the solar collection area of the SGSHPP-CHSH is 39 m², the soil heat storage in both heating systems during non-heating seasons equals to the soil heat extraction during heating seasons, effectively maintaining a balanced thermal state within the soil.

With the continuous promotion of carbon peaking and carbon neutrality goals in China, to promote the consumption of renewable energy and reduce the carbon emission of the system, this paper proposes a green certificate-carbon emissions trading interaction model based on cooperative game theory for an integrated energy system from the perspective of joint efforts of power generation, carbon sequestration, and the market mechanism. The simulation is carried out in an integrated energy system in eastern Inner Mongolia. In this model, the generation side realizes the flexible interaction between the carbon capture power plant and the renewable energy generation entity with the advantage of centralized dispatching of the integrated energy system （IES）; the carbon sequestration side realizes the low carbon economic operation of IES through carbon capture equipment and power-to-gas equipment （P2G）; the market side improves the integral economy of IES through the establishment of the green certificate-carbon emissions trading interaction mechanism, and based on the Shapley value to reasonably allocate the cooperation surplus. The results show that the proposed method can significantly reduce the system power curtailment, make full use of deep peak-regulation of thermal power units, realize the low-carbon economic operation of IES, and provide a theoretical reference for exploring the feasibility of cooperation between the existing thermal power and renewable energy projects, and exploring a new way of coal-fired power plant flexibility transformation.

In order to make thermal power units better cope with the impact on the original power grid structure under the background of rapid development of new energy sources, and improve the stability, safety and economy of thermal power unit operation, based on the current research status at home and abroad, the lithium battery-flywheel control strategy and the regional dynamic primary frequency regulation model of thermal power units are proposed, and the capacity configuration scheme of flywheel-lithium battery hybrid energy storage system under a certain energy storage capacity is studied, and the simulation verification is carried out through Matlab/Simulink, Under continuous disturbance, the frequency fluctuation degree of the system is 0.00119 pu, the fluctuation amount decreases by 30.81%, the power fluctuation decreases by 43.65%, and the actual power contribution increases by 23.17%. The results show that when the thermal power unit is disturbed by external load, the frequency regulation of hybrid energy storage auxiliary thermal power unit effectively improves the operation stability and economy of thermal power unit.

The prediction of lithium battery lifespan is of great significance for energy management and maintenance. To address challenges such as complex multidimensional time series data, long-term dependencies, and dynamic changes in characteristics during the prediction process, this paper proposes a lithium-ion battery life prediction model based on dynamic convolutional neural networks and Transformer（DCF）, covariance matrix adaptive adjustment evolution strategy（CMA-ES）, and multi-head self-attention mechanism. The DCF dynamically extracts key features from time series data, reduces data dimensionality and redundancy, and captures long-term dependencies. CMA-ES optimizes model hyperparameters to enhance the model's ability to model local features and global dependencies. The multi-head self-attention mechanism further focuses on important features and handles complex nonlinear dynamic relationships. Experimental validation is conducted using the publicly available lithium battery dataset provided by NASA. Results show that the proposed method achieves a minimum average absolute error of 0.28%, outperforming most existing methods using the same dataset. The experiments further demonstrate improvements in prediction accuracy and generalization ability, especially in long-term lifespan prediction, where the model exhibits higher precision and robustness, providing more reliable technical support for lithium battery lifespan prediction.

A sub-synchronous oscillation suppression strategy for wind power transmission systems based on multi band adaptive notch filter （MANF） is proposed. Firstly, based on the mechanism analysis of sub-synchronous oscillation, the installation position of MANF in the rotor measurement converter （RSC） of doubly fed induction generator （DFIG） is determined. Secondly, the DFT identification method optimized by the recursive group harmonic power minimization （RGPM） algorithm is used to design a sub-synchronous oscillation detector, effectively solving problems such as mode mixing. Then, using the windowed sparsity adaptive matching pursuit （SAMP） algorithm, design a MANF that can update the notch frequency in real time based on the input frequency. Finally, the effectiveness of the suppression strategy was verified by comparing with the multi notch notch filter （MNF） and the single notch adaptive notch filter （ANF）. The robustness of the suppression strategy was verified under the conditions of considering different series compensation degrees, wind power partial cutting, and thermal power replenishment power.

Aiming at the problem of strong randomness of photovoltaic power data and difficulty in accurate prediction, a combined prediction method based on temporal convolutional network （TCN）, bidirectional long short-term memory network （BiLSTM） and echo state network （ESN） is proposed. Firstly, the complete ensemble empirical mode decomposition with adaptive noise analysis （CEEMDAN） is used to decompose the power data into a series of relatively stable sub power subsequences. Then, the decomposed and reconstructed power sequence and other feature sequences are input into the TCN-BiLSTM Attention-ESN. TCN-BiLSTM Attention-ESN is applied to extract features and then spatiotemporal feature vectors are constructed. Finally, the extracted spatiotemporal feature vectors are input into ESN to obtain the prediction results. The proposed method is validated using photovoltaic power data from photovoltaic power stations in Xinjiang, China. The results showe that compared with current advanced prediction methods, the proposed method has higher prediction accuracy, which helps to increase the proportion of photovoltaic power generation and ensure the balance and operation safety of the power system.

To enhance the direct solar-thermal conversion and storage performance of sugar alcohol-based phase change materials （PCMs） and promote their large-scale application in solar thermal energy storage, Xy-in-EG （94 ℃） and Er-in-EG （120 ℃） composite PCMs with high thermal conductivity and excellent light-to-thermal conversion efficiency were prepared using mechanical mixing and melting vacuum adsorption. Xylitol and erythritol served as PCM precursors, while expanded graphite （EG） was used as a thermal conductivity additive. The application of these composite PCMs in solar-thermal direct conversion and storage was systematically investigated. The results show that with 25% EG content, the maximum thermal conductivity of the Xy-in-EG and Er-in-EG composite PCMs reached 4.39 W/（m·K） and 3.61 W/（m·K）, respectively, which is approximately an order of magnitude higher than that of the original PCMs. When the EG content was 15%, the solar spectral absorptivity of the composite PCMs reached 84.3% and 86.9%, respectively. At this concentration, the composite PCMs achieved stable encapsulation, without issues such as phase transition leakage caused by phase separation during the heating and melting phase transition processes. Furthermore, in solar-thermal direct conversion and storage experiments, the solar-thermal conversion efficiency of Xy-in-15%EG and Er-in-15%EG composite PCMs increased by~63.3% and 79.5%, respectively, compared to commercial xylitol and erythritol PCMs. The photothermal storage rate of Er-in-15%EG was 3.2 times higher than that of commercial erythritol.

The research status of photovoltaic-green roof was expounded from the following four aspects： 1） Thermal and humid environment and vegetation state on the roof; 2） Power generation efficiency of photovoltaic modules; 3） Water balance of roof; 4） Energy balance and thermal process model of roof. The results show that the interaction between PV modules and vegetation varies dynamically under different climate and spatial configuration conditions, which leads to uncertainties in the comprehensive performance of the roof. Finally, the engineering challenges faced in the application of photovoltaic-green roof and future research directions was analyzed and prospected. And it is pointed out that understanding and applying the interaction mechanism between photovoltaic modules and vegetation lays the foundation of accurate prediction and optimization of comprehensive performance of photovoltaic-green roof.

A waste heat utilization system for data centers integrating renewable energy is constructed, enabling efficient waste heat utilization. Additionally, to further reduce the system’s operating costs, a demand response mechanism based on user-side flexible load regulation is introduced to study the system’s performance under dynamic peak-shaving subsidy prices. The results show that when the subsidy prices are set to 0.149, 0.141, 0.190, 0.145 yuan/kWh respectively, the system achieves its highest net income of 266158.4 yuan. As the capacity of the heat storage device increases, the net income growth rate decreases from 10.4% to 1.6%, while renewable energy consumption shows a declining trend.

In order to improve the wind load research of photovoltaic bracket. First, the calculation principle of wind load of photovoltaic bracket of various standards and the value characteristics of related parameters were compared and analyzed. Second, the shape coefficients of the solar structure under the combination of typical inclination angles and wind directions were derived by performing a rigid model load measurement wind tunnel test on a fixed adjustable photovoltaic. Transient analysis was used to determine the photovoltaic bracket wind vibration coefficients under normal operating settings from the results of the wind tunnel tests. Finally, the wind load values and related parameters obtained by the test were compared with the reference values of each standards. The findings demonstrate that there are clear discrepancies between various standards and wind tunnel tests' wind pressure coefficients, wind vibration coefficients, and wind loads. The wind pressure coefficient of the wind tunnel test is substantially lower than the stated value of each standards, and the wind vibration coefficient result is higher than the domestic standard value of 1.0, more than 1.64-2.05 times, and there is also a considerable difference with foreign standards. In general, when a structure is upwind as opposed to leeward, the wind load and wind pressure coefficient are lower. The wind resistance design of photovoltaic bracket according to Chinese standards is radical, while the outcomes are conservative by foreign standards. More wind resistance studies are required in order to safely and rationally guide the wind resistance design of photovoltaic bracket structures because the wind load provisions in common standards are not reasonable.

Traditional photovoltaic （PV） prediction models are highly susceptible to fluctuations in meteorological data and exhibit low sensitivity to meteorological features. To address this, we propose a short-term PV output prediction method based on multi-level feature extraction using bi-directional long short-term memory （BiLSTM）, aimed at predicting PV output under various weather conditions. Firstly, meteorological factors with high correlation to PV output are selected as input features. The fuzzy C-means （FCM） clustering method is used for flexible sample division, and the Xie-Beni index is calculated to determine the optimal number of clusters, categorizing historical data into sunny, partly cloudy, cloudy, rainy, and severe weather conditions. Next, a multi-level feature extractor （MFE） comprising CNN-CBAM-TCN is constructed： convolutional neural networks （CNN） are employed for initial feature extraction, convolutional block attention module （CBAM） is used to suppress non-essential features, and temporal convolutional networks （TCN） are utilized to capture the temporal characteristics of intra-day PV output. Finally, BiLSTM is used for PV output prediction. Case studies validate the effectiveness of using the Xie-Beni index to determine the optimal number of clusters and demonstrate that this model achieves higher prediction accuracy compared to other prediction models under complex weather conditions.

This article numerically processes the grayscale histogram of visible light images of photovoltaic modules, introduces the concept of average grayscale value, and combines it with the photovoltaic experimental platform built by the school to confirm the corresponding relationship between the average grayscale value and the ash density of photovoltaic modules. Based on this conclusion, the image of a visual inspection platform for a photovoltaic power station in Inner Mongolia is identified and processed, and it is found that over time, the change in the average gray level difference between the clean side and the accumulated side is positively correlated with power generation loss. Based on the meteorological and image data collected by the experimental platform, a prediction model for the degree of ash accumulation is established using the random forest algorithm. Based on relevant meteorological factors, the average gray level of the ash accumulation board from the previous day is combined as the input variable to predict the average gray level of the ash accumulation board day by day. Different tuning parameters are set to improve the optimization robustness of the prediction algorithm.

In order to obtain a systematic understandings of the unvented filling characteristics and the influencing factors of liquid hydrogen on the ground, a two-dimensional axisymmetric numerical model was established. The influences of different inlet temperatures, inlet flow rates, initial wall temperatures, and initial pressures inside the container on the unvented filling characteristics of liquid hydrogen were investigated by using Fluent software. The Lee model and a user-defined function （UDF） were used to describe the phase change issues （evaporation, and condensation） during the unvented filling of liquid hydrogen considering the interfacial tension between gas and liquid. The results show that under normal the gravity conditions, the unvented filling process of liquid hydrogen can be divided into three phases： an initial rapid pressure rise phase, a relatively stable filling phase, and a final rapid pressure rise phase. When the saturated vapor pressure corresponding to the temperature of the added liquid is greater than the pressure inside the vessel, the initial filling will lead to flash evaporation, and the pressure rise in the initial phase is significant. The inlet flow rate affects not only the total filling time, but also the vaporization pressurization at the initial stage of unvented filling, the temperature reduction degree in the system and the compression effect of gaseous hydrogen at the later stage of filling. The initial wall temperature affects the wall boiling pressurization effect during the initial filling stage. The initial pressure has a significant impact on the pressure change during the initial filling period, and the flash evaporation does not occur in the container when the initial pressure is higher than the saturation pressure of liquid hydrogen.

The Morison equation is a widely used formula for calculating wave loads on offshore wind turbine monopile foundations. However, as monopile diameters increase, diffraction effects can no longer be ignored, and the Morison equation method becomes less accurate. This study investigates the horizontal wave load on large-diameter monopile foundations with a pile-diameter to wavelength ratio ranging from 0.15 to 0.25. To this end, a combination of physical model experiments and computational fluid dynamics （CFD） numerical simulations was employed. The variation of the maximum horizontal wave load was analyzed for monopiles of different diameters under different wave conditions. A curve depicting the values of the inertia force coefficient was given, and an empirical formula based on the Morison equation was proposed for calculating the maximum horizontal wave load on large-diameter monopile foundations.

Facing the demands of energy storage and decarbonization, the technology of methanol production from biomass coupled with wind power was studied using wheat straw and sawdust as feedstock. The performances were analyzed and evaluated including the methanol yield, the energy efficiency of storage-release processes （i.e., storage-release efficiency）, the life cycle carbon emissions, and the differences between power generation model and cogeneration mode for methanol utilization. Without the separation of CO₂ in product gas, only the one-way conversion rates of CO₂ and CO have obvious impacts on the methanol yield, ranging from 0.970 to 1.104 kg/kg. The efficiency of water electrolysis is the key parameter to improve the storage-release efficiency. Both storage-release efficiency and carbon emission intensity under the cogeneration mode have significantly advantages over those under pure power generation mode. Under the cogeneration mode, the storage-release efficiency varies from 46.1% to 58.6%, which is similar to that of the compressed air energy storage. The ranges of carbon emission intensity of the regenerated electricity and thermal energy are 37-77 g CO₂/kWh and 10-21 g CO₂/MJ, respectively, which are significantly lower than the current levels of the products. The carbon emission intensities with sawdust as feedstock are significantly lower than those with wheat straw. This technology with separation and capture of CO₂ in product gas can achieve the zero-carbon emission in the life span at the cost of decreases in the methanol yield, power rate, and storage-release efficiency.

In view of the problems of low output voltage and large environmental impact in new energy renewable systems such as photovoltaic power generation, this paper proposes a novel high-gain three-port DC-DC converter topology and control strategy. The three ports of this converter are connected to the photovoltaic power generation end, the energy storage port end, and the load end, respectively. It has advantages of low switch device stress, flexible power allocation among input sources, and sirong excitation energy recovery ability of the coupled inductor. The operating principles and steady-state conditions of the converter in various operating modes are analyzed in detail, and the performances of the circuit topology are compared with those of similar converters. According to the control requirements of the optical storage and power supply system and the power flow of the converter under different working conditions, the corresponding mode operation and switching control strategy are studied. A 300 W prototype is built for validation, and the results of the validation demonstrate the feasibility of the proposed structure.

To effectively extract valuable temporal information from photovoltaic （PV） power data and further improve the accuracy of PV power prediction, we propose an efficient channel attention mechanism （ECA）- temporal convolutional network （TCN） prediction model based on multi-factor fusion. Firstly, the maximum information coefficient （MIC） is utilized to extract relevant features of PV power. Secondly, a polynomial feature derivation method is employed to combine the features of various factors, generating high-dimensional features and facilitating feature combination. Then, the ECA module, which adaptively selects the size of one-dimensional convolutional kernels, is combined with TCN to construct the ECA-TCN prediction model, which effectively captures the temporal characteristics of PV power data. Finally, multiple models are compared through experimental evaluation. The results demonstrate that the proposed feature combination method efficiently selects PV power data features and enhances their discriminative ability. The ECA-TCN prediction model with feature combination achieves a root mean square error （RMSE） of 0.0828 kW. Compared to LSTM, LSTM-TCN, and ECA-LSTM, the RMSE of the ECA-TCN was reduces by 0.29, 0.23, and 0.13 percentage points, respectively, and exhibits the best fitting performance with an R² value of 90.74%. This model effectively improves the accuracy of PV power prediction while maintaining a high fitting performance.

To ensure the normal operation and structural safety of large-scale floating wind turbines under extreme sea conditions, extreme value estimation of structural responses is necessary to be carried out to assess the safety of structures. In this work, the OpenFast software was used to perform numerical simulations of the structural response of a 15 MW large-scale floating wind turbine in the South China Sea under a 50-year return period sea state. Combining with the Poisson approximation assumption, the average crossing rate method was employed to estimate the extreme distribution of tower base shear and bending moment for the large-scale floating wind turbine. The results show that under extreme conditions, the tower base shear force and bending moment will be significantly excited near the natural frequency of the tower's pitch motion, leading to a sharp increment of the dynamics responses. Moreover, the average upcrossing method can effectively estimate the extreme value distribution of the tower base structural response for a 15 MW floating wind turbine. Under an exceedance probability of 0.01, the estimated extreme values of tower base shear and bending moment based on 1-hour numerical simulation response samples satisfied the structural strength requirements.

The application of hydrogen production technology through water electrolysis can promote the consumption of renewable energy and enhance the flexibility of power system regulation, especially in systems with a high proportion of renewable energy. Alkaline water electrolyzers （AWEs） are widely used in the field of industrial electrolysis due to their low cost, simple structure, and mature technology. However, the low electrolysis efficiency of AWES under low load conditions makes it difficult to effectively utilize the fluctuating renewable energy power over the wide range. To address this issue, this paper firstly reveals the mechanism behind the difference in high- and low-load efficiency from the perspective of the distribution of the exciting electric field inside the electrolytic cells. An optimal power pulse-width modulation （OP-PWM） strategy is proposed to reshape the internal electric field distribution under low load conditions. A two-stage converter for pulse electrolysis has been designed. And key control parameters are analyzed. Moreover, the proposed theory is validated through an experimental platform for photovoltaic direct-drive electrolytic hydrogen production. The experimental results show that the OP-PWM strategy can significantly improve the low-load efficiency compared to the dc electrolysis mode, with an increase of 1.8 times. Under the constraint of achieving an efficiency greater than 48%, the system operation has been improved from 28%-100% to 20%-100% of rated.

In view of the phenomenon of "abandoned wind power", "abandoned photovoltaics" and "abandoned hydropower" in microgrids and the increasing demand for hydrogen load, a certain capacity of hydrogen production/storage equipment is considered to be deployed in the offsite hydrogen refueling station to centralize the consumption of surplus electricity from microgrids. Through the research of the topology and mathematical model of microgrid-hydrogen refueling station system （MHSS） containing wind power, photovoltaic, small hydropower, as well as electrolysis tank, hydrogen storage tank and reformer, a capacity optimization configuration model of hydrogen production/storage equipment with the objective of minimizing the total net present cost of the whole life cycle of the hydrogen refueling station is established, and the combined electric-hydrogen operation model for residual electricity to hydrogen-green hydrogen first （RETH-GHF） is proposed. In the example, the adaptive simulated annealing particle swarm optimization （ASAPSO） algorithm is applied to simulate the system with the practical scenario. By comparing the configuration results and cost-benefit details of offsite centralized/onsite decentralized hydrogen production schemes of microgrids, it is proved that the scheme of offsite centralized hydrogen production is more economical in the example scenario, and the economics of the two schemes of hydrogen production have been re-compared by taking into account the impact of industrial electricity price, hydrogen and natural gas price fluctuations on the total net present cost of the hydrogen refueling station throughout its life cycle, and finally, the rationality of the capacity optimization configuration model for hydrogen production/storage equipment in the hydrogen refueling station is verified with system operation of typical days in four seasons.

A mathematical model is established for generation, transmission, controllable load, and energy storage. The conditional value-at-risk is used to measure renewable energy uncertainty. A coordinated framework of “source-grid-load-storage” is proposed for the interconnected power system. A solution method is proposed to consider the information security and privacy of each regional power system. The centralized optimization model is decoupled into a distributed optimization model using inter-regional contact lines as coupling variables. The dynamic penalty parameters and prediction correction strategy are integrated into the alternating direction multiplier algorithm to improve the convergence speed. The numerical results verify the effectiveness of the risk-averse investment model based on distributed optimization.

In this paper, a forecast correction method based on machine learning is proposed, which first corrects regional errors with gridded analysis products and then corrects local errors with station observation data. The method utilizes the gridded wind speed analysis products at 100-meter height to correct the results of numerical weather prediction through deep learning algorithms, and then corrects the station wind speed forecast results with a random forest integrated forecast approach based on the observation data of wind speed at 100-meter height from the anemometer tower. An empirical analysis of a meteorological station in central Inner Mongolia shows that after correction with the proposed method, the root mean square error （RMSE） of wind speed prediction for the next 84 hours is significantly reduced from 4.25 m/s to 3.16-3.79 m/s after gridded forecast correction and to 2.81 m/s after station integrated correction. Therefore, the proposed method can effectively reduce forecast errors.