In order to improve the fuel economy of fuel cell hybrid electric vehicles during short range driving, a vehicle speed prediction model structure VBSnet based on hybrid deep neural network was constructed. This structure not only further improves the convolutional network based on the VGGNet structure, but also introduces a bidirectional long shortterm memory neural network to effectively learn the spatiotemporal dependencies of the entire vehicle speed prediction sequence. Simultaneously considering the influence of prediction time domain and input sequence length on the prediction accuracy of shortrange vehicle speed problems, Bayesian optimization hyperparameters are used to further improve the prediction accuracy of VBSNet. To address the online optimization and computational efficiency issues of energy management strategies, a multiobjective optimization based on model predictive control (MPC) energy management strategy was designed. This strategy can achieve a balance and optimization of hydrogen consumption, lithium battery state of charge (SOC) maintenance, and fuel cell utilization efficiency. Finally, under actual vehicle conditions, the proposed strategy was compared with rulebased strategies, resulting in fuel economy improvements of 7.25%, 9.94% and 19.23%, and better SOC maintenance characteristics.

Lignin nanoparticles (LNP) as the carbon precursor is used to prepare hierarchical porous carbon using template methods in this study. In comparison to the porous carbons (SLC, FLC) prepared with a single template method, the LNPbased porous carbon (FSLC) is fabricated using nanoSiO2 coupled with PluronicF127 as double templates. This approach results in a honeycomb like structure with typical mesoporous characteristics for FSLC, achieving a mesoporous ratio of up to 87%. As a supercapacitor electrode, FSLC demonstrates good electrochemical performance, with a mass specific capacitance of 250 F/g at a current density of 0.5 A/g, representing a 163% increase over mass specific capacitance (95 F/g) of the SLC. The dualtemplate method for producing highperformance porous carbon offers a novel approach for utilizing lignin in energy storage application.

To investigate the combustion characteristics of a blended fuel consisting of bituminous coal, miscellaneous wood pellets, and oil sludge, thermogravimetric analysis is conducted on the blended fuel to analyze its combustion behavior. The FWO method is employed for kinetic analysis of the blended fuel. The results show that the combustion process of bituminous coal is primarily fixed carbon combustion, while the combustion process of miscellaneous wood pellets is more complex and exhibits the best combustion performance, with a comprehensive combustion characteristic index as high as 1.93×10⁻⁷. Cofiring with miscellaneous wood pellets effectively enhances the combustion performance of bituminous coal. When bituminous coal and miscellaneous wood pellets are blended in a mass ratio of 1:4, the ignition temperature of the blended fuel decreases by 35% compared to bituminous coal alone. The comprehensive combustion characteristic index increases by 7.2% compared to miscellaneous wood pellets. The most probable function for the cofiring of bituminous coal and miscellaneous wood pellets is [ln (1x)]⁻³/⁴. When the blending ratio of miscellaneous wood pellets reaches 45%, the appropriate addition of oil sludge can improve the combustion performance of the fuel. The most probable function for the combustion reaction of the blended fuel consisting of all three components is [ln (1x)]². When the blending ratios of bituminous coal, miscellaneous wood pellets, and oil sludge are 40:40:20 and 45:45:10, respectively, the corresponding activation energies of the blended fuel are significantly lower than those of other blending ratios, and the comprehensive combustion characteristic index is approximately 1.5 times that of pure bituminous coal.

The microexplosion of emulsified fuel can promote fuel atomization and mixing, however the key to affect its microexplosion characteristic is the emulsifier. Emulsified methanoldiesel was prepared by mechanical emulsification and phacoemulsification, to study the effect of emulsifier types and content on the stability and dispersion of micro emulsified methanoldiesel. The results show the dispersion of the micro emulsified methanoldiesel can be improved by using compound emulsifiers and increasing the emulsifier content; The increase of emulsifier content can improve the microexplosion intensity, delay the initial time of microexplosion, reduce the droplet life, and reduce the droplet evaporation rate in the stabilization stage; The microexplosion intensity of methanol emulsifier content ratio of 10:5 and 10:8 was increased by 34.3% and 37.6% compared with methanol emulsifier content ratio of 10:3 respectively.

This paper proposes a method to predict the photovoltaic output based on weather state pattern recognition and SSABP, which is more accurate than traditional single models under different weather conditions. Firstly, the historical data was cleaned using the 3sigma algorithm to obtain the data that can reflect the output of photovoltaic power plants and the regularity of weather changes. Then, based on the analysis of the parameters such as irradiance, temperature, and wind speed, Gaussian mixture models were applied to classify the professional weather types and three typical generalized weather types were obtained. Furthermore, the data was used as SSABP neural network input to predict the futuristic photovoltaic power plant output. Finally, the carbon accounting method was used to calculate the carbon emission reduction of the photovoltaic power generation project. The experimental results show that through classification recognition and the optimized SSABP neural network, the mean relative errors in the prediction for the three weather types are 0.195, 0.243 and 0.310, respectively. Compared with other predication models, the relative errors are reduced by 17.8%~66.7%. In addition, the relative error between the predicted carbon dioxide emission reduction and actual value is only 3.37%. The model proposed in this work shows satisfactory prediction results.

In order to achieve clean heating in northern rural areas, this paper constructs an air source heat pump heating system based on automatic heat storage/discharge devices and compound multisurface concentrating collectors.Experimental research shows that: at the same water supply temperature, the COP of the heat pump increases gradually with the increase of the inlet air temperature of the evaporator and the intensity of solar radiation. When the heat collector and heat pump are operated together, the heat storage device can automatically store heat and automatically supplement the heat of the air entering the evaporator. The temperature rise of the air is 2~4 °C. The system heat collection and heating capacity on the experimental day were 2.29 MJ/m² and 27.99 MJ/m² respectively, the system energy efficiency ratio (SEER) was 1.84, and the daily average solar energy contribution rate was 41.2%. When the air source heat pump operates alone, the heating capacity of the heat pump on the experimental day is 8.75 MJ/m², and the SEER is 1.79. The above results show that the solar subsystem constructed in this article can automatically adjust the heat storage of the system and the heat supplement to the evaporator, improve the energy efficiency ratio of the system, reduce the difficulty of operation and maintenance, and has certain adaptability in rural areas.

In order to maximize the solar radiation yield, it's imperative to optimize the inclination angle of the solar collector. To this end, a calculation model for solar radiation on inclined surfaces was established, and the computation processes for direct and scattered radiation were separately streamlined. MATLAB was utilized to analyze and perfect the elevation angle of the collector, which led to the determination of monthly and annual optimum inclination angles. The research disclosed that the ideal annual inclination angle in Tianjin is 36.3° , marginally inferior to the local latitude. The monthly tilt angle should vary between 10~64°, with lower degrees in summer and higher in winter. Placing the collector horizontally enhances annual solar radiation by 12.4% and 17.3% at the optimal yearly and monthly inclination angles, correspondingly. Adjusting the optimum tilt angle by 5.0% is feasible when compared with using the local latitude as the tilt angle. Taking Guangzhou, Lhasa, Jinan and Changchun as examples, the annual and monthly optimum tilt angles of these regions are calculated.The comparative analysis of various cities revealed that for regions with high direct radiation proportion, refining the optimal angle on a monthly basis leads to higher energy gains.

Installing a certain capacity of flywheel energy storage system (FESS) at the grid connection of wind farms can effectively smooth the gridconnected power and improve the gridfriendliness of wind farms. To improve the power response speed of FESS and enhance the smoothing effect of wind power fluctuation while avoiding overcharge/overdischarge of FESS, this paper proposes a control strategy for FESS based on a fuzzy Kalman filter and the improved sliding mode control(SMC). The Kalman gain is adaptively adjusted according to the realtime speed and power of FESS. The difference between the filtering result and the wind farm output power is used as the input of the SMC to realize the power control of FESS. The simulation results show that the control strategy proposed in this paper has good dynamic response characteristics and the wind power can be effectively smoothed, thus meeting the requirements of grid connection. The flywheel speed is kept within the limit during the smoothing process, which extends the service life of FESS.

In this paper, in order to study the dynamics of the model turbine with driven motion at different yaw angle operating condition, a wind turbine model based on elasticity is established. Under the condition of turbulent flow, the driven motion state of the model wind turbine is realized with irregular rotation, and the wake distribution characteristics and energy characteristics of the wind turbine are studied. The results show that when the yaw angle is 0°, the wake of the driven wind turbine is obviously smaller than that of the fixed wind turbine, and the turbulent kinetic energy is greater than that of the fixed wind turbine because of the multifreedom rotational motion of the driven wind turbine. With the increase of yaw Angle, the distribution of wake and turbulent kinetic energy of wind turbine appears radial deviation, and the deviation direction is opposite to the direction of yaw Angle rotation, and the wake and turbulent kinetic energy decreases with the increase of yaw Angle. With the increase of yaw Angle, the average output power of forced and fixed wind turbines decreases, and the average output power of forced wind turbines is smaller than that of fixed wind turbines.

A transfer learningbased early fault warning method for offshore wind turbine bearings is established to address the problems of varying operating conditions of offshore wind turbines and many false alarms for early fault warning of turbine bearings. The method uses the shorttime Fourier transform to extract the timefrequency domain features of the vibration signals, which are normalised to form pre processed samples. The objective function of the convolutional autoencoder is supplemented with a support vector data description regular term and a maximum mean discrepancy regular term to constrain the feature distribution while obtaining the common features center of the bearings in normal state under different operating conditions. The Euclidean distance between the online sample features and the common feature center is calculated to construct bearing health indicator sequence, and the ADF(Augmented DickeyFuller)test is introduced to perform stationarity analysis and capture the sequence mutation points, which finally realize the early fault warning of bearings in offshore wind turbines. The validation on the XJTUSY bearing dataset showed that the proposed method has fewer false alarms, high accuracy and better detection stability.