The analysis and mastery of the inner law of the fluctuation characteristics of wind power output is conducive to improving the prediction accuracy of wind power output, thus guiding the power grid scheduling department to reasonably arrange the power generation plan and improve the economy of system operation. To characterize the probability density distribution of wind power output fluctuations, two adaptive bandwidth kernel density estimation models are developed by modifying the fixed bandwidths obtained from the empirical method and the unbiased crossvalidation method. Then, the above two models are combined and optimized, and finally the probability density distribution model of wind power output fluctuation based on hybrid adaptive kernel density estimation (HAKDE) is established. A variety of probability density distribution models were used to fit the fluctuations of wind power output at different spatial and temporal scales in a province in North China. The results show that the fitting effect of the HAKDE model is the best, which verifies the effectiveness of the HAKDE model.

In the design of highpower directdrive wind turbines, this paper proposes to design axial flux permanent magnet generators as highpower half directdriven wind turbines to address the issues of large radial dimensions and high maintenance costs in the later stage. Taking the 2 MW semi direct drive axial flux permanent magnet generator as the research object, the parameter design principle and method are studied. By exploring the influence of different pole pairs on the output performance of the generator, the optimal scheme of the generator meeting the design goal is determined. The 3D finite element method is used to analyze the air gap flux density, voltage, current waveform, output power, loss and efficiency of the generator under noload and rated load conditions, and verify the correctness of the electromagnetic design of the generator. The efficiency of the generator can reach 97.79%, which has the advantages of low loss and high efficiency. By comparing with generators in the existing research, it is concluded that the radial size of the MW half directdriven axial flux permanent magnet generator proposed in this paper is significantly reduced, and it has better performance than the existing radial flux structure highpower semi direct drive permanent magnet generator in terms of axial dimension and cogging torque.

In order to solve the problems of increasing network loss caused by bidirectional power flow and node voltage fluctuation caused by fluctuation of distributed power and load, This paper proposes a method to control power flow based on pulsewidth modulation technology of power electronic converters on both sides of Solid State Transformer. In this paper, the dynamic reactive power optimization model of active distribution network with SST is firstly established. Then, the improved multiobjective group algorithm is used to solve the control variables such as modulation Angle and modulation coefficient of the power electronic converter based on the primary and secondary sides of SST, aiming at the multitime active network loss and voltage fluctuation. Finally, the simulation model is established and compared with the active distribution network dynamic reactive power optimization method based on onload voltage regulating transformer. The results prove the superiority of the proposed method in reducing network loss and maintaining node voltage stability.

The output of wind, solar and other renewable energy sources is highly variable and stochastic, which poses a great challenge for the flexibility of the system. Therefore, a hybrid energy storage joint planning method for power systems that balances economy and flexibility is proposed. First, the flexibility requirements of the source and load sides are assessed from the perspective of power balance; then, a twolayer planning model that coordinates the flexibility retrofit of thermal power units and the hybrid energy storage is established, with three types of flexibility resources as the planning objects, namely, thermal power unit flexibility retrofit, Vanadium Redox flow Battery, and pumped storage. Using the upper and lower layer models, the scheme is iteratively optimized to obtain the optimal hybrid energy storage configuration scheme that balances economy and flexibility; secondly, an an improved whale algorithm based on inverse learning is adopted to optimize the planning model, and the validity of the model is verified by simulation. Finally, a case study is carried out using historical data from a certain region in Inner Mongolia Eastern to validate the effectiveness of the proposed method.

Modern power system has developed into cyberphysical system (CPS), which is highly integrated between power network and information network. However, advanced information technology not only improves system performance, but also introduces new security risks. With the largescale gridconnection of Electric Vehicle (EV) with mobile energy storage equipment, the absorption capacity of distribution network for new energy has been greatly improved. However, the low security and high accessibility of charging piles have further reduced the network security of distribution network. On this basis, a distributed energy management strategy based on consistency algorithm is firstly proposed in this paper, which considers the EV cluster as an energy storage device with source charge bidirectional characteristics to achieve fully distributed economic scheduling. Considering denial of service attacks and new data integrity attacks for electric vehicles, a disturbance rejection control strategy combining privacy protection protocol and isolation mechanism is proposed to achieve effective energy management and economic operation of systems under network attacks. Finally, the effectiveness of the encryption mechanism and the feasibility of the control strategy are verified by simulation.

Aiming at the problem that the complex coordination relationship of the sendingend power grid with high proportion of new energy and pumped storage combined power generation may lead to insufficient new energy consumption capacity, an energy balance optimization control method of the new energy sendingend power grid based on variable speed pumped storage is proposed. Firstly, the operating characteristics of variable speed pumped storage units are studied, and the active and reactive power output characteristics of variable speed pumped storage units based on frequency converter control are analyzed. Secondly, the variable speed pumping is used to analyze the improvement of new energy generation capacity, and the energy balance optimization model of the new energy sending end power grid is established. Then, the quantum particle swarm optimization algorithm is improved, and the energy balance optimization control algorithm of the new energy sending end power grid is proposed. Finally, with reference to the delivery mode of new energy base in a certain area of Northwest China, a typical scenario of variable speed pumped storage power station and conventional pumped storage power station cooperating with new energy is constructed. The improved new energy consumption capacity of variable speed pumped storage power station and conventional pumped storage power station is compared. The simulation results show that the energy balance optimization control method of the sending end power grid based on variable speed pumped storage can improve the utilization rate of new energy and play a significant role in improving the level of renewable energy consumption.

The traditional photovoltaic power generation system usually operates at the maximum power operation point, does not respond to the change of grid frequency, and cannot provide active power to suppress the change of grid frequency. With the increase of photovoltaic permeability, the safe and stable operation of the grid will be affected. In this paper, load shedding control is adopted to realize the response to system frequency without changing the main circuit structure, grid connection strategy of inverter and adding energy storage equipment. First, the current maximum power operation point is obtained by setting the masterslave array to achieve load shedding control. By setting the corresponding relationship between frequency and load shedding rate, the change of active power output for frequency change is achieved, and the system is provided with active power support. Finally, the effectiveness of the results is verified through simulation on the hardware in the loop simulation platform.

The doubly fed induction generator, operating in electric mode, has a certain inertia supporting capacity to the power grid, but the supporting capacity is limited, and the rotor is out of control during the lowvoltage crossing of the power grid. In this paper, a kind of rotor stability control technology is proposed when voltage sags in power system. Firstly, a flywheel is attached to the rotor shaft of a doubly fed fan to increase the inertia of the system and stabilize the speed of the rotor. Secondly, voltage compensation is used to reduce the back electromotive force of the rotor side and stabilize the rotor current when the voltage of the power grid falls, the simulation model of the system is built in MATLAB/Simulink platform, and the results show that the antimutation ability of the rotor speed of the doublyfed fan is improved by 20% by adopting the rotor stability control technology, the rotor current drop value is reduced by 45%, which verifies the validity and reliability of the proposed rotor stability control technique.

Under the background of "double carbon", distributed renewable energy and flexible resources such as energy storage and demand response develop rapidly. Virtual power plants integrate distributed resources efficiently through control technology, which improves the power generation efficiency of distributed energy. With the social capital entering the power market, different virtual power plants will belong to different investors, forming a multiagent game pattern. According to the investment preferences of investors, virtual power plants will be composed of resources with different flexibility. In order to give consideration to the interests of virtual power plant operators and virtual power plants, a twolevel masterslave game model between operators and multivirtual power plants is constructed. Considering the interaction between upper pricing and lower output, the dynamic pricing of operators and the optimal operation and scheduling of virtual power plants are studied. In the lower layer, aiming at the minimum operating cost of each virtual power plant, the optimal scheduling models of multiple virtual power plants including electric energy storage, demand response and hydrogen energy storage are established respectively. The upper layer takes the operator's profit as the goal, and combines the lower layer's output plan to dynamically formulate the purchase and sale price of virtual power plants. Particle Swarm Optimization (PSO) is used to solve the game model iteratively. Through the analysis of an example, the model can give consideration to the interests of multiagents, effectively improve the operators' income and reduce the operating cost of virtual power plants.

In order to explore the potential of fruit peel waste conversion to produce highvalue platform chemicals, the hydrothermal conversion of citrus peel to furfural (FF) and 5hydroxymethylfurfural (5HMF) was conducted in the choline chloride (ChCl)/ carboxylic acid based deep eutectic solvents (DES) system. Conclusions can be drawn from experiments in xylose and glucose: The carboxylic acid performance was ranked lactic acid > acetic acid > glycolic acid. The optimal molar ratio of choline chloride to carboxylic was 1:2. The optimal water content of solvent system was 20%. The effects of temperature and time on citrus peel conversion were studied based on the optimized operating conditions. Since the pectin component in the citrus peel was mainly converted to 5HMF in DES, the yield of the peel converted to 5HMF was significantly higher than that of FF. The maximum mass yield of 5HMF from citrus peel was 8.6% (molar yield was 43.1%). The maximum mass yield of furfural reached 2.4% (molar yield was 19.1%).