When a short circuit or grounding fault occurs in a DC distribution network, its fault characteristics are complex, and a fast and reliable fault protection method is the key to ensuring the safe operation of the DC distribution network. A fast and reliable fault protection method is an important guarantee for the safe operation of DC distribution networks. A DC distribution network pilot protection method is proposed based on the cosine similarity of the transient voltage of the current-limiting reactor. Firstly, the differences in the voltage characteristics of current-limiting reactors at both ends of the DC line are analyzed when a single line-to-ground or line-to-line faults occur inside and outside the DC distribution network, and according to the principle of cosine similarity, a fault discrimination method based on the cosine similarity of transient voltages of current-limiting reactors is proposed for the fast identification of internal/external DC fault. Then, the voltage surge ratio between the positive and negative poles of the current-limiting reactor is used to identify the type and polarity of DC faults, based on which corresponding protective measures are taken. Finally, a simulation model of a multi-terminal radiated DC distribution network is built using PSCAD/EMTDC simulation software. The effectiveness of the proposed protection method is verified. The simulation results show that the method has a certain degree of resistance to fault resistance and can better meet the rapidity and selectivity requirements of DC distribution network protection.

Geographic location of underground cable is important information in urban construction and power operation and maintenance. Ground penetrating radar is always used in detection for directly buried cable, but the influencing factors of its detection effect for underground cable is unclear. GprMax forward modeling method is used to analyze the ground penetrating radar echo image of underground cable， including buried depth, soil dielectric constant, antenna center frequency, pipeline material and cable channel structure. Furthermore an experiment is carried out to compare the ground penetrating radar image of cable, metal and PVC pipes at 0.4 GHz and 0.9 GHz antenna center frequency. Finally a field test for underground cable is carried out. There is a distinctive hyperbola in the ground penetrating radar image. The hyperbola shape is mainly influenced by burial depth and soil dielectric constant. It is difficult to distinguish different material pipeline based on the hyperbola shape. The fine sand and cover plate around underground cable can weaken the echo signal, and lead to hyperbola layering. In order to obtain good detection effect, it needs to detect the whole cable channel instead of the cable with 0.4 GHz center frequency antenna.

With the increase of voltage level and transmission power, the reactive power of the system changes more frequently, which is not conducive to arc extinguishing and seriously threatens to the safe and stable operation of transmission lines. Based on the AC/DC hybrid reactive power compensation (HRPC） line,the influence of the compensation degree of the hybrid reactive power compensation (HRPC） device on the submarine current and recovery voltage is studied. The measures of hierarchical controllable high reactance and small reactance linkage control are analyzed to suppress the submarine arc, and an improved timing control strategy for bypass circuit breakers is proposed. The results show that when the series compensation degree is 20%, the compensation degree is 70%, and the small reactance is 700 mH, the amplitude of the submarine current and recovery voltage is the smallest, and the suppression effect is the best. In addition, the control strategy of using the bypass circuit breaker to disconnect after the extinguishing of the submarine arc can further accelerate the extinguishing rate and improve the success rate of single-phase reclosing. The research results provide theoretical basis and technical support for the application of HRPC in long-distance high-capacity hybrid transmission lines.

In order to control the concentrated distribution of electric field on the surface of three-pillar insulators and inserts, a geometric-dielectric optimization method based on finite element model(FEM) combined with evolutionary algorithm (EA) is proposed. Firstly, geometric and surface dielectric parametric models of gas insulated line(GIL) three-pillar insulators are established. The maximum tangential field strength of insulator surface E_t-max, the non-uniform coefficient of tangential field strength F_t, and the maximum normal field strength of insert surface E_i-max are taken as optimization targets, and geometric-dielectric optimization calculation is carried out by cyclic FEM with evolutionary algorithm. Compared with the results before optimization, E_t-max decreases from 11.3 kV/mm to 9.47 kV/mm, F_t decreases from 1.56 to 1.34, and E_i-max decreases from 31.9 kV/mm to 13.4 kV/mm. The geometric structure of the optimized three-pillar insulator has the characteristics of dumbbell type, the distance along the surface is increased, the overall profile is a circular arc with greater curvature, and the surface dielectric gradient after optimization is distributed in a "scoop curve". The results of optimization design can provide reference for the subsequent manufacture of three-pillar insulators, and the technical route of this paper can also be used for the optimization of other insulation parts.

In order to study the electrification tendency of environment-friendly ester insulating oil, an oil flow chargeability test platform is built. By selecting two kinds of commercial ester insulating oil and mineral insulating oil, including FR3 natural ester, Midel 7131 synthetic ester and Karamay insulating oil, the influence of key factors such as oil paper aging degree, flow rate and temperature on the oil flow electrification of ester insulating oil is analyzed. The results show that the viscosity of ester insulating oil is high. Under the action of oil flow, the friction contact between oil and paper is further increased, and the charge separation at oil-paper interface is accelerated, resulting in a significantly higher electrical charge than mineral insulating oil, of which natural ester is the highest, followed by synthetic ester. With the increase of oil flow rate, the charge separation at the oil-paper interface is intensified, and the electrification tendency of ester insulating oil is enhanced. The increase of the test temperature will accelerate the ionization process of polar molecules in the insulating oil, resulting in an increase in the electrification tendency of the insulating oil. With the increase of the temperature, the viscosity of the insulating oil will decrease, resulting in the weakening of the contact friction of the oil-paper, and its electrification tendency will decrease instead. The increase of the aging degree of oil paper accelerates the formation of aging polar products in the oil, and then accelerates the dissociation of impurity molecules, resulting in the increase of the number of intrinsic ions and impurity ions in the oil, thus increasing the degree of oil flow electrification.

In order to realize fast fault section identification in distribution network with limited synchronous phase measurement unit （PMU） layout, an on-line fault section identification algorithm for distribution network with low PMU number dependence is studied. The PMU layout and area division principles are proposed, and the differences in the three-phase current between the normal area and the faulted area before and after the fault occurrence are analyzed. A multi-dimensional spatial and temporal state monitoring matrix is constructed using the collected current data, and a 2-parametric-based spectral parametric ratio coefficient is defined for dimensionality reduction of the multi-dimensional state monitoring matrix to quantify the differences in electrical state changes before and after the fault in each area, so as to achieve fast and accurate on-line fault location in the distribution network. In this paper, the effectiveness of the proposed method is verified by using IEEE 33-node distribution simulation system and 6-node actual test platform. The simulation results and actual test results show that the proposed method is less affected by the fault type, neutral grounding mode, and transition resistance under the limited PMU measurement information, and has certain noise immunity. The proposed method can quickly and accurately identify fault sections in the distribution network even in the case of asynchronous communication.

The DC bias current in the transmission line can easily cause the saturation of current transformer (CT) and the harmonic components contained in the secondary current output of CT in saturation state will lead to an increase in power energy metering error. In order to improve the accuracy of current transformer measurement under DC bias, an inverse propagation method is proposed for CT distortion current. In this method, the inverse propagation process of CT distortion current under DC bias is divided into two stages: off-grid and on-grid. In the off-grid stage, the data sample set is generated by changing the operating environment of CT, and then the saturation degree of CT is classified by random forest classification (RFC) algorithm. Finally, a simulate anneal genetic algorithm-radial basis function (SAGA-RBF) model is trained for each subclass to simulate the saturation current. In the on-grid stage, the saturation data segment of the secondary current waveform is extracted by wavelet transform, and then the saturation data segment is input into the off-line model to realize the inverse propagation of the secondary distorted current. The simulation results show that the proposed method can relike the inverse propagation of the primary side current, and improve the measurement accuracy of CT.

With the increasing growth of renewable energy sources, their decentralized, intermittent, and volatile characteristics have become prominent challenges. To address these issues, a novel solution called the virtual power plant （VPP） has been proposed. Firstly an overview of VPPs is introduced, the current research status is analyzed from three perspectives: coordinated control technology, optimal scheduling, and market trading. Furthermore, different models and algorithms for optimal scheduling and their advantages and disadvantages are analyzed. The paper also highlights the benefits of blockchain technology in distributed trading and discusses the feasibility of applying blockchain to VPP trading. Finally, future research directions in the field of virtual power plants are prospected.

With a large number of distributed photovoltaic（PV） access to the distribution network, the distribution network faces greater challenges in dealing with network reconstruction and source-load-storage uncertainty, etc. Therefore, a two-stage voltage control strategy for active distribution networks is proposed. In the first stage, the contact switch of the active distribution network is centrally controlled. The network reconstruction is carried out with the goal of minimizing the network loss in an hourly scheduling period, and a mixed integer second-order cone planning model is established to solve the problem. In the second stage, the real-time voltage control of photovoltaic and energy storage systems is carried out, and the real-time voltage control problem is converted into the Markov game process（MGP）. The multi-agent model is implemented and the offline training-online operation method is adopted. Compared with the traditional two-stage mathematical planning approach, the control strategy proposed does not rely on an accurate distribution network power flow model, has low communication requirements and a faster solution speed. Finally, the effectiveness of the proposed control strategy is verified by the improved IEEE 33-node calculation examples.

For high voltage grids containing distributed new energy sources, there are a range of adaptability problems with traditional line protection due to the variety of new energy sources and complex fault characteristics. The time domain differential equation algorithm can solve for fault distances, which has good applicability in systems containing new energy sources, but there is a "transient override" phenomenon and "voltage dead zone" problem. For this reason, a new line protection scheme is proposed. When a fault occurs on the line and it is not near the outlet of the line protection, the fault point voltage is reconstructed and isotransformed, and then substituted into the time domain differential equation algorithm, which is the isotransformed fast distance protection algorithm; when a fault occurs near the outlet of the line protection, the correlation coefficient between the measured voltage drop and the calculated voltage drop are calculated and the consistency of the two trends is used to determine the direction of the outlet fault. The line protection scheme can accurately determine line faults, regardless of the type of new energy source, capacity and fault characteristics. Simulation analysis shows the applicability of the scheme.

The existence of delay in the digital controller in active damping causes the control system to be highly susceptible to instability when the resonant frequency of the filter is shifted due to changes in the grid impedance under a weak grid. A strong robust control method of the LCL-type active power filter under the weak grid is proposed, which is based on the traditional capacitor-current-feedback active damping, and the feed-forward point is moved forward to the input of the current controller, and the structure of the current controller is improved based on which the active damping controller and the current-loop controller are combined to obtain a composite controller. On the basis of this method, the structure of the current controller is improved, and a composite controller is obtained by combining the active damping controller and the current loop controller. The stability constraints of the system are deduced based on the Juli's criterion and the Nyquist's stability criterion, in order to realize the parameter design of the composite controller and guarantee the strong robustness of the system under the weak power grid. Simulation experiments show that the LCL-type active power filter designed according to the proposed method can effectively compensate the harmonic currents in the grid and can guarantee the strong robustness of the control system to the grid impedance changes of the grid.

A method for wind power prediction based on variational autoencoder and attention-based Seq2Seq model is proposed to address the complex characteristics of multiple influencing factors, limited amount of effective data, and long time series. The method involves collecting wind tower data and corresponding continuous power values to construct a sample set. Then, variational autoencoder model is used to augment the samples, and generate sufficient samples to support the training of the prediction model. A regression analysis model is built to explore the mapping relationship between multiple monitoring indicators from the wind towers and the continuous power values. The augmented different indicators are separately input into an attention-based Seq2Seq model for indicator time series prediction. Numerical weather forecast data is used to refine the prediction results, yielding more accurate weighted prediction results for the indicators. By inputting real-time wind tower data and numerical weather forecast data into the trained weighted prediction model and regression analysis model, multi-step wind power prediction can be achieved. The proposed method is validated using actual operational data from wind farms. The results show that compared to traditional time series prediction methods, the approach based on variational autoencoder and attention-based Seq2Seq model provides more accurate wind power predictions with smaller reconstruction errors.

Distributed photovoltaic system has wide range of points, coupled with a severe lack of available data, making it difficult to detect equipment failures in a timely manner, and is prone to long-term operation with faults, reducing the life cycle power generation. An anomaly detection method of distributed photovoltaic system based on measurement data is proposed, taking advantage of the characteristic that abnormal faults in photovoltaic system will ultimately affect power generation output. Firstly, the characteristics of solar irradiance on sunny days are analyzed, and a clear sky day screening mechanism is proposed, conducting correlation analysis on different power stations, and obtaining photovoltaic power stations with high output correlation as horizontal references.Then the output curves of the tested power station on different clear days for longitudinal comparison are selected to eliminate various interference factors in abnormal detection. The output data excluding the above interference is input into the quantile regression temporal convolutional network model for training to obtain the fitting range of photovoltaic normal output, and then abnormal output of distributed photovoltaic power station is detected according to the normal output range. The simulation analysis using actual photovoltaic system data shows that the proposed method can accurately identify distributed photovoltaic systems with faults and anomalies, promoting the refined operation and maintenance of distributed photovoltaics.

In the electricity market environment, joint bidding of energy storage and wind power can deal with the uncertainty of wind power output and improve the comprehensive benefit of the wind power enterprise. In this paper, the model and algorithm of joint bidding of wind power and compressed air energy storage are studied. Firstly, by considering the operation characteristics of wind power and compressed air energy storage and the rules of electricity market, a bidding model for wind power and compressed air energy storage to participate in the energy market and frequency regulation market is established. Secondly, the proposed model is described as a multi-stage stochastic programming problem in the framework of Markov decision process, and then solved based on the stochastic dual dynamic integer programming algorithm. The algorithm realizes the period decomposition of the proposed model through Lagrange cut, which can effectively solve the problems of randomness and discreteness, and has high solving accuracy. Finally, case studies verify the effectiveness of the proposed method.

The analysis of the complexity of electric vehicle （EV） users′ charging demand and action law is the fundamental work to improve the robustness and resilience of distribution network in the context of large-scale EV integration. Existing researches made little efforts on the description of the derivative nature of EV charging demand. Firstly, starting from the origin of charging demand, a transportation power coupling network model is established based on the spatiotemporal trajectory of EV user activity travel to reveal the medium mechanism of EV in the transportation power coupling system. Secondly, from the perspective of transportation, an activity travel chain model is established considering the characteristics of EV user activities and travel behavior, and EV charging demand is predicted based on the Logit model. A logit-based model is applied to predict EV charging demand. Then, from the perspective of the power grid, an evaluation system for the impact of EV grid connection is constructed based on indicators such as voltage offset severity, maximum power transmission margin of the line, and power system line losses. Finally, based on simulation, the influence of EV charging loads in different regions on the power grid under different scenarios is analyzed with the coupling degree of transportation and electricity as the regulating variable, providing theoretical support for guiding EV users to schedule and manage orderly charging in the temporal and spatial dimensions.

The large-scale disordered integration of electric vehicles （EVs） into power grid poses many problems, such as enlarged power fluctuation for grid and increased charging cost for users. To solve these problems, this paper proposes an optimized EV charging method based on deep reinforcement learning （DRL）. Firstly, an EV ordered charging scheduling model is established aiming at minimizing the power fluctuation and user charging cost. Secondly, the EV charging behaviour is formulated as a Markov decision process （MDP） which evaluates the priority for each charging period based on load prediction information and time-of-use electricity tariff. The charging behaviour of EVs is controlled by the priority. The twin delayed deep deterministic policy gradient （TD3） algorithm is adopted to solve the MDP problem, which quickly optimizes EV ordered charging strategy. Finally, the effectivenesses of the proposed method in reducing charging cost and load fluctuation of distribution network are verified by case studies based on numerical examples.