In order to analyze the influence of harmonics on the stability of line commutated converter based high voltage direct current (LCC-HVDC) system. It is urgent to establish an accurate model of LCC-HVDC considering harmonic coupling. Based on the theory of harmonic state space (HSS), an impedance model of 12-pulse LCC is established considering frequency coupling effect and control system. The constructed AC/DC harmonic impedance model can match the sweep frequency results in a wider frequency band. Finally, the correctness of the proposed LCC-HSS impedance model is verified by comparing the PSCAD electromagnetic transient simulation results with the HSS impedance model calculation results. The LCC-HSS impedance modeling method improves the accuracy of mathematical modeling for LCC converter stations and can adapt to impedance modeling of LCC converter stations in various modes, providing a more accurate model for stability analysis and parameter optimization of LCC systems.

When an open circuit fault occurs in the internal switching components of a modular multilevel converter （MMC）, it can cause distortion of the output current, increase inter phase circulating current, and seriously endanger the safe and stable operation of the equipment. Accurately identifying the submodules where open circuit faults occur in switch devices is crucial for the stable operation of MMC systems. The fault characteristics of IGBT and freewheeling diode open circuit faults are analyzed, and an IGBT open circuit fault diagnosis method based on the ratio of capacitance voltage increment is proposed to accurately identify the submodules where IGBT open circuit faults occur. An open circuit fault diagnosis method for freewheeling diodes is proposed based on the cumulative analysis of periodic fault characteristics, utilizing the relationship between bridge arm current and switching function when an open circuit fault occurs in a diode. This method accurately identifies the submodule where the diode open circuit fault occurs. And a redundant control scheme is designed. An MMC rectifier stage model is built based on the RTDS simulation platform, and the proposed fault diagnosis strategy can accurately determine the faulty components and submodules, solving the problem of difficult diagnosis of open circuit faults in switch devices.

Polypropylene has become a hotspot in the research and development of thermoplastic cable insulation materials due to its advantages of high working temperature, excellent insulation performances and recyclability. The molded joint employs the same insulation material as the cable body as joint insulation, which can not only find a solution for the movable interface of the prefabricated joint, but also help to realize the recyclability of the whole PP cable line system. Firstly, a 10 kV PP cable molded joint is manufactured, and the crystal structure, electrical performances and mechanical properties performances of the PP cable insulation and molded joint insulation slices are studied. The results show that the PP molded joint insulation sample has higher crystallinity, larger microcrystal size, and higher melting temperature compared with the cable insulation slices. For electrical performances, the molded joint insulation has higher volume resistivity, smaller tangent of dielectric loss angle, and higher power frequency breakdown field strength. For mechanical performances, the elongation at break of the molded joint insulation is smaller and the elastic modulus is larger. These differences may be due to the different processing methods of the cable and the molded joint. The cable is cooled by water cooling, and small spherulites are formed in the rapid cooling process, while the molded joint is cooled naturally at room temperature, and the PP molecular chains are fully folded and arranged to form large crystals. The different crystalline structures of PP cable and molded joint insulation lead to the differences in electrical and mechanical performances. However, the performances of cable insulation layer and molded joint insulation layer are still consistent. The results verify the feasibility of applying the molded joint in PP cables, and provide references for the further research and development of high-voltage PP cable and molded joint.

In order to ensure the safety of the persons near the buried pipelines, the current interference of HVDC grounding electrode to buried pipelines and its personal safety distance are investigated. Firstly, CDEGS simulation software is used to establish the interference model of buried pipeline by HVDC grounding electrode, and the main factors affecting the safety distance are screened. Secondly, the safe distances of buried pipeline in different uniform soil and multilayer complex soil are studied. Finally, the safety distance conversion relationship between uniform soil and complex soil is established. A method is proposed to obtain complex soil safety distance by using maximum pipelines ground potential and utilization coefficient correction. The research results show that soil resistivity has the greatest influence on the safety distance. If the safety distance of uniform soil is less than 8 km, the safety distance of complex soil can be obtained by using a correction coefficient of 0.85. If the uniform soil safe distance is between 8 km and 12 km, the relationship between the correction coefficient and the safety distance of uniform soil satisfies power function. If the safety distance of uniform soil is greater than 12 km, a correction coefficient of 0.53 is recommended. This proposed method can provide reference for the design of HVDC grounding electrode and pipeline, and also for the site selection of HVDC grounding electrode in future.

The increasing proportion of renewable energy in the new power system brings new challenges to the security and stability of the power system. The random production simulation of the uncertainty of the renewable energy output has guiding significance for the planning and operation of the power system. Firstly, the uncertainty of new energy output and load prediction is modeled. The principal component cluster analysis method is used to process the temporal correlation of new energy output and load data, and the equivalent variable function method is combined to carry out random production simulation. Finally, the correctness of the theoretical method proposed in this paper is verified by an actual power system example in a province, which can provide new ideas and methods for stochastic production simulation of new power systems.

With the integration of a large number of high-frequency power electronic devices into the power grid system, CPU based real-time simulators are difficult to achieve smaller step size simulation calculations due to serial computing reasons. Therefore, a field programmable gate array (FPGA) based simulation method is adopted, the mixed step simulation of FPGA and CPU has gradually become a research hotspot. However, the resource consumption of FPGA has become a limiting factor for simulation scale in power systems with a large number of distributed power sources. A resource optimization scheme suitable for CPU-FPGA joint real-time simulation is proposed to address this issue, and a transient joint real-time simulation platform for power systems is built. Taking the double-fed wind power system as an example, the entire system is divided into CPU and FPGA parts, with asynchronous communication between the two parts using Gigabit Ethernet. The optimized simulation process is used within the FPGA part to reduce the usage of hardware resources. Finally, by comparing real-time simulation results with Simulink offline simulation and traditional simulation models, it is proven that adopting this approach can effectively reduce the consumption of FPGA resources while ensuring simulation accuracy.

The day-ahead power forecasting of wind cluster is one of the important foundations for provincial level or above power dispatching and control centers to make power generation plan and promote wind power consumption. Day-ahead wind power forecast （next 0 to 24 hours） is essentially to build mapping model between numerical weather prediction and actual power. Fully mining the deep mapping relationship between them is an important way to improve the accuracy of wind power prediction. In this paper, grid numerical weather prediction is used to establish wind cluster power forecasting models by using residual network, which mines the correlations between three-dimensional grid meteorological distribution and power. Case studies show that the proposed method is better than the existing mature methods in terms of progressiveness and adaptability.

The resonant grounding system has a complex operating environment and variable single-phase grounding fault occurrence status. The distribution network traditional fault section location methods based on a single characteristic quantity have their own location blind zones, and the reliability and applicability of methods are also difficult to be guaranteed. To solve the above problems, a fault section location method for distribution network based on multi-dimensional waveform difference clustering analysis without tuning is proposed. According to the waveform difference characteristics of the zero-sequence current transient and steady-state components at each detection point upstream and downstream of the fault point, the transient and steady-state waveform differences between adjacent detection points are quantified by discrete Fréchet distance algorithm, and fuzzy C-means (FCM) is used to synthesize multi-feature information to divide the normal section and fault section. The location blind zones with single feature are eliminated effectively and the accurate fault section location is realized adaptively. The simulation results of PSCAD/EMTDC and 10kV true experimental test show that the proposed method not only ensures reliability and applicability under different fault conditions, but also has high resistance resistance and noise resistance.

In order to meet the detailed simulation requirements for large-scale wind power integration into regional power grid, an equivalence method applicable to regional power grid integration with wind power is proposed. Unlike traditional dynamic equivalence methods that focus on the main grid, the equivalence of regional power grid with enriching wind power resources must not only examine the characteristics of intensive wind power integration within the regional power grid, delineate the influence scope of each wind farm, but also comprehensively consider the constraints on the source and load characteristics of the external grid at the boundary nodes. The proposed method can simplify the external grid on a large scale while maintaining consistent steady-state and transient characteristics, achieving equivalent modeling of the regional power grid. Firstly, rather than the traditional scheme of equating the grid at the access point of wind farms using short-circuit ratio and other electrical strength indexes, it is proposed to define the study area and the external area to be equated by residual voltage. Then, the regional power grid is further simplified by considering topological constraints of the regional power grid. Additionally, the equivalent model of the external system is constructed by combining an equivalent synchronous generator and composite load at the boundary node. Finally, the proposed method is applied to a practical regional power grid with enriching wind power resources. The effectiveness of the proposed method is verified by the simulation results.

Wind farms and photovoltaic new energy stations are important in constructing new power system, and their electromagnetic transient real-time simulations attract more and more attentions. Due to the complexity of the network solution and the lengthy serial logic for the new energy station, it is difficult to achieve real-time simulation on the currently mature multi-core CPU simulator. A full-topology real-time simulation modeling method for the wind farm and photovoltaic（PV） power station based on multi-core CPU is proposed. Based on the principle of electrical network equivalence, this method establishes a four-terminal equivalent model of power units, and designs a parallel simulation architecture suitable for the multi-core CPU according to the simulation process of this model to realize real-time simulation of new energy stations. The proposed station model uses the calculation for internal voltage to retain high-precision full topological information, and at the same time effectively reduces the network solution resource occupation and shortens the serial logic by reducing the order. After simulation tests, compared with the RTDS model, the model built by this method not only ensures that the accuracy error is less than 4%, but also enables a threefold increase in the scale of the power station that can be simulated in real time, effectively reducing the hardware cost of real-time simulation.

The high proportion of photovoltaic system connected to the distribution network causes the voltage to exceed the limit. The traditional reactive voltage adjustment scheme fails to work in the high impedance ratio distribution network. The technical scheme which uses the energy storage system to control voltage is widely adopted. However, the energy storage system in the existing technical scheme may cause excessive charging and discharging problems during the voltage control process. Aiming at these problems, a voltage control strategy is proposed based on the coordination of energy storage active power and on-load tap changer (OLTC）. Based on the analysis of the influence of photovoltaic connection on the voltage of distribution network, the measured voltage data are used to control the active power of energy storage, so as to realize the rapid regulation and control of the voltage level of access nodes. At the same time, the reference values of the state of charge of energy storage are added to the OLTC control link. And the OLTC tap actions are jointly controlled by the node voltage information and the state of charge of energy storage, so as to realize the regulation and control of the overall voltage level of the distribution network and avoid excessive charging and discharging of the energy storage system. Finally, the simulation results verified in IEEE 13-node system show that the proposed strategy can solve the problem of voltage exceeding the limit in the distribution network, and avoid excessive charging and discharging of the energy storage system. So the service life of the energy storage system is extended and the rational utilization of the resources in the distribution network is realized.

Aiming at the increasingly widespread application of hydrogen fuel cells in microgrid, this paper relies on the wind/photovoltaic/hydrogen fuel cell DC interconnection system in the hydrogen-power coupled DC microgrid demonstration project in Cixi, Ningbo, to study the port power coordinated control strategy of the three-port DC transformer, which is the core equipment for hydrogen fuel cell access. To shorten the simulation time, a switching period averaging model applicable to a multi-modular series-parallel multi-active bridge structure DC transformer is proposed and adopted in this paper. Moreover, an energy coordination control strategy of the fuel cell-energy storage hybrid supplying system is proposed by applying this equivalent simplified model. The control strategy can optimally adjust the power instructions of each port according to the state of charge （SOC） of energy storage, the power capacity of each port, and the dynamic response speed of the fuel cell. Thus, it can prevent the fuel cell power generation power change too quickly, avoid the overcharges and over-discharges of energy storage, and maximizes the proportion of hydrogen power generation. The simulation results using MATLAB Simulink and experiment waveforms verify that the coordinated control strategy works well in various control modes of the DC transformer. The actual power of the port with power control mode can follow the rapidly changing instruction value, and the actual voltage fluctuation of the port with voltage control mode can be controlled below 4% when the load changes suddenly.

High permeability of photovoltaic in distribution network may cause problems such as net load curve deformation, voltage deviation and network loss increase. A coordinated optimization strategy of active and reactive power of distribution network considering linear modeling of four-quadrant energy storage power regulation is proposed. Firstly, the active and reactive power regulation capability and operation characteristics of various regulatory resources in the active distribution network are analyzed, especially the operation boundary of photovoltaic and energy storage power. Based on this, a linear modeling method of four-quadrant energy storage power regulation is proposed. Then, considering the power capacity constraints and energy capacity constraints of the energy storage, the objective function and constraint conditions of the distribution network coordinated optimization model are established, which can be solved through the second-order cone algorithm. Finally, the improved IEEE 33-bus system is used to verify the effectiveness and feasibility of the proposed method.

The modern power distribution network with high penetration of photovoltaic （PV) and energy storage system （ESS) has the characteristics of large scale, numerous equipments, complex and changeable operating conditions. The traditional centralized optimization method is no longer applicable. At the same time, onload tap changer （OLTC) and parallel capacitor banks （CBs) are all traditional discrete devices, which lead the active and reactive power optimization model of distribution networks to nonconvex. However, the current general distributed algorithms are only suitable for solving convex models. In order to solve this problem, a distributed active and reactive power optimization method which considers the model nonconvexity is proposed. The distributed stochastic active and reactive power optimization （DSARPO) model is constructed to improve the operation economy of distribution network by using stochastic programming theory to characterize the uncertainty of photovoltaic active power. With consideration of the model nonconvexity, a distributed algorithm based on alternating optimization procedure （AOP) is proposed to achieve reliable solution of the model. Finally, an example is analyzed through the PG&E 69-bus system to verify the effectiveness of the proposed DSARPO method and the reliability of the distributed solution method.

Distribution network disconnection fault diagnosis is generally based on voltage or combined with voltage and current relationship, and it is assumed that three line voltages and three phase currents can be obtained, without considering the actual configuration of measuring devices. Actual overhead distribution lines are usually only configured with double PT or single PT, and can only obtain two or one line voltage, resulting in invalid of existing diagnosis methods under partial fault conditions. Firstly, the line voltage equationas under the condition of disconnection fault are established, and the feasibility of using two line voltages relation to diagnose the disconnection fault under the condition of double PT and the infeasibility under the condition of single PT are demonstrated. Secondly, a PT sliding combination power supply configuration strategy is proposed. Single PT is innovatively used for disconnection fault diagnosis, which further improves the diagnosis accuracy. Finally, PSCAD/EMTDC is used to build a medium voltage distribution line model to verify the effectiveness of the proposed power supply configuration strategy and disconnection fault diagnosis strategy.

In view of the problem that the image details are lost and the edges are blurred in the intelligent patrol inspection of electric power, resulting in the wrong target detection and recognition, a super-resolution method based on per-pixel selt-paced adversarial network （PSPA） is proposed. This method is based on the generation of adversarial network, adds multiple attention mechanisms, and restores the detailed texture through pixel by pixel comparison.The experimental results show that the super-resolution images generated by this method are not only superior to the existing algorithms in human visual system, but also 6.2 and 0.099 3 times higher than the existing algorithms in PSRN and SSIM. Then Yolov3 is applied on the super-resolution images recovered by different algorithms in the UAV transmission line insulator dataset and the power construction helmet wearing dataset. The experimental results demonstrate that the proposed method could not only decreases the residual error rate, but also improves the detection confidence as high as the high-resolution images