Abstract:

Power system is a time-varying complex system. In recent years, data-driven machine learning method has been widely used in the field of transient stability assessment of power system. However, when the power system is subjected to a large disturbance and the working condition changes, the machine learning model needs to be trained according to the new operating data. Thus, it is difficult to timely respond to transient stability assessment of the system under the new topology structure. To solve this problem, a model update mechanism is proposed in this paper, which updates the model according to different conditions. In addition, an oblique double random forest with multisurface proximal support vector machine (MPSVM) (MPDRF) model is introduced as a classifier to assess the stable state of power system. The simulation test on the New England 10-machine 39-bus system verifies the effectiveness of the proposed method. The results show that the method combined with update mechanism has high assessment performance, compared with the traditional method.

Abstract:

Transient voltage instability is one of the important factors that threaten the stability of power system. The dynamic reactive power reserve and supporting capacity of the power grid with a high proportion of renewables decrease sharply, and the control models and operation characteristics of grid-connected renewables are diverse. Thus, the reactive power voltage of the system often fluctuates rapidly after a fault occurs, which leads to a more prominent voltage stability problem. In response, a transient voltage instability identification method based on the Koopman operator is proposed in this paper to avoid power system outage accidents caused by voltage instability in time. Firstly, the Koopman operator extraction method of Hankel matrix enhanced dynamic mode decomposition (HeDMD) is proposed with short-time wide-area measurement data after fault. Secondly, the amplitude of the Koopman operator is defined, and the dominant Koopman mode is obtained in descending order. Then, based on the time domain prediction signal of the dominant Koopman mode, the maximum Lyapunov exponent (MLE) is calculated to identify transient voltage instability. Finally, the effectiveness of the proposed method is verified using the Nordic32 test system and the standard system of China Electric Power Research Institute. Compared with the traditional method, the proposed method has more advantages in accuracy and rapidity of transient voltage instability identification. The simulation experiments prove its applicability in power grid with a high proportion of renewables.

Abstract:

In recent years, with the large-scale integration of power electronic interfaced power sources such as wind power and photovoltaics into the power grid, the overall inertia level of the power grid has decreased, and the nodal inertia has shown spatial distribution differences, significantly increasing the risk of system frequency instability. It is urgent to quickly evaluate the distribution of inertia in the power grid so that dispatch and operation personnel can timely formulate effective inertia control measures. Therefore, a method for estimating the power grid nodal inertia based on maximum likelihood identification is proposed. Firstly, by using frequency and active power measurement data, the autoregressive moving average model with exogenous inputs (ARMAX) for inertia estimation is constructed. Secondly, the unknown parameters in the ARMAX model are identified via the maximum likelihood identification method. Additionally, the transfer function of active power and frequency of the node and counter are employed to determine the estimated inertia and the required minimal measurement data length. Finally, the simulation test is conducted based on the improved CEPRI-36 node system to verify the effectiveness of the proposed method.

Abstract:

To address the low detection accuracy of traveling wave heads for long-distance transmission line faults as well as the location deviation caused by the clock asynchronization of fault detection devices and the uncertainty of traveling wave velocity, this paper proposes a new fault location method for transmission lines based on improved variational mode decomposition (VMD) combined with Teager-Kaiser energy operator (TKEO) to detect the fault traveling wave head and the double-ended traveling wave location formula independent of timing and wave velocity. Firstly, the crested porcupine optimizer (CPO) is used to optimize VMD parameters. Then the line mode and zero mode components of the fault traveling wave are decomposed by VMD, and the initial wave heads of the components are detected by TKEO. Finally, according to the difference in time to reach the measuring point between the initial wave heads of the line mode and zero mode components, the proportional relationship between the fault distance and the difference distance is written, and the double-ended traveling wave location formula independent of timing and wave velocity is obtained. A 220 kV transmission line fault simulation model is built using Matlab/Simulink. The simulation results show that the method has good applicability under noise, different ground resistances, and different fault types, and the accuracy of the location results is high.

Abstract:

The safety and stable operation of power grid is the prerequisite for reliable transmission, transformation, and distribution. Therefore, when the power grid fails, it is very important to locate the fault quickly and accurately and shorten the fault time. Firstly, the information of component switching value and electrical quantity is obtained from the relevant monitoring system of the power grid. The initial decision table of relevant switching value information is formed according to the fault area, and the effective signal of electrical quantity information is extracted. Then, the rough set theory, Bayesian network, Hilbert-Huang transform (HHT), and other theories are used to calculate the component fault degree and distortion degree. Subsequently, the improved D-S evidence theory is employed to fuse the fault degree of component switching value with the distortion degree of electrical quantity. Finally, the local topology of a regional power grid is used to test the improved Bayesian network model. The simulation results show that the model can improve the diagnosis speed. The IEEE 39 node is used as an example, and it is verified that the introduction of switching value can improve the diagnostic accuracy, and data fusion reduces the uncertainty in the evaluation model.

Abstract:

To adapt the grid to the requirements of intelligentization and the dispatching and control cloud technology route, this paper proposes a relay protection setting calculation method for power grid based on distributed parallel computing. First, the cluster architecture of the Spark distributed computing platform is introduced, and the key issues of distributed parallel computing, such as load balancing, system fault tolerance, etc. are analyzed. On this basis, a computing system for relay protection setting calculation based on Spark is designed. Secondly, the extra-high-voltage power grid setting calculation in the computing system is analyzed, and the principles of protection and setting calculation are summarized. Next, to realize the preprocessing operation of the initial power grid data input to the system, data feature selection is realized by improving the monarch butterfly optimization algorithm. Finally, simulation analysis was conducted on specific instances in a region to verify the effectiveness of the system. The simulation results prove that the computing system enables the setting calculation of the power grid to accommodate intelligentization and the development of dispatching and control cloud, which can effectively increase the calculation speed and improve the reliability of power grid operation.

Abstract:

The strong variability, spatiotemporal randomness, and nonlinear controlled characteristics during faults of inverter-interfaced distributed generators (IIDGs) such as direct-drive wind turbines and photovoltaic generators pose significant challenges to the existing relay protection systems in power grids. When faults occur in high-voltage lines of substations with integration of IIDGs, the fault characteristics are more complex compared to those in the case of the access of traditional synchronous machine sources to high-voltage lines, which makes it difficult for conventional positive-sequence voltage-polarized phase-comparison distance protection to adapt to. As a result, the performance of protection for power grids deteriorates. This paper analyzes the reasons for the incorrect operation of traditional positive-sequence voltage-polarized phase-comparison distance protection due to the fault characteristics of IIDGs. Based on the electrical quantity characteristics at the protection installation point, it proposes a new method for distance protection of high-voltage lines by using power frequency phasor calculation. When the fault voltage at the protection installation point is relatively high, using the voltage at the protection installation point as the polarizing voltage for phase-comparison distance protection correctly determines the fault distance. For the low fault voltage at the protection installation point, the fault direction is determined by analyzing the phase relationship between the memory voltage at the protection installation point and the current flowing through the point. Simulation results show that the proposed distance protection method exhibits excellent performance unaffected by factors such as fault location and fault type. The proposed distance protection method is applicable to AC line protection for power grids with the integration of IIDGs, and it eliminates the issue of blind zones of protection for outlet faults.

Abstract:

The dead zone and limiting of a hydraulic turbine governor have a great influence on the characteristics of the ultra-low frequency oscillations (ULFOs), which can even lead to the oscillation of the positive damping system under large disturbance. However, the existing control methods for ULFOs mainly focus on optimizing proportional?integral?derivative (PID) control parameters of hydraulic turbine governors to improve system damping, without considering the dead zone and limiting parameters concurrently. This paper proposes a ULFO control method considering the dead zone and limiting optimization of the turbine governor and verifies its effectiveness and superiority. Firstly, in a non-smooth single hydro-generator power system (SHPS) with dead zone and limiting, the causes of continuous oscillation with constant amplitude involving dead zone and limiting are analyzed, and the influence of relevant parameters is analyzed, obtaining the qualitative demand for control parameter optimization. Then, a comprehensive evaluation index including the suppression effect of the non-smooth part, the amplification effect of the smooth part, and the performance of the primary frequency regulation is established. Additionally, the optimization parameters are solved using a particle swarm optimization algorithm, and a specific application case of the proposed method is presented. Finally, the proposed method is compared with the control method that does not consider dead zone and limiting optimization, in three simulation systems: SHPS, a four-machine two-area system, and a large-scale power grid model. The proposed method shows improved performance in primary frequency regulation and resistance to disturbance.

Abstract:

With the continuous growth of electricity demand, the high penetration level of renewable energy has become a trend in the development of distribution networks. However, the large-scale integration of distributed generators (DGs) has results in more prominent problems of power flow, congestion, and voltage fluctuations, which brings challenges to the flexible operation of distribution networks. At the same time, it is necessary to further consider the capacity of lines and the voltage deviation of terminal nodes, so that distribution networks can still retain schedulable space, that is, system flexibility. Therefore, it is necessary to rationally analyze the DG consumption capacity of distribution networks based on their actual operation status. This paper proposes a DG consumption capacity analysis model that considers system flexibility, including evaluation methods for flexibility, improvement of flexibility, and DG consumption analysis model. Through the simulation analysis of the modified IEEE 33-node test system and the construction of several scenarios, the paper verifies the effectiveness of the proposed method and the improvement of power consumption and system flexibility by soft open point.

Abstract:

Affine power flow calculation is an effective method for uncertainty analysis in power systems. With the increasing integration of distributed generation (DG), the existing Gaussian iterative affine power flow algorithm faces challenges in dealing with distribution networks that incorporate DG controlled by different strategies. To address these problems, a Gaussian iterative power flow algorithm for a distribution network that integrates an affine model for DG is proposed. Firstly, an uncertainty power flow model is established based on affine arithmetic, and a Gaussian iterative model for the affine power flow model is constructed. Secondly, the affine models for DG under different control strategies are established. By using common noise processing, reactive power correction, and voltage boundary correction, a Gauss iterative power flow algorithm for DG is proposed. Finally, through simulations on the IEEE 33-bus system, the effectiveness and reliability of the proposed algorithm are validated. The results indicate that the proposed method is capable of calculating the affine power flow in distribution networks with different DG control strategies, offering low conservatism and high computational efficiency.

Abstract:

In consideration of the problem of high-resistance grounding fault line selection and section location in asymmetric distribution networks, a fault line selection and section location method for asymmetric distribution networks based on the fast switch arc suppression device is proposed. Firstly, the basic principle of the fast switch arc suppression device is analyzed. On this basis, the electrical quantity properties of each feeder in the distribution system prior to and following the operation of the device are analyzed in depth after the high-resistance grounding fault occurs in the system. Through theoretical derivation, it is found that the fault feeder and the non-fault feeder before and after the operation of the device differ in the variation of the zero-sequence equivalent admittance. Based on this, the fault feeder identification function is established, and the line selection and section location criteria are constructed. The proposed method is verified by PSCAD/EMTDC simulation software. The results show that the proposed method can accurately identify high-resistance grounding fault feeders and sections and is not affected by the three-phase imbalance of the distribution network.

Abstract:

The access of the medium-voltage (MV) microgrid changes the radial structure of the traditional distribution network, causing the distribution network to operate in a multi-terminal power supply mode. The single-phase grounding (SPG) fault is a frequent fault type in distribution system, and multi-terminal power supply operation complicates the grounding fault characteristics of the distribution network. Thus, the traditional SPG line selection method is no longer applicable. This paper analyzes the network structure of the MV microgrid and the MV distribution network and establishes a multi-level multi-terminal power supply distribution network model with MV microgrid access. Then, the SPG fault characteristics of the multi-terminal power supply distribution system under the small current grounding mode are analyzed from the perspective of two-level bus nodes. Furthermore, based on the analysis of fault characteristics, a line selection method suitable for multi-terminal power supply networks with microgrid access is proposed, and a detailed design is carried out in terms of equipment layout and line selection task allocation, forming a relatively complete automatic line selection system. At the same time, the proposed line selection scheme is based on the two-level bus nodes, which can be analogously applied to the multi-level multi-terminal power supply distribution network. The simulation results show that the proposed line selection scheme can be effectively applied to the multi-terminal power supply distribution network with MV microgrid access.

Abstract:

Distribution lines are an integral part of modern power system, which directly influence the safety and stability of power supply. Distribution network fault location can be classified into precise fault location and fault segment location. Considering the complexity of distribution network structure, this paper proposes a fault segment location method based on Gramian angular field (GAF)-ResNet50. The one-dimensional time series is converted into a two-dimensional GAF image by the GAF algorithm, and the deeper-level fault features of the signal are extracted from the GAF image by using the framework of residual neural network so that the fault areas can be identified more accurately. To verify the effectiveness of the proposed method, the study builds an IEEE 13-node distribution network model on the MATLAB platform to generate fault data and conduct the simulation of fault segment location. The simulation results show that the proposed method can quickly and accurately locate fault segments, with a positioning accuracy of more than 98%, and has good robustness to noise.

Abstract:

The large-scale integration of distributed photovoltaics and electric vehicles exacerbates the three-phase imbalance in low-voltage distribution network (LVDN), which further leads to issues such as current in neutral wire and voltage fluctuations, preventing the simplification of three-phase line impedance identification into three independently solved single-phase networks with unified power flow. For this reason, a set of line impedance identification methods based on hierarchical deconstruction and phase recursion are proposed, addressing the issue of neutral line current imbalance using the principle of child node voltage correlation. Firstly, three layers are obtained through hierarchical deconstruction: the substation layer, branch layer, and user layer, with each layer having its internal node connected to the subnetwork. A voltage matching model is then constructed for the phase line to retrace the branch layer’s line. Next, a node power-checking model is used to estimate the node phase angle in the substation layer. The neutral line parameter is recursively derived according to the electrical vector relationship, and errors in the recursion process are minimized to obtain the line impedance identification results. Finally, the effectiveness and accuracy of this method are verified through actual system and load data.

Abstract:

With a large number of power distribution substations (transformers), three-phase imbalance is a commonly observed issue. Due to the high randomness of single-phase loads in the substation area, the amplitude ranking of three-phase voltages and the imbalance degree according to the traditional algorithm change constantly. Based on the feature analysis of the sampled data, differences in the dominant phase amplitudes of three-phase voltages in each substation area during a certain period of time are obtained, identifying the severely unbalanced substation area for manual intervention. This serves as a feasible method to balance costs and benefits. A time-series matrix of the three-phase voltages for each substation area is constructed in this paper, and a method for analyzing the dominant features of phase amplitude differences based on singular value decomposition (SVD) and constructing related indexes is proposed. The results show that the submatrix corresponding to the maximum singular value is the absolute dominant component of the original three-phase voltage time-series matrix. The ranking of the dominant phase amplitudes can be obtained from the left singular vector, and the three-phase dominant imbalance index can be constructed using the maximum singular value and the left singular vector. This method addresses the problem of constantly changing phase amplitude ranking and can determine the amplitude of each phase for any selected time period. A test case verifies the dominance and uniqueness of the proposed imbalance indexes. Based on the proposed indexes, through a rolling cycle of identification and manual intervention, the overall three-phase balance of the distribution network can be improved and maintained. The findings offer valuable practical guidance.

Abstract:

The development of intelligent distribution networks has improved the self-healing ability and recovery speed of the power grid. However, when the system encounters large-scale power outages, the recovery process becomes more complex. Therefore, in response to the problem of power interruption in damaged distribution networks, this paper proposes a two-stage recovery strategy for damaged distribution networks that takes into account cold load pick-up (CLPU), aiming to generate a power restoration plan for distribution networks with switch control actions. The first stage generates traditional recovery and distributed power-assisted island power supply recovery plans that support feeder reconfiguration. In the second stage, the optimal switching operation sequence is generated and transformed into a mixed integer linear programming (MILP) problem, which enables the damaged distribution network to quickly recover to its final configuration. Finally, a distribution network recovery strategy is simulated in a multi-feeder 1069-node power system for single- and multi-line faults. The results show that the proposed strategy can effectively generate switching operation sequences, make reasonable use of all resources to quickly recover distribution networks, and improve the recovery speed and capacity of damaged distribution networks.

Abstract:

When business expansion plans of new access loads are formulated for traditional distribution networks, their timing and controllable and interruptible loads are not considered, which can easily cause loads with the same characteristics to be concentrated in the same power source. As a result, the peak and valley values of the power point are overlapped, and thus equipment utilization and business expansion capacity are lowered. In response, this paper proposes an optimal business expansion model for distribution networks, which considers load timing and power source matching. Firstly, the paper proposes a fuzzy C-means clustering method based on the improved nutcracker optimization algorithm to obtain the temporal load clustering library. Secondly, it proposes a temporal load pattern recognition method for new installation users based on membership function. After that, the controllable and interruptible load of the power point is integrated into the above model, and an optimal model for the business expansion of the distribution network is established. Finally, simulation verification is conducted using an actual power grid as an example, demonstrating the effectiveness of the proposed method.

Abstract:

Topology information is the foundation of advanced analysis functions in distribution network, such as power flow calculation, state estimation, and fault diagnosis. Due to the inability of some nodes in the low-voltage distribution network to upload their own operational status, the existence of these implicit nodes poses a huge challenge to topology identification. This paper proposes a topology identification method for low-voltage distribution networks based on latent tree model and cluster search. Firstly, a Bayesian network with embedded implicit nodes is proposed, which is defined as a latent tree model to provide probabilistic representation for all possible low-voltage distribution network topologies. Then a cluster search algorithm is proposed to generate candidate topologies, and the accuracy of the candidate topologies is evaluated using Bayesian information criteria. Finally, simulation and experiments are conducted to demonstrate the effectiveness and robustness of the proposed method.

Abstract:

In order to improve the accuracy of switchgear fault detection, this paper proposes a fault detection algorithm for switchgear based on RFID sensors and deep learning. Firstly, RFID sensing tags are designed to collect the current signals and temperature of the switchgear. Secondly, the collected signals are subjected to deep-level feature extraction through a deep belief network (DBN), and sparse coding (SC) is integrated into the DBN to improve its detection accuracy. Finally, in order to improve the detection speed, an extreme learning machine (ELM) is used to classify and recognize the signals extracted from the features. The experimental results show that compared to other algorithms, the sparse DBN-ELM (SDBN–ELM) fault detection model proposed in the paper offers higher detection accuracy, faster recognition speed, and an accuracy rate of 99.63%.

Abstract:

Owing to the ability of both reactive power support (RPS) and grounding fault control (GFC), composite devices have attracted extensive attention. However, most existing composite devices have some disadvantages, such as high capacity and the need for additional power supply devices. To address these problems, a novel RPS and GFC composite device (RGCD) is proposed in this paper from the idea of making full use of existing station resources. Firstly, the topology and the operation principle of RGCD are introduced. The RGCD is composed of the capacitor and the arc suppression coil in the station and the multi-functional converter (MC). When the distribution network is in normal operation, most of the reactive power required by the load is compensated by the capacitor in the station, and the remaining reactive power is compensated by the MC. When the single-phase grounding fault occurs, the reactive power can still be compensated by the capacitor in the station, and the grounding fault is controlled by the arc suppression coil in the station and the MC. In brief, the capacity of the MC is decreased under different operating modes. In addition, the energy flow mechanism during GFC is analyzed in detail, and a P-Q two-phase arc suppression method based on DC-side voltage stability is proposed, which realizes the stable operation of RGCD without additional power supply devices. The simulation results demonstrate the effectiveness and feasibility of the proposed topology and regulation strategy.

Abstract:

In order to improve the utilization rate of clean energy and achieve low-carbon and economic operation of the power system, this paper constructs an optimal scheduling model of virtual power plants (VPPs) that aggregate wind power, photovoltaic power, oxygen?enriched combustion, gas turbine, waste heat recovery and utilization system, and proton exchange membrane electrolyzer (PEMEL). Firstly, the electrolyzer, methanation, and proton exchange membrane fuel cell (PEMFC) are modeled from the perspective of waste heat recovery and utilization, considering the joint operation of electrolyzers and oxygen-enriched combustion power plants. Next, from the perspective of oxygen-enriched combustion power plants, the feasibility of achieving low-carbon and economic operation is analyzed. Taking the minimization of operating costs as the objective function, a low-carbon optimal scheduling model for the VPP is constructed. Finally, in order to verify the effectiveness and feasibility of the established model, five scenarios are set up to compare and analyze the running results of the VPP. The results show that the proposed method can effectively promote the consumption of clean energy, utilize recovered waste heat, and achieve low-carbon and economic operation of VPPs.

Abstract:

In order to quantitatively assess the frequency/voltage support effects of grid-following and grid-forming stations, an evaluation index system to quantify the performance of the stations is proposed, considering the power support density and energy support density. First, based on the output response characteristics of grid-following and grid-forming controlled generating units in renewable energy stations, the characteristic quantities of power, voltage, and frequency are observed over multiple time scales in terms of voltage and frequency stability. Second, the coupling relationship between the indicators is considered comprehensively, and the base weights of the indicators are derived using the best-worst method (BWM) and the criteria importance through the inter-criteria correlation (CRITIC) method. The results of both indicator types are integrated into an improved technique for order preference by similarity to ideal solution (TOPSIS) evaluation model using the game-theory-based combinatorial assignment method to achieve a graded evaluation of the station’s grid-connected performance. Finally, the DIgSILENT simulation software is used to develop a model of the power grid in southern Xinjiang for station simulation experiments. The results show that the proposed evaluation index system can effectively compare and evaluate the grid-connected performance of grid-following and grid-forming stations from various aspects.

Abstract:

Existing studies in the planning of ultra-high power charging and switching stations lack a comprehensive depiction of user behavioral variability and stochasticity and the consideration of collaborative planning of distributed flexible resources such as photovoltaic and energy storage in the station. To this end, a two-tier siting and capacity determination method for integrated photovoltaic and energy storage charging and switching power stations involving multiple coupling factors is proposed. First, an electric vehicle charging and switching load prediction model considering user travel characteristics, temperature, and real-time road conditions is constructed. Second, to take into account user charging and switching needs and secure and economic development of distribution networks, photovoltaic and energy storage facilities are used for energy buffer, and their output is modeled in a refined manner. Additionally, a two-tier planning model for photovoltaic and energy storage charging and switching stations is constructed, with the upper model taking the optimal annualized return of charging and switching stations as the target for siting and the lower model taking the shortest distance from users to stations as the target for determining the service range of charging and switching stations. The results are fed back to the upper model, and the capacity determination is optimized in combination with prediction results. Finally, the validity of the proposed model and method is verified by taking the topology of the road network and IEEE 33 node distribution network coupling for example. The study shows that the method can make the ultra-high power charging facilities reasonably integrate with the charging and switching stations and provide theoretical and technical support for the planning of urban charging infrastructure.

Abstract:

The number of battery cells in a large-scale energy storage power station is enormous. The conventional convolutional neural networks achieve high prediction accuracy for battery capacity degradation. However, they have high demand for computational resources, which limits their application in practical battery management systems of energy storage power stations. To solve this problem, this paper proposes a battery capacity degradation prediction method based on a binary neural network. First, a lightweight model is designed by binarizing the network weights and activation functions, using the discharge capacity-voltage curve of the battery as input to output the cumulative distribution function values of key parameters. Subsequently, these parameters are solved using the bisection method and substituted into a hyperbola equation to predict the capacity degradation curve. Finally, experiments are conducted on a public lithium-ion battery dataset. The results show that under the same prediction accuracy as traditional neural network models, the proposed model reduces the number of parameters by 48.9% and improves prediction speed by 22.37%. This study reduces model computational complexity and hardware computational cost and also provides a more efficient and lightweight prediction method for battery management in large-scale energy storage power stations.

Abstract:

The rapid charging method of lithium-ion batteries for electric vehicles should also consider the influence of users’ behavior habits when considering the traditional goals such as balancing the safety, life, and charging time of the batteries. In this paper, a multi-stage constant-current charging method considering users’ behavior habits is proposed. First, Monte Carlo method is used to simulate users’ charging behaviors, and thus the state of charge (SOC) range of users’ optimal willingness to achieve is determined. Then, the dynamic equations of charging time and energy loss are established. Finally, the charging mode to obtain optimal charging current is determined by the particle swarm optimization algorithm. The results show that the proposed optimal charging strategy can reduce the maximum temperature rise of the batteries to a certain extent while improving the charging speed. It also significantly reduces charge loss, compared with the traditional constant-current and constant-voltage (CC-CV) charging method. The service life of lithium-ion batteries is indirectly improved by reducing energy loss during the charging process.

Abstract:

Modular multilevel converters (MMCs) have the problem of three-phase internal power imbalance, which easily leads to the damage of system stability and the increase of system loss. To address internal power imbalance of MMCs, this paper analyzes this problem and proposes a control method based on the injection of positive and negative sequence circulating current. The proposed method mainly includes four parts: the extraction of positive and negative sequence circulating current; the calculation of circulating current command value; the calculation of the number of input sub-modules; the acquisition of trigger signal. By injecting the extracted positive and negative sequence circulating current, this study realizes the balanced distribution of power between sub-modules of MMC and reduces switching losses. The simulation results show that the proposed method can significantly reduce the internal power imbalance of the MMC system, improve system stability, and reduce system losses.

Abstract:

The DC short-circuit fault current cannot be blocked by the modular multilevel converter (MMC) with traditional half-bridge sub-module (HBSM), which reduces the reliability of MMC based high voltage direct current (MMC-HVDC) systems. In this paper, based on two HBSM systems, a new bypass sub-module named diode clamp dual-half-bridge sub-module (DCDHBSM) with DC fault current blocking capability is proposed. Compared with other sub-modules possessing DC short-circuit current blocking capability, the proposed DCDHBSM requires fewer power devices and has lower operation loss. Moreover, a DC fault ride-through strategy suitable for DCDHBSM is designed, and sorting algorithm is utilized to balance the post-fault capacitor voltages of the sub?module. The results of simulation based on MATLAB/Simulink and physical experiment show that MMC assembled with the proposed DCDHBSM can quickly block DC fault current and realize fault ride-through.

Abstract:

The frequent occurrence of inter-turn short circuit faults of dry-type air-core series reactors in shunt capacitor sets gravely imperils the secure and steady operation of the power system. It is extremely crucial to research the variations of characteristic quantities prior to and subsequent to inter-turn short circuits in dry-type air-core series reactors. In this paper, Maxwell is used to establish the field-circuit coupling model, and the model accuracy is verified by comparing the calculated value of the analytical method with the test value of the manufacturer. On this basis, the inter-turn short-circuit model is constructed, and the electrical quantities when inter-turn insulation faults occur at different locations in different phases (inter-turn insulation aging phase and inter-turn short-circuit phase) are investigated. The results show that the absolute value of each electrical characteristic quantity increases with the change of fault location from the end to the middle in the inter-turn insulation aging phase and the inter-turn short-circuit phase. In particular, the loss factor and power factor have the largest rates of change, which can be used as electrical parameters for on-line monitoring of inter-turn short-circuit faults.

Abstract:

As critical equipment in the power system, the operation status of oil-immersed transformers directly affects the safety and stability of the power grid. Based on COMSOL, a finite element simulation software, a 10kV/400V oil-immersed transformer electromagnetic–structural multiphysics coupling calculation model is established to analyze the effect of added/unadded clamping devices on the core vibration displacement. On this basis, the vibration characteristics of the core, windings, and tank wall are analyzed, and their corresponding vibration acceleration signals are extracted. Subsequently, the Kendall and Spearman correlation coefficients are employed to calculate the correlation among the vibration signals of the core, windings, and tank wall, obtaining the optimal position for measuring vibration. Based on the Copula function, a vibration inversion model for the oil-immersed transformer is developed, which utilizes the external tank wall signals to reflect the operation status of the internal core and windings. The inversion results for the core and winding vibration acceleration are obtained under different load conditions. Finally, the model’s accuracy is evaluated using the coefficient of determination (R2) and mean absolute error. The results show that the inverse model achieves an accuracy rate of above 94%. The findings provide a reference for transformer vibration fault detection.