Abstract:

In view of the significant impact of extreme events on power system security, enhancing grid resilience has become a research hotspot. Therefore, a pre-event prevention and post-event coordinated recovery strategy considering intelligent soft open points (SOPs) and multiple emergency resources is proposed to improve the resilience of the distribution system. In the pre-event stage, to effectively reduce post-event resource scheduling time, a robust optimization model considering the uncertainty of power output is constructed based on predicted fault scenarios. A column-and-constraint generation algorithm is used to obtain pre-event deployment decisions. In the post-event stage, to minimize the curtailed load power through dynamic scheduling, a coordinated recovery strategy using intelligent SOPs and multiple emergency resources is developed, combined with pre-event deployment decisions and traffic network conditions. To address the computational complexity of the multi-source recovery model, an auxiliary induction objective function acceleration algorithm is designed to speed up the solution. Finally, the effectiveness of the proposed strategy in improving the distribution network resilience is verified using an improved IEEE 33-node distribution system case study.

Abstract:

Building a new power system dominated by new energy is the fulcrum of a new round of energy transition and revolution and an important pathway to achieving carbon peaking and carbon neutrality goals. By considering the uncertainty of wind and solar output and load fluctuations, a multi-stage robust planning method for the energy supply structure of distribution networks is proposed, incorporating the effect of carbon neutrality. First, the optimal investment capacity plan for equipment is formulated. Next, the operation plan of the distribution network is optimized under the consideration of uncertain factors, so as to minimize the daily operating cost of the distribution network. Finally, a case study of Hengqin in Zhuhai is conducted to analyze the power supply structure considering the effect of carbon neutrality, providing guidance for the development of a new power energy system for achieving carbon neutrality.

Abstract:

A high proportion of distributed energy sources connected to the grid makes the problem of insufficient system flexibility increasingly prominent, while the coupling of transmission and distribution networks further increases. First, the traditional flexibility demand model is improved based on the probability box method, and the probability box boundaries are truncated through a variable confidence interval and the conditional value-at-risk method to improve the conservatism of the probability box boundaries. Subsequently, economic models of the transmission network and the distribution network are established, respectively, and a transmission and distribution coordination model considering the risk of flexibility supply and demand imbalance is constructed. Finally, the effectiveness of the proposed method and the models is verified through a case study.

Abstract:

To reduce the system’s interruption frequency, a coordinated optimization method for load switches and circuit breakers in a distribution network is proposed, with the system’s average interruption frequency introduced as a constraint. By considering the uncertainty of distributed generation and load, a sample matrix is generated using Latin hypercube sampling, and the uncertainty problem is transformed into a deterministic problem by applying a multi-scenario analysis method. To improve the efficiency of island partitioning, a heuristic method is proposed to determine the scope of islands. Case analysis is carried out on an improved IEEE RBTS-BUS6 F4 distribution system. The results show that configuring a proper number of circuit breakers among the switches can satisfy specific interruption frequency requirements and improve the supply availability of the system. Furthermore, the optimal switch allocation scheme that considers source-load uncertainty demonstrates better robustness.

Abstract:

In response to wildfires that threaten the stable operation of distribution lines, it is important to establish a fire risk prediction model for distribution lines. However, the scarcity of data on wildfires causes sample imbalance, affecting the accuracy of the model. To this end, based on the influencing factors such as meteorological, geographic, combustible, and social factors, support vector machines and the idea of cost sensitivity are used to assign more weight to minority samples. Recursive feature elimination is used to select features that favor minority class classification. On this basis, a fire risk prediction model for distribution lines based on the random forest-adaptive boosting algorithm (RF?AdaBoost) is constructed. Finally, a 10 kV line corridor area in Xichang City, Sichuan Province, is selected to carry out an example verification. Ten-fold cross validation is used and compared with other algorithms. The results show that the recall rate of the method in this paper increases to 76.67%, which lessens the impact of sample imbalance on the model performance, reduces the misclassification of wildfires, and provides a basis for wildfire prevention and control in line corridors.

Abstract:

In view of the voltage overrun problem triggered by the unsynchronized peak output period of distributed generation (DG) and heavy load period in low-voltage distribution station areas, a phased voltage control strategy for DG in station areas based on cloud-edge collaboration is proposed. First, the historical operation data of the station area collected by the intelligent fusion terminal is received through the cloud master station, and a short-term prediction model of voltage based on spectral graph theory and graph convolution network is constructed at the cloud layer; then, the intelligent fusion terminal at the edge layer makes short-term prediction of voltage by using the voltage prediction model sent down from the cloud layer and the real-time data of the station area. The impact of the distribution of active energy on the voltage is analyzed by taking into account the relatively large sense of resistance of the low-voltage distribution network, and the grid-connected inverter of DG performs control in three stages: reactive power coordination, active power reduction, and power recovery, based on the voltage prediction values sent down from the edge layer, quickly bringing the voltage back to the normal range; when the voltage prediction value has an error with the real-time voltage, the edge layer uploads the new dataset that has an error to the cloud master station to optimize the model, realizing the collaboration of the cloud and the edge layers under the real-time voltage control strategy. Finally, the proposed control strategy is analyzed through an example of a rural low-voltage distribution station area to deal with the problem of voltage overrun in the station area and verify the effectiveness of the proposed strategy.

Abstract:

The relay protection operation requires strict technical operation and high accuracy. In response to the current challenges of high complexity and poor universality in modeling with augmented reality (AR) applications, a feature extraction edge computing framework is proposed. Firstly, the Canny algorithm is applied to perform edge analysis on the meter, pressure plate, and terminal block of the protection cabinet. To address salt-and-pepper noise in the images, a fixed-window median filtering method is employed for prepossessing. The improved Sobel operator optimizes gradients in the horizontal and vertical directions, and a set of contour boundary lines is constructed using the least squares method to handle points with significant gradient changes. Subsequently, the Harris algorithm is utilized for corner point analysis, and an improved corner point response function with fixed parameters efficiently determines corner points within the contour boundary line set. Finally, object recognition on the protection cabinet is carried out based on templates. Experimental results demonstrate that the proposed algorithm exhibits low computational complexity, fast response speed, and high accuracy, meeting the requirements for on-site operations.

Abstract:

A direct current (DC) system is key to maintaining the normal operation of a substation. To address the difficulties of detecting ground faults in DC system ring networks, a ground fault detection method combining Harris Hawk optimization (HHO) and wavelet neural network (WNN) is proposed based on the traditional low-frequency signal injection method. The proposed method first collects the status of the ring network operation by low-frequency signal injection and then applies the Mallat wavelet decomposition algorithm to decompose the initial signal of each branch into four layers. The low-frequency coefficient waveforms are obtained, and the waveform energy is calculated to construct the WNN input sample set. Finally, the HHO algorithm is used to optimize the WNN parameters, and the optimized and trained WNN is applied to the branch circuit detection of ground faults. The simulation results show that the proposed method can effectively detect ground faults of ring networks in the DC system.

Abstract:

In recent years, there have been numerous wildfires caused by various factors, seriously affecting the safe operation of overhead distribution lines. In addition to the contact between overhead lines and trees, human use of fire and lightning are also among the causes of wildfire accidents. The 10 kV overhead line is taken as the research object, and an experimental research platform for single-phase fire-contact and tree-contact faults is built. By means of a high-resistance ground fault model and a comparison in the time and frequency domains, the characteristics of the two faults are analyzed, and the change rules of leakage current flowing through the flame and flame resistance during the flame rise phase of single-phase fire-contact faults are investigated. The results show that single-phase tree-contact faults exhibit smoother amplitude variations during zero-sequence voltage rise compared to single-phase fire-contact faults. Single-phase fire-contact faults display a more pronounced low-frequency component in the zero-sequence voltage compared to single-phase tree-contact faults. During the flame rise phase of single-phase fire-contact faults, the leakage current flowing through the flame exponentially increases with time. The line-to-ground capacitive current has almost no effect on the amplitude of the leakage current flowing through the flame before an arc fault occurs.

Abstract:

The rapid development of new energy technologies will make them the main source of power generation in the future. However, the intermittent and fluctuating characteristics of new energy power generation make it highly susceptible to environmental influences. In view of insufficient power supply in urban distribution networks under extreme weather conditions, a power supply restoration strategy considering the regulation capability of flexible resources is proposed. Firstly, air conditioners and electric vehicles, which are typical adjustable flexible resources in urban distribution networks, are modeled, and their regulation capabilities are evaluated. Next, based on the evaluation results, the optimal power supply restoration path is solved to prioritize the power supply to critical loads in the distribution network. Finally, the IEEE 33-bus system proves that this method can quickly identify the optimal restoration path of power supply and effectively reduce the outage rate of critical loads, which is of great significance for improving the resilience of urban distribution networks.

Abstract:

Power supply via 10 kV lines is generally achieved in a radial branch structure, and low voltage problems are prone to occur at the end of the branch lines. The analysis of typical areas with safe power supply via 10 kV lines is the basis for the study of distribution network expansion and planning methods. The voltage drop characteristics of 10 kV lines are analyzed, and a simplified calculation method for the voltage drop along the line is derived. The power supply areas of the 10 kV distribution network are decomposed into four typical supply area models: rectangular, diamond-shaped, front-triangular, and rear-triangular. Each model uses the main lines as the central axis, with the branch lines bearing balanced loads. The boundary voltage drop of each typical area is analyzed, and the relationship among the total load, line type, and boundary voltage drop is established, thereby determining the maximum area of each type of typical power supply area under boundary voltage safety constraints. The analysis of typical power supply areas and their corresponding safe power supply capacities via 10 kV lines provides a theoretical basis for substation siting, grid layout, and line type selection in newly developed distribution areas, with both theoretical significance and engineering value.

Abstract:

Neutral and ground wiring errors on the user side are common in low-voltage transformer districts, which can cause user load currents to be converted into residual currents, resulting in frequent tripping of residual current circuit protection devices in the transformer district and their forced deactivation, thereby posing a threat to electrical safety. Since such wiring errors are present on the user side and are difficult to locate and troubleshoot, they have become a major obstacle to the deployment of residual current protection in transformer districts. This paper proposes a method for locating the user-side leakage due to wiring errors based on the significant correlation between user load currents and the residual current in the transformer districts, using correlation analysis and a gradient boosting decision tree (GBDT). The method starts with a qualitative and quantitative analysis conducted on the correlation between the residual current in the transformer district and the load current of users, with the Pearson correlation coefficient used to determine whether a causal relationship exists. A GBDT model is then constructed for the load current of each user in relation to the abnormal residual current in the transformer district, and importance scores are calculated to measure each user’s contribution to the fluctuations in residual current. This allows for the precise identification of user-side abnormalities. Experimental results demonstrate that the proposed method can accurately identify user-side abnormalities even in complex fault scenarios.

Abstract:

The load model has an important influence on the simulation analysis results of the power system, but the current load modeling methods are difficult to adapt to the diversity and time-varying characteristics of the new power system load in terms of model accuracy and computational efficiency. To address this issue, an online modeling method of the comprehensive load model for the new power system based on measurement information is proposed. Firstly, the dominant factors affecting the comprehensive load model are analyzed based on the activation subspace of parameter space. Then, according to the characteristics of the daily output curve of the load, a large number of underlying loads are clustered by using a clustering algorithm. Finally, the underlying loads are equivalently aggregated from the low voltage level to the high voltage level, and the equivalent aggregation model of the comprehensive load is constructed. The analysis results of the example show that the load model established by the proposed method can better track and record the real-time changes of the system load and more accurately reflect the actual load characteristics.

Abstract:

Short-term load forecast is an important task in power systems. However, existing studies have overlooked the spatio-temporal adjacency between multi-sequence loads. In some cases, considering this spatio-temporal adjacency can improve the prediction accuracy. To address this problem, a fully?connected graph-based graph convolutional neural network (FCGCN) is proposed. The network first encodes the multi-sequence load data into the node feature matrix of the graph, which is combined with the method of position encoding to increase the order information of the load data. The adjacency matrix of the graph is built by using the dynamic time warping (DTW) algorithm, thus forming a fully-connected spatio-temporal graph of the load data. Then, combined with the sliding window algorithm concept, the constructed fully-connected graph is divided into a series of subgraphs, and then feature extraction is accomplished for each subgraph individually using the graph convolutional neural network (GCN). Besides, in order to realize multi-perspective feature extraction on multi-source load data, FCGCN adopts a multi-branch parallel structure. The feature vectors extracted from each branch are concatenated, and different loads are predicted through a fully-connected layer. Finally, validation experiments using actual load data from a manufacturing base show that FCGCN can achieve higher prediction accuracy compared to common prediction models.

Abstract:

The siting of offshore booster stations for power collection systems has an important impact on the construction and operation costs of offshore wind farms. The siting problem of offshore booster stations is essentially an optimization problem based on the topology of the submarine cable. The key to siting optimization of the offshore booster station lies in determining the optimal submarine cable routing. A DMST?PSO algorithm, combining the dynamic minimum spanning tree (DMST) algorithm with dynamic edge weights and the particle swarm optimization (PSO) algorithm, is proposed for the siting of offshore booster stations for wind power collection systems, aiming at economic optimization. The DMST algorithm first obtains a minimum cost function with the offshore booster station’s coordinate as the independent variable. Then, the PSO algorithm is applied to identify the optimal siting by minimizing the value of this function. The example shows that the DMST?PSO algorithm can achieve economically optimal siting of offshore booster stations, with the advantages of high computational efficiency and fast convergence.

Abstract:

As the integration of renewable energy such as wind and solar power greatly increases, the challenges such as wind and solar power curtailment, voltage out-of-limit, and reverse power flow are becoming increasingly prominent in power systems. Pumped storage stands out as the most mature and cost-effective large-scale energy storage solution, which can be combined with renewable energy to help mitigate the fluctuation of renewable energy, improve the system operation flexibility, and enhance the reliability of power supply. The conventional wind-solar power storage and control methods have low efficiency in solving complex scenarios such as uncertain wind and solar power output, and the solving time is not adjustable. To address these issues, an adaptive two-stage robust scheduling method for wind-solar power storage based on an inexact column-and-constraint generation (i?C&CG) algorithm is proposed. Firstly, a model considering the operating condition conversion and power transition of pumped storage units is established. Secondly, an adaptive two-stage robust scheduling model for wind-solar power storage is developed based on polyhedral interval uncertain sets. Subsequently, the model is decomposed into master and sub-problems, and the i?C&CG algorithm is employed to derive the optimization decisions that balance the accuracy and efficiency. Finally, through case simulation, the comparison and analysis of the model solving time and error using the i?C&CG algorithm and the traditional column-and-constraint generation algorithm are conducted. It is demonstrated that the proposed two-stage scheduling method can effectively balance the robustness and economy of system operation and ensure the model solving efficiency within the allowable error range.

Abstract:

To address the current energy abandonment phenomenon of new energy sources such as wind power and photovoltaic power, an orderly charging and discharging strategy for electric vehicles (EVs) is proposed to promote wind-solar consumption. This strategy uses the vehicle-to-grid (V2G) interaction technology and aims to maximize the regional wind-solar consumption rate, minimize power load fluctuation, and maximize the power company’s electricity sales benefit in the context of mountainous cities by establishing a multi-objective charging model. The output of wind power and photovoltaic power is predicted using variational mode decomposition combined with a bidirectional long short-term memory (Bi-LSTM) network. Based on the predicted outputs, the output periods are divided, and dynamic electricity prices are set. The problem is solved using the adaptive particle swarm optimization algorithm, Yalmip + Cplex, and CVX toolbox. Case results show that when the user V2G responsiveness is 30%, 60%, and 100%, the wind-solar consumption rates are 83.73%, 89.12%, and 97.11%, respectively， power load fluctuations are decreased by 41.89%, 44.46%, and 47.32%, respectively, while ensuring the electricity sales benefits of the power company.

Abstract:

In power systems with large amounts of renewable energy and high variability, conventional scheduling methods cannot adequately accommodate the effects of the above variability. A scheduling optimization method is proposed to evaluate the optimal unit participation factor by considering the variability of solar, wind, and load demand. Both sequential and dynamic models are used, and variability and uncertainty costs are considered in the optimization process. Since the participation factor fitting function is optimized only once at the beginning of the scheduling interval, the dimensions of the proposed scheduling optimization model are the same as those of the conventional method. The simulation analysis results show that compared with that of the traditional sequential method, the cost of the proposed dynamic method is reduced by 3.6%, which verifies the effectiveness of the proposed method.

Abstract:

To address the proactivity of shared energy storage undermined by individual user-driven energy storage behavior and the diminished effectiveness and economic efficiency of day-ahead scheduling plans formulated by shared energy storage plants caused by load prediction errors, a bi-level scheduling optimization strategy is proposed for active shared energy storage communities. First, a community structure with prosumers utilizing active shared energy storage is designed. Second, a master-slave game decision-making model is constructed, where the shared energy storage power plant acts as the leader and the prosumer cluster as the follower. A genetic algorithm is used to solve the optimal day-ahead scheduling plan for the communities. Lastly, intra-day power generation and load are predicted by using Latin hypercube sampling, and an intra-day rolling scheduling model based on model predictive control (MPC) is developed. This allows the implementation of bi-level day-ahead and intra-day scheduling optimization for communities with prosumers, which enhances the proactivity of the shared energy storage plant. Experimental results show that the proposed model effectively maximizes the balanced interests of all parties in the game. Compared to that of existing user-driven shared energy storage, the economic return of the energy storage plant increases by 16.3%, which ensures the proactivity of shared energy storage.

Abstract:

The convergence speed for islanded microgrid systems using classical consensus control is slow, but the current improved consensus algorithm makes it difficult to meet the requirements of “plug and play” for distributed generation. Therefore, an improved consensus algorithm is proposed. By introducing the overall error elimination term, the information exchange between each distributed generation and the overall error elimination term acquisition are achieved without changing the original topology, and the stability of the improved consensus algorithm is verified through simulation. Subsequently, a control strategy based on the improved consensus algorithm is proposed, achieving frequency stability and precise power distribution for the islanded microgrid. Finally, an islanded microgrid simulation model is constructed to verify the effectiveness and applicability of the proposed control strategy, as well as its superiority in terms of convergence speed.

Abstract:

Due to the presence of a large number of constant power loads (CPLs) in a stand?alone DC microgrid, the damping of the DC microgrid is reduced, leading to oscillations and even significant drops in the DC bus voltage. To address this issue, an active damping method is proposed to improve system stability. The damping of the converter is increased to suppress the resonant peak by connecting a virtual resistance in parallel at the port of the energy storage converter without adding extra sensors. The method for designing the active damping parameters is provided, and the effect of the proposed parallel virtual resistance method on system stability is analyzed using the frequency stability criterion. The performance of this method is also compared with that of the conventional damping method. Experimental results verify that the parallel virtual resistance method can effectively improve the stability of a stand?alone DC microgrid system in a wide frequency band.

Abstract:

Establishing an integrated energy system (IES) is an effective means to promote the efficient utilization of energy resources. Reasonable planning and design software for IESs can provide tailored design solutions according to engineering needs. However, the lack of multi-scenario-compatible planning software remains a bottleneck for the widespread industrial application of IESs. To address this issue, a comprehensive and highly reliable IES plan (IES-PLAN) design platform has been developed to scientifically guide users in designing reliable, low-carbon, and efficient IESs. The platform offers a graphical interaction interface, allowing users to flexibly set planning scenarios and goals based on their needs. To validate the platform’s effectiveness and practicality, a simulation case study is conducted based on the “Nanjing Hospital IES.” The results demonstrate the platform’s responsiveness and planning performance under different customized requirements, achieving reduced energy costs and improved energy utilization efficiency.

Abstract:

At present, the scale of the power grid is expanding, and the amount of knowledge in power systems is increasing explosively. In order to organize, manage, and utilize mass knowledge effectively, knowledge graph technology is introduced into the field of power systems and comprehensive energy systems. Common relational databases of Oracle and structured query language (SQL) need to use tables to store data and query and analyze data through complex relationships, which is more complicated when dealing with complex relationships. The Neo4j graph database represents data as nodes and edges, which makes the correlation between entities and relationships intuitively expressed and stored, and it is especially suitable for application scenarios that need to deal with complex relationships and conduct graph analysis. Therefore, a research method for power system and comprehensive energy system knowledge graph based on the Neo4j graph database is proposed. By introducing knowledge graph technology into the power system and comprehensive energy, the power knowledge is stored in an orderly manner by using the Neo4j graph database. Then, a knowledge graph of the power field is built, and a search engine with a B/S framework is designed, realizing the interactive function between users and knowledge graph through front-end coding and handling the data related to knowledge graph through back-end coding, including data storage, query, update, and other operations. The test results show that this method can effectively improve the search efficiency of knowledge graphs of power systems and comprehensive energy systems and enhance the retrieval speed of massive data.

Abstract:

Under weak grid conditions, the interaction between the control function of the grid-connected inverter and the grid impedance will reduce the phase margin of the inverter, aggravate the harmonic oscillations, and even lead to instability of the grid-connected systems. Traditional impedance reshaping strategies for grid-connected inverters can improve the stable operation of the grid-connected system. However, with the growing integration of large shares of new energy sources and new types of loads, the grid impedance presents random variations, which undermines the effectiveness of traditional impedance reshaping strategies designed for fixed scenarios. To solve this problem, an impedance reshaping strategy based on dual harmonic injection for grid-connected inverters is proposed. First, the equivalent impedance model of an LCL-type inverter is derived using mathematical methods, and the limitation of the traditional voltage feedforward strategy is analyzed under wide variations in grid impedance. Then, by injecting dual-harmonic disturbance signals into the grid-connected inverter, real-time and accurate grid impedance information is obtained. This information is used to dynamically adjust the traditional impedance reshaping strategy, ensuring sufficient phase margin under wide variations in grid impedance and improving the adaptability of the inverter to different grid working conditions. Finally, a simulation model is built in MATLAB/Simulink to verify the correctness and effectiveness of the proposed impedance reshaping strategy under weak grid conditions.

Abstract:

To solve serious problems such as equipment shutdown caused by voltage dips due to short-circuit faults in distribution networks, a coordinated control strategy employing a distribution network static synchronous compensator (D-STATCOM) based on the voltage transient rate of change is proposed. Firstly, the control of voltage by the traditional control strategy is improved. An undervoltage module is activated when a short-circuit fault occurs in the distribution network, increasing the reactive power output of D-STATCOM to suppress the voltage dip during the fault. At the end of the fault, an overvoltage module is activated to decrease the reactive power output of D-STATCOM, thereby suppressing the post-fault voltage overshoot. On this basis, a transient fast reactive power compensation control is added to improve the dynamic reactive power regulation speed of the D-STATCOM, enabling it to supply or absorb reactive power more rapidly for voltage stabilization. The proposed coordinated control strategy is compared with the conventional control strategy in PSCAD/EMTDC, and the simulation results verify the effectiveness and correctness of the strategy.

Abstract:

To address the existing challenges in the orderly scheduling of electric vehicle (EV) charging and discharging, including the insufficient characterization of dispatchable capacity, inadequate consideration of multi-agent interests, and improper pricing mechanisms for charging stations, a two-level pricing strategy for EV charging stations based on a multi-agent Stackelberg game model is proposed. Firstly, in consideration of factors influencing charging behavior, the responsiveness of EV users is assessed, and a model for predicting the dispatchable capacity of charging stations is developed. Secondly, the game relationships among various agents are discussed. According to the EV user response willingness and the theory of locational marginal pricing, a two-level pricing strategy for charging stations based on the Stackelberg game is proposed: The outer layer model is optimized to minimize the costs of distribution network operators, and the locational marginal prices at each distribution network node are set. The inner layer model is optimized to maximize the revenue of charging stations and minimize the charging costs for EV users. The pricing strategy for charging stations is established, and the charging and discharging plans are arranged by EV users according to the electricity prices. Finally, the model is reconstructed by using duality theory and Karush-Kuhn-Tucker (KKT) conditions, and an iterative algorithm is designed to solve it. Simulation results demonstrate that the proposed pricing strategy can increase profits for charging stations while reducing costs for EV users and the distribution network, thus balancing the interests of all agents involved.

Abstract:

To achieve system load peak shaving and valley filling, as well as real-time balancing of power differences between load peaks and valleys, a multi-energy collaborative optimization method for energy supply is proposed, integrating new and conventional energy sources. Based on the different generation characteristics of wind power, photovoltaic power, hydropower, and thermal power, an optimization model is constructed for a new energy alliance. The energy block method for peak and valley loads is first applied to determine the electricity demand to be undertaken by the alliance. Through internal resource optimization allocation and multi-agent generation combination-based multi-scale cost balancing optimization, an operational decision-making method for multi-energy alliance optimization based on multi-scale dynamic cost Shapley values is proposed. This method quantifies the dynamic allocation laws of cost allocation and output proportion of each entity, and a comparison is made between the kernel method and the Shapley value method. Furthermore, a comprehensive evaluation model based on the grey relational degree-entropy weight TOPSIS method is used to select optimal power generation combinations, aiming to enhance system optimization. Case analysis demonstrates that the proposed multi-energy alliance operation mechanism effectively smooths fluctuations in new energy output, increases the internal new energy consumption ratio, and enhances economic efficiency and system stability. The research findings provide theoretical support for generation planning, cost-sharing strategies, and combination optimization within the alliance and offer practical guidance for the efficient and coordinated operation of multi-energy systems.

Abstract:

With cable equipment being widely deployed, cable faults have threatened the safe operation of the power grid. Traditional operation and maintenance work has difficulty in accurately predicting the current health status of cable insulation. To address this issue, a cable health index prediction method based on an improved genetic algorithm-back propagation (IGA-BP) neural network model is proposed. Since the rate of parameter change in underground cables varies at different aging stages, the method incorporates recent aging trend characteristics into both the fitness function and mutation operator during parameter optimization. By distinguishing individuals based on these aging characteristics, the model enhances both the efficiency of searching for a global optimum and the accuracy of predictions. Experimental results demonstrate that compared to traditional back propagation (BP) and genetic algorithm?back propagation (GA-BP) neural networks, the IGA-BP neural network improves prediction accuracy by 3.68%, achieving 99.39% accuracy in five-fold cross-validation and 95.8% accuracy in a dataset of 15 kV high-voltage cross-linked polyethylene (XLPE) underground cables. The developed model is well-suited for health index prediction as it fully accounts for the historical aging information of cables.

Abstract:

When an existing open station is converted into a terminal station arranged within the user’s premises, due to the generally compact area of the station, the reliability requirements for the grounding grid are higher, and the diagnosis and evaluation of the corrosion status of the grounding grid are becoming increasingly important. At present, the main method for diagnosing corrosion in grounding grids without excavation relies on electrical network theory analysis. However, this method is prone to having more branches than the number of observable nodes in diagnosis, leading to a high degree of uncertainty. Therefore, an improved seagull optimization algorithm is proposed, which integrates Gauss mapping and levy flight strategy on top of the traditional seagull algorithm, improving computational stability and convergence speed. A simulated grounding grid model is built, and the accuracy and reliability of the improved seagull optimization algorithm are verified through simulation calculations under various grounding grid corrosion conditions, combined with comparative analysis of other commonly used optimization algorithms. The simulation results show that the diagnostic deviation of the corrosion branch in the grounding grid using the improved algorithm is less than 5%, which is significantly lower than the other two traditional optimization algorithms, proving its ability to improve the accuracy of grounding grid corrosion diagnosis.