Power supply recovery method, device and equipment based on network construction energy storage and storage medium

Abstract

The invention relates to the field of automatic control of a power system, and discloses a power supply recovery method, device and equipment based on network construction energy storage and a storage medium, which are used for improving the efficiency and quality of power supply recovery. The power supply recovery method based on network construction energy storage comprises the following steps: acquiring operation state information of a power system in real time, positioning a fault and an affected area by using a fault detection algorithm when the fault occurs, and formulating a scheduling plan according to the fault and the affected area; globally optimizing the scheduling plan through a mixed integer linear programming algorithm to obtain a scheduling result containing power supply access point information and load point information of power supply to be recovered; and finding out a plurality of power supply paths by using a path search algorithm according to a scheduling result in combination with network topology. Screening through a multi-objective optimization algorithm to obtain an optimal recovery path; and finally, in combination with distributed energy coordination control and a micro-grid technology, executing power supply recovery according to an optimal path.

Description

Technical Field

[0001] This invention relates to the field of power system automation control, and in particular to a power supply restoration method, apparatus, equipment and storage medium based on grid-connected energy storage. Background Technology

[0002] The stable operation of the power system is crucial for ensuring social production and people's lives. However, natural disasters, equipment failures and other factors may cause power system failures, resulting in large-scale power outages. In order to quickly restore power supply and reduce the losses caused by power outages, power system restoration technology is particularly important.

[0003] Traditional power restoration methods rely heavily on direct operation by manual dispatchers and their past experience. This approach is not only time-consuming, requiring manual troubleshooting and restoration plan development, but also inefficient, as manual dispatchers often struggle to find the optimal restoration path quickly. Furthermore, reliance on experience can lead to imprecise restoration plans that fail to fully consider the complexity and real-time nature of the power system, thus compromising power quality and potentially causing voltage fluctuations, frequency instability, and other issues that disrupt users' normal electricity usage. Summary of the Invention

[0004] This invention provides a power restoration method, apparatus, equipment, and storage medium based on grid-connected energy storage to solve the problem of low power restoration efficiency and quality in the prior art.

[0005] The first aspect of this invention provides a power supply restoration method based on grid-connected energy storage, comprising: acquiring real-time operating status information of a power system; when a fault is determined to have occurred in the power system based on the real-time operating status information, locating the fault location and affected area using a fault detection algorithm to obtain fault information; formulating a corresponding scheduling plan based on the fault information; performing global optimization of the scheduling plan using a mixed-integer linear programming algorithm to obtain a scheduling result, the scheduling result including power supply access point information and load point information to be restored; based on the power supply access point and load point to be restored determined in the scheduling result, and combined with the network topology of the power system, using a path search algorithm to search for a power supply path that meets preset conditions to obtain a search result, the search result including multiple possible power supply paths and their related parameters; optimizing and selecting the search result using a multi-objective optimization algorithm to obtain an optimal restoration path; and performing power restoration execution processing based on the optimal restoration path, combined with distributed energy coordination and control technology and microgrid technology.

[0006] In one feasible implementation, the step of locating the fault location and affected area using a fault detection algorithm includes: modeling the power system using a convolutional neural network to obtain a detection model, which can learn and identify electrical signal characteristics under normal and abnormal conditions; analyzing the electrical signal characteristics using the detection model to determine suspected fault areas; transmitting high-frequency signals to the suspected fault areas using high-frequency electromagnetic wave transmission and reception technology, and locating the fault point by analyzing the time-frequency characteristics of the reflected signals; and using graph theory and complex network analysis methods to calculate the affected area based on the location of the fault point and the network topology of the power system, and displaying the fault information and affected area in real time using visualization technology.

[0007] In one feasible implementation, the step of using a mixed-integer linear programming algorithm to globally optimize the scheduling plan and obtain the scheduling result includes: constructing a MILP model containing continuous and integer variables, determining the objective function and constraints; selecting an objective MILP solver, inputting the scheduling plan into the MILP model for solving; analyzing the optimal solution output by the solver, and adjusting the MILP model parameters until a scheduling result that meets the requirements is obtained;

[0008] The objective function is:

[0009]

[0010] Where T is the power grid recovery time, α is the first weighting coefficient, β is the second weighting coefficient, and γ is the third weighting coefficient. P is the fourth weighting coefficient. i 1 and P i 2 represents the output power and input power of energy storage system i, respectively. P is the power loss of load j. e The power generation capacity of new energy sources;

[0011] The constraints are as follows: the total output of all generators and the charging and discharging power of energy storage systems in the power system must be equal to the total load demand and network losses; the voltage amplitude of each node must be maintained within the allowable range; and the active power transmitted by the line must not exceed its thermal stability limit.

[0012] In one feasible implementation, the step of using a mixed-integer linear programming algorithm to globally optimize the scheduling plan and obtain the scheduling result further includes: calculating a first weight coefficient based on a preset first quantization result using the analytic hierarchy process (AHP), wherein the calculation formula for the first weight coefficient is:

[0013]

[0014] in, The weight of the k-th primary evaluation indicator. Let be the score of the kth second evaluation indicator, and n be the total number of first evaluation indicators;

[0015] Based on the preset second quantization result, the second weighting coefficient is calculated using the entropy weighting method. The formula for calculating the second weighting coefficient is as follows:

[0016]

[0017] in, Let g be the entropy value of the g-th second evaluation index, and m be the total number of second evaluation indexes.

[0018] Based on the preset third quantification results and load classification, the third weight coefficient is calculated using grey relational analysis. The calculation formula is:

[0019]

[0020] in, Let y be the grey relational degree of the y-th load. The priority score is assigned to the y-th load, where q is the total number of loads;

[0021] Based on the preset fourth quantification result, the fourth weight coefficient is calculated using the fuzzy comprehensive evaluation method. The formula for calculating the fourth weight coefficient is as follows:

[0022]

[0023] in, Let the membership degree of the e-th fourth evaluation indicator be denoted as . Let be the weight of the e-th fourth evaluation indicator, and p be the total number of fourth evaluation indicators.

[0024] In one feasible implementation, the step of searching for a power supply path that meets preset conditions based on the power access point and the load point to be restored to power supply determined in the scheduling results, combined with the network topology results of the power system, includes: identifying the power access point and the load point to be restored to power supply based on the scheduling results; integrating the power system network topology data; constructing a graph structure model with power equipment as nodes and connecting lines as edges; and assigning weight values ​​to the edges based on comprehensive line parameters and operating status; selecting a path search algorithm adapted to the characteristics of the power system; setting the power access point as the starting point of the search and the load point as the target point; and conducting a path traversal search in the constructed graph structure model in combination with preset conditions, including line capacity, path length, and equipment reliability; and performing real-time condition judgment and filtering on the searched paths, retaining the paths that meet the preset conditions as candidate power supply paths.

[0025] In one feasible implementation, the step of optimizing the search results using a multi-objective optimization algorithm to obtain the optimal recovery path includes: defining a set of objective functions for the multi-objective optimization problem, including a path length objective function, a recovery time objective function, and an energy storage resource consumption objective function; based on the set of objective functions, iteratively searching the search results using an adaptive non-dominated sorting genetic algorithm; during the iteration process, if convergence stagnation is detected, automatically increasing the mutation probability to escape local optima until the maximum number of iterations is reached or the hypervolume of the Pareto front solution set no longer significantly improves, and outputting the optimized Pareto front; and selecting the optimal recovery path from the Pareto front based on fuzzy multi-attribute decision-making and the real-time operating status information.

[0026] In one feasible implementation, the power restoration execution process based on the optimal restoration path, combined with distributed energy coordination control technology and microgrid technology, includes: determining a power system restoration strategy based on the optimal restoration path, the restoration strategy including restoration sequence, restoration time points, and the access points and output power of each distributed energy source; using distributed energy coordination control technology, accessing distributed energy sources according to the restoration strategy and adjusting their corresponding output power; and combining microgrid technology to implement autonomous control and energy optimization scheduling within the local power grid to ensure that the distributed energy sources accessed according to the optimal restoration path can work together efficiently; the output power adjustment of the distributed energy sources must satisfy a second formula, which is:

[0027]

[0028] in, The output power of distributed energy source z after adjustment. The basic output power of distributed energy source z, This refers to the power increment adjusted according to the recovery strategy and grid conditions.

[0029] A second aspect of the present invention provides a power supply restoration device based on grid-connected energy storage, comprising: an acquisition module for acquiring real-time operating status information of a power system; a location module for locating the fault location and affected area using a fault detection algorithm when a fault is determined to have occurred in the power system based on the real-time operating status information, thereby obtaining fault information; a designation module for formulating a corresponding scheduling plan based on the fault information; an optimization module for globally optimizing the scheduling plan using a mixed-integer linear programming algorithm to obtain a scheduling result, wherein the scheduling result includes power supply access point information and load point information to be restored; a search module for searching for a power supply path that meets preset conditions based on the power supply access point and load point to be restored determined in the scheduling result, combined with the network topology of the power system, using a path search algorithm to obtain a search result, wherein the search result includes multiple possible power supply paths and their related parameters; a selection module for optimizing and selecting the search result using a multi-objective optimization algorithm to obtain an optimal restoration path; and a processing module for performing power restoration execution processing based on the optimal restoration path, combined with distributed energy coordination and control technology and microgrid technology.

[0030] In one feasible implementation, the positioning module is specifically used for: modeling the power system using a convolutional neural network to obtain a detection model, which can learn and identify electrical signal characteristics under normal and abnormal conditions; analyzing the electrical signal characteristics using the detection model to determine suspected fault areas; transmitting high-frequency signals to the suspected fault areas using high-frequency electromagnetic wave transmission and reception technology, and locating the fault point by analyzing the time-frequency characteristics of the reflected signals; and using graph theory and complex network analysis methods to calculate the affected area based on the location of the fault point and the network topology of the power system, and displaying the fault information and affected range in real time using visualization technology.

[0031] In one feasible implementation, the optimization module includes: a construction unit for constructing a MILP model containing continuous and integer variables, and determining the objective function and constraints; a solution unit for selecting an objective MILP solver, inputting the scheduling plan into the MILP model for solution; and a processing unit for analyzing the optimal solution output by the solver and adjusting the MILP model parameters until a scheduling result that meets the requirements is obtained.

[0032] The objective function is:

[0033]

[0034] Where T is the power grid recovery time, α is the first weighting coefficient, β is the second weighting coefficient, and γ is the third weighting coefficient. P is the fourth weighting coefficient. i 1 and Pi 2 represents the output power and input power of energy storage system i, respectively. P is the power loss of load j. e The power generation capacity of new energy sources;

[0035] The constraints are as follows: the total output of all generators and the charging and discharging power of energy storage systems in the power system must be equal to the total load demand and network losses; the voltage amplitude of each node must be maintained within the allowable range; and the active power transmitted by the line must not exceed its thermal stability limit.

[0036] In one feasible implementation, the optimization module further includes: a calculation unit, used to calculate a first weight coefficient based on a preset first quantization result using the analytic hierarchy process (AHP), wherein the calculation formula for the first weight coefficient is:

[0037]

[0038] in, The weight of the k-th primary evaluation indicator. Let be the score of the kth first evaluation indicator, and n be the total number of first evaluation indicators;

[0039] Based on the preset second quantization result, the second weighting coefficient is calculated using the entropy weighting method. The formula for calculating the second weighting coefficient is as follows:

[0040]

[0041] in, Let g be the entropy value of the g-th second evaluation index, and m be the total number of second evaluation indexes.

[0042] Based on the preset third quantification results and load classification, the third weight coefficient is calculated using grey relational analysis. The calculation formula is:

[0043]

[0044] in, Let y be the grey relational degree of the y-th load. The priority score is assigned to the y-th load, where q is the total number of loads;

[0045] Based on the preset fourth quantification result, the fourth weight coefficient is calculated using the fuzzy comprehensive evaluation method. The formula for calculating the fourth weight coefficient is as follows:

[0046]

[0047] in, Let the membership degree of the e-th fourth evaluation indicator be denoted as . Let be the weight of the e-th fourth evaluation indicator, and p be the total number of fourth evaluation indicators.

[0048] In one feasible implementation, the search module is specifically used for: identifying the power access point and the load point to be restored based on the scheduling results; integrating power system network topology data; constructing a graph structure model with power equipment as nodes and connecting lines as edges; and assigning weight values ​​to the edges based on comprehensive line parameters and operating status; selecting a path search algorithm adapted to the characteristics of the power system; setting the power access point as the starting point of the search and the load point as the target point; and conducting path traversal search in the constructed graph structure model in combination with preset conditions, including line capacity, path length, and equipment reliability; and performing real-time condition judgment and filtering on the searched paths, retaining paths that meet the preset conditions as candidate power supply paths.

[0049] In one feasible implementation, the selection module is specifically used to: define a set of objective functions for a multi-objective optimization problem, including a path length objective function, a recovery time objective function, and an energy storage resource consumption objective function; based on the set of objective functions, iteratively search the search results using an adaptive non-dominated sorting genetic algorithm; during the iteration process, if convergence stagnation is detected, automatically increase the mutation probability to escape local optima until the maximum number of iterations is reached or the hypervolume of the Pareto front solution set no longer improves significantly, and output the optimized Pareto front; and select the optimal recovery path from the Pareto front based on fuzzy multi-attribute decision-making and the real-time running status information.

[0050] In one feasible implementation, the processing module is specifically used for: determining a power system recovery strategy based on the optimal recovery path, the recovery strategy including recovery sequence, recovery time point, and access point and output power of each distributed energy source; utilizing distributed energy source coordination and control technology, accessing distributed energy sources according to the recovery strategy and adjusting the corresponding output power; combining microgrid technology, implementing autonomous control and energy optimization scheduling within the local power grid to ensure that distributed energy sources accessed according to the optimal recovery path can work efficiently and collaboratively; the output power adjustment of the distributed energy sources must satisfy a second formula, which is:

[0051]

[0052] in, The output power of distributed energy source z after adjustment. The basic output power of distributed energy source z, This refers to the power increment adjusted according to the recovery strategy and grid conditions.

[0053] A third aspect of the present invention provides a power restoration device based on grid-connected energy storage, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the power restoration device based on grid-connected energy storage to execute the aforementioned power restoration method based on grid-connected energy storage.

[0054] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described power restoration method based on grid-connected energy storage.

[0055] The technical solution provided by this invention involves: acquiring real-time operating status information of the power system; when a fault is determined to have occurred in the power system based on the real-time operating status information, locating the fault location and affected area using a fault detection algorithm to obtain fault information; formulating a corresponding scheduling plan based on the fault information; globally optimizing the scheduling plan using a mixed-integer linear programming algorithm to obtain a scheduling result, which includes power access point information and load point information to be restored; based on the power access point and load point to be restored determined in the scheduling result, combined with the network topology of the power system, using a path search algorithm to search for a power supply path that meets preset conditions to obtain a search result, which includes multiple possible power supply paths and their related parameters; optimizing and selecting the search result using a multi-objective optimization algorithm to obtain the optimal restoration path; and performing power restoration execution processing based on the optimal restoration path, combined with distributed energy coordination and control technology and microgrid technology. In this embodiment of the invention, by acquiring real-time operating status information of the power system, fault detection algorithms are used to quickly and accurately locate the fault location and affected area. Combined with mixed integer linear programming algorithms, the scheduling plan is globally optimized. Then, path search algorithms and multi-objective optimization algorithms are used to obtain the optimal recovery path. Finally, distributed energy coordination and control technology and microgrid technology are used to perform power restoration. This overcomes the problems of traditional power restoration methods, such as reliance on manual scheduling, long time consumption, low efficiency, and inaccurate plans. It can fully consider the complexity and real-time nature of the power system, effectively ensure power supply quality, and reduce power outage losses. Attached Figure Description

[0056] Figure 1 This is a schematic diagram of one embodiment of the power supply restoration method based on grid-connected energy storage in this invention.

[0057] Figure 2 This is a schematic diagram of another embodiment of the power supply restoration method based on grid-connected energy storage in this invention.

[0058] Figure 3This is a schematic diagram of one embodiment of the power supply restoration device based on grid-connected energy storage in this invention.

[0059] Figure 4 This is a schematic diagram of another embodiment of the power supply restoration device based on grid-connected energy storage in this invention.

[0060] Figure 5 This is a schematic diagram of one embodiment of the power restoration device based on grid-connected energy storage in this invention. Detailed Implementation

[0061] This invention provides a power restoration method, device, equipment, and storage medium based on grid-connected energy storage, which improves the efficiency and quality of power restoration through the collaboration of intelligent algorithms and energy technologies.

[0062] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0063] It is understood that the executing entity of this invention can be a power restoration device based on grid-connected energy storage, or it can be a terminal or a server; the specific implementation is not limited here. This embodiment of the invention will be described using a server as an example.

[0064] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the power restoration method based on grid-connected energy storage in this invention includes:

[0065] 101. Obtain real-time operating status information of the power system;

[0066] Real-time operating status information of the power system can be obtained by combining synchronous phasor measurement units (TPMUs) with a wide-area monitoring system (WAMS). TPMUs are deployed at key nodes of the power system, such as power plants, substations, and important loads. These TPMUs use clock signals provided by the Global Positioning System (GPS) for synchronous sampling, measuring the phasor values ​​of voltage and current in real time at a high sampling rate, including amplitude and phase angle. This high sampling rate is typically tens to hundreds of times per second. The TPMUs transmit the collected data to the WAMS master station via a high-speed communication network. The WAMS master station receives and stores the data from each TPMU, and uses data fusion and processing technologies to perform real-time analysis and integration of the phasor data from multiple nodes. This yields real-time operating status information of the entire power system, including frequency fluctuations, power flow direction, and line power flow distribution, enabling comprehensive monitoring of the dynamic processes of the power system.

[0067] 102. When a fault is determined to have occurred in the power system based on real-time operating status information, a fault detection algorithm is used to locate the fault location and the affected area to obtain fault information;

[0068] Wavelet transform is applied to analyze current and voltage waveforms to extract fault features; combined with expert systems and machine learning models, the fault features are analyzed to determine the fault type and location; based on the fault type and location, the impact area of ​​the fault on the power system is assessed.

[0069] Wavelet transform is used to perform multi-scale analysis on the collected waveform data. Through wavelet transform, fault features such as abrupt changes and harmonics in the waveform can be extracted. The extracted fault features are input into a pre-built expert system. The expert system stores a large number of power system fault cases and expert experience knowledge, which can perform preliminary analysis and judgment of fault features. At the same time, the preliminary judgment results are input into a machine learning model to determine the fault type and fault location. The machine learning model is trained using historical fault data, enabling it to autonomously learn the mapping relationship between fault features, fault type, and location. Based on the power system topology and operating parameters, combined with fault type and location information, power flow calculation and short-circuit analysis are used to assess the impact area of ​​the fault on the power system, so as to clarify the range of equipment and loads affected by the fault, thereby obtaining comprehensive fault information. The topology result of the power system refers to the graphical representation formed by the connection methods and interrelationships between various electrical devices in the power system.

[0070] 103. Develop corresponding scheduling plans based on fault information;

[0071] Based on the fault-affected areas identified in the fault information, and using load importance indicators such as load level and outage loss to screen critical loads, the power system is divided into multiple dispatch areas. Specifically, relevant data for all loads within the fault-affected areas is collected, including load level and outage loss importance indicators. Each indicator is quantified and scored according to pre-defined load importance assessment standards; for example, different load levels are assigned corresponding scores, and different intervals are divided based on the magnitude of outage loss, each with a corresponding score. The quantified scores of each indicator are then comprehensively calculated to obtain a total score for each load. Loads with the highest total scores are selected as critical loads, ranked from highest to lowest. Based on the distribution of critical loads and the electrical connection characteristics of the fault-affected areas, and considering factors such as the power system topology and power transmission paths, the power system is divided into multiple dispatch areas, ensuring that each dispatch area contains a certain number of critical loads and that the electrical connections between areas are relatively independent.

[0072] For each dispatch area, power distribution data such as rated power, current output, and ramp rate of each generator within the area are collected, along with load demand data for different time periods and industries. Electrical connection information such as tie line parameters and historical power exchange data between this area and other areas is analyzed. An intelligent optimization model based on a genetic algorithm is constructed, transforming operational constraints such as power balance, voltage limits, and line capacity limits into constraint functions in the algorithm, with the objective functions being minimizing generation costs and maximizing load power supply reliability. During the algorithm iteration process, power output and load allocation schemes are continuously optimized through operations such as selection, crossover, and mutation to generate a preliminary dispatch plan.

[0073] 104. Use a mixed-integer linear programming algorithm to perform global optimization of the scheduling plan and obtain the scheduling result, which includes power access point information and load point information to be restored.

[0074] In mixed-integer linear programming algorithms, decision variables are defined, including the output of each power source, load allocation, and switch states, with some variables being integers to meet the discreteness requirements of actual operation. An objective function is established, comprehensively considering multiple objectives such as generation cost, network loss, and load satisfaction. The multi-objective problem is transformed into a single-objective optimization problem through weighted summation. Simultaneously, a series of constraints are set, such as power balance constraints, upper and lower limits of power source output constraints, line flow constraints, and voltage constraints, to ensure that the dispatching scheme conforms to the physical characteristics and operational requirements of the power system. The constructed mathematical model is then input into a professional optimization solver, such as CPLEX or Gurobi, and the algorithms of these solvers are used to optimize the model. The algorithm automatically searches for solutions that satisfy all constraints and optimize the objective function. Based on this process, a power supply and load matching table is generated, clearly indicating which power sources supply which loads. According to the non-zero output variables of the power sources in the optimization results and network topology constraints, the power supply nodes actually participating in power supply are selected, excluding power sources that cannot be connected due to faults or capacity limitations. Then, based on the solution values ​​of the load allocation variables and preset weights, a list of load points requiring power restoration is extracted from the matching relationships. The power required by these load points can be further quantified to determine priority restoration for important loads, thus obtaining a scheduling result containing power supply access point information and load point information to be restored.

[0075] 105. Based on the power access point and load point to be restored in the scheduling results, and combined with the network topology of the power system, a path search algorithm is used to search for power supply paths that meet the preset conditions, and the search results are obtained. The search results include multiple possible power supply paths and their related parameters.

[0076] The network topology of the power system is abstracted as a graph model, where nodes represent power equipment and edges represent transmission lines. Power equipment includes substations, switches, etc. Based on the power access points and load points to be restored according to the scheduling results, the starting node and target node are determined in the graph model. Each edge in the graph is assigned a weight, which comprehensively considers factors such as line length, impedance, and reliability to reflect the quality of the power supply path. The A algorithm can be used for path search. During the search process, the cost of each node is evaluated through a heuristic function, and nodes with lower costs are prioritized for expansion. A series of preset conditions are set to filter the searched paths. These preset conditions can include the maximum allowable impedance and maximum transmission capacity of the path. When a target node is found, the path that meets the preset conditions and its related parameters are recorded, including path length, total impedance, and maximum transmittable power. The search continues to find other possible paths until all possible path combinations are traversed to obtain the search results. The search includes multiple possible power supply paths and their detailed parameters.

[0077] 106. Use a multi-objective optimization algorithm to optimize and select the search results to obtain the optimal recovery path;

[0078] The optimization objectives are defined, including path reliability, economy, and transmission efficiency. Reliability can be measured by indicators such as path failure probability and repair time. Economy considers factors such as path construction and operation / maintenance costs. Transmission efficiency is related to path impedance and transmission capacity. Each power supply path in the search results is treated as an individual to construct an initial population. During the iterative process of the genetic algorithm, new individuals are generated through selection, crossover, and mutation operations, gradually optimizing the quality of the population. Non-dominated sorting techniques are used to rank the individuals in the population, dividing them into different non-dominated levels, prioritizing individuals with higher non-dominated levels. Simultaneously, crowding distance calculation is introduced to maintain population diversity and prevent the algorithm from getting trapped in local optima. After multiple iterations, a set of non-dominated solutions, i.e., the Pareto optimal solution set, is obtained. Based on actual needs and decision-makers' preferences, the most suitable solution is selected from the Pareto optimal solution set as the optimal recovery path.

[0079] 107. Based on the optimal recovery path, and in conjunction with distributed energy coordination and control technology and microgrid technology, perform power restoration execution.

[0080] The optimal recovery path is used to determine the access location and power supply range of each distributed power source. Through communication networks and intelligent controllers, real-time operational information of each distributed power source, including output power, voltage, and frequency, is collected. Based on load demand and grid operating status, optimization algorithms are used to dynamically adjust the output of each distributed power source, ensuring power balance and stable operation among them. Simultaneously, for microgrids containing distributed power sources, microgrid technology is employed to achieve coordination with the main grid. During power restoration, the microgrid can autonomously select its operating mode, such as grid-connected or islanded operation, based on its energy reserves and load demand. When the main grid's power supply capacity is insufficient, the microgrid can independently supply power to critical loads. When the main grid returns to normal, the microgrid can smoothly reconnect to the main grid. Furthermore, during power restoration, the grid's operating status is monitored in real-time, and any abnormalities are addressed promptly to ensure the safety, reliability, and efficiency of the power restoration process.

[0081] In this embodiment of the invention, by acquiring real-time operating status information of the power system, fault detection algorithms are used to quickly and accurately locate the fault location and affected area. Combined with mixed integer linear programming algorithms, the scheduling plan is globally optimized. Then, path search algorithms and multi-objective optimization algorithms are used to obtain the optimal recovery path. Finally, distributed energy coordination and control technology and microgrid technology are used to perform power restoration. This overcomes the problems of traditional power restoration methods, such as reliance on manual scheduling, long time consumption, low efficiency, and inaccurate plans. It can fully consider the complexity and real-time nature of the power system, effectively ensure power supply quality, and reduce power outage losses.

[0082] Please see Figure 2 Another embodiment of the power restoration method based on grid-connected energy storage in this invention includes:

[0083] 201. Obtain real-time operating status information of the power system;

[0084] The execution process of step 201 is similar to that of step 101 above, and will not be described again here.

[0085] 202. When a fault is determined to have occurred in the power system based on real-time operating status information, a fault detection algorithm is used to locate the fault location and the affected area to obtain fault information;

[0086] A detection model is obtained by modeling the power system using convolutional neural networks. This model can learn and identify electrical signal characteristics under normal and abnormal conditions. The detection model is used to analyze the electrical signal characteristics to determine suspected fault areas. High-frequency electromagnetic wave transmission and reception technology is combined to transmit high-frequency signals to the suspected fault areas, and the fault point is located by analyzing the time-frequency characteristics of the reflected signals. Graph theory and complex network analysis methods are used to calculate the affected area based on the location of the fault point and the network topology of the power system. Fault information and the scope of impact are displayed in real time through visualization technology.

[0087] A power system fault detection model can be constructed using convolutional neural networks (CNNs) and trained on a large number of labeled normal and abnormal electrical signal samples. The model employs a multi-layered convolutional and pooling layer structure to extract deep features from the electrical signals layer by layer. During training, the backpropagation algorithm is used to adjust the network weights, enabling the model to learn and recognize the differences in electrical signal features under normal and abnormal states. After training, the detection model can perform real-time analysis of the input electrical signals and output a judgment result indicating whether they belong to a normal or abnormal state.

[0088] A high-frequency electromagnetic wave transmitter emits a specific frequency signal towards the suspected fault area. This high-frequency signal is reflected when it encounters the fault point during propagation, and the reflected signal is captured by a receiver. Time-frequency analysis (such as using short-time Fourier transform or wavelet transform) is performed on the reflected signal to extract its time-domain and frequency-domain characteristics, such as reflection coefficient and attenuation characteristics. Based on these characteristics, combined with the electrical parameters and model of the power system, the precise location of the fault point is calculated.

[0089] The time-frequency analysis of the reflected signal based on the Short-Time Fourier Transform (STFT) is calculated using the following formula:

[0090]

[0091] in, It is a reflected signal. It's a time delay. It is angular frequency. Representing window functions Translated along the time axis Units, It is a complex exponential term used to convert a signal from the time domain to the frequency domain.

[0092] By calculating the STFT value under different time delays and angular frequencies, the time-frequency distribution of the reflected signal can be obtained, thereby revealing the characteristics of the fault point and locating its position.

[0093] The power system is abstracted into a graph structure using graph theory, with power equipment as nodes and transmission lines as edges. After identifying the fault point, the network topology of the power system is traversed from the fault point using a graph shortest path algorithm or a breadth-first search algorithm to calculate the electrical distance or path length from each node to the fault point. Simultaneously, connectivity analysis and community detection techniques from complex network analysis methods are combined to assess the impact of the fault point on the overall connectivity of the power system, identifying the connected components formed by nodes and lines directly connected to or electrically close to the fault point. Based on the power system's operating rules and safety constraints, such as voltage stability and power balance, the areas most severely affected by the fault are further screened out; these are the affected areas.

[0094] 203. Develop corresponding scheduling plans based on fault information;

[0095] Step 203 is similar to the execution process of step 103 above, and will not be described again here.

[0096] 204. Calculate the first weighting coefficient, the second weighting coefficient, and the third weighting coefficient;

[0097] Based on the preset first quantification result, the first weight coefficient is calculated using the analytic hierarchy process (AHP). The formula for calculating the first weight coefficient is as follows:

[0098]

[0099] in, The weight of the k-th indicator. Let n be the score for the k-th indicator, and n be the total number of the first evaluation indicators.

[0100] Based on the preset second quantization result, the second weighting coefficient is calculated using the entropy weighting method. The formula for calculating the second weighting coefficient is as follows:

[0101]

[0102] in, Let g be the entropy value of the g-th indicator, and m be the total number of the second evaluation indicators.

[0103] Based on the pre-defined third quantification results and load classification, the third weight coefficient is calculated using grey relational analysis. The calculation formula is:

[0104]

[0105] in, Let y be the grey relational degree of the y-th load. The priority score is assigned to the y-th load, where q is the total number of loads;

[0106] Based on the preset fourth quantification result, the fourth weight coefficient is calculated using the fuzzy comprehensive evaluation method. The formula for calculating the fourth weight coefficient is as follows:

[0107]

[0108] in, Let the membership degree of the e-th fourth evaluation indicator be denoted as . Let be the weight of the e-th fourth evaluation indicator, and p be the total number of fourth evaluation indicators.

[0109] Collect historical power grid recovery data, including historical power grid recovery time, historical charging and discharging records of energy storage systems, historical load recovery sequence, historical power loss and characteristic data, and historical output data of new energy power generation systems.

[0110] This study analyzes the specific impact of historical power grid restoration time on system stability, user satisfaction, and economic losses. For each influencing factor, based on its characteristics and degree of impact, appropriate quantification methods and standards are used to convert these impacts into specific numerical values, thus obtaining the first quantification results. The first quantification results include the quantified values ​​of primary evaluation indicators such as system stability, user satisfaction, and economic losses. For example, by integrating parameters such as frequency and voltage and training a neural network model, restoration time is converted into a system stability index. K-means clustering analysis is used to analyze user outage tolerance, mapping restoration time to satisfaction scores. A piecewise linear function is constructed to fit the relationship between restoration time and industrial, commercial losses, and residential compensation costs. The economic loss value is calculated by combining the GDP correction coefficient. The analytic hierarchy process (AHP) is used to determine the relative importance of each primary evaluation indicator through pairwise comparisons, constructing a judgment matrix and calculating eigenvectors. After consistency testing, the weights of each primary evaluation indicator are determined, and a corresponding score is assigned to each primary evaluation indicator. This score reflects the indicator's performance in actual conditions.

[0111] Based on the analysis of historical charge and discharge records of energy storage systems, the role and efficiency of energy storage systems in the grid restoration process are analyzed, taking into account factors such as capacity, charge and discharge rate, and energy loss. Through technical specification analysis of energy storage systems, collection of actual operation data, or simulation, the impact of power differences of energy storage systems on grid restoration is quantified, resulting in a second quantification result. The second quantification result includes the quantified values ​​of second evaluation indicators such as energy storage system capacity, charge and discharge rate, and energy loss, and the entropy value of each second evaluation indicator is calculated using the entropy weight method.

[0112] Based on multiple dimensions such as historical load recovery sequence, historical power loss, and characteristic data, the impact of these dimensions on the load is quantified through methods such as questionnaires, expert consultation, and data analysis to obtain a third quantitative result. The characteristic data includes key parameters such as load type, importance, and regional distribution. The third quantitative result includes the quantitative values ​​of third evaluation indicators such as load importance, priority, sensitivity, type, historical recovery sequence, historical power loss, and user impact. The load recovery sequence in typical successful recovery cases is selected as the reference sequence, and the actual quantitative data of each load is used as the comparison sequence. The similarity between each load and the ideal state is measured by calculating the grey correlation degree between the comparison sequence and the reference sequence. At the same time, the independent scoring results of the expert group for various types of loads are integrated, and the correlation coefficient is organically combined with the expert scores through a weighted fusion algorithm to assign a corresponding priority score to each load.

[0113] Based on historical output data of the new energy power generation system, this study analyzes specific parameters such as historical output deviation rate, prediction error rate, consumption quota completion rate, and inertial response contribution during grid recovery. Through data standardization and outlier removal, the specific performance of each parameter is quantified, forming a fourth quantitative result. This fourth quantitative result includes fourth evaluation indicators such as power generation reliability, prediction accuracy, policy compliance, and system support capacity. A fuzzy comprehensive evaluation method is used to establish membership functions for each fourth evaluation indicator, transforming actual operating data into standard membership values. A time decay factor is introduced to assign higher weights to recent data, and a normalization formula is used to calculate the weights of each fourth evaluation indicator. Furthermore, the weights can be adjusted based on scores derived from new energy policy requirements and system operation needs.

[0114] 205. Construct a MILP model containing continuous and integer variables, and determine the corresponding objective function and constraints. The objective function comprehensively considers the grid recovery time, the difference in charging and discharging power of the energy storage system, the power loss of the load, and the contribution of new energy generation. The first weight coefficient, the second weight coefficient, the third weight coefficient, and the fourth weight coefficient are used to quantify the importance of grid recovery time, energy storage system power difference, load power loss, and new energy generation contribution in the objective function, respectively.

[0115] The objective function is:

[0116]

[0117] Where T is the power grid recovery time, α is the first weighting coefficient, β is the second weighting coefficient, and γ is the third weighting coefficient. P is the fourth weighting coefficient. i 1 and P i 2 represents the output power and input power of energy storage system i, respectively. P is the power loss of load j. e The power generation capacity of new energy sources;

[0118] The constraints are as follows: the total output of all generators and the charging and discharging power of energy storage systems in the power system must be equal to the total load demand and network losses; the voltage amplitude of each node must be maintained within the allowable range; and the active power transmitted by the line must not exceed its thermal stability limit.

[0119] Positive terms, such as the difference between grid recovery time and energy storage power, need to be minimized, directly corresponding to the requirements of rapid power restoration and stable operation; negative terms, such as the contribution of new energy power generation and load power loss, achieve the goal of maximization through mathematical transformation. When the objective function is minimized as a whole, negative terms prompt the algorithm to prioritize increasing the output of new energy and reducing load loss.

[0120] 206. Select the objective MILP solver, input the scheduling plan into the MILP model for solution, analyze the optimal solution and adjust the model parameters to finally obtain the scheduling result that meets the requirements;

[0121] The constructed MILP model is input into solvers such as CPLEX or Gurobi for solving. During the solution process, the weight coefficients in the objective function and constraints such as power balance, voltage amplitude, and line transmission limits are iteratively optimized to generate a matching table between power sources and loads. Based on the matching table, the actual power nodes that can be connected are screened, and power sources that cannot supply power due to faults or insufficient capacity are excluded. At the same time, a list of load points that need to be restored to power is extracted and their power requirements are quantified. Finally, a scheduling result containing power access point information and load point information that needs to be restored to power is formed.

[0122] 207. Based on the power access point and load point to be restored in the scheduling results, and combined with the network topology of the power system, a path search algorithm is used to search for power supply paths that meet the preset conditions, and the search results are obtained. The search results include multiple possible power supply paths and their related parameters.

[0123] Based on the scheduling results, the power access point and the load point to be restored are identified. The power system network topology data is integrated to construct a graph structure model with power equipment as nodes and connecting lines as edges. The edges are assigned weight values ​​based on comprehensive line parameters and operating status. A path search algorithm adapted to the characteristics of the power system is selected, with the power access point set as the starting point and the load point as the target point. Path traversal search is carried out in the constructed graph structure model in combination with preset conditions, including line capacity, path length, and equipment reliability. The searched paths are judged and filtered in real time, and paths that meet the preset conditions are retained as candidate power supply paths.

[0124] Locate the power access point and the load point to be restored, integrate the power system network topology data, and construct a graph structure model with power equipment as nodes and connecting lines as edges. In this graph structure model, assign a weight value to each edge that comprehensively considers line parameters and operating status. Line parameters include resistance R and reactance X, while operating parameters include current load rate λ and available capacity C. The formula for calculating the weight value can be:

[0125]

[0126] Where α, β, and γ are weighting coefficients determined based on system characteristics.

[0127] A path search algorithm adapted to the characteristics of the power system is selected. The power access point is set as the starting point of the search and the load point is set as the target point. In combination with preset conditions, path traversal search is carried out in the constructed graph structure model. During the search process, the quality of the path is evaluated based on the edge weight value. The smaller the weight value, the better the overall performance of the path. The searched paths are judged and filtered in real time, and the paths that meet the preset conditions are retained as candidate power supply paths, thereby realizing efficient and reliable power restoration path planning.

[0128] 208. Use a multi-objective optimization algorithm to optimize and select the search results to obtain the optimal recovery path;

[0129] Define a set of objective functions for the multi-objective optimization problem, including the path length objective function, the recovery time objective function, and the energy storage resource consumption objective function. Based on the set of objective functions, iteratively search the search results using an adaptive non-dominated sorting genetic algorithm. During the iteration process, if convergence stagnation is detected, the mutation probability is automatically increased to escape local optima until the maximum number of iterations is reached or the hypervolume of the Pareto front solution set no longer improves significantly. The optimized Pareto front is then output. Based on fuzzy multi-attribute decision-making and real-time running status information, the optimal recovery path is selected from the Pareto front.

[0130] A set of objective functions is defined, encompassing three key objective functions: path length objective function, recovery time objective function, and energy storage resource consumption objective function. The path length objective function focuses on minimizing the total path length, as shorter paths reduce energy loss during power transmission and line construction costs. The recovery time objective function aims to minimize the time required for power restoration, comprehensively considering factors such as equipment startup and line switching, with the goal of restoring power to the load as quickly as possible. The energy storage resource consumption objective function emphasizes reducing the energy consumption of the energy storage system by rationally allocating energy storage resources to reduce operating costs.

[0131] An adaptive non-dominated sorting genetic algorithm is used to iteratively search the search results. During the iteration process, the algorithm performs selection, crossover, and mutation operations based on the non-dominated sorting level and crowding distance of individuals. If convergence stagnation is detected, that is, the optimal solution set has not changed significantly for several generations, the algorithm will automatically increase the mutation probability and introduce more new individuals to escape the local optimum. The iteration will continue until the preset maximum number of iterations is reached, or the hypervolume of the Pareto front solution set no longer improves significantly.

[0132] Based on fuzzy multi-attribute decision-making and real-time operational status information, the optimal recovery path is selected from the Pareto front. Specifically, a fuzzy evaluation matrix is ​​first constructed, and the weights of each objective function and real-time operational data are comprehensively considered to calculate the fuzzy comprehensive evaluation value of each path. Then, the path with the highest evaluation value is selected as the optimal recovery path, thereby achieving efficient and reliable recovery of the power system.

[0133] The objective function for path length is:

[0134]

[0135] Where Li represents the length of the i-th segment in the path, and n is the total number of segments in the path, this function is used to minimize the total length of the path.

[0136] The recovery time objective function, used to minimize the time required for power restoration, can be expressed as: In actual calculations, Equipment startup time needs to be considered. Line switching time Factors such as The specific function form is determined based on the actual situation of the power system.

[0137] The objective function for energy storage resource consumption, used to minimize the energy consumption of the energy storage system, can be expressed as follows: ,

[0138] It can be based on the charging and discharging power of the energy storage device Integrating with time t, i.e. Where t1 and t2 are the start and end times of charging and discharging.

[0139] Hypervolume is used to measure the quality of Pareto front solution sets and can be represented as follows: Where R is the coverage area of ​​the solution set in the target space. , ,... Let n be the objective function, and n be the number of objective functions.

[0140] Calculation of fuzzy comprehensive evaluation value: Let the fuzzy evaluation matrix be... ,in, Let represent the fuzzy evaluation value of the i-th path under the j-th objective function, and let the weight vector be . ,and Then the fuzzy comprehensive evaluation value of the i-th path By comparing the paths Value, selection The path with the highest value is the optimal recovery path.

[0141] 209. Based on the optimal recovery path, and in conjunction with distributed energy coordination and control technology and microgrid technology, perform power restoration execution.

[0142] Based on the optimal recovery path, a power system recovery strategy is determined. This strategy includes the recovery sequence, recovery time points, and the connection points and output power of each distributed energy source. Distributed energy source coordination and control technology is used to connect distributed energy sources according to the recovery strategy and adjust their corresponding output power. Combined with microgrid technology, autonomous control and energy optimization scheduling are implemented within the local power grid to ensure that distributed energy sources connected according to the optimal recovery path can work efficiently and collaboratively. The output power adjustment of distributed energy sources must satisfy the second formula, which is:

[0143]

[0144] in, The output power of distributed energy source z after adjustment. The basic output power of distributed energy source z, This refers to the power increment adjusted according to the recovery strategy and grid conditions.

[0145] In this embodiment of the invention, by acquiring the real-time operating status of the power system, accurately locating faults and implementing scheduling plans, and constructing a MILP model that comprehensively considers the grid recovery time, the difference in charging and discharging power of the energy storage system, and the load power loss for optimization, the optimal recovery path is finally determined and power supply is restored, which effectively improves the efficiency and accuracy of power system fault recovery and ensures the reliability and stability of power supply.

[0146] The power restoration method based on grid-connected energy storage in the embodiments of the present invention has been described above. The power restoration device based on grid-connected energy storage in the embodiments of the present invention will be described below. Please refer to [link to relevant documentation]. Figure 3 One embodiment of the power restoration device based on grid-connected energy storage in this invention includes:

[0147] The acquisition module 301 is used to acquire real-time operating status information of the power system;

[0148] The positioning module 302 is used to locate the fault location and the affected area using a fault detection algorithm when a fault is determined to have occurred in the power system based on real-time operating status information, and to obtain fault information.

[0149] Module 303 is designated to formulate a corresponding scheduling plan based on fault information;

[0150] The optimization module 304 is used to perform global optimization of the scheduling plan using a mixed integer linear programming algorithm to obtain the scheduling result, which includes power access point information and load point information to be restored.

[0151] The search module 305 is used to search for power supply paths that meet preset conditions based on the power access point and load point to be restored in the scheduling results, combined with the network topology results of the power system, and to obtain the search results. The search results include multiple possible power supply paths and their related parameters.

[0152] The selection module 306 is used to optimize the selection of search results using a multi-objective optimization algorithm to obtain the optimal recovery path;

[0153] The processing module 307 is used to perform power restoration execution processing based on the optimal restoration path, combined with distributed energy coordination and control technology and microgrid technology.

[0154] In this embodiment of the invention, by acquiring real-time operating status information of the power system, fault detection algorithms are used to quickly and accurately locate the fault location and affected area. Combined with mixed integer linear programming algorithms, the scheduling plan is globally optimized. Then, path search algorithms and multi-objective optimization algorithms are used to obtain the optimal recovery path. Finally, distributed energy coordination and control technology and microgrid technology are used to perform power restoration. This overcomes the problems of traditional power restoration methods, such as reliance on manual scheduling, long time consumption, low efficiency, and inaccurate plans. It can fully consider the complexity and real-time nature of the power system, effectively ensure power supply quality, and reduce power outage losses.

[0155] Please see Figure 4 Another embodiment of the power restoration device based on grid-connected energy storage in this invention includes:

[0156] The acquisition module 301 is used to acquire real-time operating status information of the power system;

[0157] The positioning module 302 is used to locate the fault location and the affected area using a fault detection algorithm when a fault is determined to have occurred in the power system based on real-time operating status information, and to obtain fault information.

[0158] Module 303 is designated to formulate a corresponding scheduling plan based on fault information;

[0159] The optimization module 304 is used to perform global optimization of the scheduling plan using a mixed integer linear programming algorithm to obtain the scheduling result, which includes power access point information and load point information to be restored.

[0160] The search module 305 is used to search for power supply paths that meet preset conditions based on the power access point and load point to be restored in the scheduling results, combined with the network topology results of the power system, and to obtain the search results. The search results include multiple possible power supply paths and their related parameters.

[0161] The selection module 306 is used to optimize the selection of search results using a multi-objective optimization algorithm to obtain the optimal recovery path;

[0162] The processing module 307 is used to perform power restoration execution processing based on the optimal restoration path, combined with distributed energy coordination and control technology and microgrid technology.

[0163] Optionally, the positioning module can be specifically used for:

[0164] A detection model is obtained by modeling the power system using convolutional neural networks. This model can learn and identify electrical signal characteristics under normal and abnormal conditions. The detection model is used to analyze the electrical signal characteristics to determine suspected fault areas. High-frequency electromagnetic wave transmission and reception technology is combined to transmit high-frequency signals to the suspected fault areas, and the fault point is located by analyzing the time-frequency characteristics of the reflected signals. Graph theory and complex network analysis methods are used to calculate the affected area based on the location of the fault point and the network topology of the power system. Fault information and the scope of impact are displayed in real time through visualization technology.

[0165] Optionally, optimization module 304 includes:

[0166] Building unit 3041 is used to build a MILP model containing continuous and integer variables, and to determine the objective function and constraints.

[0167] The objective function is:

[0168]

[0169] Where T is the power grid recovery time, α is the first weighting coefficient, β is the second weighting coefficient, and γ is the third weighting coefficient. P is the fourth weighting coefficient. i 1 and P i 2 represents the output power and input power of energy storage system i, respectively. P is the power loss of load j. e The power generation capacity of new energy sources;

[0170] The constraints are as follows: the total output of all generators and the charging and discharging power of energy storage systems in the power system must be equal to the total load demand and network losses; the voltage amplitude of each node must be maintained within the allowable range; and the active power transmitted by the line must not exceed its thermal stability limit.

[0171] Solver 3042 is used to select the target MILP solver and input the scheduling plan into the MILP model for solving;

[0172] The processing unit 3043 is used to analyze the optimal solution output by the solver and adjust the MILP model parameters until a scheduling result that meets the requirements is obtained.

[0173] Optionally, optimization module 304 also includes:

[0174] The calculation unit 3044 is used to calculate the first weight coefficient based on a preset first quantization result using the analytic hierarchy process (AHP). The formula for calculating the first weight coefficient is as follows:

[0175]

[0176] in, The weight of the k-th indicator. Let n be the score for the k-th indicator, and n be the total number of the first evaluation indicators.

[0177] Based on the preset second quantization result, the second weighting coefficient is calculated using the entropy weighting method. The formula for calculating the second weighting coefficient is as follows:

[0178]

[0179] in, Let g be the entropy value of the g-th indicator, and m be the total number of the second evaluation indicators.

[0180] Based on the pre-defined third quantification results and load classification, the third weight coefficient is calculated using grey relational analysis. The calculation formula is:

[0181]

[0182] in, Let y be the grey relational degree of the y-th load. The priority score is assigned to the y-th load, where q is the total number of loads;

[0183] Based on the preset fourth quantification result, the fourth weight coefficient is calculated using the fuzzy comprehensive evaluation method. The formula for calculating the fourth weight coefficient is as follows:

[0184]

[0185] in, Let the membership degree of the e-th fourth evaluation indicator be denoted as . Let be the weight of the e-th fourth evaluation indicator, and p be the total number of fourth evaluation indicators.

[0186] Optionally, the search module 305 can be specifically used for:

[0187] Based on the scheduling results, the power access point and the load point to be restored are identified. The power system network topology data is integrated to construct a graph structure model with power equipment as nodes and connecting lines as edges. The edges are assigned weight values ​​based on comprehensive line parameters and operating status. A path search algorithm adapted to the characteristics of the power system is selected, with the power access point set as the starting point and the load point as the target point. Path traversal search is carried out in the constructed graph structure model in combination with preset conditions, including line capacity, path length, and equipment reliability. The searched paths are judged and filtered in real time, and paths that meet the preset conditions are retained as candidate power supply paths.

[0188] Optionally, selection module 306 can be specifically used for:

[0189] Define a set of objective functions for the multi-objective optimization problem, including the path length objective function, the recovery time objective function, and the energy storage resource consumption objective function. Based on the set of objective functions, iteratively search the search results using an adaptive non-dominated sorting genetic algorithm. During the iteration process, if convergence stagnation is detected, the mutation probability is automatically increased to escape local optima until the maximum number of iterations is reached or the hypervolume of the Pareto front solution set no longer improves significantly. The optimized Pareto front is then output. Based on fuzzy multi-attribute decision-making and real-time running status information, the optimal recovery path is selected from the Pareto front.

[0190] Optionally, the processing module 307 can be specifically used for:

[0191] Based on the optimal recovery path, a power system recovery strategy is determined. This strategy includes the recovery sequence, recovery time points, and the connection points and output power of each distributed energy source. Distributed energy source coordination and control technology is used to connect distributed energy sources according to the recovery strategy and adjust their corresponding output power. Combined with microgrid technology, autonomous control and energy optimization scheduling are implemented within the local power grid to ensure that distributed energy sources connected according to the optimal recovery path can work efficiently and collaboratively. The output power adjustment of distributed energy sources must satisfy the second formula, which is:

[0192]

[0193] in, The output power of distributed energy source z after adjustment. The basic output power of distributed energy source z, This refers to the power increment adjusted according to the recovery strategy and grid conditions.

[0194] In this embodiment of the invention, by acquiring the real-time operating status of the power system, accurately locating faults and implementing scheduling plans, and constructing a MILP model that comprehensively considers the grid recovery time, the difference in charging and discharging power of the energy storage system, and the load power loss for optimization, the optimal recovery path is finally determined and power supply is restored, which effectively improves the efficiency and accuracy of power system fault recovery and ensures the reliability and stability of power supply.

[0195] above Figure 3 and Figure 4 The power supply restoration device based on grid-connected energy storage in this embodiment of the invention will be described in detail from the perspective of modular functional entities. The power supply restoration equipment based on grid-connected energy storage in this embodiment of the invention will be described in detail from the perspective of hardware processing.

[0196] See Figure 5 As shown, the power restoration device based on grid-connected energy storage includes a processor 500 and a memory 501. The memory 501 stores machine-executable instructions that can be executed by the processor 500. The processor 500 executes the machine-executable instructions to implement the aforementioned power restoration method based on grid-connected energy storage.

[0197] Furthermore, Figure 5 The power restoration device based on grid-connected energy storage shown also includes a bus 502 and a communication interface 503. The processor 500, the communication interface 503, and the memory 501 are connected through the bus 502.

[0198] The memory 501 may include high-speed random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 503 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc. The bus 502 may be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 5 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0199] The processor 500 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of the processor 500 or by instructions in software form. The processor 500 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this disclosure. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this disclosure can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules may reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 501. The processor 500 reads the information in memory 501 and, in conjunction with its hardware, completes the method steps of the aforementioned embodiment.

[0200] The present invention also provides a power restoration device based on grid-connected energy storage. The computer device includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs the steps of the power restoration method based on grid-connected energy storage in the above embodiments.

[0201] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the power supply restoration method based on grid-connected energy storage.

[0202] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0203] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0204] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power supply restoration method based on networked energy storage, characterized by, The power supply recovery method based on network construction energy storage comprises: acquiring real-time operation state information of a power system; when determining that a fault occurs in the power system based on the real-time operation state information, locating a fault position and an affected area by using a fault detection algorithm to obtain fault information; formulating a corresponding dispatching plan based on the fault information; globally optimizing the dispatching plan by using a mixed integer linear programming algorithm to obtain a dispatching result, the dispatching result comprising power supply access point information and load point information to be restored; searching for power supply paths satisfying preset conditions by using a path search algorithm according to the power supply access points and the load points to be restored determined in the dispatching result and in combination with network topology results of the power system to obtain search results, the search results comprising a plurality of possible power supply paths and related parameters thereof; optimizing and selecting the search results by using a multi-objective optimization algorithm to obtain an optimal recovery path; performing a recovery execution process according to the optimal recovery path in combination with distributed energy coordination control technology and micro-grid technology.

2. The method of claim 1, wherein, The method for locating the fault position and the affected area by using the fault detection algorithm comprises: modeling the power system by using a convolutional neural network to obtain a detection model, the detection model being capable of learning and identifying electrical signal features in normal and abnormal states; analyzing the electrical signal features by using the detection model to determine a suspected fault area; in combination with high-frequency electromagnetic wave transmission and reception technology, transmitting high-frequency signals to the suspected fault area and locating a fault point by analyzing time-frequency characteristics of reflected signals; in combination with graph theory and complex network analysis methods, calculating an affected area according to the position of the fault point and a network topology structure of the power system and displaying the fault information and the affected area in real time by using a visualization technology.

3. The method of claim 1, wherein, The method for globally optimizing the dispatching plan by using the mixed integer linear programming algorithm to obtain the dispatching result comprises: constructing a MILP model comprising continuous variables and integer variables, determining an objective function and constraint conditions; selecting a target MILP solver, inputting the dispatching plan into the MILP model for solving; analyzing an optimal solution output by the solver, adjusting MILP model parameters until a dispatching result meeting requirements is obtained; the objective function is: ； Wherein, T is the power grid recovery time, a is the first weight coefficient, β is the second weight coefficient, γ is the third weight coefficient, is the fourth weight coefficient, P i 1 and P i 2 are the output power and input power of the energy storage system i respectively, is the power loss of the load j, P e is the power generation of the new energy; the constraint conditions are: the sum of all generator outputs and energy storage system charge / discharge powers in the power system must be equal to the sum of all load demands and network losses; the voltage amplitude of each node must be maintained within an allowable range; the active power transmitted by a line must not exceed its thermal stability limit.

4. The method of claim 3, wherein, The method for globally optimizing the dispatching plan by using the mixed integer linear programming algorithm to obtain the dispatching result further comprises: based on a preset first quantitative result, calculating a first weight coefficient by using an analytic hierarchy process, the calculation formula of the first weight coefficient being: ； wherein, is the weight of the kth first evaluation index, is the score of the kth first evaluation index, and n is the total number of the first evaluation indices. based on a preset second quantitative result, calculating a second weight coefficient by using an entropy weight method, the calculation formula of the second weight coefficient being: ； wherein, is the entropy value of the gth second evaluation index, and m is the total number of the second evaluation indices. Based on the preset third quantization result and the load classification condition, a third weight coefficient is calculated by using a grey correlation analysis method, and the third weight coefficient The calculation formula is: ； wherein, is the grey correlation degree of the yth load, is the priority score of the yth load, q is the total number of loads; based on a preset fourth quantitative result, calculating a fourth weight coefficient by using a fuzzy comprehensive evaluation method, the calculation formula of the fourth weight coefficient being: ； wherein, is the membership degree of the e-th fourth evaluation index, is the weight of the e-th fourth evaluation index, and p is the total number of the fourth evaluation indices.

5. The method of claim 1, wherein, The power supply access point and the load point to be restored power supply determined according to the scheduling result are combined with the network topology result of the power system, and a power supply path meeting preset conditions is searched by using a path search algorithm, including: Based on the scheduling result, the power supply access point and the load point to be restored power supply are determined, the power system network topology data is integrated, a graph structure model taking power equipment as nodes and connection lines as edges is constructed, and the weight values of comprehensive line parameters and operating states are assigned to the edges; A path search algorithm suitable for the characteristics of the power system is selected, the power supply access point is set as the starting point of the search, the load point is set as the target point, and the preset conditions are combined to carry out path traversal search in the constructed graph structure model, and the preset conditions involve line capacity, path length and equipment reliability; The searched paths are subjected to real-time condition judgment and screening, and the paths meeting the preset conditions are reserved as candidate power supply paths.

6. The method of claim 1, wherein, The search result is optimized and selected by using a multi-objective optimization algorithm to obtain an optimal restoration path, including: A set of objective functions of a multi-objective optimization problem is defined, the set of multi-objective functions includes a path length objective function, a restoration time objective function and a storage resource consumption objective function; Based on the set of objective functions, the search result is iteratively searched according to a self-adaptive non-dominated sorting genetic algorithm, in the iteration process, if convergence stagnation is detected, the mutation probability is automatically increased to jump out of the local optimum until the maximum iteration number or the hyper volume of the Pareto front solution set is no longer significantly improved, and the optimized Pareto front is output; Based on fuzzy multi-attribute decision and the real-time operating state information, an optimal restoration path is selected from the Pareto front.

7. The method of claim 1, wherein, According to the optimal restoration path, distributed energy coordination control technology and micro-grid technology are combined to perform power supply restoration execution processing, including: According to the optimal restoration path, a restoration strategy of the power system is determined, the restoration strategy includes a restoration sequence, a restoration time point and an access point and an output power of each distributed energy; Distributed energy coordination control technology is used to access the distributed energy according to the restoration strategy and adjust the corresponding output power; Autonomous control and energy optimization scheduling are implemented in the local power grid by combining micro-grid technology, so that the distributed energy accessed according to the optimal restoration path can work efficiently and cooperatively; The output power adjustment of the distributed energy needs to meet a second formula, and the second formula is: ； wherein, is the adjusted output power of the distributed energy source z, is the base output power of the distributed energy source z, is the power increment adjusted according to the recovery strategy and the grid state.

8. A power supply restoration device based on networked energy storage, characterized by, The power supply restoration device based on networked storage energy includes: An acquisition module is configured to acquire real-time operating state information of a power system; A positioning module is configured to, when it is determined that the power system has a fault based on the real-time operating state information, locate a fault position and an affected area by using a fault detection algorithm to obtain fault information; A specifying module is configured to develop a corresponding scheduling plan based on the fault information; An optimization module is configured to globally optimize the scheduling plan by using a mixed integer linear programming algorithm to obtain a scheduling result, and the scheduling result includes power supply access point information and load point information to be restored power supply; An optimization module is configured to globally optimize the scheduling plan by using a mixed integer linear programming algorithm to obtain a scheduling result, and the scheduling result includes power supply access point information and load point information to be restored power supply; The search module is configured to search for a power supply path meeting a preset condition according to the power supply access point and the load point to be restored in the scheduling result, the network topology result of the power system, and a path search algorithm, and obtain a search result including a plurality of possible power supply paths and related parameters of the power supply paths. The selection module is configured to select the search result by using a multi-objective optimization algorithm, and obtain an optimal restoration path. The processing module is configured to perform a restoration execution process according to the optimal restoration path, in combination with a distributed energy coordination control technology and a micro-grid technology.

9. A power restoration device based on networked energy storage, characterized by, The power supply restoration device based on network-constructed energy storage comprises a memory and at least one processor. The at least one processor invokes the instructions in the memory, so that the power supply restoration device based on network-constructed energy storage performs the power supply restoration method based on network-constructed energy storage according to any one of claims 1-7.

10. A computer-readable storage medium having stored thereon instructions, the computer-readable storage medium comprising: The instructions are executed by the processor to implement the power supply restoration method based on network-constructed energy storage according to any one of claims 1-7.

Images