A flexible device access method and system

Abstract

The present invention discloses a method and system for accessing a flexible device, which obtains node voltage data of a distribution network and load data of a flexible device in real time; calculates voltage distribution characteristics of the distribution network, and obtains candidate access points for flexible device access; changes the topology of the distribution network, calculates voltage stability of candidate access points under various topologies, and screens a first target access point; calculates a margin improvement value when the flexible device accesses the first target access point; if the margin improvement value is greater than a preset margin improvement threshold, the flexible device is connected to the first target access point; otherwise, the sensitivity of each node voltage in the distribution network to the candidate access point and the injected power is calculated, and the flexible device is connected to the screened second target access point. By adopting the embodiments of the present invention, the influence of flexible device access on the distribution network under different power grid topologies can be considered, and more optimal flexible device access parameters can be provided to ensure voltage stability and high reliability of the distribution network.

Description

Technical Field

[0001] The present invention relates to the technical field of flexible device configuration, and in particular to a method and system for accessing a flexible device. Background Art

[0002] In low-voltage distribution networks, flexible devices are often used to improve the flexibility and stability of the grid, and can effectively cope with load fluctuations, renewable energy access, and power quality issues. When introducing flexible devices in low-voltage distribution networks, the selection of access points has a crucial impact on the system voltage stability.

[0003] The access of flexible devices changes the power flow distribution and voltage distribution of the distribution lines, and different access locations and line topologies will lead to different cumulative effects of voltage deviation. When multiple flexible devices are connected in parallel, their coordinated control also faces challenges. On the one hand, the control objectives and constraints of different access points are not the same, making it difficult to achieve distributed autonomous regulation; on the other hand, centralized control requires a large amount of communication and computing resources, and real-time performance and reliability are difficult to guarantee. Summary of the invention

[0004] The technical problem to be solved by the present invention is how to access flexible devices to ensure the voltage stability and power supply quality of the distribution network. A method and system for accessing flexible devices are provided. The target access point is determined by introducing a margin improvement value and a sensitivity analysis method, and node voltage data and load data are collected in real time. The impact of the access of flexible devices on the distribution network under different power grid topologies can be taken into account, and more optimal flexible device access parameters can be provided to ensure the voltage stability and high reliability of the distribution network.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a method for accessing a flexible device, including:

[0006] Obtain node voltage data of the distribution network and load data of flexible equipment in real time;

[0007] Calculating voltage distribution characteristics of the distribution network according to the node voltage data, obtaining a voltage deviation area based on the voltage distribution characteristics, and using the voltage deviation area as a candidate access point for flexible device access;

[0008] According to the candidate access points, the topology of the distribution network is changed, the voltage stability of the candidate access points under each of the topologies is calculated, and according to the voltage stability, a first target access point is screened from the candidate access points;

[0009] Calculating, according to the node voltage data and the load data, a margin improvement value when the flexible device is connected to the first target access point;

[0010] If the margin improvement value is greater than a preset margin improvement threshold, connecting the flexible device to the first target access point;

[0011] If the margin improvement value is not greater than a preset margin improvement threshold, the sensitivity of each node voltage in the distribution network to the candidate access point and the injected power is calculated, a second target access point is obtained from the candidate access points according to the sensitivity, and the flexible device is connected to the second target access point.

[0012] As an improvement of the above solution, the voltage distribution characteristics of the distribution network are calculated according to the node voltage data, a voltage deviation area is obtained based on the voltage distribution characteristics, and the voltage deviation area is used as a candidate access point for flexible device access, including:

[0013] According to the geographical locations of the nodes of the distribution network, spatial classification is performed on the node voltage data to obtain a distribution network operation status data set;

[0014] According to a preset number of iterations and a preset convergence accuracy, a Newton-Raphson iterative operation is performed on the distribution network operation status data set to obtain a node voltage data matrix;

[0015] Using a density-based spatial clustering algorithm to classify the node voltage data matrix, and performing spatial mapping on the classified node voltage data matrix according to a clustering radius to obtain a voltage distribution characteristic structure diagram;

[0016] According to the voltage distribution characteristic structure diagram, a plurality of voltage characteristic regions are obtained; the voltage characteristic regions include a voltage deviation region and a voltage normal region;

[0017] According to the electrical distance between the node in the voltage deviation area and the node in the normal voltage area, a candidate access point for the flexible device to access is obtained from the voltage deviation area.

[0018] As an improvement of the above solution, the step of changing the topology of the distribution network according to the candidate access points, calculating the voltage stability of the candidate access points under each of the topologies, and selecting a first target access point from the candidate access points according to the voltage stability includes:

[0019] For each of the candidate access points, a plurality of switch combination sequences are generated based on a depth-first search; the switch combination sequences include the switch states of each node in the line;

[0020] According to the plurality of switch combination sequences, the topology structure of the distribution network is changed to generate a plurality of topology reconstruction scenarios;

[0021] Obtaining voltage stability according to a maximum voltage deviation value and a voltage fluctuation recovery time of the candidate access point in the topology reconstruction scenario;

[0022] sorting the candidate access points according to the voltage stability and a preset voltage deviation threshold;

[0023] According to the ranking result of the candidate access points, a first target access point is obtained by screening from the candidate access points.

[0024] As an improvement of the above solution, the voltage stability is obtained according to the maximum voltage deviation value and the voltage fluctuation recovery time of the candidate access point in the topology reconstruction scenario, including:

[0025] Calculating, according to the node voltage data, a voltage fluctuation value caused by topology reconstruction in the topology reconstruction scenario;

[0026] Obtaining the admittance of each node of the distribution network to obtain a node admittance matrix;

[0027] Inverting the node admittance matrix to obtain voltage fluctuation transmission decay characteristics of the candidate access point in the topology reconstruction scenario;

[0028] According to the voltage fluctuation transmission decay characteristics and the voltage fluctuation value, an exponentially weighted moving average method is used to solve the maximum voltage deviation and the recovery time;

[0029] The maximum voltage deviation and the recovery time are multiplied to obtain voltage stability.

[0030] As an improvement of the above solution, the sorting of the candidate access points according to the voltage stability and the preset voltage deviation threshold value includes:

[0031] Calculate the frequency at which the voltage stability of the candidate access point in all the topology reconstruction scenarios is less than a preset voltage deviation threshold;

[0032] The candidate access points are sorted according to the frequencies.

[0033] As an improvement of the above solution, the calculating, according to the node voltage data and the load data, the margin improvement value when the flexible device is connected to the first target access point includes:

[0034] A support vector regressor is used to calculate a load characteristic value for the load data, and a plurality of load scenarios are obtained according to the load characteristic value;

[0035] Performing four-layer wavelet decomposition on the node voltage data, and obtaining a voltage stability margin reference value corresponding to each load scenario according to the decomposition result and the load scenario;

[0036] According to a preset voltage amplitude deviation threshold and a preset recovery time threshold, performing minimum variance recursion on the node voltage data to obtain an actual value of the voltage stability margin of each node;

[0037] According to the plurality of load scenarios, a difference between the actual voltage stability margin value and the voltage stability margin reference value is calculated to obtain a margin improvement value when the flexible device is connected to the first target access point.

[0038] As an improvement of the above solution, the node voltage data is subjected to four-layer wavelet decomposition, and a voltage stability margin reference value corresponding to each load scenario is obtained according to the decomposition result and the load scenario, including:

[0039] Performing four-layer wavelet decomposition on the node voltage data to obtain an initial value of the voltage amplitude and a voltage amplitude change value;

[0040] Calculating the maximum deviation value and voltage recovery time of each load scenario according to the voltage amplitude change value and the voltage amplitude initial value;

[0041] A product operation is performed on the maximum deviation value and the voltage recovery time of each load scenario to obtain a voltage stability margin reference value corresponding to each load scenario.

[0042] As an improvement of the above solution, if the margin improvement value is not greater than a preset margin improvement threshold, the sensitivity of the voltage of each node in the distribution network to the injected power and position of the candidate access point is calculated, a second target access point is obtained from the candidate access points according to the sensitivity, and the flexible device is connected to the second target access point, including:

[0043] If the margin improvement value is not greater than a preset margin improvement threshold, simulate the flexible devices accessing the distribution network through the candidate access points one by one to obtain several access scenarios;

[0044] In each access scenario, according to the injection power data of the flexible device, a preset disturbance signal is applied to the distribution network node, and the sensitivity of the node voltage data to the injection power is obtained by using a gradient descent algorithm;

[0045] According to the sensitivity of the injected power, an electrical distance between the flexible device and a node of the distribution network is obtained;

[0046] According to the electrical distance between nodes, a node topology correlation feature matrix is ​​constructed through a graph neural network to obtain the node position sensitivity;

[0047] Performing a weighted combination on the injection power sensitivity and the node position sensitivity to obtain a comprehensive sensitivity;

[0048] According to the comprehensive sensitivity, a second target access point is obtained by screening from the candidate access points, and the flexible device is connected to the second target access point.

[0049] As an improvement of the above solution, if the margin improvement value is not greater than a preset margin improvement threshold, the sensitivity of the voltage of each node in the distribution network to the injected power and the position of the candidate access point is calculated, a second target access point is obtained from the candidate access points according to the sensitivity, and the flexible device is connected to the second target access point. The access method of the flexible device also includes:

[0050] When multiple flexible devices are connected in parallel, the flexible devices are layered according to the bus voltage levels of the flexible devices to determine the response priorities of the flexible devices;

[0051] The reactive power of the flexible device is adaptively adjusted according to the load data of the flexible device and the response priority.

[0052] The embodiment of the present invention further provides an access system for a flexible device, including:

[0053] A data acquisition module is used to obtain node voltage data of the distribution network and load data of flexible devices in real time;

[0054] A candidate access point selection module, used to calculate the voltage distribution characteristics of the distribution network according to the node voltage data, obtain a voltage deviation area based on the voltage distribution characteristics, and use the voltage deviation area as a candidate access point for flexible device access;

[0055] A first target access point screening module, configured to change the topology of the distribution network according to the candidate access points, calculate the voltage stability of the candidate access points under each of the topologies, and screen the first target access point from the candidate access points according to the voltage stability;

[0056] a margin improvement value calculation module, configured to calculate a margin improvement value when the flexible device is connected to the first target access point according to the node voltage data and the load data;

[0057] A first access module, configured to connect the flexible device to the first target access point if the margin improvement value is greater than a preset margin improvement threshold;

[0058] The second access module is used to calculate the sensitivity of each node voltage in the distribution network to the candidate access point and the injected power if the margin improvement value is not greater than a preset margin improvement threshold, screen a second target access point from the candidate access points according to the sensitivity, and connect the flexible device to the second target access point.

[0059] Compared with the prior art, the present invention discloses a method and system for accessing a flexible device, which obtains node voltage data of a distribution network and load data of a flexible device in real time; calculates voltage distribution characteristics of the distribution network according to the node voltage data, obtains a voltage deviation area based on the voltage distribution characteristics, and uses the voltage deviation area as a candidate access point for the flexible device to access; changes the topology of the distribution network according to the candidate access point, calculates the voltage stability of the candidate access point under each topology, and selects a first target access point from the candidate access points according to the voltage stability; calculates a margin improvement value when the flexible device accesses the first target access point according to the node voltage data and the load data; if the margin improvement value is greater than a preset margin improvement threshold, the flexible device is connected to the first target access point; if the margin improvement value is not greater than the preset margin improvement threshold, the sensitivity of each node voltage in the distribution network to the candidate access point and the injected power is calculated, a second target access point is obtained from the candidate access points according to the sensitivity, and the flexible device is connected to the second target access point. By adopting the embodiments of the present invention, the impact of flexible device access on the distribution network under different power grid topologies can be taken into account, and more optimal flexible device access parameters can be provided to ensure voltage stability and high reliability of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a schematic diagram of the steps of a method for accessing a flexible device provided by an embodiment of the present invention;

[0061] Figure 2 It is a structural schematic diagram of an access system for a flexible device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] In the description of the specification and claims, it is to be understood that the terms first, second, etc. in the specification and claims are only used for the purpose of describing the same technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated, nor necessarily describing the order or time sequence. The terms are interchangeable where appropriate. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features.

[0064] The embodiment of the present invention provides a method for accessing a flexible device. Figure 1 In this embodiment, the flexible device access method is specifically performed through steps S1 to S6:

[0065] S1. Obtain node voltage data of the distribution network and load data of flexible devices in real time.

[0066] It should be noted that the power flow distribution and voltage distribution of the distribution network will change after the flexible equipment is connected. By acquiring the node voltage data in real time, the data changes of the distribution network can be monitored, and the access plan of the flexible equipment can be further analyzed and optimized.

[0067] It should also be noted that during the operation of the distribution network, the load of the flexible device will fluctuate due to changes in the load of each node. When the load fluctuates, the flexible device needs to dynamically optimize its own operating conditions according to the real-time load distribution and adjust the active and reactive power output to maintain the system voltage balance. By obtaining the load data of the flexible device, the reliability of the distribution network operation can be guaranteed, and the flexible device can be further adjusted.

[0068] S2. Calculate voltage distribution characteristics of the distribution network according to the node voltage data, obtain a voltage deviation area based on the voltage distribution characteristics, and use the voltage deviation area as a candidate access point for flexible device access.

[0069] In some preferred embodiments of the present invention, a power flow calculation method is used to calculate the node voltage data to obtain the voltage amplitude and phase angle of each node, and the voltage distribution characteristics of the distribution network are obtained based on the voltage amplitude and the phase angle.

[0070] In some other preferred embodiments of the present invention, data clustering is performed based on the node voltage data to obtain voltage distribution characteristics in different regions and different time periods.

[0071] S3. According to the candidate access points, change the topology of the distribution network, calculate the voltage stability of the candidate access points under each of the topologies, and select a first target access point from the candidate access points according to the voltage stability.

[0072] It should be noted that the topology of the distribution network will change in different operating periods and under different operating conditions. It is possible that a node is closed in summer and open in winter. This is due to the change in user power demand. In addition, the access of flexible devices will also change the topology of the distribution network. By changing the topology of the distribution network, the effectiveness of the access of flexible devices can be fully verified.

[0073] S4. Calculate, according to the node voltage data and the load data, a margin improvement value when the flexible device is connected to the first target access point.

[0074] It should be noted that the margin improvement degree can be used to determine the effect of connecting flexible devices on improving voltage stability.

[0075] S5. If the margin improvement value is greater than a preset margin improvement threshold, connecting the flexible device to the first target access point.

[0076] S6. If the margin improvement value is not greater than a preset margin improvement threshold, the sensitivity of the voltage of each node in the distribution network to the candidate access point and the injected power is calculated, a second target access point is obtained from the candidate access points according to the sensitivity, and the flexible device is connected to the second target access point.

[0077] In the embodiment of the present invention, when the margin improvement value is not greater than the preset margin improvement threshold, it is considered that the improvement effect of the first target access point on voltage stability may not reach the optimal effect, so from the perspective of sensitivity analysis, the second target access point is re-screened.

[0078] In the above scheme, the target access point is determined by introducing the margin improvement value and sensitivity analysis method, and the node voltage data and load data are collected in real time. This can take into account the impact of flexible device access on the distribution network under different power grid topologies, and provide better flexible device access parameters to ensure the voltage stability and high reliability of the distribution network.

[0079] In some preferred embodiments, when the node voltage data of the distribution network and the load data of the flexible device are acquired in real time in step S1, data collection is performed at preset time intervals. In the embodiment of the present invention, the sampling period is 5 minutes.

[0080] As a preferred implementation, step S2, calculating the voltage distribution characteristics of the distribution network according to the node voltage data, obtaining a voltage deviation area based on the voltage distribution characteristics, and using the voltage deviation area as a candidate access point for flexible device access, includes:

[0081] According to the geographical locations of the nodes of the distribution network, spatial classification is performed on the node voltage data to obtain a distribution network operation status data set;

[0082] According to a preset number of iterations and a preset convergence accuracy, a Newton-Raphson iterative operation is performed on the distribution network operation status data set to obtain a node voltage data matrix;

[0083] Using a density-based spatial clustering algorithm to classify the node voltage data matrix, and performing spatial mapping on the classified node voltage data matrix according to a clustering radius to obtain a voltage distribution characteristic structure diagram;

[0084] According to the voltage distribution characteristic structure diagram, a plurality of voltage characteristic regions are obtained; the voltage characteristic regions include a voltage deviation region and a voltage normal region;

[0085] According to the electrical distance between the node in the voltage deviation area and the node in the normal voltage area, a candidate access point for the flexible device to access is obtained from the voltage deviation area.

[0086] In the above scheme, more specifically, the node voltage data is collected by the distribution network data collection device at a fixed sampling period of 5 minutes, the collected data is spatially classified and the initial data set is established according to the geographical location information of the data collection device, the collected voltage values ​​are normalized, and all normalized node voltage data are added to the time series database according to the data collection time to obtain the distribution network operation status data set. The distribution network operation status data set is read from the time series database and converted into the node injection power standard format, and the Newton-Raphson iteration is used to calculate the voltage amplitude and phase angle of all nodes in the distribution network. The iteration convergence accuracy is set to 0.0001. When the number of iterations exceeds 20 and still does not converge, the calculation step size is automatically adjusted, and the node voltage data matrix is ​​constructed according to the iterative calculation results. The voltage amplitude difference and phase angle difference between each node are calculated according to the node voltage data matrix, and the node voltage characteristics are classified based on the density-based spatial clustering algorithm. The clustering radius is set to 0.1 times the minimum electrical distance between nodes, and the clustering results are spatially mapped to obtain the voltage distribution feature structure diagram. The areas where the node voltage amplitude deviates from the rated voltage value by more than plus or minus 5% are extracted from the voltage distribution characteristic structure diagram. The voltage deviation correction priority is determined according to the actual load rate of the nodes in each area. The shortest electrical path length from each node to the node in the normal voltage area is calculated based on the shortest electrical distance path. The flexible device access point sequence is generated by sorting the path length from small to large. The flexible device access point sequence is used as the candidate access point. The point ranked higher in the sequence is considered to be more suitable as an access point.

[0087] Exemplarily, the distribution network data acquisition device obtains the AC voltage waveform through a voltage transformer, calculates the effective value of the voltage waveform according to a 5-minute sampling period, and when the voltage amplitude is 10 kV rated voltage, the sampled effective value is 9.8 kV. After normalization, a normalized voltage value of 0.98 is obtained. It is recorded that the sampling point is located at bus No. 2 of the substation in area A of the distribution network, and the sampling time is 13:25. The normalized voltage data is added to the corresponding data table in the time series database. The standard format of node injection power includes node number, injected active power, injected reactive power, and node voltage initial value. The node number corresponding to bus No. 2 of substation in area A of distribution network is N102, the injected active power is 2.5 MW, the injected reactive power is 0.8 MW, and the node voltage initial value is 0.98. The Newton-Raphson iterative calculation is based on the initial voltage value and the correction value is solved by the Jacobian matrix. The voltage correction value obtained in the first iteration is 0.005, and the voltage is updated to 0.985. When the calculation error reaches 0.0001 in the 12th iteration, the power flow calculation of node N102 is completed. The node voltage data matrix records the voltage amplitude and phase angle of each node. The voltage amplitude of node N102 is 0.985, the phase angle is -2 degrees, and the voltage amplitude of adjacent node N103 is 0.975, the phase angle is -2.5 degrees. The calculated amplitude difference is 0.01, and the phase angle difference is 0.5 degrees. Based on density space clustering, when the cluster radius is set to 0.5 kilometers, nodes N102 and N103 are divided into the same voltage characteristic area because of their similar voltage characteristics and spatial distance less than the cluster radius. The voltage distribution characteristic structure diagram shows that the voltage deviation of multiple nodes in the area where node N102 is located exceeds plus or minus 5%. The load rate of transformer T1 in this area is 85%, and the load rate of transformer T2 is 90%. The high load rate leads to an increase in the priority of voltage deviation correction. The shortest electrical path length from node N102 to the nearest voltage normal node N105 is 2.5 kilometers. The path passes through nodes N103 and N104 in turn. Nodes with an electrical distance of more than 1 kilometer on this path are selected as candidate access points for flexible devices. Feedback from maintenance of the artificial distribution line shows that the line equipment between node N102 and node N105 is in good condition, the insulation resistance value is within the normal range, the secondary side sampling data of the current transformer corresponds to the primary side current phase, and the voltage transformer sampling data is consistent with the measured bus voltage, indicating that the line equipment is operating normally, the electrical metering data is true and valid, and the calculation results of the shortest electrical path between nodes can be used to guide the site selection and layout of flexible equipment.

[0088] As a preferred implementation manner, step S3, according to the candidate access point, changing the topology of the distribution network, calculating the voltage stability of the candidate access point under each of the topologies, and selecting a first target access point from the candidate access point according to the voltage stability, includes:

[0089] For each of the candidate access points, a plurality of switch combination sequences are generated based on a depth-first search; the switch combination sequences include the switch states of each node in the line;

[0090] According to the plurality of switch combination sequences, the topology structure of the distribution network is changed to generate a plurality of topology reconstruction scenarios;

[0091] Obtaining voltage stability according to a maximum voltage deviation value and a voltage fluctuation recovery time of the candidate access point in the topology reconstruction scenario;

[0092] sorting the candidate access points according to the voltage stability and a preset voltage deviation threshold;

[0093] According to the ranking result of the candidate access points, a first target access point is obtained by screening from the candidate access points.

[0094] In some preferred embodiments, a plurality of switch combination sequences are recorded through a graph database, and a topology reconstruction scenario is generated based on historical time series data. Each topology reconstruction scenario may correspond to a typical power consumption scenario or a typical power consumption period scenario.

[0095] Further, preferably, obtaining voltage stability according to the maximum voltage deviation value and voltage fluctuation recovery time of the candidate access point in the topology reconstruction scenario includes:

[0096] Calculating, according to the node voltage data, a voltage fluctuation value caused by topology reconstruction in the topology reconstruction scenario;

[0097] Obtaining the admittance of each node of the distribution network to obtain a node admittance matrix;

[0098] Inverting the node admittance matrix to obtain voltage fluctuation transmission decay characteristics of the candidate access point in the topology reconstruction scenario;

[0099] According to the voltage fluctuation transmission decay characteristics and the voltage fluctuation value, an exponentially weighted moving average method is used to solve the maximum voltage deviation and the recovery time;

[0100] The maximum voltage deviation and the recovery time are multiplied to obtain voltage stability.

[0101] It should be noted that the node impedance matrix can be obtained by inverting the node admittance matrix, and the node impedance matrix can be used to analyze the transmission attenuation characteristics of voltage fluctuations on the line in various topology reconstruction scenarios.

[0102] Preferably, the sorting of the candidate access points according to the voltage stability and a preset voltage deviation threshold value includes:

[0103] Calculate the frequency at which the voltage stability of the candidate access point in all the topology reconstruction scenarios is less than a preset voltage deviation threshold;

[0104] The candidate access points are sorted according to the frequencies.

[0105] A specific implementation method of the above scheme is: extract the line switch state identification according to the distribution network voltage distribution characteristic structure diagram, use the graph database to record the switch combination mode, generate different switch combination sequences based on depth-first search for each candidate access point of the flexible device, determine the network reconstruction sub-scheme through the 13-speed parallel transformer tap position adjustment and the tie switch switching combination, and set the ratio change rate of each tap to 0.0125, and generate a topology reconstruction scenario matrix in combination with the candidate access point and the switch combination. According to the four seasons of spring, summer, autumn and winter, combined with the characteristics of the daily load peak, valley and flat period, read the distribution network load data of the corresponding time period from the time series database to construct the seasonal-period combined load matrix, use Newton-Raphson iteration to calculate the node voltage amplitude and phase angle data in each scenario in the topology reconstruction scenario matrix, and set the iterative convergence accuracy to 0.0001 to obtain the node state data matrix of each scenario combination. According to the node status data matrix, the line flow direction changes and line capacity margin caused by network reconstruction are calculated. The node impedance matrix is ​​obtained by inverting the node admittance matrix. The impedance matrix is ​​used to analyze the transmission attenuation characteristics of voltage fluctuations on the line under various scenarios. The exponentially weighted moving average is used to solve the cumulative impact value of voltage fluctuations and construct a voltage stability change data set. The product of the maximum voltage deviation and the recovery time is extracted from the voltage stability change data set as the voltage stability index. The voltage stability index value of each candidate access point of the flexible device is calculated for each season-time scenario combination. The access points are sorted based on the index value, and the voltage deviation threshold value is set to 0.05. The frequency of access points with voltage deviations less than the threshold value in all scenario combinations is counted to obtain the location of the first target access point.

[0106] For example, in the 10 kV busbar circuit breaker switch group in area A of the UHV distribution network, circuit breakers CB01, CB02, and CB03 are connected in a ring, and the tie switch CS01 connects CB01 and CB02. The depth-first search starts from CB01 and traverses to obtain the switch combination number 001, which includes a combination configuration of CB01 closed, CB02 disconnected, CB03 closed, and CS01 closed. The parallel transformer T1 is set to the 7th position, and the corresponding strain ratio is 1.0375. Node N102 is used as a candidate access point for the flexible device, constituting a topology reconstruction sub-scenario number S001. During the summer peak period of the four seasons, the load data shows that the load rate of the feeder where the N102 node is located reaches 92%, and the load rate of the adjacent feeder is 88%. The Newton-Raphson iteration calculates the load flow of the scenario S001. The 8th iteration error reaches 0.0001, and the voltage of the N102 node is 9.6 kV, the phase angle is -3.5 degrees, and the voltage of the adjacent node N103 is 9.55 kV, the phase angle is -3.8 degrees, which are recorded in the node status data table during the summer peak period. During the winter trough period, the feeder load rate drops to 45%, and the power flow direction of the line corresponding to the switch combination number 001 changes. The line that originally supplied power from transformer T1 to N102 is changed to reverse power to T1. The impedance from node N102 to the busbar is 0.328 ohms after inversion of the admittance matrix. The attenuation coefficient of voltage fluctuation during transmission is 0.92. The exponentially weighted calculation of the cumulative impact value of the voltage fluctuation of the N102 node is 0.023. During the spring flat period, the maximum voltage deviation of the N102 node was 0.04 per unit value. It took 0.8 seconds from the start of fluctuation to recovery to the ±5% range. The calculated voltage stability index value was 0.032. In this scenario, the N102 access point ranked third. Statistics of all scenario combinations showed that the voltage deviation of the N102 node was less than 0.05 in 75% of the scenarios, and the N105 node met the requirements in 82% of the scenarios. Node N105 was finally determined as the first target access point.

[0107] The actual operation data shows that when the peak electricity consumption in summer comes, by adjusting the gear of the parallel transformer at the N105 node to gear 8 and closing the interconnection switch CS01 at the same time, the voltage at the N102 node rises to 9.85 kV after the load is transferred, meeting the voltage quality requirements, verifying the rationality of selecting N105 as the first target access point for flexible equipment.

[0108] As a preferred implementation, step S4, calculating the margin improvement value when the flexible device is connected to the first target access point according to the node voltage data and the load data, includes:

[0109] A support vector regressor is used to calculate a load characteristic value for the load data, and a plurality of load scenarios are obtained according to the load characteristic value;

[0110] Performing four-layer wavelet decomposition on the node voltage data, and obtaining a voltage stability margin reference value corresponding to each load scenario according to the decomposition result and the load scenario;

[0111] According to a preset voltage amplitude deviation threshold and a preset recovery time threshold, performing minimum variance recursion on the node voltage data to obtain an actual value of the voltage stability margin of each node;

[0112] According to the plurality of load scenarios, a difference between the actual voltage stability margin value and the voltage stability margin reference value is calculated to obtain a margin improvement value when the flexible device is connected to the first target access point.

[0113] Further, preferably, the performing of four-layer wavelet decomposition on the node voltage data and obtaining a voltage stability margin reference value corresponding to each load scenario according to the decomposition result and the load scenario includes:

[0114] Performing four-layer wavelet decomposition on the node voltage data to obtain an initial value of the voltage amplitude and a voltage amplitude change value;

[0115] Calculating the maximum deviation value and voltage recovery time of each load scenario according to the voltage amplitude change value and the voltage amplitude initial value;

[0116] A product operation is performed on the maximum deviation value and the voltage recovery time of each load scenario to obtain a voltage stability margin reference value corresponding to each load scenario.

[0117] In some preferred embodiments, the voltage waveform data of the flexible device access node is obtained from the node voltage acquisition device, and the voltage amplitude and phase angle change characteristics are extracted by using a 4-layer wavelet decomposition. The voltage stability margin reference value is generated according to the maximum deviation of the amplitude from the initial value and the time product of the recovery to the stable interval, and the reference value and the peak period, flat period, and valley period load scene combination in the four seasons of spring, summer, autumn and winter are stored in the stability margin data table. The load operation data of each season is collected by the load measurement device, the load is normalized based on the monthly maximum load capacity, the support vector regressor of the Gaussian kernel function is used to fit the load curve, the load change slope is extracted from the fitting curve, and the load scene combination set of the peak period, flat period, and valley period in the four seasons of spring, summer, autumn and winter is obtained according to the hourly load traversal, and the load characteristic value under each scene combination is calculated. According to the commissioning records of flexible equipment, the node voltage data is obtained. Based on the preset voltage amplitude deviation threshold and recovery time threshold, the measured value of the voltage stability margin of the node is calculated using the recursive minimum variance. The initial parameters of the recursive calculation are determined by the average of the last five sampling cycles, and the measured value of the stability margin is classified according to the characteristic value of the load scenario. The voltage stability margin reference value of the corresponding scenario combination is read from the stability margin data table, and the difference sequence between the measured value of the stability margin and the reference value is calculated. The difference sequence is normalized to the interval of 0 to 1 based on the real-time data processor to obtain the margin improvement degree. If the normalized value is greater than the preset margin improvement threshold, it is determined that the access of the flexible equipment under this scenario combination significantly improves the voltage stability, and the output voltage improvement degree is quantified.

[0118] Exemplarily, the voltage acquisition device at the flexible device access point records a 10 kV AC voltage waveform with a sampling frequency of 10 kHz. The 4-layer wavelet decomposition extracts a voltage amplitude fluctuation between 10.2 kV and 9.8 kV, and the phase angle deviation is within the range of plus or minus 2 degrees. The voltage drops from the rated value of 10 kV to 9.8 kV and the recovery time is 0.5 seconds. The voltage stability margin reference value is calculated to be 0.1, and this value is recorded for the summer peak period from 13:00 to 15:00. The load measurement device shows that the maximum load capacity in July in summer is 4.5 MW, and the load value measured at 13:00 is 4.2 MW. After normalization, the load rate is 0.93. The support vector regressor fits the load curve within one hour to obtain an upward slope of 0.05 MW per hour. After traversing the 24-hour load data, it is found that 13:00 to 15:00 is a typical peak period, and the load rate generally exceeds 0.9. 9:00 to 11:00 is a secondary peak period, and the load rate is between 0.75 and 0.85, forming a summer load characteristic data set. After the flexible equipment was put into operation, the node voltage data showed that when the voltage amplitude dropped to 9.85 kV, it recovered to 9.95 kV after 0.3 seconds. The voltage deviation threshold was set to plus or minus 2% of the rated voltage, and the recovery time threshold was 0.8 seconds. The average stability margin of the last five sampling cycles was 0.06. The recursive calculation showed that the current stability margin measured value was 0.045, which was classified into the summer peak data group. The data processor compared the stability margin baseline value of 0.1 with the measured value of 0.045 in the summer peak scenario, and obtained an improvement difference of 0.055. The normalized calculation obtained an index value of 0.85, which exceeded the significance judgment threshold of 0.8, proving that the flexible equipment has a significant effect in improving voltage stability in this scenario. Similar statistics show that the load changes in spring and autumn are relatively gentle, and the stability margin improvement index value is between 0.6 and 0.7. The heating load in winter causes the voltage fluctuation to increase, and the improvement index value reaches 0.83. Monthly operation statistics show that in the peak load scenario, the voltage fluctuation amplitude is reduced by about 20% and the recovery time is shortened by about 40% after the flexible equipment is put into operation. The stability margin improvement effect increases with the increase of load factor. When the load factor exceeds 0.85, the improvement index values ​​exceed the threshold value of 0.8, indicating that the evaluation index has good adaptability to load change characteristics.

[0119] In some preferred embodiments, step S6, if the margin improvement value is not greater than a preset margin improvement threshold, calculating the sensitivity of each node voltage in the distribution network to the injected power and position of the candidate access point, screening a second target access point from the candidate access points according to the sensitivity, and connecting the flexible device to the second target access point, includes:

[0120] If the margin improvement value is not greater than a preset margin improvement threshold, simulate the flexible devices accessing the distribution network through the candidate access points one by one to obtain several access scenarios;

[0121] In each access scenario, according to the injection power data of the flexible device, a preset disturbance signal is applied to the distribution network node, and the sensitivity of the node voltage data to the injection power is obtained by using a gradient descent algorithm;

[0122] According to the sensitivity of the injected power, an electrical distance between the flexible device and a node of the distribution network is obtained;

[0123] According to the electrical distance between nodes, a node topology correlation feature matrix is ​​constructed through a graph neural network to obtain the node position sensitivity;

[0124] Performing a weighted combination on the injection power sensitivity and the node position sensitivity to obtain a comprehensive sensitivity;

[0125] According to the comprehensive sensitivity, a second target access point is obtained by screening from the candidate access points, and the flexible device is connected to the second target access point.

[0126] It should be noted that, in the embodiment of the present invention, after the first target access point is determined, the margin improvement is calculated first, that is, the voltage stability improvement is analyzed, and when the margin improvement is not greater than the preset margin improvement threshold, the sensitivity analysis is performed and the second target access point is determined. In practical applications, after the second target access point is determined, the voltage stability corresponding to the access point location and parameters can be analyzed again to select the access point at its best and obtain a distribution network with higher reliability.

[0127] In some preferred embodiments, the voltage amplitude of each node and the power injection data of the flexible device are obtained from the distribution network state database, a small disturbance signal with an amplitude of 0.5% of the rated value is applied to the injection power, the difference ratio of the node voltage and the injection power is calculated, and the sensitivity of the voltage to the power is solved by gradient descent with a learning rate of 0.01 and 50 iterations. The power transfer function between nodes is constructed based on the node admittance matrix to obtain the sensitivity data table of the influence of the power injection point on the voltage. According to the distribution network flow calculation record, the electrical distance between the current position of the flexible device and the distribution network node is read, and the node topology association feature is constructed using a three-layer graph neural network. The node feature includes three dimensions: voltage amplitude, phase angle, and power factor. The node is sorted by the response metric value of the node voltage to the power injection of the flexible device, and the position sensitivity of each node is calculated weighted according to the electrical distance to construct a position sensitivity data matrix. The preset voltage sensitivity threshold value is extracted from the distribution network status database, and a comprehensive sensitivity index is formed based on the power impact sensitivity and location sensitivity according to the weights of 0.6 and 0.4. The comprehensive sensitivity index of all nodes is compared with the preset threshold value, and the node that meets the threshold value requirements and has the highest comprehensive sensitivity is selected as the second target access point, and the voltage margin benchmark value of the access point is recorded.

[0128] For example, a 500 kVA flexible device is connected to the 10 kV busbar of the distribution network, and the injected power increases by 2.5 kWh. The voltage of the adjacent node N102 rises by 0.05 kV, the voltage of N103 rises by 0.03 kV, and the voltage of N104 rises by 0.01 kV. The calculated sensitivity of N102 to the injected power is 0.02, the sensitivity of N103 is 0.012, and the sensitivity of N104 is 0.004. The node admittance matrix shows that the power transfer attenuation coefficient is 0.6. The first layer of the graph neural network processes the basic characteristics of the nodes. The voltage of the N102 node is 0.985 per unit, the phase angle is -2 degrees, and the power factor is 0.92. The second layer integrates the information of adjacent nodes. The electrical distance between N102 and N103 is 1.2 kilometers, and the power transmission loss is 0.015. The third layer outputs the node correlation degree of 0.82, and the position sensitivity quantization value is 0.78, which is higher than the preset threshold value of 0.75. In the comprehensive sensitivity index, the power impact sensitivity of 0.02 is weighted as 0.6, and the position sensitivity of 0.78 is weighted as 0.4. The comprehensive sensitivity of the N102 node is calculated to be 0.332, the comprehensive sensitivity of the N103 node is 0.285, and the comprehensive sensitivity of the N104 node is 0.186. N102 is selected as the second target access point, and the voltage margin reference value of this point is recorded as 0.15. The recursive least squares method is used to process the sampled data in the 15-minute data window of the N102 node. The original voltage mean is 9.85 kV and the standard deviation is 0.12 kV. After the flexible equipment is connected, the voltage mean is increased to 9.92 kV, and the standard deviation is reduced to 0.08 kV. The improvement amplitude is 0.07 kV, which is higher than the threshold of 0.02 kV. The recovery time is 0.3 seconds, which is less than the threshold of 0.5 seconds, and the original access point position is maintained. Load monitoring data shows that in the daily operation of the distribution network, the voltage fluctuations caused by the start and stop of heavy-load equipment are most obvious at the N102 node, which is consistent with the results of the sensitivity analysis. When the industrial zone load starts, the voltage of the N102 node drops by 0.3 kV. After compensation by flexible equipment, the drop value is reduced to 0.15 kV, which verifies the accuracy of the sensitivity-based site selection. During operation, monitoring found that the change in network structure caused by the change in the position of the load center will cause the node sensitivity to change, triggering a re-sensitivity analysis and updating the access point selection. Multiple actual cases have shown that the flexible equipment access points determined based on sensitivity analysis can produce significant improvements when voltage anomalies occur. Statistical data show that at the selected access point, the voltage fluctuation amplitude is reduced by an average of 45%, the recovery time is shortened by 55%, and the surrounding node voltage is also significantly improved.

[0129] As a preferred implementation, the method for accessing the flexible device further includes the following steps:

[0130] When multiple flexible devices are connected in parallel, the flexible devices are layered according to the bus voltage levels of the flexible devices to determine the response priorities of the flexible devices;

[0131] The reactive power of the flexible device is adaptively adjusted according to the load data of the flexible device and the response priority.

[0132] It should be noted that when multiple flexible devices are connected in parallel, the control objectives and constraints of different access points are different, making it difficult to achieve distributed autonomous regulation. In the embodiment of the present invention, the power of the flexible device can be adaptively adjusted to maintain the dynamic balance of the distribution network voltage.

[0133] In some preferred embodiments, the active power and reactive power sampling values ​​of the load are obtained from the data acquisition unit at the flexible device access point, a time series data table is constructed according to a 15-minute sampling period, and the time series data is predicted using a long short-term memory neural network. The load change curve is generated by comparing the difference between the load prediction value and the measured value at adjacent moments, and a load characteristic vector is constructed according to the slope and fluctuation amplitude of the load change curve to obtain a characteristic index reflecting the load change trend. Calculate the reactive power adjustment amount of the flexible device according to the load characteristic vector ;in, is the adjustment coefficient calculated according to the load characteristic vector; is the maximum margin for reactive power regulation; It is the minimum margin for reactive power regulation.

[0134] Based on the ratio of the measured load value to the rated capacity, the load level is divided into 20% intervals. The maximum reactive power regulation margin corresponding to each load level is calculated through the equipment nameplate parameter curve. The load characteristic vector is smoothed by an exponential smoothing filter and mapped to the corresponding load level interval. The flexible equipment operation point is determined within the reactive power regulation margin range. The control level is divided according to the bus voltage level where the flexible equipment is located. The response priority of the flexible equipment at the same level is sorted based on the bus voltage amplitude. The 100 millisecond control time interval between devices is determined by the voltage change rate. The control instructions and operating status are transmitted between devices using the industrial bus protocol. The equipment action sequence table is constructed from the equipment response priority and control time interval. According to the equipment action sequence table, the real-time control instructions of the flexible equipment are generated, and the equipment operating parameters are obtained from the field bus. When the deviation between the operating parameters and the operating point exceeds 0.5%, the online correction is triggered. The operating parameters are corrected by a finite memory length Kalman filter. The real-time control instructions are updated according to the corrected operating parameters to achieve dynamic adjustment of the flexible equipment operating parameters.

[0135] Exemplarily, the data acquisition unit of the flexible equipment access point on the 10 kV distribution line records load data in real time. The sampling in the interval from 13:00 to 13:15 shows that the active power increases from 2.8 MW to 3.2 MW, and the reactive power increases from 0.9 MW to 1.1 MW. The long short-term memory neural network uses the data of the past four sampling cycles as input and predicts that the load active power will reach 3.5 MW at 13:30. The actual measured value is 3.45 MW. The calculated slope of the load change curve is 0.047 MW per minute, and the fluctuation amplitude is 0.65 MW. The rated capacity of the flexible equipment is 5 MVA. The current measured load value is 3.45 MW, corresponding to a load rate of 69%, which belongs to the load level range of 60%-80%. The maximum reactive power regulation margin under this load level is 1.8 MW, and the load characteristic vector after exponential smoothing filtering shows that the load continues to rise. The mapping result requires an increase in reactive power output. The reactive power output increment is determined to be 1.62 MW according to 0.9 times the maximum regulation margin. The voltage levels of bus No. 1 and bus No. 2 are both 10 kV, and they are divided into the same control level. The voltage of bus No. 1 is 9.8 kV, with a change rate of -0.02 kV / min, and the voltage of bus No. 2 is 9.85 kV, with a change rate of -0.015 kV / min. Based on the voltage amplitude, the priority response of bus No. 1 flexible equipment is determined, and the response interval is set to 100 milliseconds before triggering bus No. 2 flexible equipment, and the control instruction sequence is transmitted through the industrial bus. The fieldbus feeds back the operating parameters of the flexible equipment on bus No. 1. The reactive output is 1.58 MN, which is 2.5% different from the command value of 1.62 MN, exceeding the 0.5% threshold, triggering the Kalman filter correction. The calculated parameter correction is 0.03 MN, and the updated reactive output command value is adjusted to 1.59 MN, realizing dynamic adjustment of the operating parameters. The operating data shows that after 5 minutes of coordinated control, the voltage of bus No. 1 rebounded to 9.92 kV, the voltage of bus No. 2 rebounded to 9.95 kV, and the load voltage returned to the qualified range. The daily load curve of the distribution line shows that there are load peaks from 9:00 to 11:00 and 14:00 to 16:00 on weekdays, and the number of reactive regulation responses of flexible equipment increases accordingly. Monthly statistics show that the proportion of regulation responses during peak hours reaches 65%, which is consistent with the actual power consumption pattern, confirming the adaptability of the coordinated control mechanism based on load prediction to load changes.

[0136] Preferably, in an embodiment of the present invention, simulation calculations and actual operation tests are continuously performed to evaluate the voltage stability of flexible devices when connected to the distribution network in different scenarios, and monitor the sensitivity of node voltage to the device access location and operating parameters such as injection power, so as to make necessary corrections and improvements to the access strategy and control strategy.

[0137] Specifically, bilinear interpolation is performed on the node voltage data to obtain a voltage stability margin surface; a long short-term memory neural network is used to process the voltage stability margin surface, characteristic quantities of voltage drop depth and duration are obtained, and a voltage weak area location map is generated; a phasor acquisition device is used to record the node voltage amplitude of the voltage weak area, and a recursive least squares method is used to calculate the spatial sensitivity of the node voltage to the equipment position; a reactive power adjustment coefficient is generated according to the spatial sensitivity and the voltage weak area location map, and the control parameters of the flexible equipment are adjusted.

[0138] A flexible device access method provided by an embodiment of the present invention is adopted. By introducing a margin improvement value and a sensitivity analysis method, a target access point is determined, and node voltage data and load data are collected in real time. The impact of flexible device access on the distribution network under different power grid topologies can be considered, and more optimal flexible device access parameters can be provided to ensure voltage stability and high reliability of the distribution network.

[0139] The embodiment of the present invention provides a flexible device access system. Figure 2 The access system of the flexible device includes a data acquisition module 11, a candidate access point selection module 12, a first target access point screening module 13, a margin improvement value calculation module 14, a first access module 15 and a second access module 16, wherein:

[0140] The data acquisition module 11 is used to acquire node voltage data of the distribution network and load data of the flexible device in real time;

[0141] A candidate access point selection module 12 is used to calculate the voltage distribution characteristics of the distribution network according to the node voltage data, obtain a voltage deviation area based on the voltage distribution characteristics, and use the voltage deviation area as a candidate access point for flexible device access;

[0142] A first target access point screening module 13 is used to change the topology of the distribution network according to the candidate access points, calculate the voltage stability of the candidate access points under each of the topologies, and screen the first target access point from the candidate access points according to the voltage stability;

[0143] A margin improvement value calculation module 14, configured to calculate a margin improvement value when the flexible device is connected to the first target access point according to the node voltage data and the load data;

[0144] A first access module 15, configured to connect the flexible device to the first target access point if the margin improvement value is greater than a preset margin improvement threshold;

[0145] The second access module 16 is used to calculate the sensitivity of each node voltage in the distribution network to the candidate access point and the injected power if the margin improvement value is not greater than a preset margin improvement threshold, screen a second target access point from the candidate access points according to the sensitivity, and connect the flexible device to the second target access point.

[0146] As a preferred implementation, the candidate access point selection module 12 includes:

[0147] A spatial classification unit, used to perform spatial classification on the node voltage data according to the geographical location of the node of the distribution network, so as to obtain a distribution network operation status data set;

[0148] An iterative operation unit, used for performing Newton-Raphson iterative operation on the distribution network operation status data set according to a preset number of iterations and a preset convergence accuracy to obtain a node voltage data matrix;

[0149] A spatial mapping unit, used to classify the node voltage data matrix using a density-based spatial clustering algorithm, and perform spatial mapping on the classified node voltage data matrix according to a clustering radius to obtain a voltage distribution feature structure diagram;

[0150] A region division unit, used for obtaining a plurality of voltage characteristic regions according to the voltage distribution characteristic structure diagram; the voltage characteristic regions include a voltage deviation region and a voltage normal region;

[0151] The candidate access point selection unit is used to obtain a candidate access point for the flexible device to access from the voltage deviation area according to the electrical distance between the node in the voltage deviation area and the node in the normal voltage area.

[0152] As a preferred implementation, the first target access point screening module 13 includes:

[0153] A switch combination sequence generating unit, configured to generate a plurality of switch combination sequences based on a depth-first search for each of the candidate access points; the switch combination sequence includes a switch state of each node in the line;

[0154] A topology reconstruction scenario generating unit, used to change the topology structure of the distribution network according to the plurality of switch combination sequences, and generate a plurality of topology reconstruction scenarios;

[0155] A voltage stability calculation unit, configured to obtain voltage stability according to a maximum voltage deviation value and a voltage fluctuation recovery time of the candidate access point in the topology reconstruction scenario;

[0156] A candidate access point sorting unit, configured to sort the candidate access points according to the voltage stability and a preset voltage deviation threshold;

[0157] The first target access point screening unit is configured to screen the first target access point from the candidate access points according to the ranking result of the candidate access points.

[0158] Further, preferably, the voltage stability calculation unit is specifically used for:

[0159] Calculating, according to the node voltage data, a voltage fluctuation value caused by topology reconstruction in the topology reconstruction scenario;

[0160] Obtaining the admittance of each node of the distribution network to obtain a node admittance matrix;

[0161] Inverting the node admittance matrix to obtain voltage fluctuation transmission decay characteristics of the candidate access point in the topology reconstruction scenario;

[0162] According to the voltage fluctuation transmission decay characteristics and the voltage fluctuation value, an exponentially weighted moving average method is used to solve the maximum voltage deviation and the recovery time;

[0163] The maximum voltage deviation and the recovery time are multiplied to obtain voltage stability.

[0164] Preferably, the candidate access point ranking unit is specifically configured to:

[0165] Calculate the frequency at which the voltage stability of the candidate access point in all the topology reconstruction scenarios is less than a preset voltage deviation threshold;

[0166] The candidate access points are sorted according to the frequencies.

[0167] As a preferred implementation, the margin improvement value calculation module 14 includes:

[0168] A load scenario generating unit, configured to calculate a load characteristic value for the load data using a support vector regressor, and obtain a plurality of load scenarios according to the load characteristic value;

[0169] A voltage stability margin reference value calculation unit, used for performing four-layer wavelet decomposition on the node voltage data, and obtaining a voltage stability margin reference value corresponding to each load scenario according to the decomposition result and the load scenario;

[0170] A voltage stability margin actual value calculation unit, used to perform minimum variance recursion on the node voltage data according to a preset voltage amplitude deviation threshold and a preset recovery time threshold, to obtain the voltage stability margin actual value of each node;

[0171] The margin improvement value calculation unit is used to calculate the difference between the voltage stability margin actual value and the voltage stability margin reference value according to the plurality of load scenarios, and obtain the margin improvement value when the flexible device is connected to the first target access point.

[0172] Further, preferably, the voltage stability margin reference value calculation unit is specifically used to:

[0173] Performing four-layer wavelet decomposition on the node voltage data to obtain an initial value of the voltage amplitude and a voltage amplitude change value;

[0174] Calculating the maximum deviation value and voltage recovery time of each load scenario according to the voltage amplitude change value and the voltage amplitude initial value;

[0175] A product operation is performed on the maximum deviation value and the voltage recovery time of each load scenario to obtain a voltage stability margin reference value corresponding to each load scenario.

[0176] As a preferred implementation, the second access module 16 includes:

[0177] an access scenario generating unit, configured to simulate the flexible devices accessing the distribution network through the candidate access points one by one to obtain a plurality of access scenarios if the margin improvement value is not greater than a preset margin improvement threshold;

[0178] A first sensitivity calculation unit is used to apply a preset disturbance signal to the distribution network node in each access scenario according to the injection power data of the flexible device, and obtain the sensitivity of the node voltage data to the injection power by using a gradient descent algorithm;

[0179] an electrical distance calculation unit, configured to obtain an electrical distance between the flexible device and a node of the distribution network according to the sensitivity of the injected power;

[0180] A second sensitivity calculation unit is used to construct a node topology association feature matrix through a graph neural network according to the electrical distance between the nodes to obtain the node position sensitivity;

[0181] A comprehensive sensitivity calculation unit, used for performing a weighted combination of the injection power sensitivity and the node position sensitivity to obtain a comprehensive sensitivity;

[0182] The second target access point screening unit is configured to screen the second target access point from the candidate access points according to the comprehensive sensitivity, and connect the flexible device to the second target access point.

[0183] As a preferred implementation, the flexible device access method further includes:

[0184] A flexible device grading module is used to stratify the flexible devices according to the bus voltage levels of the flexible devices when multiple flexible devices are connected in parallel, and determine the response priority of the flexible devices;

[0185] The flexible device adaptive adjustment module is used to adaptively adjust the reactive power of the flexible device according to the load data and the response priority of the flexible device.

[0186] A flexible device access system provided by an embodiment of the present invention determines the target access point by introducing a margin improvement value and a sensitivity analysis method, and collects node voltage data and load data in real time. It can take into account the impact of flexible device access on the distribution network under different power grid topologies, and can provide more optimal flexible device access parameters to ensure voltage stability and high reliability of the distribution network.

[0187] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0188] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for accessing a flexible device, characterized in that: include: Obtain node voltage data of the distribution network and load data of flexible equipment in real time; Calculating voltage distribution characteristics of the distribution network according to the node voltage data, obtaining a voltage deviation area based on the voltage distribution characteristics, and using the voltage deviation area as a candidate access point for flexible device access; According to the candidate access points, the topology of the distribution network is changed, the voltage stability of the candidate access points under each of the topologies is calculated, and according to the voltage stability, a first target access point is screened from the candidate access points; Calculating, according to the node voltage data and the load data, a margin improvement value when the flexible device is connected to the first target access point; If the margin improvement value is greater than a preset margin improvement threshold, connecting the flexible device to the first target access point; If the margin improvement value is not greater than a preset margin improvement threshold, the sensitivity of each node voltage in the distribution network to the candidate access point and the injected power is calculated, a second target access point is obtained from the candidate access points according to the sensitivity, and the flexible device is connected to the second target access point.

2. A method for accessing a flexible device as claimed in claim 1, characterized in that: The step of calculating the voltage distribution characteristics of the distribution network according to the node voltage data, obtaining a voltage deviation area based on the voltage distribution characteristics, and using the voltage deviation area as a candidate access point for flexible device access includes: According to the geographical locations of the nodes of the distribution network, spatial classification is performed on the node voltage data to obtain a distribution network operation status data set; According to a preset number of iterations and a preset convergence accuracy, a Newton-Raphson iterative operation is performed on the distribution network operation status data set to obtain a node voltage data matrix; Using a density-based spatial clustering algorithm to classify the node voltage data matrix, and performing spatial mapping on the classified node voltage data matrix according to a clustering radius to obtain a voltage distribution characteristic structure diagram; According to the voltage distribution characteristic structure diagram, a plurality of voltage characteristic regions are obtained; the voltage characteristic regions include a voltage deviation region and a voltage normal region; According to the electrical distance between the node in the voltage deviation area and the node in the normal voltage area, a candidate access point for the flexible device to access is obtained from the voltage deviation area.

3. A method for accessing a flexible device as claimed in claim 1, characterized in that: The step of changing the topology of the distribution network according to the candidate access points, calculating the voltage stability of the candidate access points under each of the topologies, and selecting a first target access point from the candidate access points according to the voltage stability includes: For each of the candidate access points, a plurality of switch combination sequences are generated based on a depth-first search; the switch combination sequences include the switch states of each node in the line; According to the plurality of switch combination sequences, the topology structure of the distribution network is changed to generate a plurality of topology reconstruction scenarios; Obtaining voltage stability according to a maximum voltage deviation value and a voltage fluctuation recovery time of the candidate access point in the topology reconstruction scenario; sorting the candidate access points according to the voltage stability and a preset voltage deviation threshold; According to the ranking result of the candidate access points, a first target access point is obtained by screening from the candidate access points.

4. A method for accessing a flexible device as claimed in claim 3, characterized in that: The obtaining of voltage stability according to the maximum voltage deviation value and the voltage fluctuation recovery time of the candidate access point in the topology reconstruction scenario includes: Calculating, according to the node voltage data, a voltage fluctuation value caused by topology reconstruction in the topology reconstruction scenario; Obtaining the admittance of each node of the distribution network to obtain a node admittance matrix; Inverting the node admittance matrix to obtain voltage fluctuation transmission decay characteristics of the candidate access point in the topology reconstruction scenario; According to the voltage fluctuation transmission decay characteristics and the voltage fluctuation value, an exponentially weighted moving average method is used to solve the maximum voltage deviation and the recovery time; The maximum voltage deviation and the recovery time are multiplied to obtain voltage stability.

5. A method for accessing a flexible device as claimed in claim 3, characterized in that: The sorting of the candidate access points according to the voltage stability and the preset voltage deviation threshold value includes: Calculate the frequency at which the voltage stability of the candidate access point in all the topology reconstruction scenarios is less than a preset voltage deviation threshold; The candidate access points are sorted according to the frequencies.

6. A method for accessing a flexible device as claimed in claim 1, characterized in that: The calculating, according to the node voltage data and the load data, a margin improvement value when the flexible device is connected to the first target access point includes: A support vector regressor is used to calculate a load characteristic value for the load data, and a plurality of load scenarios are obtained according to the load characteristic value; Performing four-layer wavelet decomposition on the node voltage data, and obtaining a voltage stability margin reference value corresponding to each load scenario according to the decomposition result and the load scenario; According to a preset voltage amplitude deviation threshold and a preset recovery time threshold, performing minimum variance recursion on the node voltage data to obtain an actual value of the voltage stability margin of each node; According to the plurality of load scenarios, a difference between the actual voltage stability margin value and the voltage stability margin reference value is calculated to obtain a margin improvement value when the flexible device is connected to the first target access point.

7. A method for accessing a flexible device as claimed in claim 6, characterized in that: The performing four-layer wavelet decomposition on the node voltage data and obtaining a voltage stability margin reference value corresponding to each load scenario according to the decomposition result and the load scenario includes: Performing four-layer wavelet decomposition on the node voltage data to obtain an initial value of the voltage amplitude and a voltage amplitude change value; Calculating the maximum deviation value and voltage recovery time of each load scenario according to the voltage amplitude change value and the voltage amplitude initial value; A product operation is performed on the maximum deviation value and the voltage recovery time of each load scenario to obtain a voltage stability margin reference value corresponding to each load scenario.

8. A method for accessing a flexible device as claimed in claim 1, characterized in that: If the margin improvement value is not greater than a preset margin improvement threshold, calculating the sensitivity of each node voltage in the distribution network to the injected power and position of the candidate access point, screening a second target access point from the candidate access points according to the sensitivity, and connecting the flexible device to the second target access point, including: If the margin improvement value is not greater than a preset margin improvement threshold, simulate the flexible devices accessing the distribution network through the candidate access points one by one to obtain several access scenarios; In each access scenario, according to the injection power data of the flexible device, a preset disturbance signal is applied to the distribution network node, and the sensitivity of the node voltage data to the injection power is obtained by using a gradient descent algorithm; According to the sensitivity of the injected power, an electrical distance between the flexible device and a node of the distribution network is obtained; According to the electrical distance between nodes, a node topology correlation feature matrix is ​​constructed through a graph neural network to obtain the node position sensitivity; Performing a weighted combination on the injection power sensitivity and the node position sensitivity to obtain a comprehensive sensitivity; According to the comprehensive sensitivity, a second target access point is obtained by screening from the candidate access points, and the flexible device is connected to the second target access point.

9. A method for accessing a flexible device as claimed in claim 1, characterized in that: If the margin improvement value is not greater than a preset margin improvement threshold, the sensitivity of the voltage of each node in the distribution network to the injected power and the position of the candidate access point is calculated, a second target access point is obtained from the candidate access points according to the sensitivity, and the flexible device is connected to the second target access point. The access method of the flexible device also includes: When multiple flexible devices are connected in parallel, the flexible devices are layered according to the bus voltage levels of the flexible devices to determine the response priorities of the flexible devices; The reactive power of the flexible device is adaptively adjusted according to the load data of the flexible device and the response priority.

10. A flexible device access system, characterized in that: include: A data acquisition module is used to obtain node voltage data of the distribution network and load data of flexible devices in real time; A candidate access point selection module, used to calculate the voltage distribution characteristics of the distribution network according to the node voltage data, obtain a voltage deviation area based on the voltage distribution characteristics, and use the voltage deviation area as a candidate access point for flexible device access; A first target access point screening module, configured to change the topology of the distribution network according to the candidate access points, calculate the voltage stability of the candidate access points under each of the topologies, and screen the first target access point from the candidate access points according to the voltage stability; a margin improvement value calculation module, configured to calculate a margin improvement value when the flexible device is connected to the first target access point according to the node voltage data and the load data; A first access module, configured to connect the flexible device to the first target access point if the margin improvement value is greater than a preset margin improvement threshold; The second access module is used to calculate the sensitivity of each node voltage in the distribution network to the candidate access point and the injected power if the margin improvement value is not greater than a preset margin improvement threshold, screen a second target access point from the candidate access points according to the sensitivity, and connect the flexible device to the second target access point.

Images