Big data power distribution network dispatching method and system

Abstract

This invention discloses a big data distribution network scheduling method and system, belonging to the field of distribution network scheduling technology. The method includes: determining a digital twin target based on distribution network information, and establishing a corresponding digital twin based on the digital twin target; establishing a corresponding distribution network scheduling model based on each digital twin and distribution network information; obtaining the original scheduling plan, and extracting the data of each original influencing item from the original scheduling plan according to each influencing item; collecting real-time data from the distribution network scheduling model based on each influencing item to obtain the real-time influencing item data corresponding to each influencing item; processing the original influencing item data and the real-time influencing item data according to a preset matrix template to obtain the corresponding planning matrix and influence matrix; calculating the corresponding change matrix based on the planning matrix and influence matrix, and adjusting the original scheduling plan based on the change matrix to obtain the real-time scheduling plan.

Description

Technical Field

[0001] This invention belongs to the field of power distribution network dispatching technology, specifically a big data power distribution network dispatching method and system. Background Technology

[0002] To reduce the consumption of non-renewable energy, a large number of distributed generation (DG) power sources and various flexible loads are being decentralized into the distribution network. The large-scale integration of distributed renewable energy significantly impacts the system's original electrical characteristics, such as power flow distribution, voltage levels, and short-circuit capacity, making it difficult for traditional distribution networks to meet the demands of high-penetration renewable energy generation and efficient utilization in a low-carbon economy. Coordinated and optimized dispatching is a core technology and important means for distribution networks to actively manage and efficiently operate controllable resources such as distributed generation and controllable loads. The integration of these random and fluctuating resources has a significant impact on the stable operation of the power grid.

[0003] Traditional power management methods can no longer meet the operational requirements of modern power grids. To ensure the stable, safe, and efficient operation of the power system, a system capable of real-time monitoring, dispatching, and managing the distribution network is needed; based on this, the present invention provides a big data distribution network dispatching method and system. Summary of the Invention

[0004] To address the problems of the above solutions, this invention provides a big data power distribution network scheduling method and system.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A big data-driven power distribution network scheduling method, the method comprising:

[0007] Obtain distribution network information, determine the digital twin target based on the distribution network information, and establish the corresponding digital twin based on the digital twin target; establish the corresponding distribution network scheduling model based on each digital twin and the distribution network information;

[0008] Preferably, the method for determining the twin target includes:

[0009] The distribution network information is identified and matched according to the preset list of candidate targets to determine the detailed data of each candidate target;

[0010] Analyze the obtained target details data to obtain the corresponding initial values; set corresponding correction values ​​for the candidate targets;

[0011] The target evaluation value is calculated according to the formula MPG = CS + XZ; where: MPG is the target evaluation value; CS is the initial value; and XZ is the correction value.

[0012] Candidate targets whose target evaluation value is greater than the threshold X1 are marked as twin targets.

[0013] Preferably, the method for calculating the initial value includes:

[0014] Each tree-structured verification chart is set up based on the list of candidate targets. The tree-structured verification chart includes each verification item and the corresponding verification item value.

[0015] An initial evaluation model is established based on the tree-structured verification diagram and the detailed data of the candidate targets in the simulation settings;

[0016] The target details data are analyzed using the initial evaluation model to obtain the individual values; yi = Hi(x); where: yi is the individual value of the corresponding check item, i represents the corresponding check item, i = 1, 2, ..., n, and n is a positive integer; Hi(x) is the initial evaluation model;

[0017] According to the formula Calculate the corresponding initial value, where CS is the initial value and yi is the individual value of the corresponding check item.

[0018] Preferably, the expression for the initial evaluation model is: x represents the input data, indicating the details of the target to be selected; i represents the corresponding verification item, i = 1, 2, ..., n, where n is a positive integer; A indicates that the verification item requirements are met; Ci is the verification item value.

[0019] Obtain the original scheduling plan and extract the data of each original influencing item from the original scheduling plan according to the preset influencing items;

[0020] Real-time data collection is performed from the distribution network dispatch model based on each influencing factor to obtain the real-time influencing factor data corresponding to each influencing factor.

[0021] The original impact data and real-time impact data are processed according to the preset matrix template to obtain the corresponding planning matrix and impact matrix;

[0022] Calculate the corresponding change matrix based on the planning matrix and the influence matrix, and adjust the original scheduling plan based on the change matrix to obtain the real-time scheduling plan;

[0023] Preferably, the change matrix = planning matrix - influence matrix.

[0024] Preferably, the method for adjusting the original scheduling plan based on the change matrix includes:

[0025] Define a change period Δt, obtain the predicted change data within the change period Δt based on the distribution network dispatch model, and obtain the corresponding change impact data within the change period Δt based on the original dispatch plan;

[0026] Generate corresponding prediction matrices based on the data on the impact of changes and the data on predicted changes;

[0027] Identify each element value in the change matrix and mark it as the baseline element value. Extract the corresponding element value set according to the time order and the position of each baseline element value in each prediction matrix.

[0028] Generate a curve showing the change of each benchmark element value within the time period Δt based on the benchmark element value and the corresponding set of element values. The horizontal axis of the curve represents time, spanning from 0 to Δt, and the vertical axis represents the element value.

[0029] The curve function that fits each variation curve is labeled G(t);

[0030] According to the formula Calculate the corresponding element correction value;

[0031] In the formula: YA is the element correction value; Y0 is the base element value;

[0032] Replace the corresponding baseline element values ​​in the change matrix with the element correction values. After all replacements are made, mark the change matrix as the scheduling matrix.

[0033] The scheduling matrix is ​​analyzed using a pre-defined scheduling analysis model to obtain the corresponding real-time scheduling plan.

[0034] The real-time scheduling plan is simulated and verified based on the distribution network scheduling model. Once the simulation verification is successful, the corresponding real-time scheduling plan is applied.

[0035] Preferably, the method for simulating and verifying the real-time scheduling plan includes:

[0036] Simulated scheduling of real-time scheduling plans is carried out based on the digital twins in the distribution network scheduling model; the simulated scheduling process is monitored in real time to obtain corresponding monitoring data, and the obtained monitoring data is analyzed to obtain corresponding scheduling verification values;

[0037] When the scheduling verification value is not greater than the threshold X2, the simulation verification result is that the verification is passed;

[0038] When the scheduling verification value is greater than the threshold X2, the simulation verification result is a verification failure; the real-time scheduling plan is then adjusted.

[0039] Preferably, the method for analyzing the monitoring data includes:

[0040] Obtain the preset scheduling permission requirements, verify the monitoring data according to each requirement in the scheduling permission requirements, obtain the judgment result corresponding to each requirement, and set the corresponding scheduling permission value according to the obtained judgment result;

[0041] When all requirements are deemed satisfactory, the scheduling allowable value is 0; when not all requirements are deemed satisfactory, the scheduling allowable value is BN; BN is greater than the threshold X2.

[0042] An anomaly identification model is established based on historical monitoring data. The monitoring data is then analyzed using the established anomaly identification model to obtain the corresponding set of outliers.

[0043] Summing is performed on the set of outliers to obtain the corresponding cumulative value; the cumulative value is then marked as LDZ.

[0044] Then, the corresponding scheduling verification value is calculated according to the formula YUZ = LDZ + DYZ;

[0045] In the formula: YUZ is the scheduling verification value; DYZ is the scheduling allowable value.

[0046] Preferably, the outlier q = f(w);

[0047] In the formula: q represents an outlier; f(w) is the output value of the anomaly detection model;

[0048] An anomaly detection model is established based on the Isolation Forest algorithm, and the expression is: In the formula: w represents the monitoring data; f(w) represents the output data.

[0049] A big data power distribution network dispatching system includes a model module, a dispatching analysis module, and a dispatching execution module;

[0050] The model module is used to determine the twin target based on the distribution network information, establish the corresponding digital twin based on the twin target, and establish the corresponding distribution network scheduling model based on each digital twin and the distribution network information.

[0051] The scheduling analysis module is used to analyze the original scheduling plan based on the distribution network scheduling model to obtain the corresponding real-time scheduling plan.

[0052] The scheduling execution module is used to execute the real-time scheduling plan. It simulates and verifies the real-time scheduling plan based on the distribution network scheduling model. Once the simulation verification is successful, the corresponding real-time scheduling plan is applied.

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] By identifying each twin target and establishing a distribution network scheduling model based on each twin target, the system can acquire relevant data of the current distribution network in real time, facilitating scheduling adjustments and improving scheduling efficiency. Simultaneously, it can optimize the original scheduling plan. Furthermore, the obtained real-time scheduling plan can be simulated and verified using the distribution network scheduling model, ensuring the stable, safe, and efficient operation of the power system. Through real-time monitoring and scheduling of the distribution network, the system can adjust the allocation of power resources in a timely manner, avoiding excessive power load in some areas and ensuring the stable operation of the power system. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0057] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0058] like Figure 1 As shown, a big data-based power distribution network scheduling method includes:

[0059] Step SA1: Acquire distribution network information, including transmission line distribution information, power plant information, substation information, etc., connected to the distribution network; collect and summarize data using a pre-set distribution network information template; determine the digital twin targets based on the acquired distribution network information, which are the equipment, institutions, and sites that need to be digitally twinned, such as power plants and substations; establish the corresponding digital twins based on the acquired digital twin targets; and integrate the digital twins with other distribution network information to establish the corresponding distribution network dispatch model.

[0060] The methods for determining twin targets include:

[0061] Based on the distribution network information, various candidate targets are identified. These candidate targets are those that meet the requirements for establishing a digital twin, such as various power plants. Specifically, a candidate target details table can be set up according to the needs. The distribution network information is identified and matched based on the preset candidate target details table to obtain each candidate target that meets the conditions.

[0062] Detailed data for each candidate target is acquired and marked as candidate target detail data, such as the type, location, scale, and operation mode of the candidate target. Data can be collected according to a preset candidate target information template to form candidate target detail data. The acquired target detail data is analyzed to obtain the initial value corresponding to the candidate target. A corresponding correction value is set for the candidate target. The correction value is set according to the actual situation of the candidate target. If the candidate target already has a corresponding digital twin and can be retrieved and applied, the correction value is the threshold value X1. If the corresponding candidate target does not allow the establishment of a digital twin, the correction value is the opposite of the initial value. Otherwise, it is 0.

[0063] The target evaluation value is calculated based on the obtained initial and correction values. The formula for calculating the target evaluation value is: MPG = CS + XZ; where: MPG is the target evaluation value; CS is the initial value; and XZ is the correction value.

[0064] Candidate targets whose target evaluation value is greater than the threshold X1 are marked as twin targets.

[0065] The methods for calculating initial values ​​include:

[0066] Based on the candidate target details table, tree-structured verification diagrams are set up. Each tree-structured verification diagram is composed of various verification items. The first item is a type-related verification item, such as whether it is a power plant. The next level is a subdivision of the previous level, such as whether it is a thermal power plant, a wind power plant, or a solar power plant, and so on. Branches are formed based on the actual situation corresponding to the previous verification item. Each verification item is assigned a corresponding verification item value, which can be negative. These values ​​are marked at each verification item in the tree-structured verification diagram. The specific tree-structured verification diagrams and corresponding verification item values ​​are set manually. Subsequently, the tree-structured verification diagrams are continuously optimized based on the application of the digital twin. For example, if a candidate target is added as a twin target, the initial value of one of its verification items is increased accordingly. If the application of the digital twin for a certain twin target is infrequent and it is determined that a digital twin is not necessary, the initial value of one of its verification items is decreased. This process continues, and the tree-structured verification diagrams are continuously optimized based on subsequent application.

[0067] Based on the tree-structured verification diagram and a large amount of simulated target detail data, an initial evaluation model is established. The initial evaluation model is used to analyze the target detail data and determine the individual value of each verification item, such as whether it is a power plant verification item. If it is determined to be a power plant based on the target detail data, the individual value is the corresponding verification item value. If it is determined not to be a power plant based on the target detail data, the individual value of the verification item is 0.

[0068] The initial evaluation model is used to analyze the detailed data of the target to be selected and obtain the corresponding individual values; yi = Hi(x); where: yi is the individual value of the corresponding check item, i represents the corresponding check item, i = 1, 2, ..., n, n is a positive integer; Hi(x) is the initial evaluation model;

[0069] The expression for the initial evaluation model is: x represents the input data, indicating the details of the target to be selected; i represents the corresponding verification item, i = 1, 2, ..., n, where n is a positive integer; A indicates that the requirements of the verification item are met; Ci is the verification item value of the verification item.

[0070] According to the formula Calculate the corresponding initial value, where CS is the initial value.

[0071] The digital twin is created based on existing digital twin technology;

[0072] For example, a digital twin is established using a power plant as an example;

[0073] A spatial point cloud structure of the power station is built using laser sensors, while a panoramic camera collects RGB information of the surrounding environment and records the camera's pose. Using the pre-built grayscale point cloud image and the color information collected by the camera, a point cloud coloring algorithm is matched to digitally clone the physical power station, resulting in a fully covered, high-precision digital twin power station. Each point cloud device in the digital twin power station will be automatically objectified and assigned a unique identifier code to correspond to the real-world physical device. Other digital twin technologies can also be used to build digital twin power stations.

[0074] Based on each digital twin and distribution network information, a corresponding distribution network scheduling model is established. Each digital twin is integrated according to its corresponding position, and other distribution network information is modeled and supplemented using existing 3D modeling technology, and then integrated into a distribution network scheduling model.

[0075] Step SA2: Obtain the original dispatch plan, which is the dispatch plan within the previously established preset range. For example, analyze tomorrow's dispatch plan using relevant data, i.e., today's original dispatch plan, and perform analysis using the corresponding functions of the existing dispatch system; identify the impact items of the original dispatch plan, and set impact items for grid connection uncertainties such as distributed renewable energy, such as changes in power generation; the various impact items of the dispatch plan remain unchanged and are confirmed and set by professionals; extract the corresponding data from the original dispatch plan according to the set impact items to form the original impact item data for each impact item; such as the power generation of each power station;

[0076] Step SA3: Real-time identification of the collected and displayed data of each digital twin in the distribution network dispatch model. Through each digital twin, the real-time data of the corresponding target can be displayed synchronously. The collected and displayed data are extracted according to each influencing factor to form the current real-time influencing factor data.

[0077] Step SA4: Preset the corresponding matrix template according to the number of each influencing item, that is, use each influencing item to represent each element in the matrix, and then fill in the corresponding values ​​of each influencing item; each row represents one influencing item, and each element in each row is the data of different targets of the same influencing item, such as power generation, which can correspond to the power generation of thermal power plants, the power generation of wind power plants, etc.

[0078] The original and real-time impact data are transformed according to the preset matrix template to form the corresponding planning matrix and impact matrix; non-numerical data are transformed using the preset existing numerical transformation methods.

[0079] Step SA5: Calculate the corresponding change matrix based on the planning matrix and the influence matrix, and adjust the original scheduling plan according to the obtained change matrix to obtain the real-time scheduling plan;

[0080] Change matrix = Planning matrix - Influence matrix

[0081] Methods for adjusting the original scheduling plan based on the change matrix include:

[0082] Define the period of change Δt, the duration of which is set by the administrator;

[0083] Based on the prediction function of each digital twin in the distribution network dispatch model, the real-time impact data within the predicted change period Δt is predicted and marked as predicted change data; the original impact item data corresponding to the original dispatch plan within the changed period Δt is obtained and marked as changed impact data; the changed impact data and predicted change data are converted into the corresponding plan matrix and impact matrix, and then the corresponding change matrix is ​​calculated and marked as the prediction matrix, thus obtaining multiple prediction matrices within the changed period Δt.

[0084] Based on each element in the change matrix, and combined with the corresponding elements in each prediction matrix, a change curve of the element in the change period Δt is formed, with time as the horizontal axis and the corresponding element value as the vertical axis; using the existing fitting function, the curve function of each change curve is fitted and labeled G(t).

[0085] According to the formula Calculate the corresponding element correction value;

[0086] In the formula: YA is the element correction value; Y0 is the corresponding element value in the transformation matrix;

[0087] Replace the corresponding element values ​​in the change matrix with the obtained element correction values ​​to obtain the corresponding scheduling matrix;

[0088] The obtained scheduling matrix is ​​analyzed to obtain the corresponding real-time scheduling plan.

[0089] Specifically, a corresponding scheduling analysis model is established based on a CNN or DNN network. The corresponding training set is manually built and trained. The training set includes input data and output data. The input data is the original scheduling plan, the planning matrix, and the scheduling matrix. The output data is the real-time scheduling plan. The corresponding real-time scheduling plan is obtained by analyzing the successfully trained scheduling analysis model. Since neural networks are existing technology in this field, the specific building and training process will not be described in detail in this invention.

[0090] Alternatively, existing scheduling plan analysis can be used to replace the corresponding impact data and analyze the corresponding real-time scheduling plan.

[0091] Step SA6: Simulate and verify the real-time dispatch plan based on the distribution network dispatch model. Once the simulation verification is successful, apply the corresponding real-time dispatch plan.

[0092] This involves combining simulations and verifications of various digital twins within the distribution network dispatch model. The specific method is as follows:

[0093] Simulated scheduling of real-time scheduling plans is carried out based on the digital twins in the distribution network scheduling model; the simulated scheduling process is monitored in real time to obtain corresponding monitoring data; the obtained monitoring data is analyzed from the perspectives of scheduling anomalies and scheduling allowances to obtain corresponding scheduling verification values.

[0094] When the scheduling verification value is greater than the threshold X2, the verification is considered successful; otherwise, the verification is considered unsuccessful, and the real-time scheduling plan is adjusted.

[0095] Scheduling allowable angles: These refer to whether the corresponding monitoring data meets scheduling requirements, such as stability requirements, current and voltage safety requirements, etc., during the simulated real-time scheduling plan. This can be directly identified and judged based on the relevant scheduling requirements. Specifically:

[0096] Obtain the preset scheduling permission requirements, which include various requirements; perform real-time judgment on the monitoring data based on each requirement, obtain the judgment result for each requirement, and calculate the corresponding scheduling permission value based on the obtained judgment result. The scheduling permission value has only two values, 0 and BN, where BN is a numerical value and BN>X2; when all requirements are judged to be qualified, the scheduling permission value is 0; otherwise, if any requirement is judged to be unqualified, the scheduling permission value is BN.

[0097] From the perspective of scheduling anomalies: This refers to whether the system has any abnormal situations when simulating the operation of a real-time scheduling plan; a large amount of historical monitoring data after the scheduling plan has been run is obtained, and a corresponding anomaly identification model is established based on the obtained historical monitoring data. The monitoring data is analyzed through the established anomaly identification model to obtain the corresponding set of anomaly values; the anomaly value q = f(w); where q is the anomaly value and f(w) is the output value of the anomaly identification model;

[0098] The anomaly detection model is built based on the Isolation Forest algorithm, and its expression is: In the formula: w represents the input data, is the monitoring data; f(w) represents the output data;

[0099] Summing is performed on the set of outliers to obtain the corresponding cumulative value; the obtained cumulative value is then labeled as LDZ.

[0100] The corresponding scheduling verification value is calculated according to the formula YUZ = LDZ + DYZ, where YUZ is the scheduling verification value and DYZ is the scheduling allowable value.

[0101] By identifying each twin target and establishing a distribution network scheduling model based on each twin target, the system can acquire relevant data of the current distribution network in real time, facilitating scheduling adjustments and improving scheduling efficiency. Simultaneously, it can optimize the original scheduling plan. Furthermore, the obtained real-time scheduling plan can be simulated and verified using the distribution network scheduling model, ensuring the stable, safe, and efficient operation of the power system. Through real-time monitoring and scheduling of the distribution network, the system can adjust the allocation of power resources in a timely manner, avoiding excessive power load in some areas and ensuring the stable operation of the power system.

[0102] A big data power distribution network dispatching system includes a model module, a dispatching analysis module, and a dispatching execution module;

[0103] The model module is used to determine the twin target based on the distribution network information, establish the corresponding digital twin based on the twin target, and establish the corresponding distribution network scheduling model based on each digital twin and the distribution network information.

[0104] The scheduling analysis module is used to analyze the original scheduling plan based on the distribution network scheduling model to obtain the corresponding real-time scheduling plan.

[0105] The scheduling execution module is used to execute the real-time scheduling plan. It simulates and verifies the real-time scheduling plan based on the distribution network scheduling model. Once the simulation verification is successful, the corresponding real-time scheduling plan is applied.

[0106] The specific undisclosed parts in this embodiment refer to an embodiment of a big data distribution network scheduling method.

[0107] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.

[0108] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A big data-driven power distribution network scheduling method, characterized in that, The methods include: Obtain distribution network information, determine the digital twin target based on the distribution network information, and establish a corresponding digital twin based on the digital twin target; Establish corresponding distribution network scheduling models based on each digital twin and the distribution network information; Obtain the original scheduling plan, and extract the data of each original influence item in the original scheduling plan according to the preset influence items; Real-time data is collected from the power distribution network dispatch model based on each of the aforementioned influencing factors to obtain real-time influencing factor data corresponding to each of the aforementioned influencing factors. The original impact data and the real-time impact data are processed according to the preset matrix template to obtain the corresponding planning matrix and impact matrix; Calculate the corresponding change matrix based on the plan matrix and the influence matrix, and adjust the original scheduling plan based on the change matrix to obtain the real-time scheduling plan; The real-time scheduling plan is simulated and verified based on the power distribution network scheduling model. Once the simulation verification is successful, the corresponding real-time scheduling plan is applied. Change matrix = Planning matrix - Influence matrix; Methods for adjusting the original scheduling plan based on the change matrix include: Define a change period Δt, obtain the predicted change data within the change period Δt based on the power distribution network scheduling model; and obtain the corresponding change impact data within the change period Δt based on the original scheduling plan. Generate corresponding prediction matrices based on the change impact data and the predicted change data; Identify each element value in the change matrix and mark it as a baseline element value. Extract the corresponding set of element values ​​according to the time order and the position of each baseline element value in each prediction matrix. Based on the reference element values ​​and the corresponding set of element values, a change curve for each reference element value is generated within the change time period Δt. In the change curve, the horizontal axis represents time, the time span is from 0 to Δt, and the vertical axis represents the element value. The curve function that fits each variation curve is labeled G(t); According to the formula Calculate the corresponding element correction value; In the formula: YA is the element correction value; Y0 is the base element value; Replace the corresponding base element value in the change matrix with the element correction value. After all replacements are made, mark the change matrix as a scheduling matrix. The scheduling matrix is ​​analyzed using a preset scheduling analysis model to obtain the corresponding real-time scheduling plan.

2. The big data distribution network scheduling method according to claim 1, characterized in that, Methods for identifying twin targets include: The power distribution network information is identified and matched according to the preset candidate target details table to determine the details data of each candidate target; The obtained target details data are analyzed to obtain corresponding initial values; corresponding correction values ​​are set for the candidate targets. The target evaluation value is calculated according to the formula MPG=CS+XZ; where: MPG is the target evaluation value; CS is the initial value; and XZ is the correction value; The candidate targets whose target evaluation value is greater than the threshold X1 are marked as twin targets.

3. The big data distribution network scheduling method according to claim 2, characterized in that, The methods for calculating initial values ​​include: Each tree-structured verification diagram is set according to the candidate target details table. The tree-structured verification diagram includes each verification item and the verification item value corresponding to each verification item. An initial evaluation model is established based on the tree-structured verification diagram and the detailed data of the candidate targets in the simulation settings; the expression of the initial evaluation model is: ; In the formula: x is the input data, representing the details of the target to be selected; i represents the corresponding verification item, i=1, 2, ..., n, where n is a positive integer; A indicates that the verification item requirements are met; Ci is the verification item value; The target details data are analyzed using the initial evaluation model to obtain individual values; yi=Hi(x); where: yi is the individual value of the corresponding verification item, i represents the corresponding verification item, i=1,2,...,n, and n is a positive integer; Hi(x) is the initial evaluation model; According to the formula Calculate the corresponding initial value, where CS is the initial value and yi is the individual value of the corresponding check item.

4. The big data distribution network scheduling method according to claim 1, characterized in that, Methods for simulating and validating real-time scheduling plans include: The real-time scheduling plan is simulated and scheduled according to each of the digital twins in the power distribution network scheduling model; the simulated scheduling process is monitored in real time to obtain corresponding monitoring data; the obtained monitoring data is analyzed to obtain corresponding scheduling verification values. When the scheduling verification value is not greater than the threshold X2, the simulation verification result is that the verification is passed; When the scheduling verification value is greater than the threshold X2, the simulation verification result is a verification failure; the real-time scheduling plan is then adjusted.

5. The big data distribution network scheduling method according to claim 3, characterized in that, Methods for analyzing monitoring data include: Obtain preset scheduling permission requirements, verify the monitoring data according to each requirement item in the scheduling permission requirements, obtain the judgment result corresponding to each requirement item, and set the corresponding scheduling permission value according to the obtained judgment result; When all requirements are deemed satisfactory, the scheduling allowable value is 0; when not all requirements are deemed satisfactory, the scheduling allowable value is BN; BN is greater than the threshold X2. An anomaly identification model is established based on historical monitoring data. The monitoring data is then analyzed using the established anomaly identification model to obtain the corresponding set of outliers. The summation of the outlier set is performed to obtain the corresponding cumulative value; the cumulative value is then labeled as LDZ. Then, the corresponding scheduling verification value is calculated according to the formula YUZ=LDZ+DYZ; In the formula: YUZ is the scheduling verification value; DYZ is the scheduling allowable value.

6. The big data distribution network scheduling method according to claim 5, characterized in that, Outlier q=f(w); In the formula: q represents an outlier; f(w) is the output value of the anomaly detection model; An anomaly detection model is established based on the isolated forest algorithm, and the expression is: ; In the formula: w represents the monitoring data; f(w) represents the output data.

7. A big data distribution network dispatching system, characterized in that, A big data distribution network scheduling method according to any one of claims 1-6, comprising a model module, a scheduling analysis module, and a scheduling execution module; The model module is used to determine the twin target based on the distribution network information, establish a corresponding digital twin based on the twin target, and establish a corresponding distribution network scheduling model based on each digital twin and the distribution network information. The scheduling analysis module is used to analyze the original scheduling plan based on the distribution network scheduling model to obtain the corresponding real-time scheduling plan; The scheduling execution module is used to execute the real-time scheduling plan, and to simulate and verify the real-time scheduling plan based on the distribution network scheduling model. Once the simulation and verification are successful, the corresponding real-time scheduling plan is applied.

Images