A Machine Learning-Based Method for Constructing Power System Report Visualization

Abstract

This invention provides a method for constructing power system report visualization based on machine learning, belonging to the field of data processing technology. Specifically, it includes: based on the reliability of the power node data acquisition equipment, when the reliability is low, power node calculation data is obtained by using power flow calculation based on the power node's position in the power network topology; when the deviation between the power node calculation data and the actual power node data is large, the power node settlement data is used as the actual power node data; based on the power node data, the reliability of the acquisition equipment, historical power node data, weather temperature data, and historical average weather temperature data, a prediction model based on machine learning algorithms is used to obtain the accuracy of the power node data; only when the accuracy of the power node data is high, a pre-defined visualization data processing method is used to achieve visualization processing of the power node data, thereby further ensuring the reliability of the power node data.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, and in particular relates to a method for constructing power system report visualization based on machine learning. Background Technology

[0002] In recent years, with the gradual expansion of the power system, the amount and scale of data generated have increased exponentially, which has brought new challenges to further data processing and mining. The application of visualization technology in the power system has developed very rapidly since the 1990s. Due to its significant effectiveness in improving the intelligence level of the power grid, it has received high attention from power grid and power system companies worldwide. Therefore, how to combine visualization systems to achieve rapid data analysis and mining has become an urgent technical problem to be solved.

[0003] To achieve visualized processing of power system data, the authorized invention patent announcement number CN114548951B, "Visualization System and Method for Power Grid Infrastructure Ledger Based on Digital Twin Technology," processes power data in a database to obtain the individual data of each power grid infrastructure and its correlation data with other power grid infrastructures. The data visualization module generates a digital twin model based on the individual and correlation data to visually display the power data of the power grid infrastructures and their interconnections to the user. This provides an intuitive and quick way to obtain the power data of relevant power grid infrastructures. However, the following technical problems exist:

[0004] 1) Failure to identify the credibility of data based on different data sources. The credibility of data varies significantly depending on the data source. The credibility of data collection devices with different historical failure conditions also varies significantly. If the credibility of data cannot be identified, the creation of tables will be chaotic or even distorted when bad data is present.

[0005] 2) Failure to identify bad data by combining power grid topology and historical data. Based on the node in the power grid topology, the current, voltage, phase angle, active power, load, etc. of the node can be calculated by power flow calculation method. Therefore, if the two cannot be combined or only one aspect is considered, the identification efficiency and accuracy will be significantly affected.

[0006] To address the aforementioned technical problems, this invention provides a method for constructing power system report visualization based on machine learning. Summary of the Invention

[0007] To achieve the objectives of this invention, the following technical solution is adopted:

[0008] According to one aspect of the present invention, a method for constructing power system report visualization based on machine learning is provided.

[0009] A method for constructing power system report visualization based on machine learning, characterized in that it specifically includes:

[0010] S11 acquires power node data based on the data acquisition module, and determines whether the reliability of the power node data acquisition device is greater than the first reliability threshold. If yes, proceed to step S13; otherwise, proceed to step S12.

[0011] S12 Based on the position of the power node in the power network topology, the power flow calculation algorithm is used to calculate the power node calculation data. It is then determined whether the absolute value of the difference between the power node calculation data and the power node data is greater than a first absolute value threshold. If so, it is determined that there is a problem with the power node data, and the power node settlement data is used as the power node data.

[0012] S13 Based on the power node data, the reliability of the acquisition device, historical power node data, weather temperature data, and historical average weather temperature data, a prediction model based on machine learning algorithms is used to determine the authenticity of the power node data. It is determined whether the authenticity of the power node data is greater than a first authenticity threshold. If yes, proceed to step S14; otherwise, proceed to step S11 to continue acquiring the power node data. When the number of times the node is acquired is greater than the first threshold, the power node data will be constructed based on the historical power node data.

[0013] S14 Based on the power node data, a pre-defined visualization data processing method is used to realize the visualization processing of the power node data.

[0014] By setting a first confidence threshold and evaluating the confidence of the power node data acquisition equipment, the confidence of the data is identified. More complex data verification methods are used for data with lower confidence, while simpler data verification methods are used for data with higher confidence, thereby further improving the reliability of the data.

[0015] By generating data from power nodes and combining this data with the location of the power node within the power network topology, the reliability and accuracy of the power node data are further improved, and the influence of bad data is further eliminated.

[0016] By constructing the authenticity of power node data and establishing a first authenticity threshold, the authenticity of power node data can be evaluated from multiple perspectives, further ensuring the reliability and accuracy of the data generated for visualization and preventing the influence of bad data.

[0017] A further technical solution is that the power node data is any one or more of voltage, current, power, and phase angle.

[0018] A further technical solution is that the reliability is determined based on the number of failures of the data acquisition device in the past year and the quality rating of the data acquisition device manufacturer, wherein the quality rating of the manufacturer is determined based on the average service life and failure rate of the data acquisition device of the manufacturer.

[0019] By constructing credibility, we not only consider the historical data of the data acquisition equipment, but also the specific circumstances of its manufacturer, thus enabling a multi-faceted evaluation of credibility and making the credibility construction results more accurate and reliable.

[0020] A further technical solution is that the historical power node data is obtained based on the average of power node data at the same time within a week, and the historical average weather temperature data is obtained based on the average of weather temperature at the same time within a week.

[0021] A further technical solution involves the following specific steps for constructing the authenticity of the power node data:

[0022] S21 uses the absolute value of the difference between the power node data and the historical power node data as the historical error value, and determines whether the historical error value is greater than the first absolute value threshold and whether the absolute value of the difference between the weather temperature data and the historical average temperature data is less than the first temperature threshold. If yes, the authenticity of the power node data is determined to be 0. If no, proceed to step S22.

[0023] S22 constructs an input set based on the historical error values, the reliability of the acquisition equipment, weather temperature data, and historical average weather temperature data, and uses a prediction model based on intelligent algorithms to obtain the prediction result;

[0024] S23 determines the accuracy of the power node data based on the prediction results.

[0025] By setting the first absolute value threshold and the first temperature threshold, a simple method is used to filter out bad data, thereby improving the efficiency and speed of bad data identification while ensuring data reliability, and ensuring the reliability of the final construction result.

[0026] By constructing a predictive model to assess the accuracy, the technical problem of low accuracy caused by constructing empirical formulas or other methods is avoided, thereby further ensuring the efficiency and accuracy of the accuracy assessment.

[0027] A further technical solution is that the first absolute value threshold is determined based on the historical data of the power node and the data authenticity requirements.

[0028] A further technical solution involves the following specific steps for constructing the power node data based on the historical data of the power nodes:

[0029] S31 calculates power node calculation data based on the position of the power node in the power network topology using a power flow calculation algorithm;

[0030] S32 constructs a node input set based on the historical data of the power nodes, weather temperature data, and historical average weather temperature data, and inputs the node input set into the prediction model based on the PSO-BP algorithm to obtain the predicted power node data.

[0031] S33 obtains the power node data based on the power node calculation data and the average value of the predicted power node data.

[0032] By combining both calculated and predicted power node data, the system achieves an evaluation of power node data. This ensures both high efficiency and accurate evaluation of the power node data, further guaranteeing the accuracy of the final visualization results.

[0033] A further technical solution involves introducing random weights into the inertia weights of the PSO algorithm, with the weights calculated using the following formula:

[0034]

[0035] Where t is the current iteration number, ω max ω min These represent the maximum inertia weight and the minimum inertia weight, respectively. tmax is the maximum number of iterations, rand(1,2) is a random function with values ​​between 1 and 2, and K1 is a constant.

[0036] By using random weights and iteration counts to update the weights, the randomness of the optimization process is improved while avoiding the PSO algorithm from getting stuck in local optima. This further enhances the efficiency and accuracy of the optimization process.

[0037] On the other hand, a machine learning-based power system report visualization construction system is provided, which employs the aforementioned machine learning-based power system report visualization construction method, including:

[0038] Data acquisition module, credibility assessment module, power node calculation data calculation module, authenticity assessment module, bad data supplementation module, visualization processing module;

[0039] The data acquisition module is responsible for acquiring power node data;

[0040] The credibility assessment module is responsible for assessing the credibility of the power node data acquisition equipment.

[0041] The power node calculation data calculation module is responsible for calculating the power node calculation data based on the position of the power node in the power network topology and using the power flow calculation algorithm.

[0042] The authenticity assessment module is responsible for judging the authenticity of the power node data based on the power node data, the reliability of the acquisition equipment, historical power node data, weather temperature data, and historical average weather temperature data, using a prediction model based on machine learning algorithms.

[0043] The defective data supplementation module is responsible for constructing the power node data based on the historical data of the power node.

[0044] The visualization processing module is responsible for visualizing the power node data using a pre-defined visualization data processing method.

[0045] Other features and advantages will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0046] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0047] The above and other features and advantages of the present invention will become more apparent from a detailed description of exemplary embodiments thereof with reference to the accompanying drawings.

[0048] Figure 1 This is a flowchart of a machine learning-based power system report visualization construction method according to Embodiment 1. Detailed Implementation

[0049] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, they are provided so that the invention will be thorough and complete, and the concept of the exemplary embodiments will be fully conveyed to those skilled in the art. The same reference numerals in the drawings denote the same or similar structures, and therefore their detailed description will be omitted.

[0050] The terms “a,” “one,” “the,” and “the” are used to indicate the existence of one or more elements / components / etc.; the terms “including” and “having” are used to indicate an open-ended meaning of inclusion and that other elements / components / etc. may exist in addition to the listed elements / components / etc.

[0051] Example 1

[0052] To solve the above problems, according to one aspect of the present invention, such as Figure 1 As shown, a method for constructing power system report visualization based on machine learning is provided, characterized by specifically including:

[0053] S11 acquires power node data based on the data acquisition module, and determines whether the reliability of the power node data acquisition device is greater than the first reliability threshold. If yes, proceed to step S13; otherwise, proceed to step S12.

[0054] For example, the credibility of data acquisition equipment can be assessed using expert scoring or other methods, based on factors such as the number of historical failures of the equipment.

[0055] S12 Based on the position of the power node in the power network topology, the power flow calculation algorithm is used to calculate the power node calculation data. It is then determined whether the absolute value of the difference between the power node calculation data and the power node data is greater than a first absolute value threshold. If so, it is determined that there is a problem with the power node data, and the power node settlement data is used as the power node data.

[0056] For a specific example, power flow calculation, a term in electrical engineering, refers to the calculation of the distribution of active power, reactive power, and voltage in a power grid given its network topology, component parameters, and generation and load parameters. Power flow calculation determines the steady-state operating parameters of various parts of the power system based on the given grid structure, parameters, and operating conditions of generators, loads, and other components. Typically, the given operating conditions include the power of each source and load point in the system, the voltage at the hub point, the voltage at the equilibrium point, and the phase angle. The operating parameters to be determined include the voltage magnitude and phase angle of each bus node in the grid, as well as the power distribution of each branch and the power losses of the network.

[0057] S13 Based on the power node data, the reliability of the acquisition device, historical power node data, weather temperature data, and historical average weather temperature data, a prediction model based on machine learning algorithms is used to determine the authenticity of the power node data. It is determined whether the authenticity of the power node data is greater than a first authenticity threshold. If yes, proceed to step S14; otherwise, proceed to step S11 to continue acquiring the power node data. When the number of times the node is acquired is greater than the first threshold, the power node data will be constructed based on the historical power node data.

[0058] For example, historical data of power nodes can be directly used as power node data, or power node data can be constructed by combining it with power node computational data.

[0059] S14 Based on the power node data, a pre-defined visualization data processing method is used to realize the visualization processing of the power node data.

[0060] For specific examples, visualization can utilize existing reporting tools such as JReport, Excel, Crystal Reports, Smartbi, FineReport, and ActiveReports; BI analysis tools such as StyleIntelligence, BO, BIEE, ETHINK, and Yonghong Z-Suite; and domestic data visualization tools such as BDP Business Data Platform - Personal Edition, Big Data Magic Mirror, DataView, and FineBI Business Intelligence Software.

[0061] By setting a first confidence threshold and evaluating the confidence of the power node data acquisition equipment, the confidence of the data is identified. More complex data verification methods are used for data with lower confidence, while simpler data verification methods are used for data with higher confidence, thereby further improving the reliability of the data.

[0062] By generating data from power nodes and combining this data with the location of the power node within the power network topology, the reliability and accuracy of the power node data are further improved, and the influence of bad data is further eliminated.

[0063] By constructing the authenticity of power node data and establishing a first authenticity threshold, the authenticity of power node data can be evaluated from multiple perspectives, further ensuring the reliability and accuracy of the data generated for visualization and preventing the influence of bad data.

[0064] In another possible embodiment, the power node data is any one or more of voltage, current, power, and phase angle.

[0065] In another possible embodiment, the reliability is determined based on the number of failures of the data acquisition device in the past year and the quality rating of the data acquisition device manufacturer, wherein the quality rating of the manufacturer is determined based on the average service life and failure rate of the data acquisition device of the manufacturer.

[0066] For example, the formula for calculating credibility is:

[0067]

[0068] Where K2, K3, K4, and K5 are constants, C1 is the number of failures of the data acquisition equipment in the past year, and P1 is the quality rating of the data acquisition equipment manufacturer.

[0069] By constructing credibility, we not only consider the historical data of the data acquisition equipment, but also the specific circumstances of its manufacturer, thus enabling a multi-faceted evaluation of credibility and making the credibility construction results more accurate and reliable.

[0070] In another possible embodiment, the historical power node data is obtained based on the average of power node data at the same time within a week, and the historical average weather temperature data is obtained based on the average of weather temperatures at the same time within a week.

[0071] In another possible embodiment, the specific steps for constructing the authenticity of the power node data are as follows:

[0072] S21 uses the absolute value of the difference between the power node data and the historical power node data as the historical error value, and determines whether the historical error value is greater than the first absolute value threshold and whether the absolute value of the difference between the weather temperature data and the historical average temperature data is less than the first temperature threshold. If yes, the authenticity of the power node data is determined to be 0. If no, proceed to step S22.

[0073] S22 constructs an input set based on the historical error values, the reliability of the acquisition equipment, weather temperature data, and historical average weather temperature data, and uses a prediction model based on intelligent algorithms to obtain the prediction result;

[0074] For example, temperature features can be constructed based on weather temperature data and historical average weather temperature data. Specifically, they can be constructed based on the difference between the two or their similarity.

[0075] S23 determines the accuracy of the power node data based on the prediction results.

[0076] By setting the first absolute value threshold and the first temperature threshold, a simple method is used to filter out bad data, thereby improving the efficiency and speed of bad data identification while ensuring data reliability, and ensuring the reliability of the final construction result.

[0077] By constructing a predictive model to assess the accuracy, the technical problem of low accuracy caused by constructing empirical formulas or other methods is avoided, thereby further ensuring the efficiency and accuracy of the accuracy assessment.

[0078] In another possible embodiment, the first absolute value threshold is determined based on the historical data of the power node and the data authenticity requirements.

[0079] In another possible embodiment, the specific steps for constructing the power node data based on the historical power node data are as follows:

[0080] S31 calculates power node calculation data based on the position of the power node in the power network topology using a power flow calculation algorithm;

[0081] S32 constructs a node input set based on the historical data of the power nodes, weather temperature data, and historical average weather temperature data, and inputs the node input set into the prediction model based on the PSO-BP algorithm to obtain the predicted power node data.

[0082] For example, the PSO algorithm is used to optimize the initial value of the BP algorithm.

[0083] S33 obtains the power node data based on the power node calculation data and the average value of the predicted power node data.

[0084] By combining both calculated and predicted power node data, the system achieves an evaluation of power node data. This ensures both high efficiency and accurate evaluation of the power node data, further guaranteeing the accuracy of the final visualization results.

[0085] In another possible embodiment, random weights are introduced into the inertia weights of the PSO algorithm, and the formula for calculating the weights is as follows:

[0086]

[0087] Where t is the current iteration number, ω max ω min These represent the maximum inertia weight and the minimum inertia weight, respectively, t max K1 represents the maximum number of iterations, rand(1,2) is a random function with values ​​between 1 and 2, and K1 is a constant.

[0088] By using random weights and iteration counts to update the weights, the randomness of the optimization process is improved while avoiding the PSO algorithm from getting stuck in local optima. This further enhances the efficiency and accuracy of the optimization process.

[0089] Example 2

[0090] A machine learning-based power system report visualization construction system, employing the aforementioned machine learning-based power system report visualization construction method, includes:

[0091] Data acquisition module, credibility assessment module, power node calculation data calculation module, authenticity assessment module, bad data supplementation module, visualization processing module;

[0092] The data acquisition module is responsible for acquiring power node data;

[0093] The credibility assessment module is responsible for assessing the credibility of the power node data acquisition equipment.

[0094] The power node calculation data calculation module is responsible for calculating the power node calculation data based on the position of the power node in the power network topology and using the power flow calculation algorithm.

[0095] The authenticity assessment module is responsible for judging the authenticity of the power node data based on the power node data, the reliability of the acquisition equipment, historical power node data, weather temperature data, and historical average weather temperature data, using a prediction model based on machine learning algorithms.

[0096] The defective data supplementation module is responsible for constructing the power node data based on the historical data of the power node.

[0097] The visualization processing module is responsible for visualizing the power node data using a pre-defined visualization data processing method.

[0098] In this embodiment of the invention, the term "multiple" refers to two or more, unless otherwise explicitly defined. The terms "install," "connect," and "fix" should be interpreted broadly. For example, "connect" can mean a fixed connection, a detachable connection, or an integral connection. Those skilled in the art can understand the specific meaning of the above terms in this embodiment of the invention based on the specific circumstances.

[0099] In the description of the embodiments of the present invention, it should be understood that the terms "upper" and "lower" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or unit referred to must have a specific orientation or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of the present invention.

[0100] In the description of this specification, the terms "an embodiment," "a preferred embodiment," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0101] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. For those skilled in the art, the embodiments of the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of the present invention should be included within the protection scope of the embodiments of the present invention.

Claims

1. A method for constructing power system report visualization based on machine learning, characterized in that, Specifically, it includes: S11 acquires power node data based on the data acquisition module, and determines whether the reliability of the power node data acquisition device is greater than the first reliability threshold. If yes, proceed to step S13; otherwise, proceed to step S12. S12 Based on the position of the power node in the power network topology, the power flow calculation algorithm is used to calculate the power node calculation data. It is then determined whether the absolute value of the difference between the power node calculation data and the power node data is greater than a first absolute value threshold. If so, it is determined that there is a problem with the power node data, and the power node calculation data is used as the power node data. S13 Based on the power node data, the reliability of the acquisition device, historical power node data, weather temperature data, and historical average weather temperature data, a prediction model based on machine learning algorithms is used to determine the authenticity of the power node data. It is determined whether the authenticity of the power node data is greater than a first authenticity threshold. If yes, proceed to step S14; otherwise, proceed to step S11 to continue acquiring the power node data. When the number of times the node is acquired is greater than the first threshold, the power node data will be constructed based on the historical power node data. S14 Based on the power node data, a pre-defined visualization data processing method is used to realize the visualization processing of the power node data.

2. The power system report visualization construction method as described in claim 1, characterized in that, The power node data can be any one or more of voltage, current, power, and phase angle.

3. The power system report visualization construction method as described in claim 1, characterized in that, The reliability is determined based on the number of failures of the data acquisition device in the past year and the quality rating of the data acquisition device manufacturer, wherein the quality rating of the manufacturer is determined based on the average service life and failure rate of the data acquisition device of the manufacturer.

4. The power system report visualization construction method as described in claim 1, characterized in that, The historical power node data is obtained by averaging the power node data at the same time within a week, and the historical average weather temperature data is obtained by averaging the weather temperature at the same time within a week.

5. The power system report visualization construction method as described in claim 1, characterized in that, The specific steps for constructing the authenticity of the power node data are as follows: S21 uses the absolute value of the difference between the power node data and the historical power node data as the historical error value, and determines whether the historical error value is greater than the first absolute value threshold and whether the absolute value of the difference between the weather temperature data and the historical average temperature data is less than the first temperature threshold. If yes, the authenticity of the power node data is determined to be 0. If no, proceed to step S22. S22 constructs an input set based on the historical error values, the reliability of the acquisition equipment, weather temperature data, and historical average weather temperature data, and uses a prediction model based on intelligent algorithms to obtain the prediction result; S23 determines the accuracy of the power node data based on the prediction results.

6. The power system report visualization construction method as described in claim 1, characterized in that, The first absolute value threshold is determined based on the historical data of the power node and the data authenticity requirements.

7. The power system report visualization construction method as described in claim 1, characterized in that, The specific steps for constructing the power node data based on the historical power node data are as follows: S31 calculates power node calculation data based on the position of the power node in the power network topology using a power flow calculation algorithm; S32 constructs a node input set based on the historical data of the power nodes, weather temperature data, and historical average weather temperature data, and inputs the node input set into the prediction model based on the PSO-BP algorithm to obtain the predicted power node data. S33 obtains the power node data based on the power node calculation data and the average value of the predicted power node data.

8. The power system report visualization construction method as described in claim 7, characterized in that, Random weights are introduced into the inertia weights of the PSO algorithm, and the formula for calculating the weights is as follows: Where t is the current iteration number. , These represent the maximum inertia weight and the minimum inertia weight, respectively, t max The maximum number of iterations, K1 is a random function that takes values ​​between 1 and 2, and K1 is a constant.

9. A machine learning-based power system report visualization construction system, employing the machine learning-based power system report visualization construction method according to any one of claims 1-8, comprising: Data acquisition module, credibility assessment module, power node calculation data calculation module, authenticity assessment module, bad data supplementation module, visualization processing module; The data acquisition module is responsible for acquiring power node data; The credibility assessment module is responsible for assessing the credibility of the power node data acquisition equipment. The power node calculation data calculation module is responsible for calculating the power node calculation data based on the position of the power node in the power network topology and using the power flow calculation algorithm. The authenticity assessment module is responsible for judging the authenticity of the power node data based on the power node data, the reliability of the acquisition equipment, historical power node data, weather temperature data, and historical average weather temperature data, using a prediction model based on machine learning algorithms. The defective data supplementation module is responsible for constructing the power node data based on the historical data of the power node. The visualization processing module is responsible for visualizing the power node data using a pre-defined visualization data processing method.

Images