A power transaction system and method based on a space truss structure analysis model

Abstract

This invention relates to the field of power technology, and in particular to a power trading system and method based on a grid structure analysis model. The system includes: classifying electricity consumption fluctuation categories within sub-management areas; determining whether to correct the predicted electricity purchase fitting coefficient for the corresponding sub-management area based on reference values ​​during fluctuation periods; preprocessing input data to filter out abnormal data; a prediction unit for a trading volume prediction module that predicts the predicted electricity purchase for each sub-management area based on the preprocessed input data; determining whether the trading volume prediction module is functioning correctly based on the prediction difference; verifying the collected trading volume based on the predicted purchase volume; correcting the operating parameters of the trading volume prediction module when an operational anomaly is detected; and determining whether to issue an alarm message for abnormal trading data acquisition based on the re-determined prediction difference. This invention improves the reliability of acquiring trading data, thereby improving the efficiency of trading data processing.

Description

Technical Field

[0001] This invention relates to the field of power technology, and in particular to a power trading system and method based on a grid structure analysis model. Background Technology

[0002] With the continuous development of the power system, the power grid is expanding in scale and becoming increasingly complex, covering multiple sub-management areas. The electricity consumption characteristics, geographical environment, and meteorological conditions of each area are significantly different.

[0003] In different districts of large cities, due to varying industrial structures, some areas are dominated by industrial electricity consumption, while others are dominated by residential and commercial electricity consumption. In some mountainous or remote areas, complex terrain can affect the laying and maintenance of power grid lines, and variable weather conditions can also significantly impact power supply and demand. In such a complex and volatile environment, accurately predicting the electricity trading volume of each sub-management area is of paramount importance.

[0004] Currently, in terms of electricity trading volume forecasting, some existing systems mainly adopt traditional time series analysis methods, such as the ARIMA model. These methods are primarily based on historical trading volume data for forecasting, simply modeling and analyzing the volume data in chronological order.

[0005] Chinese Patent Publication No. CN115049486A discloses a method and system for electricity transaction settlement during the transition from medium- and long-term transactions to spot transactions, belonging to the power field. First, the market participants during the transition period are divided. Then, based on the electricity transaction types of each divided market participant, a settlement type and a deviation electricity settlement type are determined for each type of electricity transaction. Finally, a settlement function for each type of electricity transaction and the deviation electricity settlement type is determined. Using this settlement function, the settlement of various transaction businesses during the transition from medium- and long-term transactions to spot transactions can be calculated, and the established electricity transaction settlement system supports the settlement of various transaction businesses. However, the above technical solution has the following problems: it does not consider predicting the transaction electricity volume based on data within sub-management areas, nor does it consider verifying the collected transaction electricity volume based on the predicted purchase volume, affecting the reliability of the acquired transaction data and consequently impacting the efficiency of transaction data processing. Summary of the Invention

[0006] The purpose of this invention is to overcome the above-mentioned shortcomings and provide a power trading system and method based on a grid structure analysis model. This invention overcomes the problems in the prior art that do not consider predicting the trading volume based on data within sub-management areas, nor do they consider verifying the collected trading volume based on the predicted purchase volume, which affects the reliability of the acquired trading data and consequently the processing efficiency of the trading data.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a power trading system based on a grid structure analysis model, comprising:

[0008] The grid structure analysis module is used to determine the maintenance location and length of grid lines within the management area, which includes several sub-management areas. The data collection module is used to acquire the transaction electricity volume, meteorological data, and voltage detection units of each sub-management area; The transaction volume prediction module is connected to the data collection module and the network structure analysis module, respectively. It includes a data cleaning unit for preprocessing the input data to filter out abnormal data and a prediction unit for predicting the predicted purchase volume of each sub-management area based on the preprocessed input data. The regional division module is connected to the data collection module and the transaction volume prediction module respectively, and is used to divide the electricity consumption floating category of the sub-management area based on the voltage fluctuation characterization value. The analysis module, which is connected to the data collection module, the trading volume prediction module, and the region division module, is used to determine whether the operation of the trading volume prediction module is qualified based on the prediction difference, and when the operation of the trading volume prediction module is determined to be abnormal, to correct the operating parameters of the trading volume prediction module based on the external influence characterization value and the historical prediction fluctuation characterization value, including adjusting the screening criteria used to screen out abnormal data to the corresponding value, or adjusting the maintenance coefficient used to correct the maintenance length of the corresponding sub-management area to the corresponding value. The transaction data statistics alarm module, which is connected to the analysis module, is used to determine whether to issue an alarm message for abnormal transaction data acquisition based on the redefined prediction difference.

[0009] Preferably, the input data includes the location of the grid line maintenance, the fitted maintenance length, the historical transaction volume, wind speed, temperature, and light intensity of each sub-management area.

[0010] Preferably, the region division module is used to determine the voltage fluctuation characterization value, including: It is used to identify fluctuations based on the effective value of voltage, determine the fluctuation time interval based on the duration of the fluctuation, and determine the voltage fluctuation characterization value based on the fluctuation time interval; The region division module is used to determine the reference values ​​for fluctuation periods, including: It is used to determine the reference value for the fluctuation period based on the interval length of each fluctuation time interval.

[0011] Preferably, the area division module is used to classify the electricity fluctuation categories of sub-management areas based on voltage fluctuation characterization values, including: If the voltage fluctuation characterization value is greater than the preset voltage fluctuation characterization value, the determined individual sub-management area is classified as a high-power consumption floating category, and the prediction power purchase fitting coefficient of the corresponding sub-management area is determined based on the reference value of the fluctuation period. If the reference value for the fluctuation period is less than or equal to the preset reference value for the fluctuation period, the fitting coefficient of the predicted electricity purchase volume for a single sub-management area will be adjusted to the corresponding value based on the reference value for the fluctuation period. The decrease in the predicted electricity purchase fitting coefficient is positively correlated with the reference value during the fluctuation period; If the reference value for the fluctuation period is greater than the preset reference value for the fluctuation period, the preset prediction difference will be adjusted to the corresponding value based on the voltage fluctuation characterization value. The increase in the preset predicted difference is positively correlated with the voltage fluctuation characterization value.

[0012] Preferably, if the voltage fluctuation characterization value is less than or equal to the preset voltage fluctuation characterization value, then the individual sub-management area is identified as a weak power consumption floating category, and the transaction volume prediction module is controlled to continue to operate using the current operating parameters.

[0013] Preferably, the analysis module determines whether the transaction volume prediction module is functioning correctly based on the predicted difference, including: The absolute value of the difference between the predicted purchase volume and the transaction volume for a single sub-management area is determined as the prediction difference. If the predicted difference is greater than the preset predicted difference, the operation of the trading volume prediction module is determined to be abnormal. The operating parameters of the trading volume prediction module are corrected based on the external influence characterization value and the historical predicted volatility characterization value, including adjusting the screening criteria used to screen out abnormal data to the corresponding value, or adjusting the maintenance coefficient of the corresponding sub-management area to the corresponding value. The analysis module is used to determine the external impact characterization value and the historical predicted fluctuation characterization value; It is used to plot a concentration time-domain curve based on the acquired concentration of particulate matter in the air, so as to identify several concentration inflection point time points in the first derivative curve corresponding to the concentration time-domain curve. It is used to draw the time-domain curve of the transaction volume based on the acquired transaction volume, so as to identify several inflection points of the volume in the first derivative curve corresponding to the time-domain curve of the transaction volume. Used to determine the characterization value of external impact based on each inflection point time; The analysis module determines the historical forecast fluctuation characterization value based on the historical forecast differences.

[0014] Preferably, if the predicted difference is less than or equal to the preset predicted difference, the trading volume prediction module is deemed to be operating successfully, and the trading volume prediction module is controlled to continue operating using the current operating parameters.

[0015] Preferably, the analysis module's parameters for adjusting the trading volume prediction module based on external influence characterization values ​​and historical predicted volatility characterization values ​​include: If the external impact characterization value is less than or equal to the preset external impact characterization value, the screening criteria used to screen out abnormal data will be adjusted to the corresponding value based on the external impact characterization value. If the external influence characterization value is greater than the preset external influence characterization value, the operating parameters of the trading volume prediction module will be adjusted based on the historical predicted volatility characterization value. If the historical predicted volatility characterization value is less than or equal to the preset historical predicted volatility characterization value, the transaction data statistics alarm module is controlled to issue an alarm message for abnormal transaction data acquisition. If the historical predicted fluctuation value is greater than the preset historical predicted fluctuation value, the maintenance coefficient corresponding to a single sub-management area will be adjusted to the corresponding value based on the historical predicted fluctuation value. The reduction in the screening criteria used to remove outliers is positively correlated with the external impact characterization value. The increase in the maintenance coefficient corresponding to a single sub-management area is positively correlated with the historical predicted fluctuation value; The fitted maintenance length is the product of the maintenance coefficient and the maintenance length.

[0016] Preferably, the transaction data statistics alarm module is used to determine whether to issue an alarm message for abnormal transaction data acquisition based on the redefined prediction difference, after the operating parameters of the transaction volume prediction module have been corrected, including: If the predicted difference is less than or equal to the preset predicted difference, the trading volume prediction module is deemed to be operating successfully, and the trading volume prediction module is controlled to continue operating using the current operating parameters. If the predicted difference is greater than the preset predicted difference, an alarm will be issued to indicate an anomaly in the acquisition of transaction data.

[0017] In addition, this invention also discloses a power trading method for the power trading system based on the above-mentioned grid structure analysis model, comprising the following steps: S1. Classify electricity fluctuation categories in sub-management areas based on voltage fluctuation characterization values: When a single sub-management area is identified as a high-power-consumption floating category, the fitting coefficient of the predicted power purchase volume for the corresponding sub-management area is determined based on the reference value during the fluctuation period. S2. Preprocess the input data to filter out abnormal data; S3, the transaction volume prediction module is a prediction unit that predicts the predicted purchase volume of each sub-management area based on the pre-processed input data. The input data includes the maintenance location of the grid line, the fitted maintenance length, the historical transaction volume of each sub-management area, wind speed, temperature, and light intensity. S4. Determine whether the transaction volume prediction module is functioning correctly based on the predicted difference: When an abnormality is identified in the operation of the trading volume prediction module, the operating parameters of the trading volume prediction module are corrected based on the external impact characterization value and the historical predicted volatility characterization value. This includes adjusting the screening criteria used to filter out abnormal data to the corresponding value, or adjusting the maintenance coefficient of the corresponding sub-management area to the corresponding value. When revising the operating parameters of the trading volume prediction module, determine whether to issue an alarm for abnormal trading data acquisition based on the redefined prediction difference.

[0018] Beneficial effects of this invention: 1. This invention categorizes electricity consumption fluctuations in sub-management areas based on voltage fluctuation characterization values. These values ​​represent the degree of voltage fluctuation and reflect voltage stability. Different sub-management areas have different electricity consumption characteristics. When voltage fluctuations occur, electrical equipment within these areas may be affected and stop operating, leading to fluctuations in traded electricity volume. Therefore, voltage fluctuations affect the stability of electricity trading and the accuracy of predictions. This invention categorizes electricity consumption fluctuations based on voltage fluctuation characterization values ​​and adopts different prediction and management strategies for different areas. When the reference value for the fluctuation period is greater than a preset reference value, further analysis is performed based on the voltage fluctuation characterization values. The reference value for the fluctuation period indicates whether there is a periodic pattern in voltage fluctuations. When the reference value is less than or equal to the preset reference value, voltage fluctuations exhibit a pattern. In this case, the electricity purchase volume fitting coefficient is adjusted to correct the predicted electricity purchase volume, improving prediction accuracy and making the predicted electricity purchase volume more consistent with actual electricity consumption. When the reference value for a fluctuating period exceeds the preset reference value, voltage fluctuations become irregular, posing a significant challenge to the accuracy of predicted electricity purchases. The discrepancy between the predicted and actual purchases becomes even greater. In such cases, increasing the preset prediction discrepancy and relaxing the evaluation criteria for prediction accuracy helps adapt to situations with large voltage fluctuations and avoids issuing erroneous alarm messages. This approach enables precise classification of electricity consumption characteristics in each sub-management area, allowing for adjustments to prediction parameters and management strategies based on different categories. This improves the accuracy of predicted electricity purchases, enhances the relevance of alarm messages, and ultimately improves the efficiency of processing electricity trading data.

[0019] 2. This invention determines the operational qualification of the transaction volume prediction module based on the prediction difference. When operational anomalies occur, the operating parameters are corrected by combining external influence characterization values ​​and historical prediction fluctuation characterization values. The prediction difference is the difference between the predicted purchase volume and the actual transaction volume, reflecting the accuracy of the prediction. When there is a prediction deviation, an external influence characterization value is determined. This value characterizes the correlation between the transaction volume and external influences, indicating whether there is a direct relationship between them. When the external influence characterization value is less than or equal to a preset external influence characterization value, a correlation exists. In this case, for example, factories may need to use air purification equipment during periods of abnormal air pollution, leading to irregular high electricity consumption. The data cleaning unit may mistakenly filter out this high electricity consumption as abnormal data, resulting in inaccurate predictions. In this case, the screening criteria used to remove abnormal data are adjusted to the corresponding value based on the external influence characterization value. When the external influence characterization value is greater than the preset value, there is no direct correlation. In this case, the operating parameters of the transaction volume prediction module are corrected based on historical prediction fluctuation characterization values, which reflect the stability of historical prediction errors. When historical predicted fluctuation values ​​exceed preset historical predicted fluctuation values, prediction deviations of varying degrees frequently occur. This is because the impact of maintenance varies across different sub-management areas; the impact is greater at the ends of the power grid and less in densely populated areas. In such cases, a maintenance coefficient is introduced to correct the prediction parameters for the corresponding sub-management area, thereby improving the accuracy of predicted electricity purchases. Timely detection of operational anomalies in the transaction volume prediction module, along with adjustments to operating parameters to improve prediction accuracy and stability, further enhances the efficiency of power data processing. Attached Figure Description

[0020] Figure 1 This is a block diagram of a power trading system module based on a grid structure analysis model, according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating the steps of the power trading method based on a grid structure analysis model according to an embodiment of the present invention. Figure 3 This is a logic diagram of the region division module of the present invention, which divides the power consumption fluctuation category of sub-management regions based on the voltage fluctuation characterization value; Figure 4 This is a logic diagram showing the determination of whether the transaction volume prediction module is qualified based on the predicted difference in the analysis module of this embodiment of the invention. Detailed Implementation

[0021] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0022] Example 1: Please see Figure 1As shown, it is a block diagram of a power trading system module based on a network structure analysis model according to an embodiment of the present invention. The power trading system of the present invention includes: The regional data storage module is used to store data information of the management area, which includes several sub-management areas; The grid structure analysis module is used to determine the maintenance location and length of grid lines within the management area; The data collection module includes a power statistics unit for acquiring the transaction power of each sub-management area, a meteorological recording unit for acquiring meteorological data, and a voltage detection unit for detecting each sub-management area. The transaction volume prediction module is connected to the data collection module and the grid structure analysis module, respectively. It includes a data cleaning unit for preprocessing the input data to filter out abnormal data and a prediction unit for predicting the predicted purchase volume of each sub-management area based on the preprocessed input data. The input data includes grid line maintenance location, fitted maintenance length, historical transaction volume of each sub-management area, wind speed, temperature, and light intensity. The region division module is connected to the region data storage module, the data collection module and the transaction volume prediction module respectively. It is used to divide the electricity consumption fluctuation category of the sub-management area based on the voltage fluctuation characterization value, and when a single sub-management area is determined to be a strong electricity consumption fluctuation category, it determines whether to correct the predicted purchase volume fitting coefficient of the corresponding sub-management area based on the fluctuation period reference value. The analysis module is connected to the data collection module, the trading volume prediction module, and the region division module, respectively. It is used to determine whether the operation of the trading volume prediction module is qualified based on the prediction difference, and when the operation of the trading volume prediction module is determined to be abnormal, it corrects the operating parameters of the trading volume prediction module based on the external influence characterization value and the historical prediction fluctuation characterization value, including adjusting the screening criteria used to screen out abnormal data to the corresponding value, or adjusting the maintenance coefficient of the corresponding sub-management area to the corresponding value. The transaction data statistics alarm module, which is connected to the analysis module, is used to determine whether to issue an alarm message for abnormal transaction data acquisition based on the redefined prediction difference, after the operation parameters of the transaction volume prediction module have been corrected.

[0023] Specifically, the data stored in the regional data storage module may include terrain data, power grid line information, user information, and equipment information. Power grid line information may include the starting and ending points of the lines, line length, line material, conductor cross-sectional area, line voltage level, and tower location and type. User information may include user types, number of users, and user distribution within each sub-management area. Equipment information may include the model, capacity, and operating parameters of the power equipment, which will not be elaborated further.

[0024] Specifically, the grid structure analysis module includes a grid structure analysis model, which can determine the maintenance and blockage status of all grid lines within the management area. This is existing technology and will not be described in detail here.

[0025] Specifically, the specific structure of the electricity statistics unit is not limited, and it can be a smart meter used to obtain the transaction electricity of each sub-management area.

[0026] The specific structure of the voltage detection unit is not limited; it can be a voltage sensor capable of real-time voltage monitoring.

[0027] Specifically, the data cleaning unit calculates the mean and standard deviation of the acquired data to identify data that deviates from the mean by more than a preset multiple of the standard deviation as outliers and removes them.

[0028] The preset difference factor standard deviation is the product of the preset difference factor and the calculated standard deviation of the data.

[0029] Deviation from the mean is the absolute value of the difference between a single data point and the calculated mean of all data points.

[0030] Specifically, the prediction unit uses a Long Short-Term Memory (LSTM) network model to predict the electricity purchase volume for each sub-management area. Training data includes the maintenance location, maintenance length, wind speed, temperature, and light intensity of the power grid lines, as well as the corresponding transaction volume for each sub-management area. The model is trained using this training data. The input data for the trained LSTM network model consists of the current maintenance location of the power grid lines, the fitted maintenance length, the historical transaction volume for each sub-management area, wind speed, temperature, and light intensity. The output data of the trained LSTM network model is the predicted electricity purchase volume for each sub-management area.

[0031] In this embodiment, the Long Short-Term Memory (LSTM) network model consists of an input layer, at least two LSM hidden layers, and a fully connected output layer. The input layer receives preprocessed multi-dimensional input data, which is constructed as time-series samples. Each time-series sample contains historical data within a continuous time window. The length of this time window is set according to the business cycle of power trading forecasting, typically ranging from 24 hours to 7 days.

[0032] Specifically, for each sub-management area, the input data is constructed as follows: First, the historical electricity transaction data of each sub-management area is arranged in time series to form a historical electricity sequence. Second, the maintenance location information of the grid lines, the calculated fitted maintenance length, wind speed, temperature, and light intensity data are aligned with the same timestamps as the historical electricity sequence. The maintenance location information is converted into numerical features using one-hot encoding or embedded representation, while the fitted maintenance length, wind speed, temperature, and light intensity are used as continuous numerical features. All numerical features are normalized. Finally, for a given prediction time point, the corresponding input time series sample contains all the above feature sequences within a complete time window preceding that time point.

[0033] The first Long Short-Term Memory (LSTM) hidden layer captures short-term dependencies in the input temporal data, typically with 64 to 256 units. The second LSM hidden layer captures longer-term temporal dependency patterns, with the same or slightly fewer units as the first layer. Each LSM unit contains a forget gate, an input gate, and an output gate to control the flow of information. A dropout layer follows the hidden layers, randomly discarding some neuron outputs during training to prevent overfitting; the dropout rate is set between 0.2 and 0.5. Finally, a fully connected output layer maps the hidden layer state at the last time step to the predicted electricity scalar value for the corresponding sub-management region.

[0034] The model is trained using supervised learning. A historical dataset containing information on grid line maintenance locations, maintenance lengths, wind speeds, temperatures, light intensity, and corresponding transaction volumes is used. The training objective is to minimize the error between the model's predicted purchase volume and the actual transaction volume. The loss function is mean squared error, and the optimizer uses the Adam optimization algorithm. The initial learning rate is set between 0.001 and 0.01 and can be dynamically decayed during training. Training is performed in multiple rounds until the model's loss on the validation set no longer decreases significantly. The trained model can then receive real-time or predicted grid line maintenance locations, fitted maintenance lengths, historical electricity sequences, and meteorological data to output predicted purchase volumes for a specified future period. This predicted purchase volume will serve as the basis for subsequent analysis modules to calculate prediction discrepancies and perform data validation.

[0035] Specifically, the fitted maintenance length is the product of the maintenance degree and the maintenance coefficient of the corresponding sub-management area.

[0036] Specifically, after adjusting the fitting coefficient of the predicted purchase volume for the corresponding sub-management area, the transaction volume prediction module determines the output predicted purchase volume by multiplying the predicted purchase volume of the sub-management area output by the prediction unit with the fitting coefficient of the predicted purchase volume for the corresponding sub-management area.

[0037] Example 2: Please see Figure 2 The diagram shows a flowchart of the steps in the power trading method based on a grid structure analysis model according to an embodiment of the present invention. The power trading method of the present invention includes: S1, the electricity fluctuation categories for dividing sub-management areas based on voltage fluctuation characterization values ​​include: When a single sub-management area is identified as a high-power-consumption floating category, the fitting coefficient of the predicted power purchase volume for the corresponding sub-management area is determined based on the reference value during the fluctuation period. S2, preprocess the input data to filter out abnormal data; S3, the transaction volume prediction module is a prediction unit that predicts the predicted purchase volume of each sub-management area based on the pre-processed input data. The input data includes the maintenance location of the grid line, the fitted maintenance length, the historical transaction volume of each sub-management area, wind speed, temperature, and light intensity. S4, determining whether the transaction volume prediction module is functioning correctly based on the predicted difference includes: When an abnormality is identified in the operation of the trading volume prediction module, the operating parameters of the trading volume prediction module are corrected based on the external impact characterization value and the historical predicted volatility characterization value. This includes adjusting the screening criteria used to filter out abnormal data to the corresponding value, or adjusting the maintenance coefficient of the corresponding sub-management area to the corresponding value. When revising the operating parameters of the trading volume prediction module, determine whether to issue an alarm for abnormal trading data acquisition based on the redefined prediction difference.

[0038] Specifically, the region division module is used to determine voltage fluctuation characterization values ​​including: Used to plot the effective voltage time-domain curve based on the effective voltage values ​​at each time point within the acquired preset monitoring duration; This is used to determine the fluctuation situation when the voltage difference between the effective voltage value and the preset effective voltage value in the effective voltage time domain curve is greater than the preset voltage deviation value. Used to determine the duration interval of the fluctuation as the fluctuation time interval when the duration of the fluctuation exceeds the preset voltage deviation time; The ratio of the total duration of each fluctuation time interval to the preset monitoring duration is calculated to obtain the voltage fluctuation characterization value.

[0039] Specifically, the region division module is used to determine the reference value for the fluctuation period, including: This is used to determine the variance of the interval length of each calculated fluctuation time interval as a reference value for the fluctuation period.

[0040] Please see Figure 3The diagrams shown illustrate the logic determination of the region division module in this invention, which classifies the electricity consumption floating categories of sub-management areas based on voltage fluctuation characterization values. The region division module of this invention is used to classify the electricity consumption floating categories of sub-management areas based on voltage fluctuation characterization values, including: If the voltage fluctuation characterization value is less than or equal to the preset voltage fluctuation characterization value, then the single sub-management area will be identified as a weak power consumption floating category, and the transaction volume prediction module will be controlled to continue to run using the current operating parameters. If the voltage fluctuation characterization value is greater than the preset voltage fluctuation characterization value, the determined individual sub-management area is classified as a high-power consumption floating category, and the prediction power purchase fitting coefficient of the corresponding sub-management area is determined based on the reference value of the fluctuation period. If the reference value for the fluctuation period is less than or equal to the preset reference value for the fluctuation period, the fitting coefficient of the predicted electricity purchase volume for a single sub-management area will be adjusted to the corresponding value based on the reference value for the fluctuation period. If the reference value for the fluctuation period is greater than the preset reference value for the fluctuation period, the preset prediction difference will be adjusted to the corresponding value based on the voltage fluctuation characterization value.

[0041] Specifically, the preset voltage fluctuation characterization value is selected within the range [0.08, 0.12], and the preset fluctuation period reference value is selected within the range [0.26P0, 0.31P0], where P0 is the average duration of the interval between each fluctuation time interval. Those skilled in the art can select the preset voltage fluctuation characterization value and the preset fluctuation period reference value according to the specific situation of the management area, and can determine them based on statistical analysis of a large amount of historical data. In this embodiment, preferably, the preset voltage fluctuation characterization value is 0.12 and the preset fluctuation period reference value is 0.26P0.

[0042] Specifically, electricity consumption fluctuation categories are categorized for sub-management areas based on voltage fluctuation characterization values. These values ​​represent the degree of voltage fluctuation and reflect voltage stability. Different sub-management areas have different electricity consumption characteristics. When voltage fluctuations occur, electrical equipment within these areas may be affected and stop operating, leading to fluctuations in traded electricity volume. Therefore, voltage fluctuations affect the stability of electricity trading and the accuracy of forecasts. Based on voltage fluctuation characterization values, different forecasting and management strategies are adopted for different areas. When the reference value for the fluctuation period is greater than the preset reference value, further analysis is conducted based on the voltage fluctuation characterization values. The reference value for the fluctuation period indicates whether there is a periodic pattern in voltage fluctuations. When the reference value is less than or equal to the preset reference value, voltage fluctuations are regular. In this case, the fitting coefficient for the purchased electricity volume is adjusted to correct the predicted purchased electricity volume, improving forecast accuracy and making the predicted purchased electricity volume more consistent with actual electricity consumption. When the reference value for a fluctuating period exceeds the preset reference value, voltage fluctuations become irregular, posing a significant challenge to the accuracy of predicted electricity purchases. The discrepancy between the predicted and actual purchases becomes even greater. In such cases, increasing the preset prediction discrepancy and relaxing the evaluation criteria for prediction accuracy helps adapt to situations with large voltage fluctuations and avoids issuing erroneous alarm messages. This approach enables precise classification of electricity consumption characteristics in each sub-management area, allowing for adjustments to prediction parameters and management strategies based on different categories. This improves the accuracy of predicted electricity purchases, enhances the relevance of alarm messages, and ultimately improves the efficiency of processing electricity trading data.

[0043] Specifically, the region division module is used to adjust the predicted electricity purchase fitting coefficient of a single sub-management region to a corresponding value based on the reference value of the fluctuation period, wherein, The decrease in the predicted electricity purchase fitting coefficient is positively correlated with the reference value during the fluctuation period.

[0044] Specifically, the reference value for the fluctuation period reflects the dispersion or regularity of the voltage fluctuation time period. The smaller the reference value for the fluctuation period, the more stable the time interval of voltage fluctuation, and the more controllable the prediction error. Users can allocate the operation of electrical equipment according to the actual voltage situation. The smaller the decrease in the fitting coefficient of the predicted power purchase, the closer the prediction result is to the actual situation, and the prediction error is reduced.

[0045] In this embodiment, optionally, Compare the reference value for the fluctuation period with the comparison value for the first period and the comparison value for the second period; If the reference value during the fluctuation period is less than or equal to the comparison value during the first period, the fitting coefficient of the predicted electricity purchase volume corresponding to a single sub-management area will be adjusted to 0.94 times the initial fitting coefficient of the predicted electricity purchase volume. If the reference value for the fluctuation period is less than or equal to the comparison value for the second period and greater than the comparison value for the first period, the fitting coefficient of the predicted electricity purchase volume corresponding to a single sub-management area will be adjusted to 0.87 times the initial fitting coefficient of the predicted electricity purchase volume. If the reference value during the fluctuation period is greater than the comparison value during the second period, the fitting coefficient of the predicted electricity purchase volume corresponding to a single sub-management area will be adjusted to 0.77 times the initial fitting coefficient of the predicted electricity purchase volume. The comparison value for the first time period is 0.42B0, and the comparison value for the second time period is 0.68B0. B0 is the preset reference value for the fluctuation period.

[0046] Specifically, the region division module is used to adjust the preset predicted difference to a corresponding value based on the voltage fluctuation characterization value, wherein, The increase in the preset predicted difference is positively correlated with the voltage fluctuation characterization value.

[0047] In this embodiment, optionally, Compare the voltage fluctuation characterization value with the first preset voltage fluctuation comparison value and the second preset voltage fluctuation comparison value; If the voltage fluctuation characterization value is less than or equal to the first preset voltage fluctuation comparison value, the preset prediction difference amount will be adjusted to 1.12 times the initial preset prediction difference amount. If the voltage fluctuation characterization value is less than or equal to the second preset voltage fluctuation comparison value and greater than the first preset voltage fluctuation comparison value, then the preset prediction difference amount will be adjusted to 1.18 times the initial preset prediction difference amount. If the voltage fluctuation characterization value is greater than the second preset voltage fluctuation comparison value, the preset prediction difference amount will be adjusted to 1.26 times the initial preset prediction difference amount. The first preset voltage fluctuation comparison value is 1.34U0, and the second preset voltage fluctuation comparison value is 1.47U0, where U0 is the preset voltage fluctuation characterization value.

[0048] Specifically, the voltage fluctuation characterization value reflects the degree and frequency of voltage fluctuations. The larger the voltage fluctuation characterization value, the more severe the voltage instability in the sub-managed area, and the greater the increase in the preset prediction difference. Relaxing the evaluation criteria for determining the operational qualification of the trading volume prediction module allows the system to operate normally and evaluate prediction data even under complex voltage fluctuation environments, ensuring the stability and reliability of the power trading system.

[0049] Please see Figure 4 The diagrams shown illustrate the logic judgment of the analysis module in this invention, which determines whether the trading volume prediction module is functioning correctly based on the predicted difference. The analysis module of this invention, used to determine whether the trading volume prediction module is functioning correctly based on the predicted difference, includes: The absolute value of the difference between the predicted purchase volume and the transaction volume for a single sub-management area is determined as the prediction difference. If the predicted difference is less than or equal to the preset predicted difference, the trading volume prediction module is deemed to be operating successfully, and the trading volume prediction module is controlled to continue operating using the current operating parameters. If the predicted difference is greater than the preset predicted difference, the trading volume prediction module is determined to be malfunctioning. The operating parameters of the trading volume prediction module are then corrected based on the external influence characterization value and the historical predicted volatility characterization value. This includes adjusting the screening criteria used to filter out abnormal data to the corresponding value, or adjusting the maintenance coefficient of the corresponding sub-management area to the corresponding value.

[0050] Specifically, the preset prediction difference is selected within the range [0.07Y0, 0.11Y0], with the unit being kWh, where Y0 is the predicted electricity purchase amount. Those skilled in the art can select the preset prediction difference amount themselves. It can be understood that the division of the situation that can achieve the accuracy of the prediction is acceptable. In this embodiment, preferably, the preset prediction difference amount is 0.1Y0.

[0051] Specifically, the analysis module is used to determine the external influence characterization value and the historical predicted fluctuation characterization value, including: Used to plot concentration time-domain curves based on the concentration of particulate matter in the air at each time point within the preset analysis period; Used to plot the first derivative curve of concentration based on the concentration time domain curve, and to identify the concentration inflection point time point of each inflection point in the first derivative curve of concentration. Used to plot the transaction volume time domain curve based on the transaction volume at each time point within the preset analysis period; Used to draw the first derivative curve of electricity based on the time domain curve of the transaction electricity, and to identify the time points of each inflection point in the first derivative curve of electricity. It is used to compare the first derivative curve of concentration with the first derivative curve of charge. Taking the inflection point of concentration as the benchmark, the inflection point time interval of charge that is closest to the inflection point time interval of each inflection point of concentration is calculated in turn. The variance of the calculated time interval between each inflection point is recorded as the external influence characterization value; The analysis module is used to determine the variance of each historical prediction difference as a characterization value of historical prediction fluctuation.

[0052] Specifically, the concentration of particulate matter in the air can be defined as the PM2.5 concentration.

[0053] Specifically, the analysis module is used to adjust the operating parameters of the trading volume prediction module based on the external influence characterization value, including: If the external impact characterization value is less than or equal to the preset external impact characterization value, the screening criteria used to screen out abnormal data will be adjusted to the corresponding value based on the external impact characterization value. If the external influence characterization value is greater than the preset external influence characterization value, the operating parameters of the trading volume prediction module will be adjusted based on the historical predicted volatility characterization value. If the historical predicted volatility characterization value is less than or equal to the preset historical predicted volatility characterization value, the transaction data statistics alarm module is controlled to issue an alarm message for abnormal transaction data acquisition. If the historical predicted fluctuation value is greater than the preset historical predicted fluctuation value, the maintenance coefficient corresponding to a single sub-management area will be adjusted to the corresponding value based on the historical predicted fluctuation value. The reduction in the screening criteria used to remove outliers is positively correlated with the external impact characterization value. The increase in the maintenance coefficient corresponding to a single sub-management area is positively correlated with the historical predicted fluctuation value.

[0054] Specifically, the performance of the transaction volume forecasting module is determined based on the forecast discrepancy. In cases of operational anomalies, the operating parameters are adjusted by combining external influence characterization values ​​and historical forecast fluctuation characterization values. The forecast discrepancy is the difference between the predicted purchase volume and the actual transaction volume, reflecting the accuracy of the forecast. When there is a forecast deviation, an external influence characterization value is determined. This value represents the correlation between the transaction volume and external influences, indicating whether there is a direct relationship between them. If the external influence characterization value is less than or equal to the preset external influence characterization value, a correlation exists. In this case, for example, factories may use air purification equipment during periods of abnormal air pollution, leading to irregular high electricity consumption. The data cleaning unit may mistakenly filter out this high electricity consumption as abnormal data, resulting in inaccurate forecasts. In this situation, the screening criteria used to remove abnormal data are adjusted to the corresponding value based on the external influence characterization value. If the external influence characterization value is greater than the preset value, there is no direct correlation. In this case, the operating parameters of the transaction volume forecasting module are adjusted based on historical forecast fluctuation characterization values, which reflect the stability of historical forecast errors. When historical predicted fluctuation values ​​exceed preset historical predicted fluctuation values, prediction deviations of varying degrees frequently occur. This is because the impact of maintenance varies across different sub-management areas; the impact is greater at the ends of the power grid and less in densely populated areas. In such cases, a maintenance coefficient is introduced to correct the prediction parameters for the corresponding sub-management area, thereby improving the accuracy of predicted electricity purchases. Timely detection of operational anomalies in the transaction volume prediction module, along with adjustments to operating parameters to improve prediction accuracy and stability, further enhances the efficiency of power data processing.

[0055] Specifically, the preset external influence characterization values ​​are selected within the interval [98, 110], with units of min. 2Those skilled in the art can select preset external influence characterization values ​​and determine them through analysis and statistics of a large amount of historical data. It is understood that this allows for the classification of whether the transaction volume is closely related to external influences. In this embodiment, preferably, the preset historical predicted fluctuation characterization value is 100.

[0056] The preset historical predicted fluctuation characterization value is selected within the range [480, 520], with the unit being kWh². Those skilled in the art can select the preset historical predicted fluctuation characterization value themselves, and can determine it based on historical data analysis. It is understood that it is possible to classify whether the prediction is stable or not. In this embodiment, preferably, the preset historical predicted fluctuation characterization value is 500.

[0057] In this embodiment, optionally, Compare the external influence characterization value with the first preset association comparison value and the second preset association comparison value; If the external influence characterization value is less than or equal to the first preset correlation comparison value, the preset difference factor used to screen out abnormal data will be adjusted to 1.13 times the initial preset difference factor. If the external influence characterization value is less than or equal to the second preset correlation comparison value and greater than the first preset correlation comparison value, the preset difference factor used to screen out abnormal data will be adjusted to 1.21 times the initial preset difference factor. If the external influence characterization value is greater than the second preset correlation comparison value, the preset difference factor used to screen out abnormal data will be adjusted to 1.31 times the initial preset difference factor. The first preset correlation comparison value is 0.41G0, and the second preset correlation comparison value is 0.61G0, where G0 is the preset external influence characterization value.

[0058] Specifically, the larger the external impact characterization value, the greater the reduction in the screening criteria for outlier data. The external impact characterization value reflects the degree of correlation between traded electricity volume and external impacts. A larger external impact characterization value indicates a closer correlation between changes in traded electricity volume and external impacts. Therefore, a greater reduction in the screening criteria is necessary to retain more electricity volume data that may be related to external impacts, allowing the data to more comprehensively reflect actual electricity consumption.

[0059] In this embodiment, optionally, Compare the historical predicted fluctuation values ​​with the first preset historical prediction comparison value and the second preset historical prediction comparison value; If the historical predicted fluctuation value is less than or equal to the first preset historical predicted comparison value, the maintenance coefficient corresponding to a single sub-management area will be adjusted to 1.17 times the initial maintenance coefficient. If the historical predicted fluctuation value is less than or equal to the second preset historical predicted comparison value and greater than the first preset historical predicted comparison value, then the maintenance coefficient corresponding to a single sub-management area will be adjusted to 1.23 times the initial maintenance coefficient. If the historical predicted fluctuation value is greater than the second preset historical predicted comparison value, the maintenance coefficient corresponding to a single sub-management area will be adjusted to 1.29 times the initial maintenance coefficient. The first preset historical prediction comparison value is 1.33S0, and the second preset historical prediction comparison value is 1.61S0, where S0 is the preset historical prediction fluctuation characterization value.

[0060] Specifically, the maintenance coefficient for each sub-management area is adjusted to the corresponding value based on historical predicted fluctuation values; the increase in the maintenance coefficient for each sub-management area is positively correlated with the historical predicted fluctuation values. The greater the historical fluctuations, the greater the impact of the sub-management area's geographical location on prediction accuracy, and the greater the correction to the maintenance coefficient.

[0061] Specifically, the transaction data statistics alarm module is used to determine whether to issue an alarm for abnormal transaction data acquisition based on the newly determined prediction difference, after correcting the operating parameters of the transaction volume prediction module. This includes: If the predicted difference is less than or equal to the preset predicted difference, the trading volume prediction module is deemed to be operating successfully, and the trading volume prediction module is controlled to continue operating using the current operating parameters. If the predicted difference is greater than the preset predicted difference, an alarm will be issued to indicate an anomaly in the acquisition of transaction data.

[0062] Specifically, if the prediction discrepancy is still too large after the correction of abnormal operation is completed, there may be an anomaly in the acquisition of transaction data, and an anomaly alert should be issued in a timely manner.

[0063] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be limited to the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A power trading system based on a grid structure analysis model, characterized in that, include: The grid structure analysis module is used to determine the maintenance location and length of grid lines within the management area, which includes several sub-management areas. The data collection module is used to acquire the transaction electricity volume, meteorological data, and voltage detection units of each sub-management area; The transaction volume prediction module is connected to the data collection module and the network structure analysis module, respectively. It includes a data cleaning unit for preprocessing the input data to filter out abnormal data and a prediction unit for predicting the predicted purchase volume of each sub-management area based on the preprocessed input data. The regional division module is connected to the data collection module and the transaction volume prediction module respectively, and is used to divide the electricity consumption floating category of the sub-management area based on the voltage fluctuation characterization value. The analysis module, which is connected to the data collection module, the trading volume prediction module, and the region division module, is used to determine whether the operation of the trading volume prediction module is qualified based on the prediction difference, and when the operation of the trading volume prediction module is determined to be abnormal, to correct the operating parameters of the trading volume prediction module based on the external influence characterization value and the historical prediction fluctuation characterization value, including adjusting the screening criteria used to screen out abnormal data to the corresponding value, or adjusting the maintenance coefficient used to correct the maintenance length of the corresponding sub-management area to the corresponding value. The transaction data statistics alarm module, which is connected to the analysis module, is used to determine whether to issue an alarm message for abnormal transaction data acquisition based on the redefined prediction difference.

2. The power trading system based on a grid structure analysis model according to claim 1, characterized in that, The input data includes the location of the grid line maintenance, the fitted maintenance length, the historical transaction volume, wind speed, temperature, and light intensity of each sub-management area.

3. A power trading system based on a grid structure analysis model according to claim 2, characterized in that, The region division module is used to determine voltage fluctuation characterization values, including: It is used to identify fluctuations based on the effective value of voltage, determine the fluctuation time interval based on the duration of the fluctuation, and determine the voltage fluctuation characterization value based on the fluctuation time interval; The region division module is used to determine the reference values ​​for fluctuation periods, including: It is used to determine the reference value for the fluctuation period based on the interval length of each fluctuation time interval.

4. A power trading system based on a grid structure analysis model according to claim 3, characterized in that, The region division module is used to divide the electricity fluctuation categories of sub-management regions based on voltage fluctuation characterization values, including: If the voltage fluctuation characterization value is greater than the preset voltage fluctuation characterization value, the determined individual sub-management area is classified as a high-power consumption floating category, and the prediction power purchase fitting coefficient of the corresponding sub-management area is determined based on the reference value of the fluctuation period. If the reference value for the fluctuation period is less than or equal to the preset reference value for the fluctuation period, the fitting coefficient of the predicted electricity purchase volume for a single sub-management area will be adjusted to the corresponding value based on the reference value for the fluctuation period. The decrease in the predicted electricity purchase fitting coefficient is positively correlated with the reference value during the fluctuation period; If the reference value for the fluctuation period is greater than the preset reference value for the fluctuation period, the preset prediction difference will be adjusted to the corresponding value based on the voltage fluctuation characterization value. The increase in the preset predicted difference is positively correlated with the voltage fluctuation characterization value.

5. A power trading system based on a grid structure analysis model according to claim 4, characterized in that, If the voltage fluctuation characterization value is less than or equal to the preset voltage fluctuation characterization value, the individual sub-management area will be identified as a weak power consumption floating category, and the transaction volume prediction module will be controlled to continue to run using the current operating parameters.

6. A power trading system based on a grid structure analysis model according to claim 5, characterized in that, The analysis module is used to determine whether the trading volume prediction module is functioning correctly based on the predicted discrepancy, including: The absolute value of the difference between the predicted purchase volume and the transaction volume for a single sub-management area is determined as the prediction difference. If the predicted difference is greater than the preset predicted difference, the operation of the trading volume prediction module is determined to be abnormal. The operating parameters of the trading volume prediction module are corrected based on the external influence characterization value and the historical predicted volatility characterization value, including adjusting the screening criteria used to screen out abnormal data to the corresponding value, or adjusting the maintenance coefficient of the corresponding sub-management area to the corresponding value. The analysis module is used to determine the external impact characterization value and the historical predicted fluctuation characterization value; It is used to plot a concentration time-domain curve based on the acquired concentration of particulate matter in the air, so as to identify several concentration inflection point time points in the first derivative curve corresponding to the concentration time-domain curve. It is used to draw the time-domain curve of the transaction volume based on the acquired transaction volume, so as to identify several inflection points of the volume in the first derivative curve corresponding to the time-domain curve of the transaction volume. Used to determine the characterization value of external impact based on each inflection point time; The analysis module determines the historical forecast fluctuation characterization value based on the historical forecast differences.

7. A power trading system based on a grid structure analysis model according to claim 6, characterized in that, If the predicted difference is less than or equal to the preset predicted difference, the trading volume prediction module is deemed to be operating successfully, and the trading volume prediction module is controlled to continue operating using the current operating parameters.

8. A power trading system based on a grid structure analysis model according to claim 7, characterized in that, The analysis module is used to adjust the operating parameters of the trading volume prediction module based on the external influence characterization value and the historical predicted volatility characterization value, including: If the external impact characterization value is less than or equal to the preset external impact characterization value, the screening criteria used to screen out abnormal data will be adjusted to the corresponding value based on the external impact characterization value. If the external influence characterization value is greater than the preset external influence characterization value, the operating parameters of the trading volume prediction module will be adjusted based on the historical predicted volatility characterization value. If the historical predicted volatility characterization value is less than or equal to the preset historical predicted volatility characterization value, the transaction data statistics alarm module is controlled to issue an alarm message for abnormal transaction data acquisition. If the historical predicted fluctuation value is greater than the preset historical predicted fluctuation value, the maintenance coefficient corresponding to a single sub-management area will be adjusted to the corresponding value based on the historical predicted fluctuation value. The reduction in the screening criteria used to remove outliers is positively correlated with the external impact characterization value. The increase in the maintenance coefficient corresponding to a single sub-management area is positively correlated with the historical predicted fluctuation value; The fitted maintenance length is the product of the maintenance coefficient and the maintenance length.

9. A power trading system based on a grid structure analysis model according to claim 8, characterized in that, The transaction data statistics alarm module is used to determine whether to issue an alarm message for abnormal transaction data acquisition based on the redefined prediction difference, after correcting the operating parameters of the transaction volume prediction module. include, If the predicted difference is less than or equal to the preset predicted difference, the trading volume prediction module is deemed to be operating successfully, and the trading volume prediction module is controlled to continue operating using the current operating parameters. If the predicted difference is greater than the preset predicted difference, an alarm will be issued to indicate an anomaly in the acquisition of transaction data.

10. A power trading method for a power trading system based on a grid structure analysis model as described in any one of claims 1-9, characterized in that, Includes the following steps: S1. Classify electricity fluctuation categories in sub-management areas based on voltage fluctuation characterization values: When a single sub-management area is identified as a high-power-consumption floating category, the fitting coefficient of the predicted power purchase volume for the corresponding sub-management area is determined based on the reference value of the fluctuation period. S2. Preprocess the input data to filter out abnormal data; S3, the transaction volume prediction module is a prediction unit that predicts the predicted purchase volume of each sub-management area based on the pre-processed input data. The input data includes the maintenance location of the grid line, the fitted maintenance length, the historical transaction volume of each sub-management area, wind speed, temperature, and light intensity. S4. Determine whether the transaction volume prediction module is functioning correctly based on the predicted difference: When an abnormality is identified in the operation of the trading volume prediction module, the operating parameters of the trading volume prediction module are corrected based on the external impact characterization value and the historical predicted volatility characterization value. This includes adjusting the screening criteria used to filter out abnormal data to the corresponding value, or adjusting the maintenance coefficient of the corresponding sub-management area to the corresponding value. When the operating parameters of the trading volume prediction module are corrected, an alarm message for abnormal trading data acquisition is issued based on the redefined prediction difference.

Images