Electric energy metering error correction method and system for charging facility under multiple working conditions

Abstract

The invention discloses an electric energy metering error correction method and system for a charging facility under multiple working conditions. The method comprises the following steps: screening features and constructing cross features according to correlation coefficients and mutual information of various historical working condition parameter data; establishing an error correction model according to the cross characteristics and the error parameter data, and updating and optimizing the error correction model through data collected in real time in a plurality of set fusion time recording periods and historical working condition parameter data; measuring the electric energy metering value in real time, predicting an error correction value through the updated and optimized error correction, and correcting the measured electric energy metering value according to the predicted error correction value; and setting a monitoring period, and calculating a working condition health index value according to the predicted error correction value in the period to judge whether the working condition is abnormal. According to the invention, through real-time error correction, model dynamic updating and working condition health monitoring, the electric energy metering precision of the charging facility under a complex working condition and the stability of system operation are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric energy metering, and particularly relates to a method and system for correcting the electric energy metering error under multiple working conditions of a charging facility. Background Technique

[0002] With the popularization of electric vehicles, charging facilities have become an important part of the power grid system and transportation infrastructure; the metering accuracy of electric energy of charging facilities is directly related to the charging costs of users and the revenues of operators; however, in actual use, the metering accuracy of charging facilities is vulnerable to various factors, such as environmental temperature, electromagnetic interference, equipment aging, etc.; in addition, with the increase in the usage frequency and complexity of charging equipment, the risk of accumulation of metering errors is also increasing, which may lead to a decline in user experience and economic disputes.

[0003] CN118604440A proposes a method for reducing the output electric energy metering error of a DC charging pile, belonging to the technical field of DC charging piles, including: collecting data such as current, voltage, electric energy metering, load rate, etc. of the charging pile in real time; performing multivariate mixed decomposition on the current and voltage data to obtain stable components and variable components; establishing functional relationships between voltage stable components, input voltage variable components, input harmonics, load rate, etc. and current variable components and voltage variable components, and establishing a corresponding coefficient matrix; calculating the error compensation amounts of current and voltage based on the coefficient matrix; applying the error compensation amounts to the original current and voltage data to obtain corrected data, and recalculating the electric energy metering data corrected by errors based on this; however, this solution mainly corrects the error compensation of current and voltage and does not comprehensively consider the comprehensive influence of multiple working conditions. Summary of the Invention

[0004] The purpose of the present invention is to propose an electric energy metering error correction system for a charging facility under multiple working conditions in view of the current deficiencies.

[0005] The present invention adopts the following technical solutions.

[0006] The first aspect of the present invention proposes a method for correcting the electric energy metering error under multiple working conditions of a charging facility, which is characterized by including the following contents:

[0007] Obtain historical working condition parameter data from the historical database; the historical working condition parameter data includes historical electric energy metering values, historical external environment parameters, and historical error parameter data;

[0008] Extract the characteristics of each historical working condition parameter data, and screen the characteristics and construct cross-characteristics according to the correlation coefficients and mutual information of the characteristics of various historical working condition parameter data, and perform standardization processing on the cross-characteristics;

[0009] An error correction model is established based on the cross - features after standardization processing, and the error correction model is updated and optimized through the working condition parameter data collected in real - time within a set number of fusion time recording periods.

[0010] Measure the electric energy measurement value and the working condition parameter data in real - time according to the set sampling frequency. Input the working condition parameter data measured at each sampling point into the updated and optimized error correction model to predict the error correction value, and correct the measured electric energy measurement value according to the predicted error correction value.

[0011] Preferably, extracting the features of each historical working condition parameter data specifically includes:

[0012] The features extracted from the historical electric energy measurement value include charging current, charging voltage, charging power, load rate, total historical measured charging electric energy, and charging efficiency.

[0013] The features of the historical external environment parameters include: ambient temperature, ambient humidity, electromagnetic interference intensity around the charging facility, grid voltage fluctuation, geographical location climate, and wind speed.

[0014] The features of the historical error parameters include: theoretical charging electric energy, actual measured electric energy, and error change rate.

[0015] Preferably, screening the features and constructing cross - features according to the correlation coefficients and mutual information of the features of various historical working condition parameter data specifically includes:

[0016] Use a rolling window to calculate the correlation coefficients between all pairs of features of the historical electric energy measurement value, historical external environment parameters, and historical error parameters. If the correlation coefficient between two features exceeds the set correlation coefficient threshold, then delete any one of the features.

[0017] For all the features of the historical electric energy measurement value, historical external environment parameters, and historical error parameters after deleting the features whose correlation coefficients exceed the set correlation coefficient threshold, calculate the mutual information between each feature of the historical electric energy measurement value and historical external environment parameters and each feature of the historical error parameters, and calculate the mutual information between all the features of the historical electric energy measurement value and historical external environment parameters under the condition that the features of various historical error parameters are known.

[0018] Use the sum of the mutual information between one feature of the historical electric energy measurement value and one feature of the historical external environment parameter and one feature of the historical error parameter minus the mutual information between the corresponding historical electric energy measurement value and historical external environment parameter features under the condition that the feature of the corresponding historical error parameter is known as the comprehensive mutual information of the three features; select the top N comprehensive mutual information with the largest values, and fuse the features of the historical electric energy measurement value, historical external environment parameter, and historical error parameter corresponding to each selected comprehensive mutual information to form cross - features.

[0019] Preferably, the correlation coefficient threshold is set to 0.8.

[0020] Preferably, the features of the historical power metering values, historical external environment parameters, and historical error parameters corresponding to each selected comprehensive mutual information are fused to form cross features, specifically as follows:

[0021] Normalize the features of the historical power metering values, historical external environment parameters, and historical error parameters corresponding to each selected comprehensive mutual information;

[0022] The cross feature F is:

[0023]

[0024] where exp(·) is the exponential function of e; E, C, and F are the features of the normalized historical power metering values, historical external environment parameters, and historical error parameters, respectively; tanh(·) is the hyperbolic tangent function; r EC and r CF and r EF are the correlation coefficients between E and C, C and F, and E and F, respectively; I ECF is the comprehensive mutual information.

[0025] Preferably, an error correction model is established based on the standardized cross features, and the error correction model is updated and optimized by the working condition parameter data collected in real time within a set number of fusion time recording periods, specifically as follows:

[0026] An error correction model is trained based on the standardized cross features and the corresponding true error values as the original training set, and this model is a convolutional neural network;

[0027] The working condition parameter data collected in real time within a set number of fusion time recording periods is input into the error correction model to obtain all predicted error correction values; after each fusion time recording period, calculate all loss functions of each fusion time recording period, and multiply them by the set recent data weight;

[0028] And extract the historical working condition parameter data in the original training set that is the same as the working condition parameter data in one fusion time recording period and the corresponding true error values, and also calculate all their loss functions, and multiply them by the difference between 1 and the set recent data weight; add the two weighted loss functions as the final loss function of each fusion time recording period, and update the error correction model.

[0029] Preferably, the recent data weight is specifically:

[0030] When performing the first iteration, the value of the weight α(1) of the recent data is the set value, and the weight α(k) of the recent data for the iteration number k is as follows:

[0031]

[0032] ΔE(k) = |E(k) - E(k - 1)|

[0033] where γ is the sensitivity parameter; ΔE(k) is the error change rate at the current iteration number k, k ≥ 2; ΔE threshold is the preset error change threshold; E(k) is the average value of all error correction values in the training set at the current iteration number k; E(k - 1) is the average value of all error correction values in the training set at the previous iteration of the current iteration number k.

[0034] Preferably, the method further includes judging whether there is an abnormality in the charging facility condition based on the error correction value, so as to give an early warning. Specifically: modifying the set monitoring period, obtaining the predicted error correction values of each sampling point within the monitoring period at the end of each monitoring period, and calculating the condition health index value within each monitoring period according to these error correction values; comparing the obtained condition health index value with the set health threshold, and when the condition health index value is greater than the health threshold, it indicates that there is an abnormality in the condition health and an early warning is given;

[0035] Calculating the condition health index value within the monitoring period, the formula is:

[0036]

[0037] where Q is the condition health index value within the monitoring period; E avg is the average value of the error correction values of each sampling point within the detection period; γ is the set smoothing parameter; ΔE corr is the change trend parameter of the error correction value within the monitoring period.

[0038] Preferably, the change trend parameter of the error correction value within the monitoring period is specifically

[0039] ΔE corr The calculation formula is:

[0040]

[0041] where is the cumulative change amount of the error correction value within the monitoring period; n is the total number of sampling points within the monitoring period, and E corr (t) is the error correction value at the t-th time point within the monitoring period, and the larger t is, the closer the time point it is located to the current time.

[0042] The second aspect of the present invention proposes an electric energy metering error correction system for a charging facility under multiple operating conditions according to the method described in the first aspect of the present invention, including a data acquisition module, a feature extraction and fusion module, an error correction model establishment module, an error correction model update module, and an error correction module, characterized in that:

[0043] Data acquisition module: Obtain historical operating condition parameter data from the historical database; the historical operating condition parameter data includes historical electric energy metering values, historical external environment parameters, and historical error parameter data;

[0044] Feature extraction and fusion module: Extract the features of each historical operating condition parameter data, screen features and construct cross features according to the correlation and importance of the features of various historical operating condition parameter data, and perform standardization processing on the cross features;

[0045] Error correction model establishment module: Establish an error correction model based on the standardized cross features, and update and optimize the error correction model through the operating condition parameter data collected in real time within a set number of fusion time recording periods;

[0046] Error correction module: Measure the electric energy metering value and the operating condition parameter data in real time according to the set sampling frequency, input the operating condition parameter data measured at each sampling point into the updated and optimized error correction model to predict the error correction value, and correct the measured electric energy metering value according to the predicted error correction value.

[0047] The beneficial effects of the present invention are as follows. Compared with the prior art, the present invention quantifies the influence of operating condition factors in the electric energy metering process by establishing an error correction model and performing feature extraction, screening, and fusion on historical electric energy metering values, historical external environment parameters, and historical error parameter data, and corrects the electric energy metering error in the real-time charging process through the error correction model to ensure the accuracy and consistency of the final electric energy data; the model is updated and optimized by fusing the data within the time recording period with the data in the historical database, ensuring that the model can not only quickly adapt to the operating condition changes in the recent charging process but also maintain long-term robustness and generalization ability, improving the stable performance of the model under different operating conditions and environmental conditions; the weights of the data within the time recording period of the model and the data in the historical database during the training process are dynamically adjusted by the error change rate. When the error fluctuates greatly, the model gives priority to relying on the weights of recent data; when the error changes smoothly, it relies more on the long-term data in the historical database, thereby balancing the flexibility and stability of the model. Description of the Drawings

[0048] Figure 1 It is a framework diagram of the system of the present invention;

[0049] Figure 2 It is a flowchart of the method of the present invention. Detailed implementation manners

[0050] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. The embodiments described in this application are only a part of the embodiments of the present invention, rather than all embodiments. Based on the spirit of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0051] As Figure 1 shown, Embodiment 1 of the present invention proposes a method for correcting the power metering error under multiple working conditions of a charging facility, which is characterized by including the following contents:

[0052] Obtain historical working condition parameter data from the historical database; the historical working condition parameter data includes historical power metering values, historical external environment parameters, and historical error parameter data;

[0053] Extract the features of each historical working condition parameter data, and screen the features and construct cross features according to the correlation coefficients and mutual information of the features of various historical working condition parameter data, and perform standardization processing on the cross features;

[0054] Establish an error correction model based on the standardized cross features, and update and optimize the error correction model through the working condition parameter data collected in real time within a set number of fusion time recording periods;

[0055] Measure the power metering value and the working condition parameter data in real time according to the set sampling frequency, input the working condition parameter data measured at each sampling point into the updated and optimized error correction model to predict the error correction value, and correct the measured power metering value according to the predicted error correction value.

[0056] Specifically, according to the predicted error correction value E corr correct the measured power metering value W actual as follows:

[0057] W corrected = W actual - E corr

[0058] wherein, W corrected is the corrected power metering value;

[0059] Preferably, the extraction of the features of each historical working condition parameter data is specifically:

[0060] The features extracted from the historical power metering value include charging current, charging voltage, charging power, load rate, total historical metered charging power, and charging efficiency;

[0061] The characteristics of historical external environment parameters include: ambient temperature, ambient humidity, electromagnetic interference intensity around charging facilities, grid voltage fluctuation, geographical location climate, wind speed;

[0062] The characteristics of historical error parameters include: theoretical charging electric energy, actual measured electric energy, error change rate.

[0063] Preferably, the screening of features and construction of cross - features according to the correlation coefficients and mutual information of the characteristics of various historical working condition parameter data are specifically as follows:

[0064] Use a rolling window to calculate the correlation coefficients between all pairs of the characteristics of historical electric energy measurement values, historical external environment parameters, and historical error parameters. If the correlation coefficient between two characteristics exceeds the set correlation coefficient threshold, then delete any one of the two characteristics;

[0065] For all the characteristics of historical electric energy measurement values, historical external environment parameters, and historical error parameters after deleting the characteristics whose correlation coefficients exceed the set correlation coefficient threshold, calculate the mutual information between each characteristic of historical electric energy measurement values and historical external environment parameters and each characteristic of historical error parameters, and calculate the mutual information between all the characteristics of historical electric energy measurement values and historical external environment parameters under the condition of knowing the characteristics of various historical error parameters;

[0066] Use the sum of the mutual information between a characteristic of a historical electric energy measurement value and a characteristic of a historical external environment parameter and a characteristic of a historical error parameter minus the mutual information between the corresponding historical electric energy measurement value and historical external environment parameter characteristics under the condition of knowing the characteristic of the corresponding historical error parameter as the comprehensive mutual information of the three characteristics; select the top N largest comprehensive mutual information, and fuse the characteristics of historical electric energy measurement values, historical external environment parameters, and historical error parameters corresponding to each selected comprehensive mutual information to form cross - features.

[0067] Preferably, the correlation coefficient threshold is set to 0.8.

[0068] Preferably, the fusing of the characteristics of historical electric energy measurement values, historical external environment parameters, and historical error parameters corresponding to each selected comprehensive mutual information to form cross - features is specifically as follows:

[0069] Normalize all the characteristics of historical electric energy measurement values, historical external environment parameters, and historical error parameters corresponding to each selected comprehensive mutual information;

[0070] The cross - feature F is:

[0071]

[0072] where exp(·) is the exponential function with base e; E, C, and F are the features of the normalized historical electricity consumption measurement value, historical external environment parameters, and historical error parameters, respectively; tanh(·) is the hyperbolic tangent function; r EC , r CF , r EF are the correlation coefficients between E and C, between C and F, and between E and F, respectively; I ECF is the comprehensive mutual information.

[0073] Preferably, an error correction model is established based on the cross features after normalization, and the error correction model is updated and optimized by the working condition parameter data collected in real time within a set number of fusion time recording periods, specifically:

[0074] An error correction model is trained based on the cross features after normalization and the corresponding true error values as the original training set, and this model is a convolutional neural network;

[0075] All predicted error correction values are obtained by inputting the working condition parameter data collected in real time within a set number of fusion time recording periods into the error correction model; after each fusion time recording period, all loss functions of each fusion time recording period are calculated and multiplied by the set recent data weight;

[0076] Specifically, the form of the loss function is:

[0077]

[0078] where J total is the loss function value; m is the number of training samples; the error correction value of the i-th sample predicted by the model; E (i) is the true error value of the i-th sample; by calculating the gradient of the loss function through multiple iterations and using the gradient descent method to continuously adjust the weight and bias parameters of the model, the loss function value is gradually reduced, thereby completing the training and optimization of the model.

[0079] And the historical working condition parameter data in the original training set that is the same as the working condition parameter data in one fusion time recording period and the corresponding true error values are extracted, and all their loss functions are also calculated and multiplied by the difference between 1 and the set recent data weight; the sum of the two weighted loss functions is used as the final loss function for each fusion time recording period to update the error correction model.

[0080] The form of the final loss function is:

[0081] J′ total = α(k)·J(D recent,k )+(1 - α(k))·J(D long,k );

[0082] Among them, J′ total is the final loss function value after update and optimization, α(k) represents the weight of recent data at the current iteration number k, and J(D recent,k ) represents the loss function value obtained from the training samples within the latest fusion time record period at the current iteration number k, and J(D long,k ) represents the loss function value obtained from the training samples in the historical database at the current iteration number k.

[0083] Preferably, the weight of recent data is specifically:

[0084] The value of the weight of recent data α(1) at the first iteration is a set value, and the weight of recent data α(k) at the iteration number k is:

[0085]

[0086] ΔE(k) = |E(k) - E(k - 1)|

[0087] where γ is the sensitivity parameter; ΔE(k) is the error change rate at the current iteration number k, k ≥ 2; ΔE threshold is the preset error change threshold; E(k) is the average value of all error correction values in the training set at the current iteration number k; E(k - 1) is the average value of all error correction values in the training set at the previous iteration of the current iteration number k.

[0088] The method further includes judging whether there is an abnormality in the working condition health of the charging facility according to the error correction value, so as to give an early warning. Specifically: modify the set monitoring period, and after obtaining the predicted error correction values of each sampling point within the monitoring period at the end of each monitoring period, calculate the working condition health index value within each monitoring period according to these error correction values; compare the obtained working condition health index value with the set health threshold, and when the working condition health index value is greater than the health threshold, it indicates that there is an abnormality in the working condition health and give an early warning;

[0089] It should be noted that the working condition health index refers to the working condition health index of the electric vehicle charging facility. Although the calculation of this index is mainly based on the electric energy metering error, the electric energy metering error can reflect the health state of the charging facility to a certain extent. For example, when the sensor of the charging facility ages, the voltage fluctuates abnormally, or there is a fault in the power supply system, the electric energy metering error may show abnormal fluctuations or deviate from the normal range for a long time. In addition, if the electric energy metering error of this charging facility deviates from the normal range for a long time and shows an increasing trend, this may indicate potential problems in the internal electric energy metering system or power supply system of the equipment, thus affecting the normal operation of the charging facility; therefore, calculating the working condition health index based on the electric energy metering error helps to identify possible abnormal conditions of the charging facility.

[0090] Preferably, the formula for calculating the working condition health index value within the monitoring period is as follows:

[0091]

[0092] where Q is the working condition health index value within the monitoring period; E avg is the average value of the error correction values at each sampling point within the detection period; γ is the set smoothing parameter; ΔE corr is the change trend parameter of the error correction value within the monitoring period.

[0093] Preferably, the change trend parameter of the error correction value within the monitoring period is specifically

[0094] ΔE corr The calculation formula is:

[0095]

[0096] where is the cumulative change amount of the error correction value within the monitoring period; n is the total number of sampling points within the monitoring period, and E corr (t) is the error correction value at the t-th time point within the monitoring period. The larger t is, the closer the time point it represents is to the current time.

[0097] As Figure 1 shown, Embodiment 2 of the present invention proposes an electric energy metering error correction system under multiple working conditions of a charging facility according to the method described in Embodiment 1 of the present invention, including a data acquisition module, a feature extraction and fusion module, an error correction model establishment module, an error correction model update module, and an error correction module, characterized in that:

[0098] Data acquisition module: Obtain historical working condition parameter data from the historical database; the historical working condition parameter data includes historical electric energy metering values, historical external environment parameters, and historical error parameter data;

[0099] Feature extraction and fusion module: Extract the features of each historical working condition parameter data, screen features and construct cross features according to the relevance and importance of the features of various historical working condition parameter data, and perform standardization processing on the cross features;

[0100] Error correction model establishment module: Establish an error correction model based on the standardized cross features, and update and optimize the error correction model through the working condition parameter data collected in real time within a set number of fusion time recording periods;

[0101] Error correction module: Measure the electric energy measurement value and operating condition parameter data in real time according to the set sampling frequency, input the operating condition parameter data measured at each sampling point into the updated and optimized error correction model to predict the error correction value, and correct the measured electric energy measurement value according to the predicted error correction value.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: It is still possible to modify the specific implementation manners of the present invention or make equivalent replacements, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A method for correcting the electric energy metering error under multiple working conditions of a charging facility, characterized in that It includes the following contents: Obtain historical operating condition parameter data from the historical database; the historical operating condition parameter data includes historical power metering values, historical external environment parameters, and historical error parameter data; Extract the characteristic data of each historical operating condition parameter data, and screen features and construct cross features according to the correlation coefficients and mutual information of the characteristic data of various historical operating condition parameter data, and perform standardization processing on the cross features; Establish an error correction model based on the standardized cross features, and update and optimize the error correction model through the operating condition parameter data collected in real time within a set number of fusion time recording periods; Measure the power metering value and the operating condition parameter data in real time according to the set sampling frequency, input the operating condition parameter data measured at each sampling point into the updated and optimized error correction model to predict the error correction value, and correct the measured power metering value according to the predicted error correction value.

2. The method for correcting the power metering error under multiple operating conditions of a charging facility according to claim 1, wherein: The extraction of the characteristic data of each historical operating condition parameter data is specifically: The characteristic data extracted from the historical power metering value includes charging current, charging voltage, charging power, load rate, total historical metered charging power, and charging efficiency; The characteristic data of the historical external environment parameters includes: ambient temperature, ambient humidity, electromagnetic interference intensity around the charging facility, grid voltage fluctuation, wind speed; The characteristic data of the historical error parameters includes: theoretical charging power, actual measured power, error change rate.

3. The method for correcting the power metering error under multiple operating conditions of a charging facility according to claim 2, wherein: The screening of features and construction of cross features according to the correlation coefficients and mutual information of the characteristic data of various historical operating condition parameter data is specifically: Use a rolling window to calculate the correlation coefficients between all pairs of characteristic data of the historical power metering value, historical external environment parameters, and historical error parameters. If the correlation coefficient between two characteristic data exceeds the set correlation coefficient threshold, delete any one of the characteristic data; For all the characteristic data of the historical power metering value, historical external environment parameters, and historical error parameters after deleting the characteristic data whose correlation coefficient exceeds the set correlation coefficient threshold, calculate the mutual information between each characteristic data of the historical power metering value and historical external environment parameters and each characteristic data of the historical error parameters, and calculate the mutual information between all the characteristic data of the historical power metering value and historical external environment parameters under the condition of knowing the characteristic data of various historical error parameters; The sum of the mutual information between the feature data of a historical electricity metering value and the feature data of a historical external environment parameter and the feature data of a historical error parameter is subtracted by the mutual information between the corresponding historical electricity metering value and the feature data of the historical external environment parameter when the feature data of the corresponding historical error parameter is known, and the result is used as the comprehensive mutual information of the three feature data; the top N comprehensive mutual information values are selected, and the feature data of the historical electricity metering value, the historical external environment parameter, and the historical error parameter corresponding to each selected comprehensive mutual information are fused to form cross features.

4. The method for correcting the electricity metering error under multiple working conditions of a charging facility according to claim 3, wherein: The correlation coefficient threshold is set to 0.

8.

5. The method for correcting the electricity metering error under multiple working conditions of a charging facility according to claim 3, wherein: The fusing of the feature data of the historical electricity metering value, the historical external environment parameter, and the historical error parameter corresponding to each selected comprehensive mutual information to form cross features is specifically as follows: The feature data of the historical electricity metering value, the historical external environment parameter, and the historical error parameter corresponding to each selected comprehensive mutual information are all normalized. The cross feature F is: Among them, exp(·) is the exponential function with base e; E, C, and F are the characteristic data of the normalized historical electricity measurement value, historical external environment parameter, and historical error parameter respectively; tanh(·) is the hyperbolic tangent function; r EC , r CF , r EF are the correlation coefficients of E and C, C and F, and E and F respectively; I ECF is the comprehensive mutual information.

6. The method for correcting the electricity metering error under multiple working conditions of a charging facility according to claim 1, wherein: Establishing an error correction model based on the cross features after standardization processing, and updating and optimizing the error correction model through the working condition parameter data collected in real time within a set number of fusion time recording periods, specifically as follows: Training an error correction model using the cross features after standardization processing and the corresponding true error values as the original training set, and this model is a convolutional neural network; Inputting the working condition parameter data collected in real time within a set number of fusion time recording periods into the error correction model to obtain all predicted error correction values; After each fusion time recording period, calculate all loss functions of each fusion time recording period, and multiply them by the set weight of recent data; And extract the historical working condition parameter data in the original training set that is the same as the working condition parameter data in one fusion time recording period and the corresponding true error values, also calculate all their loss functions, and multiply them by the difference between 1 and the set weight of recent data; add the two weighted loss functions as the final loss function of each fusion time recording period, and update the error correction model.

7. The method for correcting the electricity metering error under multiple working conditions of a charging facility according to claim 6, wherein: The weight of recent data is specifically as follows: The value of the weight of recent data α(1) in the first iteration is the set value, and the weight of recent data α(k) for the iteration number k is: ΔE(k) = |E(k) - E(k - 1)| where γ is the sensitivity parameter; ΔE(k) is the error change rate at the current iteration number k, k ≥ 2; ΔE threshold is the preset error change threshold; E(k) is the average value of all error correction values in the training set at the current iteration number k; E(k - 1) is the average value of all error correction values in the training set at the previous iteration of the current iteration number k.

8. The method for correcting the electricity metering error under multiple working conditions of a charging facility according to claim 1, wherein: The method further includes determining whether there is an abnormality in the operating condition health of the charging facility according to the error correction value, so as to give an early warning. Specifically: modify the set monitoring period, and after obtaining the predicted error correction values of each sampling point within the monitoring period at the end of each monitoring period, calculate the operating condition health index value within each monitoring period according to these error correction values; compare the obtained operating condition health index value with the set health threshold. When the operating condition health index value is greater than the health threshold, it indicates that there is an abnormality in the operating condition and an early warning is given; Calculate the operating condition health index value within the monitoring period, and the formula is: Among them, Q is the value of the working condition health index within the monitoring period; E avg is the mean value of the error correction values of each sampling point within the detection period; γ is the set smoothing parameter; ΔE corr is the variation trend parameter of the error correction value within the monitoring period.

9. The method for correcting the electric energy metering error under multiple operating conditions of a charging facility according to claim 8, characterized in that: The variation trend parameter of the error correction value within the monitoring period is specifically ΔE corr The calculation formula is as follows: Among them, is the cumulative change amount of the error correction value within the monitoring period; n is the total number of sampling points within the monitoring period, and E corr (t) is the error correction value at the t-th time point within the monitoring period. The larger t is, the closer the time point it is located at is to the current time.

10. A system for correcting the electric energy metering error under multiple operating conditions of a charging facility according to the method described in any one of claims 1-9, including a data acquisition module, a feature extraction and fusion module, an error correction model establishment module, an error correction model update module, and an error correction module, characterized in that: Data acquisition module: Obtain historical operating condition parameter data from the historical database; the historical operating condition parameter data includes historical electric energy metering values, historical external environment parameters, and historical error parameter data; Feature extraction and fusion module: Extract the feature data of each historical operating condition parameter data, and screen the features and construct cross features according to the correlation coefficients and mutual information of the feature data of various historical operating condition parameter data, and perform standardization processing on the cross features; Error correction model establishment module: Establish an error correction model according to the standardized cross features, and update and optimize the error correction model through the operating condition parameter data collected in real time within a set number of fusion time recording periods; Error correction module: Measure the electric energy metering value and the operating condition parameter data in real time according to the set sampling frequency, input the measured operating condition parameter data of each sampling point into the updated and optimized error correction model to predict the error correction value, and correct the measured electric energy metering value according to the predicted error correction value.

Images