An improved SVR-based electric vehicle state of charge data completion method

Abstract

The application discloses an electric vehicle state of charge data completion method based on an improved SVR model, and comprises the following steps: 1, cleaning and deleting abnormal data from existing data, and dividing the overall data sample into fast charging data sample and slow charging data sample; 2, predicting the state of charge (SOC) of the battery before charging of the slow charging data sample based on the fast charging data sample and the improved SVR model; 3, correcting the predicted value of the SOC before charging of the slow charging data sample by using physical constraints, and calculating the SOC after charging of the slow charging based on the corrected SOC before charging of the slow charging, so as to complete the state of charge data completion. The application can solve the problems of missing electric vehicle charging state data and difficulty in effective application, and can complete data completion by combining SVR with physical correction, thereby providing support for improving the quality of electric vehicle charging data.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle charging data applications, and specifically to a method for completing electric vehicle state-of-charge data based on an improved SVR. Background Technology

[0002] Influenced by commuting patterns and time-of-use pricing, electric vehicle charging demand exhibits a significant temporal and spatial concentration during evenings and holidays, easily overlapping with residential electricity consumption peaks. Simultaneously, the coexistence of fluctuations in distributed renewable energy output, time-varying price signals, and user travel uncertainties makes the coordinated scheduling of power sources, grids, loads, and storage, as well as load forecasting, more reliant on high-quality charging operation data. In real-world scenarios, slow-charging stations generally lack or have inconsistent battery state of charge (SOC) data before and after charging: on the one hand, the reporting granularity of slow-charging stations and vehicles is low, and measurement channels are diverse; on the other hand, the slow charging process at home or at slow-charging stations is lengthy, with slow power changes and is affected by external factors, resulting in omissions, errors, and noise in SOC records. Existing completion methods often rely on fixed thresholds or simple linear interpolation, making it difficult to characterize nonlinear relationships; some methods directly use end-to-end models, but ignore the physically feasible region, easily leading to anti-physics predictions such as overcharging and over-rate. These problems limit the accuracy and reliability of applications such as load forecasting, settlement verification, and demand response. Summary of the Invention

[0003] To overcome the shortcomings of the existing technologies, this invention proposes a method for completing electric vehicle state-of-charge data based on an improved SVR. The aim is to transfer the nonlinear mapping of the improved SVR model learned on fast charging samples to slow charging samples when the SOC is missing or unreliable. Combined with necessary physical consistency constraints, the method can completely and reliably complete the SOC before and after charging of the slow charging samples, thereby providing support for improving the quality of electric vehicle charging data.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0005] The present invention provides a method for completing the state-of-charge data of electric vehicles based on an improved SVR model, characterized by the following steps:

[0006] Step 1: Construct the dimension as A × Y Fast charging characteristic matrix of electric vehicle users and dimensions D × Y Slow charging characteristic matrix of electric vehicle users ,in, Indicates the first a The first electric vehicle user yThe charging pile features fast charging characteristics, and with the first a electric vehicle users y The fast charging feature sequence of the charging pile is used as the first... a A fast charging sample , Indicates the first d The first electric vehicle user y The charging pile features slow charging characteristics, and with the first d electric vehicle users Y The characteristic sequence of slow charging at the charging pile is used as the first... d A slow charging sample , , , ; A This represents the number of fast-charging samples from electric vehicle users. D This represents the number of slow charging samples from electric vehicle users. Y This indicates the number of features in the electric vehicle user charging sample;

[0007] The construction dimension is 1× A Set of SOC characteristics of electric vehicle users before fast charging and a dimension of 1× A SOC feature set of electric vehicle users after fast charging ,in, This represents the SOC (State of Charge) characteristics of the a-th electric vehicle user before fast charging. This represents the SOC (State of Charge) characteristic of the a-th electric vehicle user after fast charging;

[0008] Step 2: Use the improved SVR model to... To make predictions, thereby obtaining The corresponding predicted SOC feature set before slow charging ,in, Indicates the first d Predicted SOC features before slow charging for individual electric vehicle users;

[0009] Step 3: Based on the capacity upper limit constraint, the minimum SOC constraint under full charge, and the growth rate constraint, respectively... Physical corrections are performed to obtain the complete set of SOC features before slow charging for electric vehicle users. ;in, Indicates the first d SOC characteristics of an electric vehicle user before slow charging after correction;

[0010] based on Calculate the SOC feature set of electric vehicle users after slow charging. ;in, Indicates the first d The SOC characteristics of an electric vehicle user after slow charging are completed.

[0011] The method for completing the state-of-charge data of electric vehicles based on an improved SVR model, as described in this invention, is characterized in that step one is performed as follows:

[0012] Step 1.1: Obtain all The charging power and current sets of each electric vehicle user are preprocessed to obtain the preprocessed charging power feature set. and the preprocessed current feature set ,in, Indicates the first i The preprocessed charging power characteristics of an electric vehicle user Indicates the first i Current characteristics of an electric vehicle user after preprocessing Z This represents the total number of electric vehicle users, and Z = A + D ;

[0013] Get all The remaining electric vehicle users Y - Two charging pile charging datasets were merged and preprocessed to obtain a preprocessed charging pile charging feature set, which was then compared with... and Together constitute Z OK Y Charging pile feature matrix and with the first i electric vehicle users Y The charging characteristic sequence of the charging pile is used as the first charging pile charging feature sequence. i One charging sample ;in, Indicates the first i The first electric vehicle user y Charging characteristics of individual charging stations ;

[0014] Step 1.2, if Current characteristic set , then it means For alternating current characteristics, and If the samples are classified as slow charging samples, Current characteristic set , then it means It is a direct current characteristic, and will The samples are divided into fast charging samples; thus, the fast charging feature matrix is ​​obtained accordingly. and slow charging feature matrix ;

[0015] Will The corresponding set of charging capacity features is denoted as the fast charging capacity feature set. ,in, Indicates the first i The fast charging power characteristics of individual electric vehicle users;

[0016] Will The corresponding set of charging capacity characteristics is denoted as the slow charging capacity characteristic set. ,in, Indicates the first d The charging characteristics of slow charging power consumption for individual electric vehicle users;

[0017] Will middle The corresponding set of slow charging time features is denoted as ,in, Indicates the first d Characteristics of slow charging time for individual electric vehicle users;

[0018] Step 1.3, Obtain The corresponding pre-charging SOC features and post-charging SOC features are used to construct a 1× A Set of SOC characteristics of electric vehicle users before fast charging and a dimension of 1× A SOC feature set of electric vehicle users after fast charging ,in, Indicates the first a The state-of-charge (SOC) characteristics of an electric vehicle user before fast charging. Indicates the first a The state of charge (SOC) characteristics of an electric vehicle user after fast charging.

[0019] Furthermore, step two is carried out as follows:

[0020] Step 2.1, for AX Standardization was performed to obtain the standardized fast charging feature matrix for electric vehicle users. ,in, Indicates the first a The first standardized electric vehicle user y Each charging station features fast charging capabilities and, with XA The Middle a Standardized electric vehicle users Y The fast charging feature sequence of the charging pile is used as the first... a A fast charging sample ;

[0021] based on AX Using equation (1), we obtain the first... d The first standardized electric vehicle user y The characteristics of slow charging at individual charging stations Thus, the standardized slow charging feature matrix of electric vehicle users is obtained. ;

[0022] (1)

[0023] In formula (1): Represents the standardized first generation of all electric vehicle users y The mean of the fast charging characteristic sequence, Represents the standardized first generation of all electric vehicle users y Standard deviation of fast charging characteristic sequences;

[0024] Step 2.2, Calculation XA The Middle y Fast charging feature sequence and FAST The absolute value of the Pearson correlation coefficient This allows us to obtain the mean of all fast charging feature sequences. ;

[0025] Step 2.3, Calculation XA The Middle a A fast charging sample The sum of squares of the Euclidean norms is used to obtain all fast charging samples. The sum of squares of the Euclidean norms and the reciprocal of their mean are used as the global scale. ;

[0026] Step 2.4: Calculate using equation (2) XA The Middle y Scale of fast charging feature sequence :

[0027] (2)

[0028] Step 2.5: Calculate using equation (3) XA The Middle a A fast charging sample With the A fast charging sample kernel matrix values ​​between Thus obtain XA kernel matrix value :

[0029] (3)

[0030] In equation (3), Indicates the first The first standardized electric vehicle user y Each charging station features fast charging capabilities.

[0031] The first number is calculated using equation (4). d A slow charging sample With the a A fast charging sample kernel matrix values ​​between :

[0032] (4)

[0033] Step 2.6: Calculate the first step using equation (5). a A fast charging sample tolerance :

[0034] (5)

[0035] In equation (5): express The fluctuation coefficient, express sparsity coefficient, Indicates the baseline tolerance. Indicates the lower limit of tolerance. Indicates the upper limit of tolerance. Indicates the fusion weights;

[0036] Step 2.7: Construct the objective function using equation (6) And solve for it to obtain The value of is denoted as . ;

[0037] (6)

[0038] In equation (6): and Representing a dimension of 1× A The two dual variables, and , , express The a-th element in express The a-th element; Indicates transpose;

[0039] Step 2.8: Calculate the first step using equation (7). d SOC characteristics of an electric vehicle user before slow charging :

[0040] (7)

[0041] In equation (7): Indicates the baseline offset;

[0042] like Then let , Then let ; thereby obtaining The corresponding predicted SOC feature set before slow charging .

[0043] Furthermore, step three is carried out as follows:

[0044] Step 3.1: Calculate the first step using equation (8). a A fast charging sample Battery capacity Thus, the battery capacity set is obtained. And obtain the minimum battery capacity from it. Maximum battery capacity Average battery capacity :

[0045] (8)

[0046] Step 3.2, using equation (9) to... Perform the first layer of correction to obtain the... d The first-level corrected SOC characteristics of an electric vehicle user before slow charging. :

[0047] (9)

[0048] Step 3.3, using equation (10) to... Perform a second-level correction to obtain the first... d The SOC characteristics of an electric vehicle user before slow charging after second-level correction. :

[0049] (10)

[0050] Step 3.4, using equation (11) to... Perform the third layer correction to obtain the... d SOC characteristics of an electric vehicle user before slow charging. :

[0051] (11)

[0052] Step 3.5: Calculate the first number using equation (12). d SOC characteristics after slow charging for individual electric vehicle users ;

[0053] (12).

[0054] The present invention provides an electronic device, including a memory and a processor, characterized in that the memory is used to store a program that supports the processor in executing the electric vehicle state-of-charge data completion method, and the processor is configured to execute the program stored in the memory.

[0055] The present invention discloses a computer-readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, performs the steps of the electric vehicle state of charge data completion method.

[0056] Compared with existing technologies, the beneficial effects of this invention are reflected in:

[0057] 1. This invention addresses the problems of missing SOC, high noise, and inaccurate data acquisition in charging data. It utilizes an improved SVR model to first learn a nonlinear mapping on data samples with complete information, and then transfers the learning to the missing samples for prediction. This effectively improves the stability and consistency of the completion results, reduces errors, and provides more reliable input for subsequent load forecasting, settlement analysis, and strategy formulation.

[0058] 2. This invention addresses the potential for purely data-driven model predictions to violate physical laws, such as SOC exceeding 1 or unreasonable charging rates. Without altering the core model, it introduces a consistency correction between capacity boundaries and slow charging rate limits, aligning the algorithm output with engineering realities. This correction process uses a combination of charging capacity and duration with capacity statistics as a basis, automatically limiting unreasonable increments and boundary violations, thereby significantly improving data reliability and practical value. Attached Figure Description

[0059] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0060] In this embodiment, a method for completing the state-of-charge data of electric vehicles based on an improved SVR is described, such as... Figure 1 As shown, it includes:

[0061] Step 1: Construct the dimension as A × Y Fast charging characteristic matrix of electric vehicle users and dimensions D ×Y Slow charging characteristic matrix of electric vehicle users ,in, Indicates the first a The first electric vehicle user y The charging pile features fast charging characteristics, and with the first a electric vehicle users y The fast charging feature sequence of the charging pile is used as the first... a A fast charging sample , Indicates the first d The first electric vehicle user y The charging pile features slow charging characteristics, and with the first d electric vehicle users Y The characteristic sequence of slow charging at the charging pile is used as the first... d A slow charging sample , , , ; A This represents the number of fast-charging samples from electric vehicle users. D This represents the number of slow charging samples from electric vehicle users. Y This indicates the number of features in the electric vehicle user charging sample;

[0062] The construction dimension is 1× A Set of SOC characteristics of electric vehicle users before fast charging and a dimension of 1× A SOC feature set of electric vehicle users after fast charging ,in, This represents the SOC (State of Charge) characteristics of the a-th electric vehicle user before fast charging. This represents the SOC (State of Charge) characteristic of the a-th electric vehicle user after fast charging;

[0063] Step 1.1: Obtain all The charging power and current sets of each electric vehicle user are preprocessed to obtain the preprocessed charging power feature set. and the preprocessed current feature set ,in, Indicates the first i The preprocessed charging power characteristics of an electric vehicle user Indicates the first i Current characteristics of an electric vehicle user after preprocessing Z This represents the total number of electric vehicle users, and Z = A + D ;

[0064] Get all The remaining electric vehicle users Y - Two charging pile charging datasets were merged and preprocessed to obtain a preprocessed charging pile charging feature set, which was then compared with... and Together constitute Z OK Y Charging pile feature matrix and with the first i electric vehicle users Y The charging characteristic sequence of the charging pile is used as the first charging pile charging feature sequence. i One charging sample ;in, Indicates the first i The first electric vehicle user y Charging characteristics of individual charging stations .

[0065] Step 1.2, if Current characteristic set , then it means For alternating current characteristics, and If the samples are classified as slow charging samples, Current characteristic set , then it means It is a direct current characteristic, and will The samples are divided into fast charging samples; thus, the fast charging feature matrix is ​​obtained accordingly. and slow charging feature matrix ;

[0066] Will The corresponding set of charging capacity features is denoted as the fast charging capacity feature set. ,in, Indicates the first i The fast charging power characteristics of individual electric vehicle users;

[0067] Will The corresponding set of charging capacity characteristics is denoted as the slow charging capacity characteristic set. ,in, Indicates the first d The charging characteristics of slow charging power consumption for individual electric vehicle users;

[0068] Will middle The corresponding set of slow charging time features is denoted as ,in, Indicates the first d Characteristics of slow charging time for individual electric vehicle users.

[0069] Step 1.3, Obtain The corresponding pre-charging SOC features and post-charging SOC features are used to construct a 1× A Set of SOC characteristics of electric vehicle users before fast charging and a dimension of 1× A SOC feature set of electric vehicle users after fast charging ,in, Indicates the first a The state-of-charge (SOC) characteristics of an electric vehicle user before fast charging. Indicates the first a The state of charge (SOC) characteristics of an electric vehicle user after fast charging.

[0070] Step 2: Use the improved Support Vector Regression (SVR) model to... To make predictions, thereby obtaining The corresponding predicted SOC feature set before slow charging ,in, Indicates the first d The improved SVR model can make the important features in the fast charging samples of electric vehicle users more sensitive and improve the prediction accuracy. At the same time, the nonlinear relationship learned from the fast charging samples of electric vehicle users will be transferred to the slow charging to make up for the data gap of the SOC of electric vehicle users before slow charging.

[0071] Step 2.1, for AX Standardization was performed to obtain the standardized fast charging feature matrix for electric vehicle users. ,in, Indicates the first a The first standardized electric vehicle user y Each charging station features fast charging capabilities and, with XA The Middle a Standardized electric vehicle users Y The fast charging feature sequence of the charging pile is used as the first... a A fast charging sample ;

[0072] based on AX Using equation (1), we obtain the first... d The first standardized electric vehicle user y The characteristics of slow charging at individual charging stations Thus, the standardized slow charging feature matrix of electric vehicle users is obtained. ;

[0073] (1)

[0074] In formula (1): Represents the standardized first generation of all electric vehicle users y The mean of the fast charging characteristic sequence, Represents the standardized first generation of all electric vehicle users y Standard deviation of the fast charging characteristic sequence.

[0075] Step 2.2: Calculate using equation (2) XA The Middle y Fast charging feature sequence and FAST The absolute value of the Pearson correlation coefficient This allows us to obtain the mean of all fast charging feature sequences. :

[0076] (2)

[0077] In formula (2): express FAST The mean;

[0078] Step 2.3, Calculation XA The Middle a A fast charging sample The sum of squares of the Euclidean norms is used to obtain all fast charging samples. The sum of squares of the Euclidean norms and the reciprocal of their mean are used as the global scale. As shown in equation (3):

[0079] (3)

[0080] Step 2.4: Calculate using equation (4) XA The Middle y Scale of fast charging feature sequence :

[0081] (4)

[0082] Step 2.5: Calculate using equation (5) XA The Middle a A fast charging sample With the A fast charging sample kernel matrix values ​​between Thus obtain XA kernel matrix value :

[0083] (5)

[0084] In equation (5), Indicates the first The first standardized electric vehicle user y Each charging station features fast charging capabilities.

[0085] The first number is calculated using equation (6). d A slow charging sample With the a A fast charging sample kernel matrix values ​​between :

[0086] (6)

[0087] Step 2.6: Calculate the first step using equation (7). a A fast charging sample tolerance :

[0088] (7)

[0089] In equation (7): express The fluctuation coefficient, express sparsity coefficient, Indicates the baseline tolerance. Indicates the lower limit of tolerance. Indicates the upper limit of tolerance. This indicates the fusion weight.

[0090] Step 2.7: Construct the objective function using equation (8) And solve for it to obtain The value of is denoted as . ;

[0091] (8)

[0092] In equation (8): and Representing a dimension of 1× A The two dual variables, and , , express The a-th element in express The a-th element; This indicates transpose.

[0093] Step 2.8: Calculate the first step using equation (9). d SOC characteristics of an electric vehicle user before slow charging :

[0094] (9)

[0095] In equation (9): Indicates the baseline offset;

[0096] like Then let , Then let By tightening the upper and lower limits, unrealistic prediction results can be effectively prevented, thereby obtaining... The corresponding predicted SOC feature set before slow charging .

[0097] Step 3: Based on the capacity upper limit constraint, the minimum SOC constraint under full charge, and the growth rate constraint, respectively... Physical corrections are performed to obtain the complete set of SOC features before slow charging for electric vehicle users. ;in, Indicates the first d The corrected SOC characteristics of an electric vehicle user before slow charging; the purpose of physical correction is to correct the predicted SOC characteristics after slow charging in step two, so that the results are more in line with the actual situation and more accurate.

[0098] based on Calculate the SOC feature set of electric vehicle users after slow charging. ;in, Indicates the first d The SOC characteristics of an electric vehicle user after slow charging are completed.

[0099] Step 3.1: Calculate the first step using equation (10). a A fast charging sample Battery capacity Thus, the battery capacity set is obtained. And obtain the minimum battery capacity from it. Maximum battery capacity Average battery capacity :

[0100] (10)

[0101] Step 3.2, using equation (11) to... Perform the first layer of correction to eliminate obvious out-of-bounds errors, and obtain the second layer. d The first-level corrected SOC characteristics of an electric vehicle user before slow charging. :

[0102] (11)

[0103] Step 3.3, using equation (12) to... A second layer of correction is performed to unify the baseline under full-fill conditions, resulting in the third... d The SOC characteristics of an electric vehicle user before slow charging after second-level correction. :

[0104] (12)

[0105] Step 3.4, using equation (13) to... Perform a third-level correction, and based on the rate constraint, suppress the occurrence of non-slow charging results to obtain the... d SOC characteristics of an electric vehicle user before slow charging. :

[0106] (13)

[0107] Step 3.5: Calculate the first... using equation (14). d SOC characteristics after slow charging for individual electric vehicle users ;

[0108] (14)

[0109] By completing the State of Charge (SOC) of electric vehicle users before and after slow charging, the charging status data completion operation is completed. This method can effectively retain a small number of slow charging samples, improve the quality of charging data samples, and provide a foundation for subsequent prediction and scheduling work.

[0110] In this embodiment, an electronic device includes a memory and a processor. The memory stores a program that supports the processor in executing the above-described method, and the processor is configured to execute the program stored in the memory.

[0111] In this embodiment, a computer-readable storage medium stores a computer program, which is executed by a processor to perform the steps of the above method.

[0112] In summary, this invention first addresses the problem of missing charging state values ​​in multi-source heterogeneous charging data. It first divides the data into fast-charging and slow-charging samples based on DC or AC attributes. Secondly, it proposes a SOC feature data completion method based on an improved SVR, outputting the predicted SOC of the slow-charging samples before charging, thus achieving high-precision estimation of the missing SOC feature data. Finally, based on consistency constraints such as capacity limits, minimum SOC at full charge termination, and rate limits, the predicted pre-charging SOC is physically corrected, and the post-charging SOC is calculated by combining the charging amount and capacity, forming paired state completion results. This method ensures algorithm accuracy and robustness while avoiding inverse physics scenarios, providing a reliable data foundation and decision-making basis for improving data quality and subsequent electric vehicle load prediction and grid-side coordinated scheduling.

Claims

1. A method for completing the state-of-charge data of electric vehicles based on an improved SVR model, characterized in that, The procedure is as follows: Step 1: If the charging feature sequence of Y charging piles for the i-th electric vehicle user... Current characteristic set , then it means For alternating current characteristics, and If the samples are classified as slow charging samples, Current characteristic set , then it means It is a direct current characteristic, and will The samples are divided into fast charging samples, thereby constructing a fast charging feature matrix of electric vehicle users with dimension A×Y. and the slow charging feature matrix of electric vehicle users with dimension D×Y. ,in, This represents the fast charging characteristics of the y-th charging pile for the a-th electric vehicle user, and uses the sequence of fast charging characteristics of the y-th charging piles for the a-th electric vehicle user as the fast charging sample for the a-th charging pile. , This represents the slow charging characteristics of the y-th charging pile for the d-th electric vehicle user, and uses the sequence of slow charging characteristics of the y-th charging piles for the d-th electric vehicle user as the d-th slow charging sample. , , , A represents the number of fast charging samples of electric vehicle users, D represents the number of slow charging samples of electric vehicle users, and Y represents the number of features of electric vehicle user charging samples. Construct a 1×A feature set of SOC (State of Charge) features for electric vehicle users before fast charging. And a 1×A set of SOC features for electric vehicle users after fast charging. ,in, This represents the SOC (State of Charge) characteristics of the a-th electric vehicle user before fast charging. This represents the SOC (State of Charge) characteristic of the a-th electric vehicle user after fast charging; Step 2: Use the improved SVR model to... To make predictions, thereby obtaining The corresponding predicted SOC feature set before slow charging ,in, This represents the predicted SOC feature before slow charging for the d-th electric vehicle user. Step 2.1: Standardize AX to obtain the standardized fast charging feature matrix for electric vehicle users. ,in, This represents the standardized fast-charging feature of the y-th charging pile for the a-th electric vehicle user, and uses the standardized fast-charging feature sequence of the y-th charging piles for the a-th electric vehicle user in XA as the fast-charging sample. ; Based on AX, the standardized slow charging characteristics of the y-th charging pile for the d-th electric vehicle user are obtained using equation (1). Thus, the standardized slow charging feature matrix of electric vehicle users is obtained. ; (1) In formula (1): Let represent the mean of the y-th column of the standardized fast charging feature sequence for all electric vehicle users. This represents the standard deviation of the y-th column of the fast charging feature sequence after standardization for all electric vehicle users. Step 2.2: Calculate the absolute value of the Pearson correlation coefficient between the y-th column of the fast charging feature sequence in XA and FAST. This allows us to obtain the mean of all fast charging feature sequences. ; Step 2.3: Calculate the a-th fast charging sample in XA. The sum of squares of the Euclidean norms is used to obtain all fast charging samples. The sum of squares of the Euclidean norms and the reciprocal of their mean are used as the global scale. ; Step 2.4: Calculate the scale of the y-th column of the fast charging feature sequence in XA using equation (2). : (2) Step 2.5: Calculate the a-th fast charging sample in XA using equation (3). With the A fast charging sample kernel matrix values ​​between Thus, the kernel matrix value of XA is obtained. : (3) In equation (3), Indicates the first The fast charging characteristics of the yth charging pile after standardization for electric vehicle users. The d-th slow charging sample is calculated using equation (4). Compared with the a-th fast charging sample kernel matrix values ​​between : (4) Step 2.6: Calculate the a-th fast charging sample using equation (5). tolerance : (5) In equation (5): express The fluctuation coefficient, express sparsity coefficient, Indicates the baseline tolerance. Indicates the lower limit of tolerance. Indicates the upper limit of tolerance. Indicates the fusion weights; Step 2.7: Construct the objective function using equation (6) And solve for it to obtain The value of is denoted as . ; (6) In equation (6): and Let A represent two dual variables of dimension 1×A, and , , express The a-th element in express The a-th element; Indicates transpose; Step 2.8: Calculate the SOC characteristics of the d-th electric vehicle user before slow charging using equation (7). : (7) In equation (7): Indicates the baseline offset; like Then let , Then let ; thereby obtaining The corresponding predicted SOC feature set before slow charging ; Step 3: Based on the capacity upper limit constraint, the minimum SOC constraint under full charge, and the growth rate constraint, respectively... Physical corrections are performed to obtain the complete set of SOC features before slow charging for electric vehicle users. ;in, This represents the corrected SOC (State of Charge) characteristic of the d-th electric vehicle user before slow charging; based on Calculate the SOC feature set of electric vehicle users after slow charging. ;in, This represents the SOC feature after slow charging for the d-th electric vehicle user.

2. The method for completing the state-of-charge data of electric vehicles based on an improved SVR model according to claim 1, characterized in that, Step one is to proceed as follows: Step 1.1: Obtain all The charging power and current sets of each electric vehicle user are preprocessed to obtain the preprocessed charging power feature set. and the preprocessed current feature set ,in, This represents the preprocessed charging capacity characteristics of the i-th electric vehicle user. Let Z represent the preprocessed current characteristics of the i-th electric vehicle user, and let Z represent the total number of electric vehicle users, where Z = A + D. Get all The charging data sets of the remaining Y-2 charging piles for each electric vehicle user are merged and preprocessed to obtain the preprocessed charging pile charging feature set. and Together they form a Z-row, Y-column feature matrix for charging piles And the charging feature sequence of Y charging piles for the i-th electric vehicle user is used as the i-th charging sample. ;in, This represents the charging characteristics of the y-th charging station for the i-th electric vehicle user. ; Step 1.2, The corresponding set of charging capacity features is denoted as the fast charging capacity feature set. ,in, This represents the fast charging power characteristics of the i-th electric vehicle user; Will The corresponding set of charging capacity characteristics is denoted as the slow charging capacity characteristic set. ,in, This represents the slow charging power characteristics of the d-th electric vehicle user; Will middle The corresponding set of slow charging time features is denoted as ,in, This represents the slow charging time characteristic of the d-th electric vehicle user; Step 1.3, Obtain The corresponding pre-charging SOC features and post-charging SOC features are used to construct a 1×A set of pre-charging SOC features for electric vehicle users during fast charging. And a 1×A set of SOC features for electric vehicle users after fast charging. ,in, This represents the SOC (State of Charge) characteristics of the a-th electric vehicle user before fast charging. This represents the SOC (State of Charge) characteristic of the a-th electric vehicle user after fast charging.

3. The method for completing the state-of-charge data of electric vehicles based on an improved SVR model according to claim 2, characterized in that, Step three is to proceed as follows: Step 3.1: Calculate the a-th fast charging sample using equation (8). Battery capacity Thus, the battery capacity set is obtained. And obtain the minimum battery capacity from it. Maximum battery capacity Average battery capacity : (8) Step 3.2, using equation (9) to... Perform the first-level correction to obtain the SOC characteristics of the d-th electric vehicle user before slow charging after the first-level correction. : (9) Step 3.3, using equation (10) to... Perform a second-level correction to obtain the SOC characteristics of the d-th electric vehicle user before slow charging after the second-level correction. : (10) Step 3.4, using equation (11) to... Perform a third-level correction to obtain the SOC feature before slow charging for the d-th electric vehicle user after completion. : (11) Step 3.5: Calculate the SOC characteristics of the d-th electric vehicle user after slow charging using equation (12). ; (12)。 4. An electronic device, comprising a memory and a processor, characterized in that, The memory is used to store a program that supports the processor in executing the electric vehicle state-of-charge data completion method according to any one of claims 1-3, and the processor is configured to execute the program stored in the memory.

5. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is run by the processor, it performs the steps of the electric vehicle state-of-charge data completion method as described in any one of claims 1-3.

Images