Vehicle battery detection method and system based on charging pile, and charging pile

Abstract

The invention provides a vehicle battery detection method and system based on a charging pile and the charging pile, and relates to the technical field of new energy automobiles, and the method comprises the steps: obtaining a vehicle identification code and vehicle charging data based on the communication between a vehicle charging stage and the charging pile; according to the vehicle identification code, obtaining shared charging data of the same type as the vehicle battery; according to the vehicle charging data and the shared charging data, predicting to obtain a predicted SOC value of the vehicle battery; and comparing the predicted SOC value with the read SOC value read at the corresponding moment to obtain an SOC relative estimation error. According to the invention, the accurate prediction of the SOC can be realized, and the estimation error of the battery management system for the SOC can be obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy vehicles, and in particular to a vehicle battery detection method and system based on a charging pile, and a charging pile. Background Art

[0002] Battery testing for new energy vehicles is crucial to ensure the safety and stability of batteries during normal use. The battery management system (BMS) of new energy vehicles will measure the measurable parameters of the battery, such as voltage, current, temperature, etc. in real time during operation, and estimate the unmeasurable internal states, such as the battery state of charge (SOC) and health state (SOH). The battery management system usually combines a variety of methods and technologies, using the physical model of the battery, experimental data and various algorithms to estimate the internal state of the battery, among which SOC is the main state to be estimated, and then other internal states can be estimated through SOC. In related technologies, the estimation of states such as SOC is usually implemented in the battery management system. However, on the one hand, it is difficult for ordinary users to read these state parameters. Manufacturers may only provide some parameters for users to query, and there is a possibility that manufacturers provide false data in order to promote product stability, which makes it difficult for users to intuitively grasp the battery status; on the other hand, the battery management system may be affected by various factors and its own computing power in the estimation, and there is a defect of low estimation accuracy. Summary of the invention

[0003] In order to address the deficiencies in the prior art, the purpose of the present invention is to provide a vehicle battery detection method, system and charging pile based on a charging pile, so as to realize the detection of the power battery at the charging pile end, which helps the user to intuitively understand the battery status.

[0004] In order to achieve the above object, according to some embodiments, a first aspect of the present invention provides a vehicle battery detection method based on a charging pile, comprising:

[0005] Based on the communication between the vehicle and the charging pile during the charging phase, the vehicle identification code and vehicle charging data are obtained;

[0006] According to the vehicle identification code, obtain the shared charging data of the same battery type as the vehicle;

[0007] According to the vehicle charging data and shared charging data, the predicted SOC value of the vehicle battery is predicted;

[0008] The predicted SOC value is compared with the read SOC value at the corresponding moment to obtain the relative SOC estimation error.

[0009] Preferably, a shared data set of the vehicle is constructed based on the vehicle charging data; based on the shared data set of the vehicle and the shared data sets of other vehicles with the same battery type as the vehicle, a SOC prediction model is used to predict the predicted SOC value of the vehicle battery.

[0010] Preferably, the SOC prediction model includes three CNN layers and two MLP layers connected in sequence, which are used to obtain a compact feature representation of the input data and generate a single-value prediction result.

[0011] Preferably, the SOC prediction model is deployed synchronously on the charging pile and the cloud. When the error between the prediction results of the cloud and the charging pile exceeds a set value, the model parameter file on the cloud is updated to the charging pile.

[0012] Preferably, it also includes constructing a vehicle-specific data set based on the vehicle identification code and the vehicle charging data, and predicting the remaining charging time using a charging remaining time prediction model based on the vehicle-specific data set.

[0013] Preferably, clearing the exclusive data set according to the number of charging times specifically includes:

[0014] If the number of charging times is less than the set number and the vehicle is not charged within the first set time, the exclusive data set of the vehicle corresponding to the vehicle identification code is cleared;

[0015] If the charging times exceed the set times and the vehicle is not charged within the second set time, the exclusive data set of the vehicle corresponding to the vehicle identification code is cleared; wherein the second set time is calculated according to the charging times.

[0016] A second aspect of the present invention provides a vehicle battery detection system based on a charging pile, comprising:

[0017] A data acquisition module is configured to acquire a vehicle identification code and vehicle charging data based on communication with a charging pile during a vehicle charging phase;

[0018] A shared charging data acquisition module is configured to acquire shared charging data of the same battery type as the vehicle according to the vehicle identification code;

[0019] A predicted SOC value acquisition module is configured to predict a predicted SOC value of a vehicle battery according to the vehicle charging data and the shared charging data;

[0020] The comparison module is configured to compare the predicted SOC value with the read SOC value read at the corresponding moment to obtain a relative SOC estimation error.

[0021] Preferably, it also includes a charging remaining time prediction module, which is configured to construct a vehicle-specific data set based on the vehicle identification code and the vehicle charging data, and use a charging remaining time prediction model to predict the charging remaining time based on the vehicle-specific data set.

[0022] According to a third aspect of the present invention, a charging pile is provided, comprising a user mainboard having a computer program stored therein, wherein the user mainboard executes the computer program to complete the steps of the above-mentioned vehicle battery detection method based on the charging pile.

[0023] Preferably, the charging pile further includes a display module, and the display module is used to display the remaining charging time predicted by the user mainboard when the charging pile is connected to the vehicle to charge the vehicle.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] The present invention provides a vehicle battery detection method, system and charging pile based on a charging pile. A battery detection function is integrated at the charging pile end. When the vehicle is charging, the battery can be detected according to charging data, and the SOC of the battery can be estimated, so as to evaluate the state and performance of the battery, and then generate a detection report for the user to view.

[0026] The present invention constructs a data set uniquely corresponding to the vehicle identification code based on the vehicle identification code at the charging pile end, including a shared data set and an exclusive data set, wherein the shared data set is used to predict the SOC value, and the data in the shared data set of different vehicles with the same battery type can be shared, thereby improving the accuracy of SOC prediction; the exclusive data set is dedicated to predicting the remaining charging time of the vehicle. As the number of charging times by the user increases, the prediction result will be more accurate, and the remaining charging time of the vehicle will be displayed to the user in real time at the charging pile end, which can facilitate the user to intuitively judge the battery charging status outside the vehicle, thereby improving the user experience.

[0027] The present invention designs an intelligent data management strategy, which determines the user data clearing time according to the number of times the user charges, so as to regularly clear the user data that has been inactive for a long time, thereby improving the operation efficiency.

[0028] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0030] Figure 1Schematic diagram of the method in Embodiment 1 of the present invention;

[0031] Figure 2 It is the charging flow chart;

[0032] Figure 3 It is a schematic diagram of the shared data parameter combination;

[0033] Figure 4 It is a schematic diagram of the SOC prediction model;

[0034] Figure 5 It is a schematic diagram of exclusive data parameter combination;

[0035] Figure 6 This is a schematic diagram of the remaining charging time prediction model;

[0036] Figure 7 It is the flow chart of authentication communication;

[0037] Figure 8 Communication flow chart for parameter configuration stage;

[0038] Fig. 9 This is the communication flow chart between BMS and charging pile during the charging stage;

[0039] Fig.10 This is the communication flow chart for the charging end stage;

[0040] Fig.11 This is a schematic diagram of the overall charging process. DETAILED DESCRIPTION

[0041] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0042] like Figure 1 As shown, the embodiment of the present invention first provides a vehicle battery detection method based on a charging pile, comprising:

[0043] Based on the communication between the vehicle and the charging pile during the charging phase, the vehicle identification code and vehicle charging data are obtained;

[0044] According to the vehicle identification code, obtain the shared charging data of the same battery type as the vehicle;

[0045] According to the vehicle charging data and shared charging data, the predicted SOC value of the vehicle battery is predicted;

[0046] The predicted SOC value is compared with the read SOC value at the corresponding moment to obtain the relative SOC estimation error.

[0047] At present, the industry usually predicts the SOC value by obtaining the initial value of SOC through the open circuit voltage OCV characteristics, and then estimates the SOC using the ampere-hour integration method. The open circuit voltage needs to obtain the voltage of the battery that needs to be stationary for a period of time, and for ternary batteries, the corresponding relationship of OCV is strictly monotonic, and the OCV difference of different SOCs is relatively obvious; however, for lithium iron phosphate batteries, there is a fairly long SOC interval that is very flat, and the same OCV value corresponds to multiple SOC points, so using the OCV-SOC correspondence to estimate the SOC has problems such as long standing time and insufficient accuracy. Inaccurate SOC estimation may cause battery damage, mileage errors, reduced user experience and safety hazards. Therefore, this embodiment proposes to use a neural network model at the charging pile end, and with the convenience of charging pile data collection, extract multiple features of the same type of batteries for accurate prediction, thereby improving the accuracy of SOC prediction.

[0048] The method of this embodiment is applied to the charging pile end. On the basis of the traditional battery detection based on the vehicle's own battery management system BMS, a battery detection function is integrated on the charging pile to evaluate the accuracy of the BMS's estimation of the battery status, and the results are sent to the user so that the user can fully understand the performance and status of the battery. In addition, this embodiment also realizes the prediction of the remaining charging time at the charging pile end.

[0049] In this embodiment, the overall process of charging pile and battery charging is as follows: Figure 2 As shown in the figure, it includes the charging handshake stage, the charging parameter configuration stage, the charging stage and the charging termination stage. The charging stage mainly includes the pre-charging, constant current charging, constant voltage charging and charging termination processes.

[0050] Pre-charging: For batteries that have not been used for a long time or have very low power, a pre-charging stage may be performed first, charging with a smaller current to avoid damage to the battery.

[0051] Constant current charging: The BMS sets a predefined charging current value and the charger starts charging the battery at a constant current. The battery voltage gradually rises, but the charging current remains constant.

[0052] Constant voltage charging: Once the battery voltage reaches the preset threshold, the charging mode switches from constant current to constant voltage. The charger begins to adjust the output to maintain the battery voltage at a level close to the rated voltage while reducing the charging current. During the constant voltage stage, the battery continues to charge until the charging current drops to a preset termination current value.

[0053] Charge termination: When the charge current drops to the termination current threshold, the BMS considers the battery to be nearly full and stops the charging process to prevent overcharging.

[0054] This embodiment is based on the solution provided by the charging pile, including the calculation and evaluation of the state of charge SOC and other parameters based on the charging data of the constant current and constant voltage charging stages after charging is completed, as well as the prediction of the remaining charging time during the charging process. Figure 3-Figure 11 Provide detailed explanation.

[0055] The charging pile has a pre-installed SOC prediction model. Figure 4 As shown, the model in this embodiment adopts a customized convolutional neural network (CNN) architecture, called DATA-CNN, which is designed to effectively capture and predict data patterns. The SOC prediction model consists of 3 CNN layers and 2 MLP layers. The model starts with a convolution layer (conv1) with a 1×1 convolution kernel and 16 output channels. This layer extracts preliminary features from the input data while maintaining the spatial dimension; the features output from the first layer are then processed by the second convolution layer (conv2), which uses a 2×2 convolution kernel, 32 output channels, and a stride of 1 to capture more complex features by reducing the spatial dimension; the final convolution layer (conv3) further refines the feature extraction process based on the input data. The convolution kernel is 2×2 in dimension, with 64 output channels and no padding, generating a compact feature representation; the output of the final convolutional layer is flattened into a vector of size 64×1×1 and input into a fully connected layer (fc1) with 128 neurons, introducing nonlinearity and allowing the network to learn complex patterns; the output of the fully connected layer is then passed to the final output layer (fc2), reducing the dimension to a single value for prediction; the network uses the ReLU activation function between layers to introduce nonlinearity, thereby facilitating the learning of complex relationships in the data. This architecture is designed to balance simplicity and effectiveness, providing a powerful framework for data prediction tasks. The model is used to accurately predict SOC based on shared data.

[0056] The charging pile also has a preset charging remaining time prediction model, such as Figure 6 As shown in the figure, the charging remaining time prediction model also uses DATA-CNN, and the overall structure is similar to the SOC prediction model. The difference is that due to the different input data of the two, the input of the SOC prediction model is 3*3 with a total of 9 features, such as Figure 3 As shown in Figure 2, the input of the remaining charging time prediction model is 4*3 with a total of 12 features, such as Figure 5 As shown, the final convolution layer (conv3) of the charging remaining time prediction model adopts a 3×2 convolution kernel, while the SOC prediction model adopts a 2×2 convolution kernel.

[0057] Specifically, the charging remaining time prediction model consists of 3 CNN layers and 2 MLP layers. The model starts with a convolution layer (conv1) with a 1×1 convolution kernel and 16 output channels. This layer extracts preliminary features from the input data while maintaining the spatial dimension. The features output from the first layer are then processed by the second convolution layer (conv2), which uses a 2×2 convolution kernel, 32 output channels, and a stride of 1 to capture more complex features by reducing the spatial dimension. The final convolution layer (conv3) further refines the feature extraction process and extracts the features based on the input data. The input data is flattened to a 3×2 convolution kernel, 64 output channels, and no padding to generate a compact feature representation; the output of the final convolutional layer is flattened into a vector of size 64×1×1 and input into a fully connected layer (fc1) with 128 neurons, introducing nonlinearity and allowing the network to learn complex patterns; the output of this fully connected layer is then passed to the final output layer (fc2), reducing the dimension to a single value for prediction; the network uses the ReLU activation function between layers to introduce nonlinearity, thereby facilitating the learning of complex relationships in the data.

[0058] A data set is built for each vehicle identification code. The collected data is first cleaned and preprocessed to ensure the quality of the data. Then the data is transmitted to the cloud and integrated into the existing data set of the corresponding identification code. For each identification code, that is, each vehicle, a shared data set and a dedicated data set are constructed, where the shared data set is used to realize SOC prediction, and the dedicated data set is used to realize the remaining charging time prediction. The data in the shared data set includes battery type, charging current, charging voltage, charging power, battery capacity, battery internal resistance, power battery temperature, battery health state (SOH), and the remaining number of charging cycles; the data in the dedicated data set includes battery type, battery charging state, charging current, charging voltage, charging power, battery capacity, state of charge (SOC), power battery temperature, battery health state (SOH), the remaining number of charging cycles, battery internal resistance, and the number of charging cycles that have been performed. Each time charging is performed, the collected data is added to the data set of the corresponding vehicle identification code to enrich the data set and improve the prediction accuracy. For the prediction of SOC, the data in the shared data set of the same type of battery can be used for prediction to further improve the prediction accuracy.

[0059] Battery testing mainly evaluates the performance and status of the battery based on various indicators during the charging process. The most important of these is to evaluate the SOC estimated by the BMS. The predicted SOC value of the battery is obtained based on the SOC prediction model, and compared with the read SOC value from the BMS to obtain the relative estimation error. Different methods are used for new and old charging piles. New charging piles can directly integrate data acquisition functions, SOC prediction, error calculation, and remaining charging time prediction functions into the charging pile motherboard. For old charging piles that have already been built, a CAN reading device can be added to send the read data information to the cloud server for calculation, and the results can be pushed to the user-side APP. By calculating the remaining charging time and comprehensively evaluating the performance and status of the battery, a series of problems caused by inaccurate SOC estimation can be prevented.

[0060] The prediction of the remaining charging time is mainly in the charging stage. According to the vehicle identification code (VI N and EVI N), a unique data set belonging to the vehicle with this identification code is built to accurately predict the remaining charging time.

[0061] Figure 7 The following is the authentication communication flow chart. Authentication generally occurs during the charging handshake startup phase. The electric vehicle will send the vehicle identification code to the charging pile. When the authentication is successful, the charging pile will search / create a shared data set and a dedicated data set unique to the vehicle identification code based on the vehicle identification code, collect information such as the charging type (constant voltage or constant current charging), charging voltage, charging current, power battery temperature, SOC, SOH, and power battery status during charging, and collect the charging time and SOC after charging are completed and save them in the data set as the training features of the neural network, and train to obtain a prediction model unique to the vehicle identification code. As the number of times the vehicle is charged increases, the charging pile will be more accurate in predicting the remaining charging time. In order to save cloud storage space, an intelligent data management strategy is designed for users who do not frequently use charging services to regularly clean up user data that has not been active for a long time. Specifically, the data management strategy is divided into two stages according to the number of times the user charges. The first stage is when the number of times the user charges is lower than the set number. At this time, the number of times the user charges is relatively small, and the data set corresponding to the vehicle is more likely to be cleared. If the vehicle is not charged at the designated charging pile within the first set time (for example, two consecutive weeks), the data set created based on the user's vehicle will be automatically cleared; when the number of times the user charges exceeds the set number, it enters the second stage. In the second stage, there is a specific correspondence between the number of times the user charges and the time to clear the user data. As the number of times the user charges increases, the time interval for clearing the user data (the second set time) increases accordingly, and the data set is less likely to be cleared, and the corresponding vehicle's remaining charging time prediction will be more accurate.

[0062] In the second stage, the user charging times and the second set time have the following functional correspondence:

[0063] T(C)=T0×(1+k) C ;

[0064] Among them, C represents the number of times the user charges, T represents the time interval for clearing user data (the second set time), T0 represents the basic time interval, and the first set time is used here, and k is a constant not greater than 1, representing the growth factor.

[0065] like Fig. 9 As shown in the figure, during the charging process, the BMS of the vehicle will continuously communicate with the charging pile and send the charging information to the charging pile with a message cycle of 250ms. The vehicle is charged according to the preset of the charging pile. During the charging process, the battery management system of the vehicle will cyclically send the total charging status of the battery, the charging demand of the battery, the battery status information, voltage, temperature and SOC and other information. After the charging pile receives this information, it uses this information as the input of the charging prediction model to obtain the predicted remaining charging time in real time. The remaining charging time can be displayed in real time on the charging pile during the charging process and pushed to the user-side APP.

[0066] In addition, in the early stages of charging when the amount of data is relatively small, a parameter file will be trained based on the collected data and deployed locally to the charging pile. The charging pile will use the model to make predictions based on this parameter file, and the cloud will also make predictions synchronously, and then compare them. When the error between the two exceeds 3%, the cloud will update the new model parameter file locally to the charging pile.

[0067] The calculation and evaluation of the state of charge SOC and other parameters are performed after charging is completed. Fig.10 As shown in the figure, when charging is completed, that is, when the charging current reaches the cut-off current, the BMS will send a statistical message of the entire charging process to the charging pile. For the evaluation of SOC and battery performance and indicators, analysis and evaluation are performed based on the communication message between the charging pile and the battery.

[0068] After charging is completed, according to the charging message, a certain time point in the constant current charging stage is selected to predict the SOC. In the constant current charging stage, the battery is charged with a constant current, and the relationship between the battery voltage and the SOC is relatively stable, which is suitable for SOC prediction. The predicted SOC value obtained by using the SOC prediction model is SOC1, and the estimated SOC value of the BMS read at this time is SOC0, so the relative error is:

[0069]

[0070] The relative estimation error of BMS for SOC is obtained. According to the requirements for the cumulative error of SOC estimation of BMS for different types of electric vehicles specified in the Technical Requirements for Battery Management Systems for Electric Vehicles:

[0071] The cumulative error of SOC estimation of lithium-ion power battery management system should not exceed 15%;

[0072] The cumulative error of SOC estimation of the NiMH power battery management system should not exceed 20%;

[0073] Pure electric vehicles and plug-in hybrid electric vehicles: The cumulative error of BMS for SOC estimation should not exceed 5%.

[0074] The prediction of SOC is the same as the prediction of the remaining charging time. In the early stage of charging when the amount of user data is relatively small, a parameter file will be trained based on the collected data and deployed locally to the charging pile. The charging pile will use the model to make predictions based on this parameter file, and the cloud will also make predictions synchronously, and then compare them. When the error between the two exceeds 3%, the cloud will update the new model parameter file to the local charging pile.

[0075] It is determined whether the SOC error estimated by the BMS exceeds a specified value. If the error exceeds the specified value, the BMS error data is sent to the user.

[0076] For the evaluation of other parameters, according to the messages in the charging process, the voltage, current and temperature changes in the whole charging process are analyzed and obtained, mainly for the evaluation of the maximum and minimum values ​​of the parameters; the balance state of the battery cells, the voltage balance of the battery pack, the maximum and minimum values ​​of the temperature values ​​detected by the temperature probes of the rechargeable energy storage subsystem and the fault codes in the charging process are analyzed and compared with the specified values ​​to obtain abnormal information, and then a comprehensive evaluation of the battery stability in the charging process is carried out.

[0077] After the above steps are completed, a battery test report can be generated and pushed based on the evaluation results. A battery test report is generated based on the statistical message after charging and the above evaluation indicators. For new charging piles, the above functions can be integrated into the mainboard; for old charging piles that have already been built, a CAN reading device is added to the electric pile to read the communication message between the charging pile and the BMS.

[0078] The method provided in this embodiment is intended to predict the remaining charging time, evaluate the estimation accuracy of the vehicle's BMS system for the battery's SOC, and evaluate the parameter values ​​of other battery indicators during the charging process. In the normal charging scenario of the vehicle, the remaining charging time can be predicted in real time and the battery performance and status can be comprehensively evaluated. After charging is completed, a report on the battery is generated and sent to the user. The user can view the battery performance and status through the report to achieve a comprehensive evaluation of the vehicle battery.

[0079] The embodiment of the present invention further provides a vehicle battery detection system based on a charging pile, comprising:

[0080] A data acquisition module is configured to acquire a vehicle identification code and vehicle charging data based on communication with a charging pile during a vehicle charging phase;

[0081] A shared charging data acquisition module is configured to acquire shared charging data of the same battery type as the vehicle according to the vehicle identification code;

[0082] A predicted SOC value acquisition module is configured to predict a predicted SOC value of a vehicle battery according to the vehicle charging data and the shared charging data;

[0083] The comparison module is configured to compare the predicted SOC value with the read SOC value read at the corresponding moment to obtain a relative SOC estimation error.

[0084] In some embodiments, a charging remaining time prediction module is also included, which is configured to construct a vehicle-specific data set based on the vehicle identification code and the vehicle charging data, and use a charging remaining time prediction model to predict the charging remaining time based on the vehicle-specific data set.

[0085] It should be noted here that the various modules in this embodiment correspond one-to-one to the steps of the aforementioned method embodiment, and the specific implementation process is the same, which will not be repeated here.

[0086] Some embodiments further provide a charging pile, the charging pile comprising a user mainboard, the user mainboard storing a computer program, the user mainboard executing the computer program to complete the steps of the above-mentioned vehicle battery detection method based on the charging pile. The charging pile also includes a display module, the display module is used to display the remaining charging time predicted by the user mainboard when the charging pile is connected to the vehicle to charge the vehicle.

[0087] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A vehicle battery detection method based on a charging pile, characterized in that: include: Based on the communication between the vehicle and the charging pile during the charging phase, the vehicle identification code and vehicle charging data are obtained; According to the vehicle identification code, obtain the shared charging data of the same battery type as the vehicle; The predicted SOC value of the vehicle battery is predicted based on the vehicle charging data and the shared charging data; The predicted SOC value is compared with the read SOC value at the corresponding moment to obtain the relative SOC estimation error.

2. A vehicle battery detection method based on a charging pile as claimed in claim 1, characterized in that: A shared data set of the vehicle is constructed based on the vehicle charging data; based on the shared data set of the vehicle and the shared data sets of other vehicles with the same battery type as the vehicle, a predicted SOC value of the vehicle battery is predicted using a SOC prediction model.

3. A vehicle battery detection method based on a charging pile as claimed in claim 2, characterized in that: The SOC prediction model includes three CNN layers and two MLP layers connected in sequence, which are used to obtain a compact feature representation of the input data and generate a single-value prediction result.

4. A vehicle battery detection method based on a charging pile as claimed in claim 2, characterized in that: The SOC prediction model is deployed synchronously on the charging pile and the cloud. When the error between the prediction results of the cloud and the charging pile exceeds the set value, the model parameter file on the cloud is updated to the charging pile.

5. A vehicle battery detection method based on a charging pile as claimed in claim 1, characterized in that: It also includes constructing a vehicle-specific data set based on the vehicle identification code and vehicle charging data, and predicting the remaining charging time using a charging remaining time prediction model based on the vehicle-specific data set.

6. A vehicle battery detection method based on a charging pile as claimed in claim 5, characterized in that: Clear exclusive data sets based on charge times, including: If the number of charging times is less than the set number and the vehicle is not charged within the first set time, the exclusive data set of the vehicle corresponding to the vehicle identification code is cleared; If the charging times exceed the set times and the vehicle is not charged within the second set time, the exclusive data set of the vehicle corresponding to the vehicle identification code is cleared; wherein the second set time is calculated according to the charging times.

7. A vehicle battery detection system based on a charging pile, characterized in that: include: A data acquisition module is configured to acquire a vehicle identification code and vehicle charging data based on communication with a charging pile during a vehicle charging phase; A shared charging data acquisition module is configured to acquire shared charging data of the same battery type as the vehicle according to the vehicle identification code; A predicted SOC value acquisition module is configured to predict a predicted SOC value of a vehicle battery according to the vehicle charging data and the shared charging data; The comparison module is configured to compare the predicted SOC value with the read SOC value read at the corresponding moment to obtain a relative SOC estimation error.

8. A vehicle battery detection system based on a charging pile as claimed in claim 7, characterized in that: It also includes a charging remaining time prediction module, which is configured to construct a vehicle-specific data set based on the vehicle identification code and vehicle charging data, and use a charging remaining time prediction model to predict the charging remaining time based on the vehicle-specific data set.

9. A charging pile, characterized in that: The charging pile includes a user mainboard, in which a computer program is stored. The user mainboard executes the computer program to complete the steps of the method as described in any one of claims 1 to 6.

10. A charging pile as claimed in claim 9, characterized in that: The charging pile also includes a display module, which is used to display the remaining charging time predicted by the user motherboard when the charging pile is connected to the vehicle to charge the vehicle.

Images