Charging pile communication self-repairing control system and self-repairing method

Abstract

The invention relates to the technical field of charging pile communication control systems, in particular to a charging pile communication self-repairing control system and a self-repairing method, and the charging pile communication self-repairing control system is based on a CNN and LSTM deep learning model. Local features (such as an abnormal data packet format) and a time sequence dependency relationship (such as periodic interference) of communication data can be captured at the same time, the diagnosis accuracy is improved by 30%-50% compared with that of a traditional method, and misjudgment and missed judgment are remarkably reduced; secondly, through the fault self-repairing strategy module, the system can automatically reload configuration according to software protocol errors and adjust communication parameters according to electromagnetic interference, the average repairing time is shortened to be within 30 minutes from 4 hours, and the requirement for manual intervention is greatly reduced. Secondly, the deep learning model supports online updating, normal and abnormal modes under a new communication protocol can be learned, and it is ensured that the system still keeps high stability during protocol iteration or environment change.

Description

Technical Field

[0001] The present invention relates to the technical field of charging pile communication control systems, and in particular, to a charging pile communication self-repair control system and a self-repair method. Background Art

[0002] During the charging process of electric vehicles, the stability of vehicle-pile communication directly affects the charging efficiency and user experience. In the prior art, the following problems generally exist in the vehicle-pile communication system of charging piles: First, the fault detection method relies on simple communication status judgment (such as timeout retry), and it is difficult to accurately identify the fault type (such as line short circuit, protocol conflict, or electromagnetic interference); Second, the fault repair highly depends on manual troubleshooting. Maintenance personnel need to check the hardware connection, protocol configuration, etc. one by one, with an average time consumption of up to 4 hours, resulting in the long-term shutdown of charging piles and high operation and maintenance costs. In addition, the traditional system lacks adaptability and cannot adapt to the update of diverse communication protocols, resulting in frequent new protocol compatibility problems. With the diversification of vehicle-pile communication protocols, the limitations of the existing system become more prominent. Summary of the Invention

[0003] The purpose of the present invention is to provide a charging pile communication self-repair control system and a self-repair method to improve the problem that the above traditional system lacks adaptability and cannot adapt to the update of diverse communication protocols, resulting in frequent new protocol compatibility problems.

[0004] To achieve the above purpose, the embodiments of the present application provide the following technical solutions: On the one hand, the embodiments of the present application provide a charging pile communication self-repair control system, and the system includes: a data acquisition module, which is used to collect communication data in real time through sensors deployed at the communication interface between the charging pile and the electric vehicle. The communication data includes communication status, data transmission rate, the number of error data packets, and communication protocol interaction information, and performs cleaning and normalization processing on the communication data; a fault diagnosis module, which is composed of a convolutional neural network and a long short-term memory network. The convolutional neural network is used to extract local features of the preprocessed communication data, and the long short-term memory network is used to analyze the time series dependence relationship and output the fault type diagnosis result; a fault self-repair module, which performs repair operations according to the fault type diagnosis result: if it is a software communication fault, it automatically reloads the communication protocol stack or adjusts the protocol parameter configuration; if it is a hardware communication fault, it triggers an alarm and switches to the standby communication line; if it is an electromagnetic interference fault, it adjusts the communication line or enhances the anti-interference measures.

[0005] Optionally, the convolutional neural network in the fault diagnosis model includes 3 convolutional layers and 3 pooling layers, which are used to extract local features of different scales, and the long short-term memory network includes 2 layers of networks, which are used to capture long-term time series dependence relationships; In response to protocol conflict faults, the fault self-repair module automatically compares the differences between the current communication protocol and the preset protocol library, and dynamically updates the protocol configuration parameters.

[0006] Optionally, it further includes a communication coordination module, which is used to coordinate the operation processes of the data acquisition and preprocessing module, the deep learning fault diagnosis model, and the fault self-repair strategy module, communicate with the charging control module of the charging pile and the management platform, upload fault information in real time and receive remote instructions, and send instructions to the charging control module to reduce the charging power when the communication is unstable until the communication resumes stability.

[0007] In a second aspect, this embodiment provides a self-repair method based on a charging pile communication self-repair control system. The method includes: In response to a user order instruction, the data acquisition module sends a communication request to the BMS of the current new energy vehicle, so that the BMS feeds back the vehicle information of the current new energy vehicle. The vehicle information includes the vehicle model and the vehicle battery information. The data acquisition module sends the vehicle information to the fault diagnosis module, so that the fault diagnosis module detects whether there is a charging power curve consistent with the vehicle model in the charging power preset operation library. If not, it determines that there is a communication anomaly with the BMS, generates a communication anomaly instruction that the vehicle model cannot be matched, and sends the communication anomaly instruction to the fault self-repair module, so that the fault diagnosis model retrieves the vehicle battery information and generates a charging power curve matching the vehicle model based on the vehicle battery information; and establishes a new communication protocol with the BMS based on the charging power curve, so that the charging pile charges the new energy vehicle according to the charging power curve.

[0008] Optionally, generating a charging power curve matching the vehicle model based on the vehicle battery information includes: Analyze the vehicle battery information to obtain the single-cell voltage, single-cell temperature, and battery type, and based on the pressure difference between the single-cell voltage and the standard voltage, construct a pressure difference distribution map corresponding to the battery module. Determine the thermal resistance coefficient of the single cell based on the battery type. The thermal resistance coefficient is provided by the battery supplier of the new energy vehicle and is burned into the vehicle battery information. Sort multiple single-cell voltages based on the voltage value size, and then construct a voltage amplitude curve. Select multiple basic voltage cells by framing with a preset fluctuation slope. The starting point of the fluctuation slope coincides with the starting point of the voltage amplitude curve, and starting from the starting point of the voltage amplitude curve, select multiple consecutive single-cell voltages with voltage values below the set fluctuation slope, and mark the selected single-cell voltages as basic voltage cells. Calculate the first average voltage of multiple basic voltage monomers, construct multiple voltage threshold ranges based on the difference between the first average voltage and the standard voltage, and construct a monomer distribution population corresponding to each voltage threshold range in the differential pressure distribution map; Construct a core aggregation area in each monomer distribution population based on a distance-based clustering algorithm. The core aggregation areas corresponding to multiple monomer distribution populations do not overlap with each other. Calculate the second average voltage corresponding to each core aggregation area, and calculate the difference from the standard voltage, the number of monomers in the core aggregation area, and the monomer distribution shape in the core aggregation area, and construct a time-heat accumulation map of the core aggregation area during the charging process; Construct a temperature cloud map based on the monomer temperature, and superimpose the time-heat accumulation map on the temperature cloud map to obtain the first battery heat change cloud map of the battery module during the charging process. Calculate the dynamic heat dissipation distribution map of the first battery heat change cloud map based on the current ambient temperature and battery type. Obtain the second battery heat change cloud map of the battery module during the charging process by superimposing the first battery heat change cloud map and the corresponding dynamic heat dissipation distribution map. Develop a corresponding charging power curve based on the second battery heat change cloud map, and input the vehicle model, charging power curve, differential pressure distribution map, and ambient temperature into the big data storage server as training data for the neural network model for later charging power matching.

[0009] Optionally, calculating the dynamic heat dissipation distribution map of the first battery heat change cloud map based on the current ambient temperature and battery type includes: Construct an ambient temperature field based on the current ambient temperature, and perform normalization processing on the first battery heat change cloud map and the ambient temperature field in sequence; Extract the spatio-temporal local features of the first battery heat change cloud map after normalization processing through two 3D convolutional layers, and capture the global dependence relationship in the time dimension through a multi-head self-attention mechanism to generate time series enhanced features; Bilinearly interpolate the normalized ambient temperature field to the resolution corresponding to the first battery heat change cloud map, and perform channel splicing with the time series enhanced features to form a fusion feature; Input the fusion feature into a preset physical constraint model, and then generate the corresponding dynamic heat dissipation distribution map according to the time series.

[0010] The physical constraint model constructs the heat dissipation amount of a local area in the following way: Extract the ambient temperature channel in the fusion feature, generate a temperature sensitivity coefficient matrix, and multiply the temperature sensitivity coefficient matrix and the convolutional kernel weights element by element to obtain the initial heat dissipation amount; Extract the battery temperature in the fusion features, adjust the direction consistency threshold based on the temperature difference between the battery temperature and the ambient temperature, calculate the heat dissipation amount after gradient correction through a loss function based on the direction consistency threshold, then construct multiple heat dissipation distribution maps frame by frame based on the heat dissipation amount after gradient correction, and further form a dynamic heat dissipation distribution map.

[0011] Optionally, formulate a corresponding charging power curve based on the second battery heat change cloud map, including: Construct a first safety temperature reference surface based on a preset safe charging temperature threshold, and reduce the preset charging power at a corresponding moment based on the volume of the part of the second battery heat change cloud map protruding from the first safety temperature reference surface at a certain moment, so as to obtain a corrected preset charging power, and further reduce the entire corrected preset charging power based on the current voltage consistency fluctuation map to obtain a charging power curve.

[0012] Optionally, the construction method of the voltage consistency fluctuation map is: Calculate the basic voltage through a weighted algorithm based on the mean voltage and the number of monomers of a monomer distribution population corresponding to the lowest voltage threshold range and the mean voltage corresponding to the pressure difference distribution map, and then construct a basic voltage base; Fit the basic voltage base to the pressure difference distribution map, so that the part of the basic voltage base is deducted from the pressure difference distribution map to obtain a voltage consistency fluctuation map, and the voltage consistency fluctuation map is used to characterize the voltage floating conditions of multiple monomers in the battery module.

[0013] In a third aspect, an embodiment of the present application provides a partial discharge type identification device for high-voltage electrical equipment, and the device includes a memory and a processor.

[0014] The memory is used to store a computer program; the processor is used to implement the steps of the above-mentioned charging pile communication self-repair control system when executing the computer program.

[0015] In a fourth aspect, an embodiment of the present application provides a medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned charging pile communication self-repair control system are implemented.

[0016] The beneficial effects of the present invention are: The charging pile communication self-repair control system of the present invention, based on the deep learning model of CNN and LSTM, can simultaneously capture the local features of communication data (such as abnormal data packet formats) and temporal dependence relationships (such as periodic interference), and the diagnostic accuracy rate is increased by 30% to 50% compared with traditional methods, significantly reducing misjudgments and missed judgments; Secondly, through the fault self-repair strategy module, the system can automatically reload the configuration for software protocol errors, adjust communication parameters for electromagnetic interference, shorten the mean time to repair from 4 hours to within 30 minutes, and significantly reduce the need for manual intervention.

[0017] Secondly, the deep learning model supports online updates and can learn normal and abnormal patterns under new communication protocols, ensuring high stability of the system even when the protocol iterates or the environment changes.

[0018] Other features and advantages of the present invention will be described in the subsequent specification, and in part, will be obvious from the specification, or can be understood by implementing the embodiments of the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in the written specification, claims, and drawings. Brief Description of the Drawings

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0020] Figure 1 It is a schematic flow chart of a self-repair method for charging pile communication faults described in the embodiments of the present invention. Detailed Embodiments

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and illustrated in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0022] It should be noted that similar reference numerals or letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the present invention, terms such as "first" and "second" are only used for differential description and cannot be understood as indicating or implying relative importance.

[0023] Embodiment 1: This embodiment provides a charging pile communication self-repair control system, and the system includes: A data acquisition module, which is used to collect communication data in real time through sensors deployed at the communication interface between the charging pile and the electric vehicle. The communication data includes communication status, data transmission rate, the number of error data packets, and communication protocol interaction information, and performs cleaning and normalization processing on the communication data; A fault diagnosis module, which is composed of a convolutional neural network and a long short-term memory network. The convolutional neural network is used to extract local features of the preprocessed communication data, and the long short-term memory network is used to analyze time series dependence relationships and output the diagnosis result of the fault type. The convolutional neural network in the fault diagnosis model includes 3 convolutional layers and 3 pooling layers, which are used to extract local features of different scales. The long short-term memory network includes 2 layers of networks, which are used to capture long-term time series dependence relationships. Secondly, the fault self-repair module automatically compares the differences between the current communication protocol and the preset protocol library for protocol conflict faults, and dynamically updates the protocol configuration parameters (for example, in the case of no complete charging protocol for the corresponding vehicle model, constructing a corresponding charging power curve based on specific parameters of the vehicle battery, etc.); A fault self-repair module, which performs repair operations according to the fault type diagnosis result: if it is a software communication fault, it automatically reloads the communication protocol stack or adjusts the protocol parameter configuration; if it is a hardware communication fault, it triggers an alarm and switches to a backup communication line; if it is an electromagnetic interference fault, it adjusts the communication line or enhances anti-interference measures; A communication coordination module, which is used to coordinate the operation processes of the data acquisition and preprocessing module, the deep learning fault diagnosis model, and the fault self-repair strategy module, and communicate with the charging control module and the management platform of the charging pile, upload fault information in real time and receive remote instructions, and when the communication is unstable, send instructions to the charging control module to reduce the charging power until the communication resumes stability.

[0024] For the charging pile communication self-repair control system described in this embodiment, on the one hand, based on the deep learning model of CNN and LSTM, it can simultaneously capture the local features of communication data (such as abnormal data packet formats) and time series dependence relationships (such as periodic interference), and the diagnostic accuracy is improved by 30% - 50% compared with traditional methods, significantly reducing misjudgment and missed judgment. Secondly, through the fault self-repair strategy module, the system can automatically reload the configuration for software protocol errors and adjust communication parameters for electromagnetic interference, shortening the average repair time from 4 hours to within 30 minutes, greatly reducing the need for manual intervention. At the same time, the deep learning model supports online update and can learn normal and abnormal patterns under new communication protocols, ensuring that the system still maintains high stability when the protocol iterates or the environment changes.

[0025] Embodiment 2: As Figure 1As shown in the figure, this embodiment is based on Embodiment 1 and is used to elaborate a specific self-repair method for communication faults, including: Before elaborating on the principle, a brief introduction to the background is given: New energy vehicle models are updated extremely fast, and the corresponding charging protocols are also diverse, which increases the adaptation difficulty of general-purpose charging piles. This embodiment provides a method for constructing a new charging power curve based on the basic battery module information sent by the BMS in the absence of a complete charging protocol for the corresponding vehicle model (lacking a preset charging power curve), thereby repairing communication faults.

[0026] Step S100: In response to a user order instruction, the data acquisition module sends a communication request to the BMS of the current new energy vehicle, so that the BMS feeds back the vehicle information of the current new energy vehicle. The vehicle information includes the vehicle model and the vehicle battery information. Step S200: The data acquisition module sends the vehicle information to the fault diagnosis module, so that the fault diagnosis module detects whether there is a charging power curve consistent with the vehicle model in the charging power preset operation library. If not, it is determined that there is a communication anomaly with the BMS, and a communication anomaly instruction with the vehicle model not matching is generated, and the communication anomaly instruction is sent to the fault self-repair module, so that the fault diagnosis model retrieves the vehicle battery information and generates a charging power curve matching the vehicle model based on the vehicle battery information; and a new communication protocol is established with the BMS based on the charging power curve, so that the charging pile charges the new energy vehicle according to the charging power curve.

[0027] The generation of the charging power curve matching the vehicle model in Step S200 includes: Step S210: Analyze the vehicle battery information to obtain the single-cell voltage, single-cell temperature, and battery type, and based on the pressure difference between the single-cell voltage and the standard voltage, construct a pressure difference distribution map corresponding to the battery module (a three-dimensional map composed of multiple pressure difference columns distributed in a matrix), and determine the thermal resistance coefficient of the single cell based on the battery type. The thermal resistance coefficient is provided by the battery supplier of the new energy vehicle and is burned into the vehicle battery information for later calculation of the heat generation of the single cell during charging. Step S220: Sort multiple single - cell voltages based on the magnitude of the voltage values, and then construct a voltage amplitude curve. Select multiple basic voltage cells by enclosing them within a preset fluctuation slope. The starting point of the fluctuation slope coincides with the starting point of the voltage amplitude curve. Starting from the starting point of the voltage amplitude curve, select multiple consecutive single - cell voltages whose values are below the preset fluctuation slope, and mark the selected single - cell voltages as basic voltage cells. The basic voltage cells are those with lower voltage values and are used as the lower limit of the voltage threshold range for subsequent population division. Different from the conventional method of using the lowest single - cell voltage as the lower limit of the voltage threshold range, this method ensures the quantity basis of low - voltage cells and makes the division of multiple voltage threshold ranges more reasonable; Step S230: Calculate the first average voltage of multiple basic voltage cells, construct multiple voltage threshold ranges based on the difference between the first average voltage and the standard voltage, and construct a single - cell distribution population corresponding to each voltage threshold range in the differential pressure distribution diagram; Step S240: Construct a core aggregation area in each single - cell distribution population based on a distance - based clustering algorithm (in actual use, the performance degradation of batteries in a battery module shows a regional distribution. That is, after a battery's performance is abnormal, its surrounding batteries will also be affected in the long - term use process, resulting in a certain degree of performance degradation, which appears as multiple peak groups in the differential pressure distribution diagram). The core aggregation areas corresponding to multiple single - cell distribution populations do not overlap with each other. Calculate the second average voltage corresponding to each core aggregation area, and construct a time - heat accumulation diagram of the core aggregation area during the charging process based on the difference between the second average voltage and the standard voltage, the number of single cells in the core aggregation area, and the shape of the single - cell distribution in the core aggregation area; Compared with the conventional construction of the heat accumulation of each battery cell during charging, in this embodiment, the instantaneous heat generation is calculated based on the average voltage and thermal resistance coefficient of the core aggregation area through Ohm's law, and the cumulative heat is calculated through time integration to obtain the time - heat accumulation corresponding to this area. Compared with the conventional method of calculating the time - heat accumulation diagram corresponding to the entire battery module based on a single cell, the method described in this embodiment greatly reduces the computing power requirement and shortens the abnormal repair response time of the charging pile; Step S250: Construct a temperature cloud map based on the monomer temperature, and overlay the time-heat accumulation map on the temperature cloud map, thereby obtaining the first battery heat change cloud map during the charging process of the battery module. Calculate the dynamic heat dissipation distribution map of the first battery heat change cloud map based on the current ambient temperature and battery type. Obtain the second battery heat change cloud map during the charging process of the battery module by overlaying the first battery heat change cloud map and the corresponding dynamic heat dissipation distribution map. Develop a corresponding charging power curve based on the second battery heat change cloud map, and input the vehicle model, charging power curve, differential pressure distribution map, and ambient temperature into the big data storage server as training data for the neural network model for later charging power matching.

[0028] Secondly, calculating the dynamic heat dissipation distribution map of the first battery heat change cloud map based on the current ambient temperature and battery type in step S250 includes: Step S251: Construct an ambient temperature field based on the current ambient temperature, and perform normalization processing on the first battery heat change cloud map and the ambient temperature field in sequence to facilitate feature recognition and extraction by the later convolutional neural network. Step S252: Extract the spatio-temporal local features of the first battery heat change cloud map after normalization processing through two 3D convolutional layers, and capture the global dependency relationship in the time dimension through the multi-head self-attention mechanism to generate time series enhanced features. Step S253: Bilinearly interpolate the normalized ambient temperature field to the resolution corresponding to the first battery heat change cloud map (the pixels in the resolution correspond to facilitate later channel splicing. It should be noted that the resolution of the ambient temperature field here is not exactly the same as that of the first battery heat change cloud map, and there is a certain scaling ratio between them, that is, the resolution of the first battery heat change cloud map is an integer multiple of the ambient temperature field - preferably 4 times), so as to splice the channels of the time series enhanced features to form a fused feature. Step S254: Input the fused feature into a preset physical constraint model, and then generate the corresponding dynamic heat dissipation distribution map according to the time series.

[0029] The physical constraint model constructs the heat dissipation amount of the local area in the following way: Step S2541: Extract the ambient temperature channel in the fused feature, generate a temperature sensitivity coefficient matrix, and multiply the temperature sensitivity coefficient matrix and the convolutional kernel weights element by element to obtain the initial heat dissipation amount. Step S2542: Extract the battery temperature in the fused feature, adjust the heat loss direction consistency threshold based on the temperature difference between the battery temperature and the ambient temperature, calculate the gradient-corrected heat dissipation amount through the loss function based on the direction consistency threshold, and then construct multiple heat dissipation distribution maps frame by frame based on the gradient-corrected heat dissipation amount to form a dynamic heat dissipation distribution map.

[0030] Formulating a corresponding charging power curve based on the second battery heat change cloud map described in step S250 includes: Step S255: Construct a first safety temperature reference surface based on a preset safe charging temperature threshold, and reduce the preset charging power at a corresponding moment based on the volume of the part of the second battery heat change cloud map protruding from the first safety temperature reference surface at a certain moment, so as to obtain a corrected preset charging power, and further reduce the entire corrected preset charging power based on the current voltage consistency fluctuation map to obtain a charging power curve.

[0031] Among them, the construction method of the voltage consistency fluctuation map described in step S255 is: Step S2551: Calculate a basic voltage through a weighted algorithm by combining the mean voltage and the number of monomers of a monomer distribution population with the lowest voltage threshold range and the mean voltage corresponding to the differential pressure distribution map, and then construct a basic voltage base; Step S2552: Fit the basic voltage base to the differential pressure distribution map so that the differential pressure distribution map deducts the part of the basic voltage base to obtain a voltage consistency fluctuation map, and the voltage consistency fluctuation map is used to characterize the voltage floating conditions of multiple monomers in the battery module.

[0032] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A communication self-repair control system for a charging pile, characterized in that, The system includes: A data acquisition module, which is used to collect communication data in real time through sensors deployed at the communication interfaces between charging piles and electric vehicles. The communication data includes communication status, data transmission rate, the number of error data packets, and communication protocol interaction information, and performs cleaning and normalization processing on the communication data; A fault diagnosis module, which is composed of a fusion of a convolutional neural network and a long short-term memory network. The convolutional neural network is used to extract local features of the preprocessed communication data, and the long short-term memory network is used to analyze time series dependence relationships and output fault type diagnosis results; A fault self-repair module, which performs repair operations according to the fault type diagnosis results: if it is a software communication fault, it automatically reloads the communication protocol stack or adjusts protocol parameter configurations; if it is a hardware communication fault, it triggers an alarm and switches to a backup communication line; if it is an electromagnetic interference fault, it adjusts the communication line or enhances anti-interference measures.

2. The charging pile communication self-repair control system according to claim 1, wherein The convolutional neural network in the fault diagnosis model includes 3 convolutional layers and 3 pooling layers, which are used to extract local features of different scales, and the long short-term memory network includes 2 layers of networks, which are used to capture long-term time series dependence relationships; For protocol conflict faults, the fault self-repair module automatically compares the differences between the current communication protocol and the preset protocol library and dynamically updates the protocol configuration parameters.

3. The charging pile communication self-repair control system according to claim 1, wherein, It also includes a communication coordination module, which is used to coordinate the operation processes of the data acquisition and preprocessing module, the deep learning fault diagnosis model, and the fault self-repair strategy module, communicate with the charging control module of the charging pile and the management platform, upload fault information in real time and receive remote instructions, and when the communication is unstable, send instructions to the charging control module to reduce the charging power until the communication returns to stability.

4. A self-repair method for the charging pile communication self-repair control system according to claim 2, characterized in that, The method includes: In response to a user order instruction, the data acquisition module sends a communication request to the BMS of the current new energy vehicle, and then the BMS feeds back the vehicle information of the current new energy vehicle. The vehicle information includes vehicle model and vehicle battery information; The data acquisition module sends the vehicle information to the fault diagnosis module, so that the fault diagnosis module detects whether there is a charging power curve consistent with the vehicle model in the charging power preset operation library. If not, it determines that there is a communication anomaly with the BMS, generates a communication anomaly instruction with an unmatched vehicle model, and sends the communication anomaly instruction to the fault self-repair module, so that the fault diagnosis model retrieves the vehicle battery information and generates a charging power curve matching the vehicle model based on the vehicle battery information; and establishes a new communication protocol with the BMS based on the charging power curve, so that the charging pile charges the new energy vehicle according to the charging power curve.

5. The self-healing method according to claim 4, characterized in that, Generating a charging power curve matching the vehicle model based on the vehicle battery information includes: Analyzing the vehicle battery information to obtain the single-cell voltage, single-cell temperature, and battery type, constructing a differential pressure distribution map corresponding to the battery module based on the pressure difference between the single-cell voltage and the standard voltage, determining the thermal resistance coefficient of the single cell based on the battery type, and the thermal resistance coefficient is provided by the battery supplier of the new energy vehicle and burned into the vehicle battery information; Sort multiple single-cell voltages based on the magnitude of the voltage value, and then construct a voltage amplitude curve. Based on a preset fluctuation slope, select multiple basic voltage single cells. The starting point of the fluctuation slope coincides with the starting point of the voltage amplitude curve. Starting from the starting point of the voltage amplitude curve, select multiple single-cell voltages with consecutive voltage values below the preset fluctuation slope, and mark the selected single-cell voltages as basic voltage single cells; Calculate the first average voltage of multiple basic voltage single cells, construct multiple voltage threshold ranges based on the difference between the first average voltage and the standard voltage, and construct a single-cell distribution population corresponding to each voltage threshold range in the differential pressure distribution map; Construct a core aggregation area in each single-cell distribution population based on a distance-based clustering algorithm. The core aggregation areas corresponding to multiple single-cell distribution populations do not overlap with each other. Calculate the second average voltage corresponding to each core aggregation area, and construct a time-heat accumulation map of the core aggregation area during the charging process based on the difference between the second average voltage and the standard voltage, the number of single cells in the core aggregation area, and the shape of the single-cell distribution in the core aggregation area; Construct a temperature cloud map based on the single-cell temperature, and overlay the time-heat accumulation map on the temperature cloud map to obtain the first battery heat change cloud map of the battery module during the charging process. Calculate the dynamic heat dissipation distribution map of the first battery heat change cloud map based on the current ambient temperature and battery type. Obtain the second battery heat change cloud map of the battery module during the charging process by overlaying the first battery heat change cloud map and the corresponding dynamic heat dissipation distribution map. Develop a corresponding charging power curve based on the second battery heat change cloud map, and enter the vehicle model, charging power curve, differential pressure distribution map, and ambient temperature into the big data storage server as training data for the neural network model for later charging power matching.

6. The self-healing method according to claim 5, characterized in that, Calculate the dynamic heat dissipation distribution map of the first battery heat change cloud map based on the current ambient temperature and battery type, including: Construct an ambient temperature field based on the current ambient temperature, and perform normalization processing on the first battery heat change cloud map and the ambient temperature field in sequence; Extract the spatio-temporal local features of the normalized first battery heat change cloud map through two 3D convolutional layers, and capture the global dependence relationship in the time dimension through the multi-head self-attention mechanism to generate time-series enhanced features; Interpolate the normalized ambient temperature field bilinearly to the resolution corresponding to the first battery heat change cloud map, and perform channel splicing with the time-series enhanced features to form a fused feature; Input the fused feature into a preset physical constraint model, and then generate the corresponding dynamic heat dissipation distribution map according to the time series; The physical constraint model constructs the heat dissipation amount of the local area in the following way: Extract the ambient temperature channel in the fused feature, generate a temperature sensitivity coefficient matrix, and multiply the temperature sensitivity coefficient matrix and the convolutional kernel weights element by element to obtain the initial heat dissipation amount; Extract the battery temperature in the fusion features, adjust the heat loss direction consistency threshold based on the temperature difference between the battery temperature and the ambient temperature, calculate the heat dissipation after gradient correction through a loss function based on the direction consistency threshold, and then construct multiple heat dissipation distribution maps frame by frame based on the heat dissipation after gradient correction, thereby constituting a dynamic heat dissipation distribution map.

7. The self-healing method according to claim 6, characterized in that Formulate a corresponding charging power curve based on the second battery heat change cloud map, including: Construct a first safety temperature reference plane based on a preset safe charging temperature threshold, and reduce the preset charging power at a corresponding moment based on the volume of the part of the second battery heat change cloud map protruding from the first safety temperature reference plane at a certain moment, thereby obtaining a corrected preset charging power, and further reduce the entire corrected preset charging power based on the current voltage consistency fluctuation map to obtain a charging power curve.

8. The self-healing method according to claim 7, wherein The construction method of the voltage consistency fluctuation map is as follows: Calculate the basic voltage through a weighted algorithm based on the mean voltage and the number of monomers of a single-body distribution population corresponding to the lowest voltage threshold range and the mean voltage corresponding to the pressure difference distribution map, and then construct a basic voltage base; Fit the basic voltage base to the pressure difference distribution map so that the pressure difference distribution map deducts the part of the basic voltage base to obtain a voltage consistency fluctuation map, and the voltage consistency fluctuation map is used to characterize the voltage floating conditions of multiple monomers in the battery module.

Images