A method and system for protecting multi-gun data acquisition in charging piles

Abstract

This invention relates to the field of charging pile data acquisition and protection technology, specifically disclosing a method and system for protecting multi-gun data acquisition in charging piles. The method includes the following steps: S1, synchronously acquiring operating data from multiple charging guns; S2, preprocessing the acquired operating data; S3, dynamically allocating the charging power of each charging gun in stages based on battery state of charge and individual cell voltage data; S4, constructing a layered protection system, providing layered protection for the transmission layer, access layer, storage layer, and monitoring layer; S5, constructing an early warning and response mechanism according to the severity of anomalies, and providing graded early warning and handling for abnormal situations. This invention, employing the above-mentioned method and system for protecting multi-gun data acquisition in charging piles, solves the technical problems of poor synchronization of multi-gun data acquisition, weak data protection, and fixed power allocation in supercharging piles, providing efficient and safe technical support for new energy vehicle charging facilities.

Description

Technical Field

[0001] This invention relates to the field of charging pile data acquisition and protection technology, and in particular to a method and system for protecting multi-gun data acquisition from charging piles. Background Technology

[0002] The global new energy vehicle industry is developing rapidly. Multi-gun supercharging piles, as the core infrastructure for replenishing energy for new energy vehicles, have the characteristics of high power and multiple gun positions, but the existing technology has significant shortcomings:

[0003] At the data acquisition level, most supercharging stations can only collect basic parameters such as current and voltage, lacking key data such as battery SOH and individual cell voltage. The traditional sequential acquisition method results in poor data synchronization among multiple charging stations, with acquisition cycles as long as 5-10 seconds, which cannot support the dynamic power allocation requirements of 60kW-600kW. Outdoor electromagnetic interference and other factors cause data packet loss rates to be as high as [missing information]. - This seriously affects charging efficiency and the accuracy of battery warnings.

[0004] In terms of data protection, existing technologies lack end-to-end encrypted transmission mechanisms, posing a high risk of leakage of sensitive information such as user battery data and charging records; the lack of a fine-grained access control system makes unauthorized access easy to occur; data storage lacks off-site backup mechanisms, making it easy for hardware failures to lead to data loss, and it is difficult to resist network threats such as DDoS attacks against supercharging stations and battery data tampering.

[0005] In terms of fault response, due to incomplete data collection and lack of abnormal monitoring capabilities, it is impossible to provide early warning of mechanism-related faults such as battery overcharging. Repairs are mostly carried out passively after the fault occurs, resulting in low availability of supercharging stations and poor user charging experience.

[0006] Therefore, there is an urgent need for a supercharging pile technology solution that combines efficient multi-gun data acquisition, dynamic power allocation, and comprehensive data protection. Summary of the Invention

[0007] The purpose of this invention is to provide a method and system for protecting multi-gun data acquisition in charging piles, aiming to solve the problems of poor synchronization of multi-gun data acquisition, weak data protection, and fixed power distribution in supercharging piles.

[0008] To achieve the above objectives, the present invention provides a method and system for protecting multi-gun data acquisition in charging piles, comprising the following steps:

[0009] S1. Synchronously collect operating data from multiple charging guns;

[0010] S2. Perform preprocessing operations on the collected runtime data;

[0011] S3. Based on the battery state of charge and individual cell voltage data, dynamically allocate the charging power of each charging gun in stages.

[0012] S4. Construct a layered protection system, providing layered protection for the transport layer, access layer, storage layer, and monitoring layer;

[0013] S5. Establish an early warning and response mechanism based on the severity of the anomaly, and conduct graded early warning and handling of abnormal situations.

[0014] Preferably, in S1, the hardware architecture includes a main control module, a data acquisition module, a communication module, and a power supply module. The data acquisition module synchronously acquires the basic parameters and key data of multiple charging guns. The data acquisition module includes a Hall current sensor, a voltage divider voltage sensor, a negative temperature coefficient temperature sensor, and a battery data acquisition module.

[0015] The basic data includes charging current and leakage current collected based on Hall current sensors, charging voltage after proportional voltage division collected based on voltage divider sensors, and temperatures of the multi-channel charging gun head, the internal printed circuit board (PCB) of the charging pile, and the battery interface collected based on negative temperature coefficient temperature sensors.

[0016] Key data includes the battery state of charge obtained from the battery data acquisition module. Battery health status and single cell voltage data .

[0017] Preferably, in S1, a high-precision time synchronization mechanism is used to synchronously collect multi-channel charging gun operation data, specifically as follows:

[0018] The main control module obtains Coordinated Universal Time (UTC) through the Global Positioning System / BeiDou dual-mode positioning module and synchronizes the UTC to the DS3231 real-time clock (RTC) of each acquisition channel every 10 minutes.

[0019] If the RTC of a certain acquisition channel deviates from the Coordinated Universal Time (UTC) by more than 0.5ms, the main control module will perform active calibration; if the GPS / BeiDou signal is weak, the main control module will automatically switch to the NTP server.

[0020] Preferably, in S1, the data acquisition scheduling based on the priority queue includes the following steps:

[0021] S11. The collected operational data is divided into three levels: P1 level emergency safety data, P2 level core operational data, and P3 level auxiliary monitoring data. P1 level emergency safety data includes overcurrent signals, overvoltage signals, leakage current, and battery overcharge warning signals. P2 level core operational data includes charging current, charging voltage, charging power, and battery state of charge. And individual battery voltage data; P3 level auxiliary monitoring data includes charging chamber gun temperature, temperature of the internal printed circuit board (PCB) and battery interface of the charging pile, and battery health status. and charging time;

[0022] S12. Construct three priority queues: Q1 corresponds to P1 level emergency safety data, Q2 corresponds to P2 level core operation data, and Q3 corresponds to P3 level auxiliary monitoring data.

[0023] S13. The main control module adopts a preemptive scheduling strategy, prioritizing the processing of data in the Q1 queue. When data is triggered in the Q1 queue, the current data acquisition tasks of the Q2 and Q3 queues are immediately interrupted, and the data acquisition and transmission of the Q1 queue is started. When the Q1 queue is idle, the data in the Q2 queue is processed. When the Q2 queue is idle, the data in the Q3 queue is processed.

[0024] Preferably, in S3, the main control module determines the battery state of charge. and individual unit voltage data The charging power of each charging gun is dynamically allocated in stages, specifically:

[0025] During the low-charge fast charging phase, i.e., the battery's state of charge... When the voltage difference of a single cell When the battery is determined to be in a low charge but healthy state, indicating excellent battery consistency, a safe fast charging window is activated, allocating the maximum power. ;

[0026] When the voltage difference of individual cells If a slight unevenness risk is detected in the battery, indicating generally poor battery consistency, then equalization should be prioritized over fast charging, limiting the power to a certain level. ,power For maximum power is of ;

[0027] When the voltage difference of individual cells If the battery is deemed to have an overcharge risk, indicating poor battery consistency, the power will be immediately limited to a certain level. ,power For maximum power is of ;

[0028] During the equalization charging phase, i.e., the battery's state of charge... When the voltage difference of a single cell This indicates excellent battery consistency, according to the battery's state of charge. Interval dynamic matching power:

[0029] When the battery is in state of charge At that time, power distribution ;

[0030] When the battery is in state of charge At that time, power distribution , Maximum power of ;

[0031] When the voltage difference of individual cells When the battery consistency is generally poor, the power is reduced by a certain level to force equalization.

[0032] When the battery is in state of charge At that time, from power Reduced to power ;

[0033] When the battery is in state of charge At that time, from power Reduced to power ;

[0034] When the voltage difference of individual cells This indicates poor battery consistency, limiting power to [specific value]. within, Maximum power of ;

[0035] During the high charge phase, i.e., the battery state of charge... When the voltage difference of a single cell This indicates excellent battery consistency, according to the battery's state of charge. Gradually reduce to power Controlling the full charge process:

[0036] When the battery is in state of charge Distribute power ,power Maximum power of ;

[0037] When the battery is in state of charge At that time, power distribution ;

[0038] When the battery is in state of charge At that time, allocate the minimum power. Trickle charging continues until the battery management system sends a full charge signal, at minimum power. Maximum power of ;

[0039] When the voltage difference of individual cells When the battery consistency is generally poor, the system will immediately switch to the lowest power setting. Forced slow charging equalization;

[0040] When the voltage difference of individual cells When this occurs, it indicates poor battery consistency, prompting an emergency stop to reduce charging power to zero and triggering overcharge protection.

[0041] Preferably, in S4, a layered protection system is constructed, with separate protection layers for the transport layer, access layer, storage layer, and monitoring layer, specifically as follows:

[0042] Transport layer encryption uses an improved 256-bit Galois / counter mode algorithm of the Advanced Encryption Standard for data encryption and message authentication;

[0043] The access layer constructs a fine-grained permission control policy based on role-based access control. Users are first divided into five levels of roles: system administrator, operation and maintenance administrator, power operator, ordinary user and visitor. When a user initiates an operation request, the permission takes effect after three steps of verification: identity authentication, permission query and operation judgment.

[0044] The storage layer employs a dual-layer storage approach combining local encrypted storage with off-site cloud backup for data storage and backup, and uses a hash algorithm for backup verification.

[0045] The monitoring layer constructs an abnormal behavior monitoring strategy based on an improved convolutional neural network-bidirectional long short-term memory network model.

[0046] Preferably, in S4, the improved convolutional neural network-bidirectional long short-term memory network model includes an input layer, a convolutional neural network feature extraction layer, a bidirectional long short-term memory network temporal modeling layer, and an output layer:

[0047] Input layer input The feature vector includes current change rate, voltage fluctuation value, battery state of charge change rate, single cell voltage difference, temperature change, power change, key request count, and percentage of abnormal commands; the feature data is arranged according to time step. Seconds to divide, each sample contains Each time step forms The feature matrix;

[0048] The convolutional neural network feature extraction layer consists of two convolutional layers and one pooling layer. The first convolutional layer has a 3×3 kernel size, 32 kernels, and a stride of 1, used to extract local features. The pooling layer uses max pooling with a 2×2 kernel size and a stride of 2, used to reduce feature dimensionality. The second convolutional layer has a 2×2 kernel size, 64 kernels, and a stride of 1, used to extract deep features. Both the first and second convolutional layers use the same padding method and the ReLU activation function.

[0049] The temporal modeling layer of the bidirectional long short-term memory network includes one hidden layer with 128 neurons, used to capture the temporal dependencies of features; the forget gate, input gate, and output gate activation functions of the bidirectional long short-term memory network units all adopt the Sigmoid function, and the cell state update adopts hyperbolic tangent tanh activation.

[0050] The output layer uses the Sigmoid activation function to output a binary classification result of 0 or 1, where 0 represents normal and 1 represents abnormal. The loss function is the binary cross-entropy.

[0051] Preferably, in S5, a three-level early warning and response mechanism is constructed according to the severity of the anomaly, including Level 1 Emergency Early Warning, Level 2 Important Early Warning, and Level 3 General Early Warning, specifically:

[0052] Level 1 emergency warnings include data transmission interruptions caused by equipment failure or severe network attacks; Level 2 important warnings are for abnormal data or suspicious access by multiple attempts to log in from different Internet Protocol (IP) addresses; and Level 3 general warnings are for abnormal basic parameters or low-frequency abnormal operations.

[0053] Level 1 emergency alerts are sent to operations and maintenance administrators via SMS, platform pop-ups, and telephone notifications; Level 2 important alerts are sent via SMS and platform pop-ups; Level 3 general alerts are only displayed in the management platform alert logs; all alert information includes the time of the anomaly, the type of anomaly, the device number involved, and the suggested handling solution.

[0054] Once an alert is triggered, if it is a Level 1 emergency alert, the charging circuit will be immediately cut off to ensure the safety of personnel and equipment, and fault data will be recorded at the same time; if it is a Level 2 important alert, the abnormal channel data transmission will be suspended and switched to the backup channel; if it is a Level 3 general alert, the abnormal data will be marked, and after the operation and maintenance personnel verify and process it, the processing result will be entered into the system, and the alert status will be automatically lifted.

[0055] The present invention also provides a charging pile multi-gun data acquisition and protection system, including: a main control module, an acquisition module, a communication module and a power supply module. The power supply module is electrically connected to the main control module, the acquisition module, the communication module and the layered protection module respectively, and is used to provide stable power supply and integrate overvoltage protection components, overcurrent protection components and overtemperature protection components.

[0056] The data acquisition module includes a Hall current sensor, a voltage divider sensor, a negative temperature coefficient temperature sensor, and a battery data acquisition module. These are used to simultaneously acquire basic parameters and key data from multiple charging guns. Basic parameters include charging current, leakage current, charging voltage, and the temperatures of the charging gun heads, the internal printed circuit board (PCB) of the charging pile, and the battery interfaces. Key data includes the battery's state of charge. Battery health status and single cell voltage data ;

[0057] The communication module has a built-in TCP / IP protocol and communicates with the main control module through the Universal Asynchronous Receiver / Transmitter (UART) interface to realize data transmission.

[0058] Preferably, the main control module adopts a microcontroller unit with built-in flash memory and static random access memory. It communicates with the Hall current sensor and the voltage divider sensor through the serial peripheral interface SPI bus, communicates with the negative temperature coefficient temperature sensor and the DS3231 real-time clock RTC through the integrated circuit I2C bus, and communicates with the battery data acquisition module through the controller area network CAN bus.

[0059] The main control module is equipped with a UBLOX NEO-7M dual-mode positioning module and an NTP server switching unit to achieve high-precision time synchronization of multi-channel charging gun operation data; it is also equipped with a priority queue scheduling unit to preemptively schedule the collected data according to emergency safety level, core operation level, and auxiliary monitoring level; and it is equipped with a power dynamic allocation unit to allocate the charging power of each charging gun in stages according to the battery state of charge and individual battery voltage data.

[0060] The main control module is also equipped with a three-level early warning response unit, which is used to trigger emergency cut-off, backup channel switching or abnormal data marking operations according to the severity of the abnormality, and send early warning notifications through the communication module.

[0061] Therefore, the present invention employs the above-mentioned method and system for protecting multi-gun data acquisition in charging piles, and the beneficial effects are as follows:

[0062] (1) This invention relies on GPS / BeiDou dual-mode time synchronization and priority queue scheduling mechanism to realize efficient collection of multi-dimensional data from multiple charging guns, dynamically adjust the charging power in combination with the real-time status of the battery, cover a wide power range, and have a fast power response speed, effectively improving the efficiency of the charging pile and solving the limitation of fixed power of traditional supercharging piles.

[0063] (2) The present invention constructs a multi-layer protection system, which achieves full interception of network attacks through improved encryption algorithms, adopts encryption design for local storage to prevent physical disassembly and leakage, and accelerates data recovery through multi-center backup in the cloud, thereby minimizing the risk of data leakage.

[0064] (3) Based on the improved neural network model, the present invention realizes early warning of faults, accurately identifies potential equipment problems, significantly reduces the frequency of supercharging pile faults and maintenance frequency; at the same time, it reduces abnormal interruptions during the charging process, shortens the user waiting time, reduces the complaint rate, and comprehensively improves the energy replenishment experience of new energy vehicles.

[0065] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0066] Figure 1 This is an overall flowchart of an embodiment of a multi-gun data acquisition and protection method for charging piles according to the present invention;

[0067] Figure 2 This is a detailed flowchart of an embodiment of a multi-gun data acquisition and protection method for charging piles according to the present invention;

[0068] Figure 3 This is an overall system block diagram of an embodiment of a charging pile multi-gun data acquisition and protection system of the present invention. Detailed Implementation

[0069] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0070] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0071] like Figure 1 As shown, a method and system for protecting multi-gun data acquisition in charging piles is characterized by the following steps:

[0072] S1. Synchronously collect operating data from multiple charging guns.

[0073] First, such as Figure 3 As shown, the hardware architecture includes a main control module, a data acquisition module, a communication module, and a power supply module. The data acquisition module synchronously collects basic parameters and key data from multiple charging guns. The data acquisition module includes a Hall current sensor, a voltage divider sensor, a negative temperature coefficient temperature sensor, and a battery data acquisition module.

[0074] The basic data includes charging current and leakage current data acquired using the ACS758LCB-500B Hall current sensor, with a range of 0-500A, linearity of 0.1%, and a response time of less than 3 microseconds. Based on a voltage divider sensor, using a high-precision resistor network (R1=999) R2=1 (With a resistance accuracy of 0.01%), the charging voltage of 0-1000V is divided at a ratio of 1:1000 to output an analog signal of 0-1V. After conversion by an analog-to-digital converter, the actual voltage value is calculated, and the measurement accuracy is [not specified]. To avoid the voltage divider resistors overheating and affecting accuracy, R1 is a 2W power-type metal film resistor, and a 0.1W resistor is connected in parallel at the output of the voltage divider network. Ceramic capacitor filtering.

[0075] Based on the NCP18XH103F03RC type negative temperature coefficient temperature sensor, its range is ,exist Within the range, accuracy By connecting a fixed resistor R0=10 in series It forms a voltage divider circuit, which converts temperature changes into voltage changes, enabling real-time monitoring of the temperature of multiple charging gun heads, the internal printed circuit board (PCB) of the charging pile, and the battery interface.

[0076] Key data includes the battery state of charge obtained from the battery data acquisition module. Battery health status and single cell voltage data The acquisition frequency is synchronized with the charging current / voltage; the battery data acquisition module communicates with the new energy vehicle battery management system through the controller area network bus.

[0077] In the hardware architecture of this invention, the modules are electrically connected through a glass fiber epoxy resin FR-4 printed circuit board. The core of the main control module adopts an STM32H743ZI2 microcontroller unit, based on the ARM Cortex-M7 core, with a main frequency of 480MHz. It supports single-precision / double-precision floating-point operations and has built-in 2MB flash memory (for storing programs and configuration parameters) and 1MB static random access memory (for temporarily storing acquired data). The communication module has built-in TCP / IP protocol, and the power supply module integrates overvoltage protection, overcurrent protection, and overtemperature protection components.

[0078] The microcontroller unit integrates a 16-bit high-precision analog-to-digital converter with a sampling rate of 2 megasamples per second, supporting 16 channels of analog signal input and capable of parallel processing of multi-channel data. Simultaneously, the microcontroller unit possesses multiple interfaces to achieve high-speed data interaction with various acquisition and communication modules. Specifically, it communicates with Hall current sensors and voltage divider sensors via the SPI bus, with negative temperature coefficient temperature sensors and real-time clocks via the integrated circuit bus, with the battery data acquisition module via the controller area network bus, and with the communication module via the universal asynchronous transceiver (UART) interface.

[0079] In this step, a high-precision time synchronization mechanism is used to synchronously collect operating data from multiple charging guns, specifically as follows:

[0080] The main control module simultaneously receives signals from the Global Positioning System (GPS) and BeiDou satellites via the UBLOX NEO-7M Global Positioning System / BeiDou dual-mode positioning module to obtain Coordinated Universal Time (UTC). It outputs a Coordinated Universal Time signal every 10 minutes. After receiving the standard time, the microcontroller synchronizes the UTC to the DS3231 real-time clock (RTC) of each acquisition channel via the integrated circuit bus as a time synchronization reference.

[0081] During calibration, the microcontroller calculates the deviation between the real-time clock of each channel and the reference time. If the RTC of a certain acquisition channel deviates from the world standard time UTC by more than 0.5ms, the main control module sends a calibration command to perform active calibration and adjusts the real-time clock count to ensure that the acquisition time synchronization accuracy of multiple charging guns is less than or equal to 1ms.

[0082] If the GPS / BeiDou signal is weak (such as in an underground parking lot scenario), the microcontroller unit in the main control module will automatically switch to the National Time Service Center's Network Time Protocol (NTP) server, obtain the network time every 30 seconds, and compensate and calibrate the real-time clocks of each channel to avoid the accumulation of time deviations.

[0083] Furthermore, this invention also uses a priority queue for data acquisition scheduling to adapt to the real-time requirements of different data from multiple charging piles, including the following steps:

[0084] S11. The collected operational data is divided into three levels: P1 level emergency safety data, P2 level core operational data, and P3 level auxiliary monitoring data. P1 level emergency safety data includes overcurrent signals, overvoltage signals, leakage current, and battery overcharge warning signals, requiring real-time response. P2 level core operational data includes charging current, charging voltage, charging power, and battery state of charge. Individual battery voltage data requires high-frequency acquisition. P3-level auxiliary monitoring data includes the charging chamber gun temperature, the temperature of the internal printed circuit board (PCB) and battery interface of the charging pile, and the battery health status. Charging time can be collected at low frequencies.

[0085] S12. Construct three priority queues in the microcontroller unit of the main control module: Q1 corresponds to P1 level emergency safety data, Q2 corresponds to P2 level core operation data, and Q3 corresponds to P3 level auxiliary monitoring data. The length of each queue is set to 10 data frames, 20 data frames, and 30 data frames, respectively. Each data frame includes a timestamp, parameter value, and checksum information.

[0086] S13. The microcontroller unit in the main control module adopts a preemptive scheduling strategy, prioritizing the processing of Q1 queue data:

[0087] When data is triggered in queue Q1, the current data acquisition tasks of queues Q2 and Q3 are immediately interrupted, and data acquisition and transmission in queue Q1 are started, with an acquisition cycle of 100 milliseconds. When queue Q1 is idle, data in queue Q2 is processed, with an acquisition cycle of 500 milliseconds. When queue Q2 is idle, data in queue Q3 is processed, with an acquisition cycle of 1 second.

[0088] During the scheduling process, the main control module controls the collection interval of each queue through timer TIM2 to avoid data conflicts. When the data volume of a certain queue reaches the maximum length, the earliest non-critical data is automatically discarded and an alarm is triggered. At the same time, an overflow log is recorded to facilitate maintenance personnel to check whether there is a fault in the collection module.

[0089] S2. Preprocessing of the collected runtime data includes:

[0090] The analog signals output by each sensor in S1 are processed by an RC low-pass filter circuit to remove high-frequency noise with a cutoff frequency of 1kHz. Then, the current or voltage signal is amplified to the range suitable for the analog-to-digital converter (ADC) by an operational amplifier. The temperature signal is boosted, and then random noise is removed using a Kalman filter before being converted into a digital signal by the ADC. For missing data, when the missing rate is less than or equal to... When linear interpolation is used for filling, the noise covariance of the Kalman filtering process is... Observation noise covariance .

[0091] During analog-to-digital conversion, the microcontroller unit controls the 16-bit analog-to-digital converter to sample the preprocessed analog signal. The sampling rate is set according to the data priority: 10kHz for P1 level, 5kHz for P2 level, and 1kHz for P3 level. The sampled data is temporarily stored in the analog-to-digital converter's data register.

[0092] After the microcontroller reads the data from the analog-to-digital converter, it uses the Cyclic Redundancy Check (CRC) 32 algorithm for verification. The generated polynomial is 0xEDB88320. If the verification passes, the actual value of each parameter is calculated according to the acquisition principle of each parameter. If the verification fails, re-acquisition is triggered. The number of re-acquisitions is less than or equal to 3. If it still fails, it is marked as "invalid data" and a fault log is recorded (the log includes the fault time, channel number and number of failures).

[0093] The calculated valid data is marked in the format of timestamp (millisecond level), gun number (1-12) and parameter type, and temporarily stored in the microcontroller's built-in static random access memory. It supports circular storage and automatically overwrites the oldest data when the memory is full.

[0094] S3, the main control module determines the battery state of charge. and individual unit voltage data The charging power of each charging gun is dynamically allocated in stages, specifically:

[0095] like Figure 2 As shown, during the low-charge fast charging phase, i.e., the battery's state of charge... When the voltage difference of a single cell When the battery is determined to be in a low charge but healthy state, indicating excellent battery consistency, a safe fast charging window is activated, allocating the maximum power. The present invention takes .

[0096] When the voltage difference of individual cells If a slight unevenness risk is detected in the battery, indicating generally poor battery consistency, then equalization should be prioritized over fast charging, limiting the power to a certain level. ,power For maximum power is of .

[0097] When the voltage difference of individual cells If the battery is deemed to have an overcharge risk, indicating poor battery consistency, the power will be immediately limited to a certain level. ,power For maximum power is of .

[0098] During the equalization charging phase, i.e., the battery's state of charge... When the voltage difference of a single cell This indicates excellent battery consistency, according to the battery's state of charge. Interval dynamic matching power:

[0099] When the battery is in state of charge At that time, power distribution .

[0100] When the battery is in state of charge At that time, power distribution , Maximum power of .

[0101] When the voltage difference of individual cells When the battery consistency is generally poor, the power is reduced by a certain level to force equalization.

[0102] When the battery is in state of charge At that time, from power Reduced to power .

[0103] When the battery is in state of charge At that time, from power Reduced to power .

[0104] When the voltage difference of individual cells This indicates poor battery consistency, limiting power to [specific value]. within, Maximum power of .

[0105] During the high charge phase, i.e., the battery state of charge... When the voltage difference of a single cell This indicates excellent battery consistency, according to the battery's state of charge. Gradually reduce to power Controlling the full charge process:

[0106] When the battery is in state of charge Distribute power ,power Maximum power of .

[0107] When the battery is in state of charge At that time, power distribution .

[0108] When the battery is in state of charge At that time, allocate the minimum power. Trickle charging continues until the battery management system sends a full charge signal, at minimum power. Maximum power of .

[0109] When the voltage difference of individual cells When the battery consistency is generally poor, the system will immediately switch to the lowest power setting. Forced slow charging and equalization.

[0110] When the voltage difference of individual cells When this occurs, it indicates poor battery consistency, prompting an emergency stop to reduce charging power to zero and triggering overcharge protection.

[0111] S4. Construct a layered protection system, such as Figure 2 As shown, layered protection is implemented for the transport layer, access layer, storage layer, and monitoring layer, specifically as follows:

[0112] ① Transport layer encryption uses an improved 256-bit Galois / counter mode algorithm of the Advanced Encryption Standard (AES) for data encryption and message authentication. Specifically, it introduces chaotic mapping and message authentication mechanisms on top of the standard AES 256-bit algorithm.

[0113] A key seed is generated using the ATECC608A hardware security module HSM. The key seed is then subjected to chaotic perturbation using a logistic chaotic mapping to generate a 256-bit dynamic key. The key rotation cycle is 24 hours. The nonlinear characteristics of the chaotic mapping make the key unpredictable.

[0114] During transport layer encryption, the Galois / counter mode is used for encryption mode optimization and message authentication. During encryption, the counter starts from an initial value (generated by a timestamp and random number) and increments by 1 for each 128-bit data block encrypted, ensuring that the same plaintext outputs different ciphertexts. During message authentication, a 128-bit message authentication code is calculated using Galois field multiplication. The receiving end verifies whether the message authentication code matches to determine if the data has been tampered with. In practical applications, the Advanced Encryption Standard (AES) 256-bit Galois / counter mode algorithm can be implemented in the microcontroller unit using a hardware acceleration module STM32H743ZI2 with a built-in Advanced Encryption Standard (AES) hardware accelerator. This achieves an encryption rate of up to 50Mbps, meeting the real-time data transmission requirements of multiple receivers.

[0115] ② The access layer constructs fine-grained permission control policies based on role-based access control, including the following steps:

[0116] Sa1. Users are divided into 5 role levels: system administrator, operations and maintenance administrator, power operator, regular user, and visitor. Specifically:

[0117] Grant system administrators full permissions, including user management, permission configuration, data viewing, and system configuration functions. User management includes adding, deleting, or modifying user information; permission settings include adjusting permissions at each level; data viewing includes viewing all charging pile operation data and sensitive user data; and system configuration functions include modifying the data collection cycle and encryption parameters.

[0118] Operations administrators are assigned operational permissions, but not access to sensitive user data. These permissions include device monitoring, fault handling, and data export functions. Device monitoring is used to view the operating data of charging piles in the assigned area, fault handling is used to remotely restart charging piles or calibrate sensors, and data export is used to export non-sensitive operating data.

[0119] The power operator is assigned operating permissions, but not configuration modification permissions. These permissions include charging pile start / stop control, charging parameter settings, and basic data viewing. Charging parameter settings include setting the charging power limit, and basic data viewing includes viewing the current and voltage data of charging piles in this area.

[0120] Regular users are assigned personal permissions, but they do not have access to other user data or device data. They can only view their own charging records, initiate charging requests, and change their personal passwords. Their own charging records include charging time, charging amount, and charging cost.

[0121] Visitors are only granted query permissions, with no data viewing or operation permissions; they can only view the real-time availability status of charging stations.

[0122] Sa2. When a user initiates an operation request, a three-step verification process is performed: identity authentication, permission query, and operation judgment. Specifically:

[0123] A triple authentication mechanism is used for identity verification. Users enter a username and password. The password is encrypted and stored using the key derivation function 2 algorithm with 10,000 iterations. The salt value is randomly generated. After successful verification, an SMS verification code is sent to the user's bound mobile phone. After entering the verification code, the user enters the permission query through device fingerprint verification. Only after all three verifications are passed can the user enter the permission query stage.

[0124] When querying permissions, the system queries a preset permission list based on the user's role. The permission list exists in the form of a two-dimensional table of operations and resources. The permission list is stored in an encrypted database and is encrypted using AES-128.

[0125] The system determines whether the user's requested operation is within the permission list. If it is, the operation is executed, and the operation log is recorded, including the operation time, user ID, operation content, and terminal information. Otherwise, the operation is rejected, an "insufficient permissions" message is returned, and an abnormal operation log is recorded. The abnormal operation log is retained for one year and cannot be tampered with.

[0126] Sa3. Based on business needs, the permissions of each role can be dynamically adjusted. System administrators modify permissions through the "application, approval, and activation" process: submit a permission adjustment application, explain the reason and scope of the adjustment, have it cross-approved by two other system administrators, and the permission will take effect immediately after approval. At the same time, a permission change log will be generated for easy auditing and traceability.

[0127] ③ The storage layer adopts a two-tier storage method that combines local encrypted storage with off-site cloud backup for data storage and backup, and uses a hash algorithm for backup verification.

[0128] When local storage is encrypted, the local charging station uses an embedded multimedia card flash memory to store the collected data, including sensitive data and general data. Sensitive data includes user payment information and battery health status, while general data includes charging current and charging voltage.

[0129] The embedded multimedia card has a built-in encryption controller. Before the collected data is written to the embedded multimedia card, the data is classified and encrypted. Sensitive data is encrypted using AES-128. The key is generated by the hardware security module and is independent of the transmission key. The encrypted data is stored in partitions according to time and serial number for easy and fast retrieval.

[0130] The embedded multimedia card is equipped with an access password, which is dynamically generated by the hardware security module and updated every 72 hours. Only the microcontroller unit can access the stored data after the password is verified, preventing data theft after physical disassembly.

[0131] The off-site cloud backup adopts a three-center backup architecture, with one cloud storage node deployed in each center. The nodes are physically isolated from each other, and the network is connected via dedicated lines. It uses a combination of full backup and incremental backup. A full backup is performed at 00:00 every day to back up all local storage data, and an incremental backup is performed every hour to back up only the data added in the previous hour. The backup data is encrypted with AES-256 before transmission, and the transmission process is secured through a virtual private network tunnel (Internet Protocol Security).

[0132] Meanwhile, the backup data is verified for integrity every 24 hours by comparing the hash value of the backup data with the hash value of the local data. If they do not match, a new backup is performed. In case of local data loss, such as an embedded multimedia card failure, the nearest backup node can be selected for recovery. The recovery process supports resuming interrupted downloads, with a recovery time of 1GB. 10 minutes. Normal operational data is retained for 1 year, and sensitive user data is retained for 3 years. After the expiration, the data will be automatically de-identified, user identification information will be deleted, and only anonymous operational data will be retained.

[0133] ④ The monitoring layer constructs an abnormal behavior monitoring strategy based on an improved convolutional neural network-bidirectional long short-term memory network model. This improved model includes an input layer, a convolutional neural network feature extraction layer, a bidirectional long short-term memory network temporal modeling layer, and an output layer. The input layer... The feature vector includes current change rate, voltage fluctuation value, battery state of charge change rate, single cell voltage difference, temperature change, power change, key request count, and percentage of abnormal commands; the feature data is arranged according to time step. Seconds to divide, each sample contains Each time step forms The characteristic matrix of .

[0134] The convolutional neural network feature extraction layer consists of two convolutional layers and one pooling layer. The first convolutional layer has a kernel size of 3×3, a number of 32, and a stride of 1, and is used to extract local features. The pooling layer uses max pooling with a kernel size of 2×2 and a stride of 2, and is used to reduce the feature dimension. The second convolutional layer has a kernel size of 2×2, a number of 64, and a stride of 1, and is used to extract deep features. Both the first and second convolutional layers use the same padding method and the ReLU activation function.

[0135] The temporal modeling layer of the bidirectional long short-term memory network includes one hidden layer with 128 neurons. The discarding technique (dropout coefficient 0.25) is used to prevent overfitting and to capture the temporal dependencies of features. The forgetting gate, input gate, and output gate activation functions of the bidirectional long short-term memory network units all use the Sigmoid function, and the cell state update uses hyperbolic tangent tanh activation.

[0136] The output layer uses the Sigmoid activation function to output a binary classification result of 0 or 1, where 0 represents normal and 1 represents abnormal. The loss function is the binary cross-entropy.

[0137] In S5, a three-tiered early warning and response mechanism is constructed based on the severity of the anomaly, such as... Figure 2 As shown, there are three levels of alert: Level 1 Emergency Alert, Level 2 Important Alert, and Level 3 General Alert. Specifically:

[0138] Level 1 emergency alerts include data transmission interruptions caused by equipment failure (such as overcurrent or overvoltage) or severe network attacks (such as distributed denial-of-service attacks), requiring immediate attention. Level 2 critical alerts are for data anomalies (such as current fluctuations exceeding normal ranges). Suspicious access attempts involving multiple login attempts from IP addresses located in different areas must be processed within one hour. Level 3 general alerts involve abnormal basic parameters or low-frequency abnormal operations (one insufficient access request) and must be processed within 24 hours.

[0139] Warning methods: Level 1 emergency warnings are notified to the operations and maintenance administrators via SMS, platform pop-up, and telephone notification; Level 2 important warnings are notified via SMS and platform pop-up; Level 3 general warnings are only displayed in the management platform warning log; all warning information includes the time of the anomaly, the type of anomaly, the number of the device involved, and the suggested handling solution.

[0140] Response Process: Upon triggering an alert, the system automatically executes an initial response. If it is a Level 1 emergency alert, the charging circuit is immediately cut off to ensure the safety of personnel and equipment, and fault data is recorded simultaneously. If it is a Level 2 important alert, data transmission on the abnormal channel is suspended, and the system switches to the backup channel. If it is a Level 3 general alert, the abnormal data is marked, and after verification and processing by maintenance personnel, the processing results are entered into the system, and the alert status is automatically lifted.

[0141] like Figure 3 As shown, a multi-gun data acquisition and protection system for charging piles includes: a main control module, an acquisition module, a communication module, and a power supply module. The power supply module is electrically connected to the main control module, the acquisition module, the communication module, and the layered protection module, respectively, and is used to provide stable power supply and integrate overvoltage protection components, overcurrent protection components, and overtemperature protection components.

[0142] The data acquisition module includes a Hall current sensor, a voltage divider sensor, a negative temperature coefficient temperature sensor, and a battery data acquisition module. These are used to simultaneously acquire basic parameters and key data from multiple charging guns. Basic parameters include charging current, leakage current, charging voltage, and the temperatures of the charging gun heads, the internal printed circuit board (PCB) of the charging pile, and the battery interfaces. Key data includes the battery's state of charge. Battery health status and single cell voltage data .

[0143] The communication module has a built-in TCP / IP protocol and communicates with the main control module through the Universal Asynchronous Receiver / Transmitter (UART) interface to realize data transmission.

[0144] The main control module uses a microcontroller unit with built-in flash memory and static random access memory. It communicates with the Hall current sensor and the voltage divider sensor through the serial peripheral interface SPI bus, communicates with the negative temperature coefficient temperature sensor and the DS3231 real-time clock RTC through the integrated circuit I2C bus, and communicates with the battery data acquisition module through the controller area network CAN bus.

[0145] The main control module is equipped with a UBLOX NEO-7M dual-mode positioning module and an NTP server switching unit to achieve high-precision time synchronization of multi-channel charging gun operation data; it is also equipped with a priority queue scheduling unit to preemptively schedule the collected data according to emergency safety level, core operation level, and auxiliary monitoring level; and it is equipped with a power dynamic allocation unit to allocate the charging power of each charging gun in stages according to the battery state of charge and individual battery voltage data.

[0146] The main control module is also equipped with a three-level early warning response unit, which is used to trigger emergency cut-off, backup channel switching or abnormal data marking operations according to the severity of the abnormality, and send early warning notifications through the communication module.

[0147] Example 1

[0148] This embodiment selects a 12-gun DC supercharging pile with a total power of 720kW and a maximum power of 600kW per gun. The hardware deployment steps are as follows:

[0149] Main control and positioning module installation: Fix the STM32H743ZI2 development board (equipped with a 480MHz main frequency microcontroller, 1MB static random access memory, and 2MB flash memory) to the internal control compartment of the charging pile and connect it to the power module via DuPont wires; connect the UBLOX NEO-7M global positioning system / BeiDou dual-mode positioning module to the microcontroller through the universal asynchronous transceiver interface (TX=PA9, RX=PA10), install the module antenna on the top of the charging pile (ensuring unobstructed satellite signal reception), and seal it with waterproof glue.

[0150] Data Acquisition Module Deployment: Data acquisition units are configured for each of the 12 charging guns: the current acquisition unit uses an ACS758LCB-500B Hall sensor, connected in series in the positive circuit of the charging gun; the output signal is conditioned by an operational amplifier and then connected to the analog-to-digital converter pins (PA0-PA11) of the microcontroller unit; the voltage acquisition unit uses a voltage divider resistor network (R1=999). R2=1 The battery data acquisition unit is connected in parallel between the positive and negative terminals of the charging gun, and its output signal is connected to the analog-to-digital converter pins (PB0-PB11) of the microcontroller unit. The temperature acquisition unit uses a negative temperature coefficient thermistor (NCP18XH103F03RC), which is attached to the inner wall of the charging gun head and the key position of the printed circuit board inside the charging pile. It is connected to the microcontroller unit through the integrated circuit bus (SDA=PC9, SCL=PC10). The battery data acquisition unit interfaces with the vehicle battery management system through the controller area network bus (CAN_RX=PD0, CAN_TX=PD1).

[0151] Communication and security module installation: Connect the EC200S-CN multimode communication module to the microcontroller expansion board via the Mini PCIe (Mini Peripheral Component Interconnect) interface. Use an industrial-grade 5G mobile data SIM card (supporting nationwide roaming). Fix the antenna to the side of the charging pile away from high-voltage components to avoid electromagnetic interference. Simultaneously connect the DP83848 Ethernet module via an RJ45 Ethernet interface. Connect the ATECC608A hardware security module to the microcontroller via the integrated circuit bus (SDA=PD2, SCL=PD3). The module is powered by 3.3V and encapsulated with a metal shield. Install an emergency stop button on the charging pile casing, connecting it to the microcontroller's general-purpose input / output pins. Power module installation: A switching power supply outputs 5V / 3A and 12V / 2A DC voltages to power each module.

[0152] The software system adopts an embedded real-time operating system (RTOS) combined with a cloud management platform architecture. An embedded program based on the FreeRTOS is burned into the STM32H743ZI2 microcontroller unit. The program includes a driver module (implementing drivers for sensors, communication modules, and hardware security modules, supporting interface protocols such as serial peripheral interfaces / integrated circuit buses / universal asynchronous transceivers / controller area networks), a data acquisition and scheduling module (integrating a priority queue acquisition algorithm to control analog-to-digital converter sampling and data processing according to P1 / P2 / P3 level data priorities), an encryption module (porting an improved advanced encryption standard 256-bit Galois / counter mode algorithm and RSA-2048 algorithm to achieve data encryption and key management), an early warning module (embedding a lightweight version of an improved convolutional neural network-bidirectional long short-term memory network model to support real-time anomaly detection), and a power allocation module (implementing dynamic power adjustment from 60kW to 600kW). The program is compiled into a .bin file using STM32CubeIDE and burned into the microcontroller unit's flash memory using a J-Link emulator. After burning, verification is performed (comparing the file hash values ​​before and after burning).

[0153] When building the cloud platform, an Alibaba Cloud server was selected, configured with 4 cores, 8GB of memory, and a 100GB SSD to deploy the cloud management platform. The platform is developed based on the Spring Boot framework and includes a data receiving module for receiving encrypted data uploaded by charging piles via message queue telemetry transmission protocol; a decryption and storage module for calling the cloud hardware security module service to decrypt the data and store the decrypted data in a relational database MySQL; a monitoring and early warning module for displaying the real-time operating status of charging piles, receiving early warning information and pushing it to the maintenance personnel's application (App); and a permission management module for implementing role-based access control for user permission control. The cloud platform is deployed using Docker containerization, uses Nginx as a reverse proxy, and is configured with a Secure Sockets Layer (SSL) certificate and Transport Layer Security (TLS) 1.3 protocol to ensure the security of Hypertext Transfer Security (HTTPS) communication.

[0154] Hardware debugging was performed, and a multimeter was used to measure the power supply voltage of each module to ensure that the 5V / 12V / 3.3V voltages were stable and the ripple was minimal. The sensor output signal amplitude was observed to be 0-3.3V using an oscilloscope, and there was no obvious noise.

[0155] Software debugging involved using a J-Link emulator to monitor the task scheduling of the free real-time operating system, ensuring that data acquisition tasks were executed correctly according to priority, and that P1-level task response times were maintained. 100 milliseconds; Test the hardware security module's key generation function to verify whether the key length and format conform to the Advanced Encryption Standard 256-bit standard; Test the power distribution module to verify the response time of switching from 60kW to 600kW. 500 milliseconds.

[0156] Network debugging:

[0157] Communication testing: Test data was sent via Ethernet and 5G mobile communication networks respectively to test the data transmission rate (Ethernet rate). 10Mbps, the network speed of fifth-generation mobile communication technology. 100Mbps) and packet loss rate (continuous 24-hour test, packet loss rate) Data consistency test: Compare the local storage data of the charging pile with the cloud storage data, and randomly select 1000 data entries for verification to check data consistency. Access Control Test: Log in to the cloud platform using different role accounts to verify the effectiveness of access control; Multi-gun Concurrency Test: Simultaneously start 12 charging guns to simulate full-load operation, with each gun's charging power dynamically switching between 60kW and 600kW, for 72 hours, monitoring data acquisition stability; Anomaly Injection Test: Inject abnormal scenarios such as 600A overcurrent, 1100V overvoltage, and distributed denial-of-service attacks (10,000+ requests per second) into the system to test the early warning response time and protection effectiveness.

[0158] This embodiment uses a public charging station in a highway service area, deploying 10 12-gun supercharging piles and collecting actual operational data over a period of 3 months to verify the technical effectiveness.

[0159] Data acquisition and power distribution performance: Data synchronization accuracy of 12 charging guns ≤ 0.8 milliseconds, current acquisition error Voltage acquisition error Power distribution response time In 450 milliseconds, it achieved free switching between 60kW and 600kW, improving the utilization rate of the pile body. .

[0160] No data breaches or cyberattacks occurred during operation, and all 10 simulated distributed denial-of-service attacks and 5 attempts to tamper with battery data were successfully tested. Interception; Local storage data was physically disassembled for testing (removal of embedded multimedia card flash memory), and the data could not be read due to encryption protection; In the cloud backup data recovery test, 1GB of data took 8 minutes to recover.

[0161] The abnormal behavior monitoring model identified 23 instances of battery overcharging and equipment malfunction warnings. Thanks to the timely handling by maintenance personnel, no charging pile downtime occurred, and the failure rate was lower than that of traditional supercharging piles. User charging interruption rate has decreased from that of traditional technologies. Down to User complaint rate decreased Charging wait time is shortened .

[0162] Therefore, the present invention adopts the above-mentioned method and system for data acquisition and protection of multi-gun charging piles, which effectively solves the pain points of poor synchronization of multi-gun data acquisition, weak data protection, and fixed power distribution of supercharging piles. It takes into account efficient data acquisition, dynamic power adjustment and comprehensive protection, and can be widely used in scenarios such as highway service areas, providing strong support for the safe and efficient operation of new energy vehicle charging facilities.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A charging pile multi-gun data acquisition protection method, characterized in that, The method comprises the following steps: S1, synchronously collecting multi-path charging gun operation data; S2, pre-processing the collected operation data; S3, dynamically allocating charging power of each charging gun in stages according to battery state of charge and single battery voltage data; S4, constructing a layered protection system to protect the transmission layer, access layer, storage layer and monitoring layer in layers; S5, constructing an early warning and response mechanism according to the severity of the abnormality to grade the early warning and processing of abnormal conditions; In S1, the operation data collection scheduling is based on a priority queue, comprising the following steps: S11, the collected operation data is divided into three levels, P1 level emergency safety data, P2 level core operation data and P3 level auxiliary monitoring data; wherein, the P1 level emergency safety data includes overcurrent signal, overvoltage signal, leakage current and battery overcharge warning signal; the P2 level core operation data includes charging current, charging voltage, charging power, battery state of charge and monomer battery voltage data; the P3 level auxiliary monitoring data includes charging cavity gun body temperature, charging pile internal printed circuit board PCB and battery interface temperature, battery health status and charging duration; S12, constructing three priority queues Q1 corresponding to P1 level emergency safety number, Q2 corresponding to P2 level core operation data, and Q3 corresponding to P3 level auxiliary monitoring data; S13, the main control module adopts a preemptive scheduling strategy, and processes the Q1 queue data first. When the Q1 queue has data triggering, the current data collection tasks of the Q2 and Q3 queues are interrupted immediately, and the Q1 data collection and transmission are started. When the Q1 queue is idle, the Q2 queue data is processed. When the Q2 queue is idle, the Q3 queue data is processed.

2. The charging pile multi-gun data acquisition protection method of claim 1, wherein, In S1, the hardware architecture is built, including a main control module, a collection module, a communication module and a power module. The basic parameters and key data of the multi-path charging gun are synchronously collected through the collection module. The collection module includes a Hall current sensor, a voltage divider type voltage sensor, a negative temperature coefficient temperature sensor and a battery data collection module. The basic data includes charging current and leakage current collected based on the Hall current sensor, charging voltage collected based on the voltage divider type voltage sensor after proportional voltage division, and the temperature of the multi-path charging gun head, the charging pile internal printed circuit board (PCB) and the battery interface collected based on the negative temperature coefficient temperature sensor. Key data includes battery state of charge , battery state of health , and cell voltage data acquired based on the battery data acquisition module.

3. The charging pile multi-gun data acquisition protection method according to claim 2, characterized in that, In S1, a high-precision time synchronization mechanism is used for synchronous collection of multi-path charging gun operation data. Specifically: The main control module obtains the universal time coordinated (UTC) through a global positioning system / Beidou dual-mode positioning module, and synchronizes the UTC to the DS3231 real-time clock (RTC) of each gun collection channel every 10 minutes. If the RTC of a certain collection channel deviates from the UTC by more than 0.5 ms, the main control module performs active calibration. If the global positioning system / Beidou signal is weak, the main control module will automatically switch to an NTP server.

4. The charging pile multi-gun data acquisition protection method of claim 3, characterized in that, In S3, the master module dynamically allocates charging power of each charging gun according to battery state of charge and single cell voltage data in stages, specifically: In the low power fast charging phase, i.e. battery state of charge When the single battery voltage difference is less than 0.1V, it is determined that the battery is in low power and the state is healthy, indicating that the battery consistency is good, at this time the safe fast charging window is adopted, and the maximum power is allocated ; When the single cell voltage difference is less than 0.1V, it is determined that the battery has a slight risk of unevenness, indicating that the battery consistency is general, at which time equalization is prior to fast charging, and the power is limited to the power , the power is the maximum power of ; When the single cell voltage difference is greater than 0.1V, it is determined that the battery has overcharge risk, indicating poor battery consistency, and the power is urgently limited to the power , the power is the maximum power of ; In the middle of the equalization charging phase, i.e. the battery state of charge When the single battery voltage difference is less than 0.05V, it indicates that the battery consistency is good, and the battery state of charge Dynamic matching power in the interval: When the battery state of charge power is allocated ; When the battery state of charge is less than 50%, the power , is maximum power . ; When the single cell voltage difference is greater than 0.1V, it indicates that the battery consistency is generally, and the power is down a gradient, and the equalization is forced: When the battery state of charge power power ; When the battery state of charge power power ; When the cell voltage difference is greater than 0.1 V, it indicates poor cell consistency, and the power is limited to within 0.1 V, the maximum power is ; In the high state of charge phase, i.e. battery state of charge When the cell voltage difference is less than 0.1 V, it indicates that the battery consistency is good, and the battery state of charge is gradually reduced to the power control full process: when the battery state of charge , allocate power , power is maximum power of ; When the battery state of charge power is allocated ; When the battery state of charge is less than the minimum power trickle charging is performed until the battery management system sends a full charge signal, the minimum power is the maximum power of ; When the single cell voltage difference is less than 0.1V, it indicates that the battery consistency is generally, directly jump to the lowest power , forced to slow charging equalization; When the cell voltage difference is greater than 0.1V, it indicates poor cell consistency, emergency stop charging, reduce power to zero, trigger overcharge protection.

5. The charging pile multi-gun data acquisition protection method according to claim 4, characterized in that, In S4, a layered protection system is constructed to protect the transmission layer, access layer, storage layer and monitoring layer in layers. Specifically: The transmission layer encryption uses an improved 256-bit Galois / counter mode algorithm of the advanced encryption standard (AES) for data encryption and message authentication; The access layer constructs a fine-grained permission control strategy based on role-based access control. First, the users are divided into five roles: system administrator, operation and maintenance administrator, power operator, ordinary user and visitor. When a user initiates an operation request, the identity authentication, permission query and operation judgment are performed in three steps, and the permission takes effect after the verification; The storage layer uses a double-layer storage method of local encrypted storage combined with cloud off-site backup for data storage and backup, and performs backup verification through a hash algorithm; The monitoring layer constructs an abnormal behavior monitoring strategy based on an improved convolutional neural network-long short-term memory network model.

6. The charging pile multi-gun data acquisition protection method according to claim 5, characterized in that, In S4, the improved convolutional neural network-bi-directional long short-term memory network model comprises an input layer, a convolutional neural network feature extraction layer, a bi-directional long short-term memory network time sequence modeling layer and an output layer: Input layer input V characteristic vector, including current rate of change, voltage fluctuation value, battery state of charge rate of change, single battery voltage difference, temperature change amount, power change, key request times and abnormal instruction proportion; characteristic data is divided by time step Second, each sample contains Time step, forming Characteristic matrix; The convolutional neural network feature extraction layer comprises two convolutional layers and one pooling layer, the first convolutional layer has a convolution kernel size of 3*3, a number of 32 and a step of 1, and is used for extracting local features; the pooling layer adopts maximum pooling, has a pooling kernel size of 2*2 and a step of 2, and is used for reducing the feature dimension; the second convolutional layer has a convolution kernel size of 2*2, a number of 64 and a step of 1, and is used for extracting deep features; the padding mode of the first convolutional layer and the second convolutional layer is same, and the activation function adopts a linear rectifier function ReLU; The bi-directional long short-term memory network time sequence modeling layer comprises one hidden layer, and the number of neurons is 128, which is used for capturing the time sequence dependence of the features; the activation functions of the forgetting gate, the input gate and the output gate of the bi-directional long short-term memory network unit all adopt Sigmoid functions, and the cell state update adopts a hyperbolic tangent tanh activation; The output layer adopts a Sigmoid activation function, outputs a binary classification result of 0 or 1, 0 represents normal, 1 represents abnormal, and the loss function adopts binary cross entropy.

7. The charging pile multi-gun data acquisition protection method of claim 6, wherein, In S5, a three-level early warning and response mechanism is constructed according to the severity of the anomaly, including first-level emergency warning, second-level important warning and third-level general warning, specifically: The first-level emergency warning includes data transmission interruption caused by device failure or serious network attack, the second-level important warning includes data anomaly or suspicious access of multiple attempts of login of foreign Internet Protocol IP, and the third-level general warning includes basic parameter anomaly or low-frequency abnormal operation; The first-level emergency warning adopts a three-way notification mode of short message, platform pop-up window and telephone notification to notify the operation and maintenance administrator; the second-level important warning adopts a short message and platform pop-up window mode; the third-level general warning only displays the warning log in the management platform; all warning information includes abnormal time, abnormal type, involved device number and suggested processing scheme; After the warning is triggered, if it is the first-level emergency warning, the charging circuit is immediately cut off to ensure the safety of personnel and equipment, and the fault data is recorded; if it is the second-level important warning, the data transmission of the abnormal channel is suspended and switched to the standby channel; if it is the third-level general warning, the abnormal data is marked, and after the operation and maintenance personnel check and handle, the processing result is input into the system, and the warning state is automatically released.

8. A charging pile multi-gun data acquisition protection system for implementing the charging pile multi-gun data acquisition protection method according to any one of claims 1-7, characterized in that, It comprises a main control module, a collection module, a communication module and a power module, the power module is electrically connected with the main control module, the collection module, the communication module and the layered protection module, and is used for providing stable power supply and integrating overvoltage protection elements, overcurrent protection elements and overtemperature protection elements; The acquisition module includes a Hall current sensor, a voltage divider type voltage sensor, a negative temperature coefficient temperature sensor and a battery data acquisition module, which are used to synchronously acquire basic parameters and key data of multiple charging guns; the basic parameters include charging current, leakage current, charging voltage and temperatures of the heads of the multiple charging guns, internal printed circuit boards (PCBs) of the charging piles and battery interfaces; the key data includes state of charge , state of health and single battery voltage data of the battery. The communication module is built-in TCP / IP protocol, communicates with the main control module through a universal asynchronous receiver transmitter UART interface, and is used for realizing data transmission.

9. The charging pile multi-gun data acquisition protection system according to claim 8, characterized in that, The master control module adopts a micro control unit with built-in flash memory and static random access memory, communicates with the Hall current sensor and the voltage divider type voltage sensor through a serial peripheral interface (SPI) bus, communicates with the negative temperature coefficient temperature sensor and a DS3231 real-time clock (RTC) through an integrated circuit (I2C) bus, and communicates with the battery data acquisition module through a controller area network (CAN) bus. The master control module is configured with a UBLOX NEO-7M dual-mode positioning module and an NTP server switching unit, which are used to realize high-precision time synchronization of the operating data of multiple charging guns; the master control module is also configured with a priority queue scheduling unit, which is used to preemptively schedule the collected data according to the emergency safety level, the core operation level, and the auxiliary monitoring level; and the master control module is further configured with a power dynamic allocation unit, which is used to allocate the charging power of each charging gun in stages according to the state of charge of the battery and the voltage data of the single battery. The master control module is also configured with a three-level early warning response unit, which is used to trigger emergency shutdown, backup channel switching, or abnormal data marking operation according to the severity of the abnormality, and send a warning notification through the communication module.

Images