Storage battery automatic bridging method and automatic bridging device based on battery data detection

Abstract

The invention provides a storage battery automatic bridging method and device based on battery data detection, and the method comprises the following steps: obtaining the battery position information of each storage battery unit in a storage battery pack, and collecting the battery parameters of each storage battery unit; preprocessing battery parameters, and extracting static parameters and dynamic parameters; constructing a characteristic fuzzy relation matrix based on the battery characteristic parameters, and performing fuzzy comprehensive evaluation on the storage battery unit according to the characteristic fuzzy relation matrix; inputting the abnormal battery characteristic parameters of the abnormal storage battery units into a preset storage battery fault identification model, and identifying fault storage battery units in the abnormal storage battery units; and if the storage battery fault identification model identifies the faulty storage battery unit, generating an automatic bridging task of the automatic bridging device, and controlling the automatic bridging device to realize automatic bridging processing on the faulty storage battery unit through the automatic bridging task. According to the invention, the fault storage battery can be automatically detected and bridged, and the efficiency is high.

Description

Technical Field

[0001] The invention belongs to the technical field of battery fault detection and processing, and in particular relates to a battery automatic jumper method and an automatic jumper device based on battery data detection. Background Art

[0002] In the power system, smart substations use battery packs as emergency power supplies for key equipment to ensure that the system can maintain normal operation when the main power supply fails. Battery packs require long-term management and maintenance. If a single battery in the battery pack fails, the capacity and output voltage of the entire battery pack will be reduced, resulting in the battery pack being unable to meet the normal power supply needs of substation equipment in an emergency. Therefore, it is necessary to perform jumper operations on single failed batteries in the battery pack in a timely manner. By jumpering the failed unit in time, it can ensure that the battery pack continues to operate at a near-normal performance level and ensure its reliability as a backup power supply.

[0003] Traditional battery management mainly relies on manual inspection and manual operation, which is inadequate for large-scale, distributed battery systems. First, the frequency and coverage of manual inspection are limited, making it difficult to detect potential faults in a timely manner. Even if a faulty single battery is detected in the battery pack, manual jumper operation is required, but manual operation is easily affected by human factors and there is a risk of misoperation. Summary of the invention

[0004] The present invention provides a battery automatic jumper method and an automatic jumper device based on battery data detection, so as to solve the problems of low efficiency and high safety risk in manual inspection and jumpering of faulty batteries.

[0005] In a first aspect, the present invention provides a battery automatic crossover method based on battery data detection, which is applied to an intelligent substation deployed with a battery pack and an automatic crossover device, wherein an intelligent sensor network is installed on the battery pack, the automatic crossover device comprises an insulating shell and a mobile chassis, the insulating shell is fixedly arranged on the top of the mobile chassis, the automatic crossover device further comprises a main control module, a crossover protection module and a crossover execution module, the main control module is respectively connected to the mobile chassis, the crossover protection module and the crossover execution module, the crossover protection module is electrically connected to the crossover execution module, the main control module is used to receive an automatic crossover task and generate an automatic crossover instruction according to the automatic crossover task, the mobile chassis responds to the automatic crossover instruction and moves to the location of the faulty battery unit, the crossover execution module responds to the automatic crossover instruction and completes the automatic crossover processing of the faulty battery unit, the crossover protection module responds to the automatic crossover instruction and protects the crossover execution module during the automatic crossover processing;

[0006] The method comprises the following steps:

[0007] Obtaining the battery location information of each battery unit in the battery pack, and collecting the battery parameters of each battery unit through an intelligent sensor network;

[0008] Preprocessing battery parameters, extracting static parameters and dynamic parameters from the preprocessed battery parameters, and integrating the static parameters and dynamic parameters into battery characteristic parameters;

[0009] A characteristic fuzzy relationship matrix is ​​constructed based on the battery characteristic parameters, and a fuzzy comprehensive evaluation is performed on the battery unit according to the characteristic fuzzy relationship matrix to obtain a battery status score for each battery unit;

[0010] Marking a battery cell whose battery status score is lower than a preset score threshold as an abnormal battery cell;

[0011] The abnormal battery characteristic parameters of the abnormal battery cell are input into a preset battery fault identification model, and the faulty battery cell in the abnormal battery cell is identified by the battery fault identification model, wherein the battery fault identification model is constructed based on a support vector machine;

[0012] If the battery fault identification model identifies a faulty battery cell, an automatic bridging task of the automatic bridging device is generated in combination with the battery position information and static parameters of the faulty battery cell, and the automatic bridging device is controlled by the automatic bridging task to realize automatic bridging processing of the faulty battery cell.

[0013] Optionally, the method further comprises the following steps:

[0014] If the battery fault identification model fails to identify the faulty battery unit, all battery characteristic parameters are converted into multi-dimensional characteristic time series data;

[0015] Combining long short-term memory network and graph convolutional network to build a battery life prediction model;

[0016] Input the multi-dimensional feature time series data into the battery life prediction model, and generate a battery impact association graph of the battery cells based on the graph network construction module in the battery life prediction model and through a data-driven method. The graph network construction module is composed of a long short-term memory network layer and a self-attention mechanism layer. The graph nodes in the battery impact association graph are all battery cells, and the graph node edges in the battery impact association graph represent the physical connection relationship or electrical coupling relationship between the battery cells.

[0017] The high-dimensional impact features are extracted from the battery impact association graph through the feature extraction module in the battery life prediction model. The feature extraction module includes a graph convolution layer, a graph domain conversion layer, a long short-term memory network layer, and a graph domain inverse conversion layer.

[0018] The high-dimensional influencing features are reduced in dimension into low-dimensional influencing features by using the dimensionality reduction prediction module in the battery life prediction model. The life prediction estimate is generated according to the low-dimensional influencing features through the fully connected layer in the dimensionality reduction prediction module and the low-dimensional influencing features are reconstructed. The reconstructed low-dimensional influencing features are back-propagated to the graph network construction module so that the graph network construction module performs residual fusion on the subsequently input multi-dimensional feature time series data and the reconstructed low-dimensional influencing features.

[0019] Determine an aged battery cell with the lowest life prediction estimate value among all abnormal battery cells, and if the life prediction estimate value of the aged battery cell is lower than a preset service life threshold, mark the aged battery cell as a faulty battery cell;

[0020] An automatic bridging task is generated in combination with the battery position information and static parameters of the faulty battery unit, and an automatic bridging device is controlled by the automatic bridging task to realize automatic bridging processing of the faulty battery unit.

[0021] Optionally, preprocessing the battery parameters, extracting static parameters and dynamic parameters from the preprocessed battery parameters, and integrating the static parameters and dynamic parameters into battery characteristic parameters includes the following steps:

[0022] The median filtering method based on the adaptive window mechanism is used to process the abnormal values ​​in the battery parameters, and the multiple interpolation method is used to fill the missing values ​​in the battery parameters to obtain the preprocessed target battery parameters;

[0023] According to the preset parameter selection rules, static parameters and dynamic parameters are extracted from the target battery parameters, and both the static parameters and the dynamic parameters are time series data;

[0024] After time alignment of the static parameters and the dynamic parameters, for each parameter timestamp, the static parameters and the dynamic parameters belonging to the parameter timestamp are merged into a multi-dimensional feature vector;

[0025] The multi-dimensional feature vectors of all parameter timestamps are integrated into battery feature parameters.

[0026] Optionally, static parameters include open circuit voltage data, internal resistance data and operating temperature data, and dynamic parameters include capacity decay rate, charging curve characteristics, discharge curve characteristics, charging efficiency, self-discharge rate, coulomb efficiency, internal resistance growth rate and cycle life decay rate.

[0027] Optionally, constructing a characteristic fuzzy relationship matrix based on battery characteristic parameters, performing fuzzy comprehensive evaluation on the battery cells according to the characteristic fuzzy relationship matrix, and obtaining a battery status score of each battery cell includes the following steps:

[0028] The parameter weights of static parameters and dynamic parameters are calculated by combining the analytic hierarchy process and the entropy weight method respectively;

[0029] The step function is used as the membership function of the static parameter, the S-type function is used as the membership function of the dynamic parameter, and the characteristic fuzzy relationship matrix is ​​constructed based on the battery characteristic parameters and combined with the parameter weights and membership functions;

[0030] The battery cells are subjected to fuzzy comprehensive evaluation according to the characteristic fuzzy relationship matrix to obtain the battery status score of each battery cell.

[0031] In a second aspect, the present invention further provides an automatic crossover device according to the first aspect, the device comprising an insulating shell and a mobile chassis, the insulating shell is fixedly arranged on the top of the mobile chassis, the insulating shell is divided into a plurality of isolation chambers, a cable sealing sleeve is arranged at a cable crossing point between any two isolation chambers, each isolation chamber is equipped with an independent grounding system, and the mobile chassis is used to serve as a basic support for the insulating shell and also to drive the insulating shell to move;

[0032] The device also includes a main control module, a power management module, a crossover protection module and a crossover execution module. The main control module, the power management module and the crossover protection module are respectively deployed in different isolation chambers in the insulating shell according to different voltage levels. The power management module is respectively connected to the main control module, the crossover protection module and the crossover execution module, and supplies power to the main control module, the crossover protection module and the crossover execution module. The crossover execution module is arranged on the outside of the insulating shell.

[0033] The main control module is connected to the mobile chassis, the jumper protection module and the jumper execution module respectively. The jumper protection module is electrically connected to the jumper execution module. The main control module is used to receive the automatic jumper task and generate an automatic jumper instruction according to the automatic jumper task. The mobile chassis responds to the automatic jumper instruction and moves to the location of the faulty battery unit. The jumper execution module responds to the automatic jumper instruction and completes the automatic jumper processing of the faulty battery unit. The jumper protection module responds to the automatic jumper instruction and protects the jumper execution module during the automatic jumper processing.

[0034] Optionally, the main control module includes a central processing unit, a positioning unit and a communication unit, the positioning unit and the communication unit are both connected to the central processing unit, the positioning unit is used to obtain real-time position information of the automatic jumper device, the communication unit is used to receive the automatic jumper task, the central processing unit is used to parse the battery position information and static parameters of the faulty battery unit from the automatic jumper task, the central processing unit is used to generate automatic jumper instructions in combination with the real-time position information, the battery position information and the static parameters, the automatic jumper instructions include movement instructions, protection module configuration instructions and jumper execution instructions that are executed in sequence, the movement instructions are generated in combination with the real-time position information and the battery position information and using a path planning algorithm, the protection module configuration instructions are generated based on the static parameters, and the communication unit is used to send the movement instructions to the mobile chassis, send the protection module configuration instructions to the jumper protection module, and send the jumper execution instructions to the jumper execution module.

[0035] Optionally, the jumper protection module includes an array switching controller, a relay array, a diode array and a drive circuit, the relay array includes a plurality of solid-state relays connected in parallel, the diode array includes a plurality of Schottky diodes connected in parallel, the relay array is used to reduce the current impact of the automatic jumper processing process, the diode array is used to prevent reverse current and voltage spikes during the automatic jumper processing process, the relay array and the diode array constitute a jumper protection circuit which is electrically connected to the jumper execution module to form a jumper loop, the jumper protection circuit is electrically connected to the array switching controller through the drive circuit, the array switching controller is connected to the communication unit, the array switching controller is used to receive the protection module configuration instructions, and adjust the parallel number of solid-state relays in the relay array according to the protection module configuration instructions, and adjust the parallel number of Schottky diodes in the diode array.

[0036] Optionally, the automatic jumper device also includes a jumper monitoring module, which is deployed in an isolated chamber in the insulating shell. The power management module is connected to the jumper monitoring module and supplies power to the jumper monitoring module. The jumper monitoring module includes a voltage monitoring unit, a current monitoring unit, a temperature monitoring unit and a data processing unit. The voltage monitoring unit, the current monitoring unit and the temperature monitoring unit are all connected to the data processing unit. The data processing unit is connected to the main control module. The voltage monitoring unit and the current monitoring unit are used to monitor the jumper voltage data and the jumper current data of the jumper circuit during the automatic jumper processing. The temperature monitoring unit is used to monitor the device temperature data of the automatic jumper device. The data processing unit is used to monitor the data changes of the jumper voltage data, the jumper current data and the device temperature data. When the jumper voltage data and the jumper current data suddenly change or the device temperature data exceeds a preset temperature threshold, the data processing unit is also used to send an alarm message to the main control module. The main control module is also used to generate a jumper interrupt instruction according to the alarm information, and send the jumper interrupt instruction to the jumper execution module, so that the jumper execution module interrupts the automatic jumper processing of the faulty battery unit.

[0037] Optionally, the jumper execution module includes a microprocessor, a visual perception unit, a servo motor, a linear drive and a robotic arm. The robotic arm is a six-degree-of-freedom articulated structure. A battery clamp is provided at the end of the robotic arm. The visual perception unit, the servo motor and the linear drive are all connected to the microprocessor. The microprocessor is connected to the communication unit. The microprocessor is used to receive jumper execution instructions and generate image acquisition instructions according to the jumper execution instructions. The visual perception unit responds to the image acquisition instructions and collects battery image information of the faulty battery cell. The microprocessor is also used to identify and locate the jumper interface on the faulty battery cell through image recognition technology, and generate robotic arm operation instructions based on the position of the jumper interface. The servo motor and the linear drive respond to the robotic arm operation instructions at the same time, and coordinately control the robotic arm to connect the battery clamp to the jumper interface.

[0038] The beneficial effects of the present invention are:

[0039] The present invention collects the battery parameters of each unit in the battery pack in real time through an intelligent sensor network, and combines preprocessing, feature extraction and fuzzy comprehensive evaluation techniques to achieve accurate evaluation of the battery status. This data-driven approach greatly improves the accuracy and timeliness of fault detection, and can identify potential problem units at an early stage. Secondly, by constructing a battery fault identification model based on a support vector machine, the system can intelligently distinguish between abnormal states and actual faults, reducing the probability of false alarms and missed reports. This not only improves the accuracy of fault diagnosis, but also reduces unnecessary maintenance costs. On the other hand, the present invention seamlessly combines fault identification with automatic cross-connection processing. Once a faulty unit is identified, the system can automatically generate a cross-connection task and control the automatic cross-connection device to execute. This fully automated processing flow significantly improves the response speed, reduces the need for manual intervention, and thus reduces the risk of human operational errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a schematic diagram of the structure of an automatic jumper device in one embodiment of the present application.

[0041] Figure 2 This is a schematic diagram of command transmission of a main control unit in one embodiment of the present application.

[0042] Figure 3 This is a schematic diagram of the module structure of some modules in the automatic jumper device in one embodiment of the present application.

[0043] Figure 4 It is a flowchart of a method for automatic battery jumpering based on battery data detection in one embodiment of the present application.

[0044] Figure 5 This is a schematic diagram of the model structure of a battery life prediction model in one implementation manner of the present application.

[0045] Description of reference numerals:

[0046] 1. Insulating shell; 11. Isolation chamber; 2. Mobile chassis; 3. Main control module; 4. Power management module; 5. Jumper protection module; 6. Jumper execution module; 7. Jumper monitoring module; 8. Jumper circuit. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0048] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0049] The present invention discloses an automatic cross-connection device, referring to Figure 1 The automatic jumper device of the present invention adopts a unique partition isolation design concept, the core of which is to ensure the safety and reliability of the equipment through the combined structure of the insulating shell 1 and the mobile chassis 2. Specifically, the insulating shell 1 is made of high-strength engineering plastic material (such as modified PC / ABS material), which not only has excellent insulation performance, but also can withstand mechanical shocks in daily operation. The interior of the shell is carefully divided into multiple independent isolation chambers 11, and this design is inspired by the explosion-proof design concept of power equipment. Each isolation chamber 11 is equipped with an independent grounding system, using a copper grounding busbar, and the grounding resistance is not greater than 4 ohms, so that even if a chamber fails, it will not affect other chambers. Special cable sealing sleeves are installed at the cable crossings between the chambers. This sealing sleeve adopts a three-layer structure design: the outer layer is a metal protective layer, the middle layer is a silicone sealing layer, and the inner layer is a flame-retardant insulating layer, which can effectively prevent interference and fault propagation between different voltage levels. The mobile chassis 2 adopts an aluminum alloy frame structure and is equipped with four omnidirectional wheels. Each wheel is equipped with an independent drive motor and steering motor to achieve flexible all-round movement. The load-bearing capacity of the chassis is not less than 200kg, the maximum moving speed can reach 1m / s, and the positioning accuracy can reach ±5mm. This structural design not only ensures the safety of equipment operation, but also greatly improves the mobility and adaptability of the equipment. In practical applications, such as in a standard battery room, the device can flexibly move through narrow passages and accurately locate the faulty battery, while ensuring that each functional module is always in a safe isolation state.

[0050] like Figure 1As shown, the automatic jumper device adopts a modular design concept, mainly including four core modules: main control module 3, power management module 4, jumper protection module 5 and jumper execution module 6. The importance of this modular design lies in the realization of functional decoupling and partition management of safety levels. As the energy center of the entire system, the power management module 4 adopts a multi-level power supply architecture design. It contains a main power supply unit (usually a 24V lithium battery pack with a capacity of not less than 100Ah) and a multi-channel DC-DC conversion circuit. The conversion circuit can provide outputs of multiple voltage levels: for example, providing a stable 5V power supply for the main control module 3, a 12V working voltage for the jumper protection module 5, and a 24V driving voltage for the jumper execution module 6. Each output is equipped with overvoltage protection, overcurrent protection and short-circuit protection circuits, and the response time of the protection circuit is less than 10ms. The arrangement of each module in the insulating housing 1 follows the electromagnetic compatibility design principle. The main control module 3 is usually deployed in the middle isolation chamber 11, away from the power part that may generate strong electromagnetic interference. The power management module 4 is located in the lower layer, which is convenient for heat dissipation and wiring. The jumper protection module 5 is arranged in the upper chamber close to the jumper execution module 6 to reduce the control signal transmission distance and improve the response speed. The communication between all modules adopts optical fiber isolation technology, and the communication rate can reach 10Mbps to ensure the reliability of signal transmission. The power cord adopts double-layer shielded cable, with a braided shielding layer on the outer layer and an aluminum foil shielding layer on the inner layer, and the shielding effect is better than 60dB. The wiring of all cables is managed by special cable troughs and kept at a distance of no less than 15cm from the signal line.

[0051] In actual application scenarios, for example, when the system needs to cross-connect a group of faulty batteries, the power management module 4 will first perform a power assessment to ensure that the system has sufficient energy reserves. If the power consumption of a module is detected to be abnormal (exceeding 120% of the rated value), the power management module 4 will immediately start the protection mechanism, first issuing a warning signal, and if the situation persists, it will cut off the power supply to the corresponding module. This multi-level protection mechanism ensures that the system can operate safely and stably under various working conditions. The effect of this modular design is mainly reflected in three aspects:

[0052] Improved safety: By combining physical isolation and electrical isolation, faults can be effectively prevented from spreading;

[0053] Easy maintenance: Each module can be maintained and replaced independently, greatly reducing maintenance time and cost;

[0054] System reliability: Even if a module fails, other modules can still maintain normal operation, improving the system's fault tolerance.

[0055] In one embodiment, referring to Figure 2, the main control module 3, as the brain of the automatic cross-connection device, adopts a layered architecture design, and is composed of a complete intelligent decision-making system consisting of a central processing unit, a positioning unit, and a communication unit. The central processing unit adopts a high-performance ARM architecture, with a main frequency of not less than 1.2GHz, a memory capacity of not less than 2GB, and a real-time operating system to ensure the real-time processing of instructions. The positioning unit integrates an indoor UWB positioning system and an inertial navigation module, wherein the UWB positioning accuracy can reach ±10cm, the gyroscope accuracy of the inertial navigation module is better than 0.1° / h, and the accelerometer accuracy is better than 0.1mg. The communication unit supports Wi-Fi 6, industrial Ethernet, and 4G / 5G communications to ensure the reliability of data transmission. In the actual working process, the communication unit first receives the JSON format task data sent by the automatic cross-connection system, including the faulty battery ID, location coordinates, specification parameters and other information. The central processing unit then parses these data to extract the three-dimensional coordinate position information of the battery, the battery model, the rated voltage and other static parameters, and the task priority information.

[0056] Based on the analysis results, the main control module 3 generates three types of core instructions through the central processor: 1. Mobile instructions: Path planning is performed through the A* algorithm and a segmented path point coordinate sequence is generated; 2. Protection module configuration instructions: Protection level and threshold are set based on battery parameters; 3. Crossover execution instructions configure specific operation parameters. Finally, instructions are distributed to each execution unit through different communication protocols, among which the mobile chassis 2 uses the real-time Ethernet protocol with a cycle of 20ms; the protection module uses the CAN bus with a rate of 500kbps; and the execution module uses the EtherCAT protocol with a synchronization cycle of 1ms. Taking a typical 48V battery crossover task as an example, after the main control module 3 receives the battery position coordinates and parameters, it will plan the optimal path to avoid obstacles, limit the protection current to less than 100A, and generate an accurate robot arm operation sequence. This intelligent control system can achieve a comprehensive positioning accuracy of ±5cm, and the task response delay does not exceed 200ms. At the same time, it ensures operational safety through multiple protection mechanisms and can adapt to the crossover requirements of various types of batteries.

[0057] In one embodiment, when the mobile chassis 2 receives a movement instruction, the built-in positioning system of the mobile chassis 2 first starts the GPS module, and combines the RTK technology to achieve centimeter-level positioning accuracy. At the same time, the inertial measurement unit starts working, providing accurate posture information through the accelerometer and gyroscope. The chassis' obstacle avoidance sensor system, such as a laser radar, an ultrasonic sensor array, or a stereo vision camera, starts scanning the surrounding environment. The electric wheel hub starts moving according to the planned path, and each wheel hub can independently control the speed and direction to achieve precise motion control. During the movement, the obstacle avoidance system continues to work, and if an obstacle is detected, the path will be adjusted in real time. During the entire navigation process, the device will continue to communicate with the main control module 3, report the current position and status, and receive possible path update instructions. When the device arrives near the target position, it will start the precise positioning program, and use the visual recognition system (such as a target detection algorithm based on deep learning) to accurately identify the position of the faulty battery to ensure that the positioning error is less than 1 cm.

[0058] In one embodiment, referring to Figure 3 , the cross-connect protection module 5 adopts the design concept of "multiple redundancy and dynamic adjustment" to achieve all-round protection through array device configuration and intelligent control strategy. The array switching controller adopts a 32-bit microcontroller with a main frequency of not less than 100MHz, an integrated 12-bit ADC, a sampling rate of not less than 1MSPS, supports multiple communication interfaces, and a response time of less than 1ms. The relay array adopts a 10-channel parallel solid-state relay structure. The rated current of a single relay is 100A, the switching time is less than 100μs, the overheat protection is built-in, the withstand voltage is not less than 1000V, and the typical value of the on-state resistance is 0.5mΩ. The diode array uses 12-channel parallel Schottky diodes. The rated current of a single diode is 80A, the forward voltage drop does not exceed 0.4V, the reverse breakdown voltage is 1000V, the reverse recovery time is less than 50ns, and the operating temperature range is -55℃ to +175℃. The drive circuit adopts an optocoupler isolation solution, the drive voltage is 12V, the signal rise and fall times are less than 1μs, and the isolation withstand voltage is 2500VRMS.

[0059] In this embodiment, if Figure 3 As shown, a relay array and a diode array are combined and a series-parallel hybrid topology is adopted to construct a jumper protection circuit. Specifically, each route of the relay array is composed of two solid-state relays connected in series to form a dual control point. The diode array adopts a ladder parallel structure. The 12 Schottky diodes are divided into three groups, each group of 4 is connected in parallel, and the current is balanced between the three groups through the current-sharing resistor. The two arrays are connected in series in the order of "relay array-diode array-current-sharing network" to form a complete protection circuit. In terms of the connection with the jumper execution module 6, a twisted shielded cable (with a cross-sectional area of ​​not less than 50mm) is used. 2), connected to the battery clamp at the end of the robot arm through a specially designed high-voltage connector (rated voltage 1500V, rated current 800A). The connector adopts a quick locking mechanism to ensure reliable connection and easy maintenance during operation.

[0060] When the system starts the jumper operation, the relay array is first turned on step by step, with a delay of 50μs between each stage. This soft start method reduces the surge current. The diode array is always in the on state during the whole process, responsible for preventing reverse current and voltage spikes. The current-sharing network ensures that the current of the parallel branches is evenly distributed to avoid local overload. In practical applications, the jumper circuit 8 can withstand a maximum continuous working current of 800A, and the transient overcurrent capacity reaches 1200A (duration does not exceed 100ms), and it has an insulation withstand voltage of 5000V. Through the optimized circuit design, the voltage drop of the system is controlled within 0.8V, and the energy efficiency conversion rate exceeds 99%, which meets the needs of high-power jumper operations.

[0061] The jumper protection module 5 dynamically adjusts the number of parallel channels according to the load conditions: 3-4 channels are enabled for light load (<200A), 5-7 channels are enabled for medium load (200A-500A), and more than 8 channels are enabled for heavy load (>500A). The protection strategy includes current surge protection, reverse protection and overvoltage protection, which are implemented through relay staggered conduction (interval 50μs) and real-time monitoring mechanisms. Taking the 48V / 500Ah battery jumper as an example, the system will configure 8 parallel channels according to the expected maximum current of 600A, and ensure safety through step-by-step conduction and real-time monitoring. The module can control the current rise rate within 2000A / ms, the overcurrent response time is less than 100μs, has a reverse voltage blocking capability of 1000V, a response time of less than 50ns, a leakage current of less than 1mA, and can control the device temperature rise within 40℃.

[0062] In this embodiment, if Figure 3As shown, the jumper monitoring module 7 serves as the "neural perception system" of the entire automatic jumper device, ensuring the safety and reliability of the jumper process through multi-dimensional real-time monitoring and data analysis. The module adopts a distributed sensing architecture, including a voltage monitoring unit, a current monitoring unit, a temperature monitoring unit and a data processing unit. The voltage monitoring unit adopts high-precision differential sampling technology and uses a precision operational amplifier for signal conditioning. The sampling accuracy reaches 16 bits, the range can cover 0-1000V, the sampling rate is not less than 10kHz, and it has the characteristics of a common mode rejection ratio greater than 100dB, which can effectively avoid electromagnetic interference in industrial environments. The current monitoring unit integrates a dual detection solution of Hall sensor and shunt. The Hall sensor has a range of 0-1000A, an accuracy better than 0.1%, and a response bandwidth of 100kHz. The shunt is made of manganese-copper alloy with a temperature coefficient of less than 20ppm / ℃, and can maintain stable measurement accuracy in the range of -40℃ to 85℃. The temperature monitoring unit is equipped with multiple thermocouples and PT100 temperature sensors to achieve real-time monitoring of the temperature of key components. The temperature measurement range is -50℃ to 200℃, the accuracy is better than ±0.5℃, and the sampling period is 100ms.

[0063] The data processing unit adopts a dual-core processor architecture with a main frequency of 1GHz, equipped with a hardware floating-point unit, and a real-time multi-tasking operating system to ensure the real-time and reliability of data processing. In actual applications, when the device performs a cross-connection process on a group of faulty batteries, the monitoring module will collect and analyze the various parameters in the cross-connection circuit 8 in real time: the voltage mutation detection adopts a sliding window algorithm with a window length of 100ms, and triggers an alarm when the voltage change rate exceeds the set threshold (such as 10V / ms); the current monitoring adopts oversampling technology, and identifies current harmonics through 1024-point FFT analysis, and issues an early warning when the total harmonic distortion exceeds 5%; the temperature monitoring adopts a fuzzy control strategy, and dynamically adjusts the early warning threshold according to the temperature change trend. All collected data are transmitted to the main control module 3 in real time through the CAN bus, with a communication rate of 500kbps, to ensure that the main control module 3 can respond to abnormal situations in a timely manner. When an abnormality is detected, the main control module 3 will complete the safe interruption of the cross-connection operation within 50ms, including disconnecting the relay, retracting the robotic arm, and a series of actions. This all-round monitoring and rapid response mechanism ensures the safety of the cross-connection operation and effectively prevents equipment damage and safety accidents caused by abnormal electrical parameters. Through a large amount of actual application data statistics, the reliability of the monitoring system reaches 99.99%, the false alarm rate is less than 0.1%, and the missed alarm rate is close to zero, providing a strong guarantee for the safe operation of the automatic cross-connection device.

[0064] In this embodiment, if Figure 3As shown, the crossover execution module 6 is the execution terminal of the automatic crossover device, integrating intelligent visual perception and precision mechanical control system. The module adopts a six-degree-of-freedom articulated mechanical arm structure, and realizes precise crossover operation through the coordinated work of a microprocessor, a visual perception unit, a servo motor and a linear drive. The microprocessor adopts an industrial-grade ARM architecture, with a main frequency of 1.5GHz, equipped with a real-time operating system, and an integrated machine vision algorithm library and motion control module. The visual perception unit is equipped with a 2-megapixel industrial camera with a frame rate of 60fps and a matching LED structured light source, which can work stably in complex lighting environments. Image processing adopts a deep learning target detection algorithm, supports the recognition of multiple crossover interfaces, and the recognition accuracy rate reaches 98%, and the average recognition time is less than 100ms. The mechanical arm adopts a modular design, and each joint is equipped with a high-precision harmonic reducer (accuracy better than 1 arc minute) and a 17-bit absolute encoder, and the repeat positioning accuracy reaches ±0.02mm. The terminal battery clamp adopts an adaptive clamping mechanism, equipped with a torque sensor and an anti-slip clamping surface, and the clamping force can be continuously adjusted within the range of 0-500N. The servo motor adopts a permanent magnet synchronous motor with a rated power of 200W, a maximum speed of 3000rpm, and is equipped with a 20-bit incremental encoder. The linear drive adopts a ball screw drive with a stroke of 300mm, a maximum thrust of 1000N, and a positioning accuracy of 0.01mm.

[0065] In actual operation, after receiving the jumper execution instruction, the vision system first collects the battery image and performs in-depth analysis to extract the three-dimensional position information of the jumper interface. The system then plans the optimal robot arm motion trajectory, taking into account obstacle avoidance, speed planning and posture optimization, and guides the robot arm to accurately approach the target through the coordinated control of the servo motor and linear drive. The end effector adaptively adjusts the clamping force according to the force feedback information to ensure a reliable connection. The average time for the entire jumper process does not exceed 30 seconds, and the positioning success rate reaches 99.5%. Through the closed-loop force and position control strategy, mechanical damage to the battery that may be caused during operation is effectively avoided. The execution module can adapt to various types of battery jumper interfaces, and the working range covers a spherical space with a radius of 800mm. The load capacity reaches 10kg, which meets the needs of most field applications.

[0066] The present invention also discloses a battery automatic cross-connection method based on battery data detection, which is applied to a smart substation, in which a battery pack and Figures 1 to 3 In the automatic jumper device shown in FIG, a smart sensor network is installed on the battery pack. Figure 4 FIG. 1 is a flow chart of a method for automatically bridging a battery based on battery data detection in one embodiment. It should be understood that although Figure 4The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 4 At least part of the steps in the above method may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps. Figure 4 As shown, the present invention discloses a battery automatic cross-connection method based on battery data detection, which specifically includes the following steps:

[0067] S101. Obtain the battery location information of each battery unit in the battery pack, and collect the battery parameters of each battery unit through an intelligent sensor network.

[0068] Among them, the key information of each battery cell is obtained through the intelligent sensor network deployed in the battery pack. The sensor network consists of distributed nodes, each of which is equipped with a high-precision position sensor and a multi-parameter acquisition module. The position sensor uses UWB positioning technology to achieve centimeter-level positioning accuracy in indoor environments, and calculates the three-dimensional coordinates (x, y, z) of each battery cell through the TDOA (time difference of arrival) algorithm. The battery parameter acquisition module integrates a voltage sampling unit (sampling accuracy 0.1mV, range 0-100V), a current sampling unit (accuracy 0.1A, range ±1000A), a temperature sensing unit (accuracy 0.1℃, range -20℃ to 80℃) and an internal resistance measurement unit (measurement range 0.1mΩ-1000mΩ). All sensor data are transmitted to the data concentrator in real time via a wireless network (using a Mesh network topology) with a transmission cycle of 100ms. Taking a 48V battery system as an example, it contains 24 2V battery cells. The location information of each cell can be expressed as Pi(xi,yi,zi), where i is the cell number (1-24). The deployment of this sensor network significantly improves the automation of data collection, and the collection efficiency is increased by 95% compared with the manual method. The data accuracy reaches 99.9%, providing a reliable data basis for subsequent analysis. The influence of electromagnetic interference is taken into account in the process of collecting physical parameters, and digital filtering technology is used to eliminate power frequency interference, and the signal-to-noise ratio is better than 60dB.

[0069] S102. Preprocess the battery parameters, extract static parameters and dynamic parameters from the preprocessed battery parameters, and integrate the static parameters and dynamic parameters into battery characteristic parameters.

[0070] Among them, the collected battery parameters are preprocessed and feature extracted to construct a battery characteristic parameter system. In the preprocessing stage, the original data is first detected and processed for outliers. The preprocessed data is divided into two categories: static parameters and dynamic parameters. The three indicators of open circuit voltage, internal resistance and operating temperature are selected as static parameters because these parameters can directly reflect the instantaneous state of the battery: the open circuit voltage reflects the battery charge state and electrochemical activity, the internal resistance characterizes the power characteristics and internal structure integrity of the battery, and the operating temperature is directly related to the electrochemical reaction rate and safety. The selection of dynamic parameters is based on the battery aging mechanism and performance degradation characteristics: the capacity decay rate reflects the decrease in the utilization rate of active substances, the characteristics of the charge and discharge curves reflect the degree of electrochemical polarization, the charging efficiency and coulomb efficiency characterize the energy conversion performance, the self-discharge rate reflects the degree of side reactions, and the internal resistance growth rate and cycle life decay rate directly indicate the aging state. Specifically, the static parameters include:

[0071] Open circuit voltage data (VOC): measured when the battery is in open circuit state, with a measurement accuracy of 0.1mV.

[0072] Internal resistance data (R): measured by AC impedance method, frequency range 0.1Hz-1kHz.

[0073] Operating temperature data (T): The temperature at multiple points on the battery surface is measured using a thermocouple array.

[0074] Dynamic parameters include:

[0075] Capacity decay rate (Cd): Where Cn is the current capacity and C0 is the initial capacity.

[0076] Charging curve characteristics: including charging platform voltage (Vc), charging time (tc), and charging voltage rise rate (δVc / δt).

[0077] Discharge curve characteristics: including discharge platform voltage (Vd), discharge time (td), and discharge voltage drop rate (δVd / δt).

[0078] Charging efficiency (ηc): Where Eout is the released energy and Ein is the input energy.

[0079] Self-discharge rate (Sr): Monthly capacity loss percentage.

[0080] Coulombic efficiency (ηQ): The ratio of the output charge to the input charge.

[0081] Internal resistance growth rate (Rr): Where Rn is the current internal resistance and R0 is the initial internal resistance.

[0082] Cycle life decay rate (Lr): The ratio of the current number of cycles to the rated number of cycles.

[0083] Combine static and dynamic parameters into a feature vector:

[0084] F=[VOC, R, T, Cd, Vc, tc, δVc / δt, Vd, td, δVd / δt, eta, Sr, etaQ, Rr, Lr].

[0085] Through this systematic parameter preprocessing and feature extraction method, the feature vector can fully reflect the static characteristics and dynamic performance degradation trend of the battery, providing a quantitative basis for subsequent fuzzy judgment. The feature extraction accuracy of this method reaches 98%, and the data reliability is improved by 85%.

[0086] S103. Construct a characteristic fuzzy relationship matrix based on the battery characteristic parameters, perform fuzzy comprehensive evaluation on the battery cells according to the characteristic fuzzy relationship matrix, and obtain a battery status score of each battery cell.

[0087] Among them, a characteristic fuzzy relationship matrix is ​​constructed based on the extracted battery characteristic parameters and a fuzzy comprehensive evaluation is performed. The process adopts a three-layer fuzzy evaluation system. The division of the evaluation factor set follows the two-dimensional evaluation system of "static characteristics-dynamic characteristics". Static characteristics reflect the instantaneous state of the battery, which is timely and intuitive; dynamic characteristics reflect the performance change trend of the battery, which is predictive and comprehensive. This division ensures both the real-time nature of the evaluation and the foresight of the evaluation. First, the evaluation factor set U = {U1, U2, U3} is established, where U1 is a static feature subset, including open circuit voltage, internal resistance data, and operating temperature data; U2 is a dynamic feature subset I, including capacity attenuation rate, charge and discharge curve characteristics, and charging efficiency; U3 is a dynamic feature subset II, including self-discharge rate, coulomb efficiency, internal resistance growth rate, and cycle life attenuation rate. The evaluation level set V = {excellent, good, general, poor, fault}. The weight distribution is determined by the analytic hierarchy process (AHP) and entropy weight method, followed by the first-level fuzzy comprehensive evaluation calculation and the second-level fuzzy comprehensive evaluation calculation, and finally the evaluation result vector is obtained, which indicates the membership of the battery status to each level. Finally, the battery status score is calculated according to the membership.

[0088] S104. Mark the battery cells whose battery status scores are lower than a preset score threshold as abnormal battery cells.

[0089] Among them, this step mainly realizes the mark identification of abnormal battery cells, which is completed by comparing the battery status score with the preset score threshold. In this embodiment, the preset score threshold is determined by statistical methods combined with expert experience, and specifically includes three levels of thresholds: warning threshold (75 points), early warning threshold (65 points) and fault threshold (55 points). The theoretical basis for threshold setting is as follows:

[0090] The determination of the warning threshold (75 points) is based on a large amount of historical data statistics, using the 3σ principle, and the calculation formula is:

[0091] T warning =μ-2σ

[0092] Where μ is the mean of normal battery scores (about 90 points), and σ is the standard deviation (about 7.5 points). When the score is lower than this threshold, it indicates that the battery performance begins to decline significantly.

[0093] The warning threshold (65 points) is determined based on the battery performance inflection point analysis:

[0094] T alert =μ-3σ

[0095] At this point, the battery is in a stage of accelerated degradation and requires special attention.

[0096] The failure threshold (55 points) corresponds to the minimum availability standard of the battery:

[0097] T fault =μ-4σ

[0098] Batteries below this threshold can no longer meet normal usage requirements.

[0099] The anomaly labeling process can be achieved by establishing a hierarchical labeling matrix M:

[0100]

[0101] Where i is the battery pack number and j is the unit number.

[0102] Record three consecutive scoring results and construct a time series feature vector:

[0103] V t =[Score t , Score t-1 , Score t-2 ]

[0104] Calculate the rate of change of rating:

[0105]

[0106] If any of the following conditions are met, the battery cell is marked as abnormal:

[0107] 1. The current score is lower than the fault threshold: Score<55.

[0108] 2. The score is lower than the warning threshold and the rate of decline exceeds the threshold: Score < 65 and δ_t < -2 points / day.

[0109] 3. The scores are lower than the warning threshold for three consecutive times: min(V_t)<75.

[0110] S105. Inputting abnormal battery characteristic parameters of the abnormal battery cells into a preset battery fault identification model, and identifying faulty battery cells among the abnormal battery cells through the battery fault identification model.

[0111] Among them, a battery fault identification model is constructed based on support vector machine to achieve accurate identification of faulty battery cells in abnormal battery cells. The selection of support vector machine model is based on the following considerations: it has strong small sample learning ability and generalization ability, can effectively handle high-dimensional nonlinear classification problems, and has relatively low computational complexity. The construction process of the fault identification model is as follows:

[0112] 1. Feature vector construction:

[0113] The input features include two categories: static parameters and dynamic parameters, with a total of 11 indicators:

[0114] X=[VOC,R,T,Cd,Vc c urve,Vd c urve,η c , Sr, η Q ,Rr,Lr]

[0115] Feature normalization:

[0116]

[0117] 2. Construction of training sample set:

[0118] Collect 1000 groups of typical fault samples, including: capacity attenuation fault: 300 groups; internal resistance increase fault: 250 groups; self-discharge increase fault: 200 groups; charge and discharge performance degradation fault: 150 groups; other types of fault: 100 groups. Sample label: y = {1: fault, 0: non-fault}

[0119] 3. Kernel function selection:

[0120] Using radial basis kernel function (RBF):

[0121]

[0122] The kernel parameters are optimized using the grid search method: the penalty parameter C range is: [2^-5, 2^15]; the kernel parameter γ range is: [2^-15, 2^3].

[0123] 4.SVM model training:

[0124] Objective function:

[0125]

[0126] Constraints:

[0127]

[0128] ξ i ≥0, i=1,2,...,n

[0129] Solve to obtain the optimal separating hyperplane:

[0130]

[0131] 5. Model optimization strategy: use PSO algorithm to optimize SVM parameters; feature importance analysis and feature selection; sample balancing processing (SMOTE algorithm); ensemble learning improvement (Bagging method).

[0132] 6. Fault identification process: input abnormal battery feature vector X_test; preprocessing: standardization, dimensionality reduction; model prediction:

[0133] y pred =f(X test )

[0134] Confidence calculation:

[0135]

[0136] S106. If the battery fault identification model identifies a faulty battery cell, an automatic bridging task of the automatic bridging device is generated in combination with the battery location information and static parameters of the faulty battery cell, and the automatic bridging device is controlled by the automatic bridging task to implement automatic bridging processing of the faulty battery cell.

[0137] Among them, this step uses the automatic jumper device to complete the automatic jumper processing of the faulty battery unit. According to the fault identification result and the battery position information, the automatic jumper task T = [ID, Position, Spec, Interface, Priority] is generated, which is received by the main control module and converted into the standard execution instruction format: Task = {header, parameters, timestamp}. The main control module sends the motion instruction through the mobile chassis control interface, uses the improved A* algorithm to plan the optimal path (obstacle avoidance accuracy ±5cm), and controls the mobile chassis to carry the entire device to the faulty battery location. During the movement, the power management module monitors the power supply status of each part of the system in real time to ensure stable operation (voltage fluctuation <±2%). After reaching the specified position, the main control module activates the jumper execution module, which calculates the motion trajectory of the robot arm based on the inverse kinematics algorithm (positioning accuracy ±1mm), and establishes a communication link with the jumper protection module. The jumper protection module starts the protection circuit constructed by the relay array and the diode array, and sets the protection threshold (overcurrent protection 800A, overvoltage protection 50V). The robotic arm of the jumper execution module accurately approaches the faulty battery interface under visual guidance (alignment accuracy <0.5mm) to complete the mechanical connection. During the entire process, all modules are deployed in an insulating shell (insulation withstand voltage >5000V), and real-time data interaction is carried out through the CAN bus (communication delay <10ms). The jumper operation adopts a soft start method, and the current climb rate is controlled within 100A / s. The electrical parameters and mechanical parameters are monitored throughout the process, and the protection mechanism is automatically triggered when an abnormality occurs. The solution achieves precise positioning of the faulty battery (error <10mm), safe jumpering (success rate >99%) and efficient processing (average time <3min), and has a complete safety protection mechanism (fault response time <1ms).

[0138] In one embodiment, the method for automatic battery jumpering based on battery data detection further includes the following steps:

[0139] If the battery fault identification model fails to identify the faulty battery unit, all battery characteristic parameters are converted into multi-dimensional characteristic time series data;

[0140] Combining long short-term memory network and graph convolutional network to build a battery life prediction model;

[0141] Input the multi-dimensional feature time series data into the battery life prediction model, and generate a battery impact association graph of the battery cells based on the graph network construction module in the battery life prediction model and through a data-driven method. The graph network construction module is composed of a long short-term memory network layer and a self-attention mechanism layer. The graph nodes in the battery impact association graph are all battery cells, and the graph node edges in the battery impact association graph represent the physical connection relationship or electrical coupling relationship between the battery cells.

[0142] The high-dimensional impact features are extracted from the battery impact association graph through the feature extraction module in the battery life prediction model. The feature extraction module includes a graph convolution layer, a graph domain conversion layer, a long short-term memory network layer, and a graph domain inverse conversion layer.

[0143] The high-dimensional influencing features are reduced in dimension into low-dimensional influencing features by using the dimensionality reduction prediction module in the battery life prediction model. The life prediction estimate is generated according to the low-dimensional influencing features through the fully connected layer in the dimensionality reduction prediction module and the low-dimensional influencing features are reconstructed. The reconstructed low-dimensional influencing features are back-propagated to the graph network construction module so that the graph network construction module performs residual fusion on the subsequently input multi-dimensional feature time series data and the reconstructed low-dimensional influencing features.

[0144] Determine an aged battery cell with the lowest life prediction estimate value among all abnormal battery cells, and if the life prediction estimate value of the aged battery cell is lower than a preset service life threshold, mark the aged battery cell as a faulty battery cell;

[0145] An automatic bridging task is generated in combination with the battery position information and static parameters of the faulty battery unit, and an automatic bridging device is controlled by the automatic bridging task to realize automatic bridging processing of the faulty battery unit.

[0146] In this embodiment, it is first necessary to perform time series conversion on the characteristic parameters of all batteries. This process can be understood as creating a data table that changes over time. Specifically, a 30-day sliding time window is used, and data is recorded once a day, and the window moves back 1 day each time. For each battery cell, 11 key parameters are collected: open circuit voltage VOC (volts), internal resistance R (milliohms), and operating temperature T (degrees Celsius) as static parameters; capacity decay rate Cd (percentage), charging curve characteristics Vc (slope and inflection point of the voltage-time curve), discharge curve characteristics Vd (platform characteristics of the voltage-time curve), charging efficiency ηc (percentage), self-discharge rate Sr (percentage / month), coulomb efficiency ηQ (percentage), internal resistance growth rate Rr (percentage / month), and cycle life decay rate Lr (percentage / thousand times) as dynamic parameters. These raw data are standardized, and then the db4 wavelet transform is used to reduce the noise of the data to remove measurement noise and environmental interference. Finally, a [30×11]-dimensional feature matrix is ​​obtained, which not only retains the time-varying characteristics of the parameters, but also reflects the correlation between the parameters.

[0147] Reference Figure 5The battery life prediction model adopts a hybrid architecture of LSTM and GCN. This structural design fully considers the timing characteristics and spatial correlation of battery performance degradation. The LSTM network is responsible for processing timing features and includes an input layer, two hidden layers, and an output layer. The input layer receives 11 feature parameters, each hidden layer contains 128 neurons, and uses a tanh activation function. The core of LSTM lies in its special gating mechanism: the forgetting gate ft determines which historical information is discarded, the input gate it controls the input amount of new information, and the output gate ot determines how much information is output. The mathematical expressions of these three gates are:

[0148] f t =σ(W f ·[h t-1 , x t ]+b f )

[0149] i t =σ(W i ·[h t-1 , x t ]+b i )

[0150] o t =σ(W o ·[h t-1 , x t ]+b o )

[0151] The GCN network is used to process the spatial correlation between battery cells. The entire network is trained using the Adam optimizer, with an initial learning rate of 0.001 and a decay of 50% every 50 rounds. The loss function combines mean square error and L2 regularization:

[0152]

[0153] The graph network building module generates a battery impact correlation graph through a data-driven approach, which first uses an LSTM layer to process time series features. The LSTM layer contains 128 hidden units, and the state update of each unit follows:

[0154]

[0155] h t =o t ⊙tanh(c t )

[0156] Where ⊙ represents element-by-element multiplication. The processed features are input into the self-attention mechanism layer, which calculates the association weights between different battery cells. The self-attention calculation process is as follows: First, the input feature X is transformed through three different linear transformations to obtain the query matrix Q, key matrix K and value matrix V, all of which have dimensions [n×d], where n is the number of battery cells and d is the feature dimension. Then the attention score is calculated: Here the scaling factor Used to prevent gradient vanishing.

[0157] The construction of the association graph comprehensively considers two types of relationships: physical connection relationships (such as series-parallel topology) and electrical coupling relationships (such as uneven current distribution, temperature field coupling, etc.). For physical connections, the adjacency matrix is ​​directly constructed based on the circuit topology; for electrical coupling, the weight of the edge is determined based on the parameter correlation coefficient:

[0158]

[0159] In the final association graph G = (V, E), the node set V contains all battery cells, the edge set E represents the association relationship between cells, and the edge weight combines the physical connection strength and the degree of electrical coupling: W ij =α·W phy +(1-α)·W ele , where α is the balance coefficient (usually 0.6). This graph structure representation method can not only intuitively reflect the interaction mechanism within the battery pack, but also support subsequent graph neural network operations. Practice shows that the association recognition accuracy of this method reaches 96%, and it can effectively capture more than 90% of significant influence relationships. The space complexity is O(n^2) and the time complexity is O(n^2d).

[0160] The feature extraction module extracts high-dimensional impact features from the battery impact association graph through a four-layer network structure. First, the graph convolution layer updates the features of each node, and the calculation formula is:

[0161] H (l+1) =σ(D -1 / 2 AD -1 / 2 H (l) W (l) )

[0162] A is the adjacency matrix (indicating the connection relationship between battery cells), D is the degree matrix (indicating the number of connections for each battery cell), H is the node feature matrix, and W is the learnable weight matrix. Two layers of graph convolution are used, the first layer outputs 64-dimensional features, the second layer outputs 32-dimensional features, and the activation function is ReLU. Next, the spectral domain conversion layer maps the features to the frequency domain, using Chebyshev polynomial expansion:

[0163]

[0164] Where Λ is the eigenvalue of the graph Laplacian matrix, and K is 5. This transformation can capture global features and periodic patterns in the graph structure. The transformed features are input into the LSTM layer, which contains 64 hidden units to extract temporal dependencies. Finally, the features are mapped back to the spatial domain through the graph domain inverse transformation layer to maintain the physical interpretability of the features. The inverse transformation uses the inverse transformation of the fast Fourier transform (FFT): x = F -1 X, where X is the frequency domain feature and x is the spatial domain feature. The entire feature extraction process can be expressed as: out =LSTM(FFT -1 (GCN(FFT(H in )))). This multi-level feature extraction method ensures accurate extraction of local features (spatial accuracy > 95%) and effectively captures global features (temporal correlation recognition rate > 92%). The output of the module is a 256-dimensional high-dimensional feature vector that contains the complete performance characteristics of the battery cell.

[0165] The dimensionality reduction prediction module processes and predicts high-dimensional influencing features. First, a variational autoencoder is used to compress the 256-dimensional high-dimensional features into 32-dimensional low-dimensional features. The encoding process uses a three-layer fully connected network:

[0166] Z=f enc (H)=W3·ReLU(W2·ReLU(W1·H+b1)+b2)+b3

[0167] Where H is the input high-dimensional feature, and Z is the generated low-dimensional feature. To ensure the effectiveness of feature compression, KL divergence loss is introduced: L KL =D KL (q(z|x)||p(z)), where q(z|x) is the posterior distribution and p(z) is the prior distribution (standard normal distribution). Then, a three-layer fully connected network is used to generate the lifespan prediction estimate:

[0168] Life pred =W6·ReLU(W5·ReLU(W4·Z+b4)+b5)+b6

[0169] At the same time, the low-dimensional features are reconstructed, and the structure of the reconstructed network is symmetrical with the encoding network:

[0170] Z recon =f dec (Z)=W9·ReLU(W8·ReLU(W7·Z+b7)+b8)+b9

[0171] The reconstruction loss uses the mean square error: The reconstructed low-dimensional features are back-propagated to the graph network building module through residual connections to achieve feature fusion: Where α is the adaptive fusion coefficient, the initial value is set to 0.5, and it is dynamically adjusted during the training process: The total loss function of the entire module is: L total =λ1L pred +λ2L KL +λ3L recon , where λ1=0.5, λ2=0.3, λ3=0.2.

[0172] The life prediction results of all abnormal battery cells are evaluated and screened. First, a multi-level evaluation standard system is established: For each battery cell, its comprehensive health index is calculated:

[0173]

[0174] Among themLife pred,i To predict remaining life expectancy, Life nom is the rated cycle life (usually 2000 times), Cap i is the current capacity, Cap nom is the rated capacity, R i is the current internal resistance, R nom is the initial internal resistance, weight coefficients w1=0.5, w2=0.3, w3=0.2. The preset service life threshold adopts a piecewise function:

[0175]

[0176] Then sort all abnormal units by predicted life value: Life sort =sort({Life pred,i}), select the minimum value as the aging unit: Life min =min({Life pred,i}). If the condition is met: Life min <T life , the unit is marked as a faulty unit. The marking result is stored in a structured data format, and the replacement priority is calculated according to the following formula: This judgment method has high accuracy (fault warning accuracy reaches 97%) and pre-instantaneity (potential faults can be warned 8-12 days in advance), and the misjudgment rate is less than 2%. At the same time, through the dynamic evaluation of the health index, it can effectively avoid misjudgments caused by instantaneous fluctuations (anti-interference ability is improved by 70%). In addition, this method has high computational efficiency (single evaluation time <100ms) and low storage overhead (data volume per unit <1KB), which is suitable for online real-time monitoring applications.

[0177] In one embodiment, preprocessing battery parameters, extracting static parameters and dynamic parameters from the preprocessed battery parameters, and integrating the static parameters and dynamic parameters into battery characteristic parameters includes the following steps:

[0178] The median filtering method based on the adaptive window mechanism is used to process the abnormal values ​​in the battery parameters, and the multiple interpolation method is used to fill the missing values ​​in the battery parameters to obtain the preprocessed target battery parameters;

[0179] According to the preset parameter selection rules, static parameters and dynamic parameters are extracted from the target battery parameters, and both the static parameters and the dynamic parameters are time series data;

[0180] After time alignment of the static parameters and the dynamic parameters, for each parameter timestamp, the static parameters and the dynamic parameters belonging to the parameter timestamp are merged into a multi-dimensional feature vector;

[0181] The multi-dimensional feature vectors of all parameter timestamps are integrated into battery feature parameters.

[0182] In this embodiment, the abnormal values ​​and missing values ​​of the battery parameters are processed. First, the median filter based on the adaptive window is used to process the abnormal values. The adaptive window mechanism dynamically adjusts the window size according to the degree of data fluctuation: s =max(3,ceil(2σ t )), where σ t is the standard deviation of the current period. For each data point in the time series x(t), calculate its median within the window: med (t) = median(x(tk:t+k)). Then the outliers are determined based on the 3σ criterion: if |x(t)-x med (t)|>3σ t , then x(t) is marked as an outlier. For the detected outliers, local weighted regression is used for correction: The weight coefficient uses the Gaussian kernel function: Bandwidth parameter h adaptive adjustment: h = 0.9σ t ·n -0.2 For missing values, multiple imputation is used. First, a Markov chain Monte Carlo (MCMC) model is constructed to estimate the conditional distribution of missing data, and then multiple (usually 5) random sampling fills are performed, each filling is based on the previous result. The final filling value is the weighted average of multiple results.

[0183] The pre-processed battery parameters are classified and extracted based on the preset parameter selection rules. The parameter selection follows the principle of "three highs and one low": high signal-to-noise ratio (signal-to-noise ratio>20dB), high stability (coefficient of variation<0.1), high diagnostic value (information gain>0.3) and low acquisition cost. For each type of parameter, a standardized scoring function is established. The sampling period of static parameters is usually 5 minutes, and the calculation period of dynamic parameters is 30 minutes. All parameters are stored in time series. To ensure the temporal continuity of the data, linear interpolation is performed on data points that exceed the sampling period. The time alignment process adopts a multi-level sliding window mechanism. First, all parameters are initially aligned according to the minimum sampling period (5 minutes). For each time window W(t), the time labels of static parameters and dynamic parameters are mapped to a unified time grid. For parameters with different sampling frequencies, a piecewise linear interpolation method is used for time point matching: To ensure the accuracy of interpolation, a time constraint is introduced: when the time interval between adjacent sampling points exceeds the preset threshold τ (usually 30 minutes), local polynomial regression is used for filling. The aligned parameters are merged into a multi-dimensional feature vector at each timestamp t.

[0184] The multi-dimensional feature vectors of all parameter timestamps are integrated into complete battery feature parameters and organized in a hierarchical matrix structure. First, the basic feature matrix is ​​constructed: base =[X(t1), X(t2), ..., X(t n )] T , where n is the total number of timestamps. To capture the correlation between parameters, the feature correlation matrix is ​​calculated:

[0185] Based on the correlation matrix, a dynamic weight adjustment mechanism is introduced:

[0186] Where λ is the weight adjustment parameter (default value is 1.5), |C i | is the average correlation between the i-th parameter and other parameters, and the feature vector is weighted and integrated. In order to retain the time series characteristics, the time difference matrix is ​​constructed:

[0187] ΔF=[X(t2)-X(t1),X(t3)-X(t2),...,X(t n )-X(t n-1 )]

[0188] At the same time, the sliding statistical features are calculated, including the mean μ(t), standard deviation σ(t), skewness s(t) and kurtosis k(t), with a window size of 12 time points (1 hour). The final feature parameters are composed of the following components:

[0189] F total ={F base,ΔF,[μ(t),σ(t),s(t),k(t)],C,w(t)}

[0190] In order to improve the usability of data, the characteristic parameters are standardized: The integrated feature parameters have the characteristics of multi-scale (time scale, parameter scale) and multi-level (original features, statistical features, and correlation features).

[0191] In one embodiment, a characteristic fuzzy relationship matrix is ​​constructed based on battery characteristic parameters, and a fuzzy comprehensive evaluation is performed on the battery cells according to the characteristic fuzzy relationship matrix to obtain a battery status score of each battery cell, including the following steps:

[0192] The parameter weights of static parameters and dynamic parameters are calculated by combining the analytic hierarchy process and the entropy weight method respectively;

[0193] The step function is used as the membership function of the static parameter, the S-type function is used as the membership function of the dynamic parameter, and the characteristic fuzzy relationship matrix is ​​constructed based on the battery characteristic parameters and combined with the parameter weights and membership functions;

[0194] The battery cells are subjected to fuzzy comprehensive evaluation according to the characteristic fuzzy relationship matrix to obtain the battery status score of each battery cell.

[0195] In this embodiment, the principle of parameter selection is first explained. The three indicators of open circuit voltage, internal resistance and operating temperature are selected as static parameters because these parameters can directly reflect the instantaneous state of the battery: open circuit voltage reflects the battery charge state and electrochemical activity, internal resistance characterizes the power characteristics and internal structural integrity of the battery, and operating temperature is directly related to the electrochemical reaction rate and safety. The selection of dynamic parameters is based on the battery aging mechanism and performance degradation characteristics: the capacity decay rate reflects the decline in the utilization rate of active substances, the characteristics of the charge and discharge curve reflect the degree of electrochemical polarization, the charging efficiency and coulomb efficiency characterize the energy conversion performance, the self-discharge rate reflects the degree of side reactions, and the internal resistance growth rate and cycle life decay rate directly indicate the aging state.

[0196] The parameter weights are determined by combining the analytic hierarchy process (AHP) and the entropy weight method. Calculation process of the analytic hierarchy process:

[0197] Construct the judgment matrix A using the 1-9 scaling method:

[0198] For static parameters:

[0199] Calculate eigenvalues ​​and eigenvectors: A1x = λx;

[0200] Solve the maximum eigenvalue λmax and the corresponding eigenvector, and get the weight vector after normalization:

[0201] W1=[0.539, 0.297, 0.164];

[0202] For dynamic parameters, construct an 8×8 judgment matrix:

[0203]

[0204] The weight vector is obtained by solving the same problem:

[0205] W2 = [0.251, 0.146, 0.146, 0.127, 0.082, 0.127, 0.060, 0.060] Entropy weight method calculation process:

[0206] Normalize the raw data: Calculate the information entropy of the jth indicator:

[0207] in:

[0208] Calculate weights: Get the entropy weight method weight vectors W1 entropy and W2entropy.

[0209] Combination weight calculation:

[0210] The AHP weight and entropy weight are combined using the product method: Finally, the static parameter combination weight is obtained: W1 = [0.45, 0.35, 0.20].

[0211] Dynamic parameter combination weights:

[0212] W2=[0.23, 0.15, 0.15, 0.12, 0.10, 0.12, 0.07, 0.06]

[0213] Next, we construct the membership function. The static parameter membership function (step function) is as follows:

[0214] Open circuit voltage VOC:

[0215]

[0216] Internal resistance R:

[0217]

[0218] Operating temperature T:

[0219]

[0220] The dynamic parameter membership function (S-type function) is as follows:

[0221] For capacity decay rate Cd:

[0222]

[0223] Among them, a=5%, b=7.5%, c=10%.

[0224] Other dynamic parameters use similar S-type functions, and the parameter values ​​are determined according to specific indicators. Next, the fuzzy relationship matrix is ​​constructed:

[0225] Static parameter fuzzy relationship matrix R1:

[0226]

[0227] Dynamic parameter fuzzy relationship matrix R2:

[0228]

[0229] Next, we will conduct fuzzy comprehensive evaluation, including static parameter evaluation:

[0230] Dynamic parameter evaluation:

[0231] Overall judgment: Where W = [0.4, 0.6] is the weight vector of static and dynamic features.

[0232] Finally, the status score is calculated: Where v = [100, 80, 60, 40, 20] is the rating value vector.

[0233] A person skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of protection of the present application is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of different aspects of one or more embodiments of the present application as above, which are not provided in detail for the sake of simplicity.

[0234] One or more embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the present application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of the present application should be included in the protection scope of the present application.

Claims

1. A battery automatic jumper method based on battery data detection, characterized in that: The invention is applied to a smart substation with a battery pack and an automatic jumper device deployed, wherein a smart sensor network is installed on the battery pack, the automatic jumper device comprises an insulating shell and a mobile chassis, the insulating shell is fixedly arranged on the top of the mobile chassis, the automatic jumper device also comprises a main control module, a jumper protection module and a jumper execution module, the main control module is respectively connected to the mobile chassis, the jumper protection module and the jumper execution module, the jumper protection module is electrically connected to the jumper execution module, the main control module is used to receive an automatic jumper task and generate an automatic jumper instruction according to the automatic jumper task, the mobile chassis responds to the automatic jumper instruction and moves to the location of the faulty battery unit, the jumper execution module responds to the automatic jumper instruction and completes the automatic jumper processing of the faulty battery unit, the jumper protection module responds to the automatic jumper instruction and protects the jumper execution module during the automatic jumper processing; The method comprises the following steps: Obtaining the battery location information of each battery unit in the battery pack, and collecting the battery parameters of each battery unit through an intelligent sensor network; Preprocessing battery parameters, extracting static parameters and dynamic parameters from the preprocessed battery parameters, and integrating the static parameters and dynamic parameters into battery characteristic parameters; A characteristic fuzzy relationship matrix is ​​constructed based on the battery characteristic parameters, and a fuzzy comprehensive evaluation is performed on the battery unit according to the characteristic fuzzy relationship matrix to obtain a battery status score for each battery unit; Marking a battery cell whose battery status score is lower than a preset score threshold as an abnormal battery cell; The abnormal battery characteristic parameters of the abnormal battery cell are input into a preset battery fault identification model, and the faulty battery cell in the abnormal battery cell is identified by the battery fault identification model, wherein the battery fault identification model is constructed based on a support vector machine; If the battery fault identification model identifies a faulty battery cell, an automatic bridging task of the automatic bridging device is generated in combination with the battery position information and static parameters of the faulty battery cell, and the automatic bridging device is controlled by the automatic bridging task to realize automatic bridging processing of the faulty battery cell.

2. The method for automatically bridging a battery based on battery data detection according to claim 1, characterized in that: The method further comprises the steps of: If the battery fault identification model fails to identify the faulty battery unit, all battery characteristic parameters are converted into multi-dimensional characteristic time series data; Combining long short-term memory network and graph convolutional network to build a battery life prediction model; Input the multi-dimensional feature time series data into the battery life prediction model, and generate a battery impact association graph of the battery cells based on the graph network construction module in the battery life prediction model and through a data-driven method. The graph network construction module is composed of a long short-term memory network layer and a self-attention mechanism layer. The graph nodes in the battery impact association graph are all battery cells, and the graph node edges in the battery impact association graph represent the physical connection relationship or electrical coupling relationship between the battery cells. The high-dimensional impact features are extracted from the battery impact association graph through the feature extraction module in the battery life prediction model. The feature extraction module includes a graph convolution layer, a graph domain conversion layer, a long short-term memory network layer, and a graph domain inverse conversion layer. The high-dimensional influencing features are reduced in dimension into low-dimensional influencing features by using the dimensionality reduction prediction module in the battery life prediction model. The life prediction estimate is generated according to the low-dimensional influencing features through the fully connected layer in the dimensionality reduction prediction module and the low-dimensional influencing features are reconstructed. The reconstructed low-dimensional influencing features are back-propagated to the graph network construction module so that the graph network construction module performs residual fusion on the subsequently input multi-dimensional feature time series data and the reconstructed low-dimensional influencing features. Determine an aged battery cell with the lowest life prediction estimate value among all abnormal battery cells, and if the life prediction estimate value of the aged battery cell is lower than a preset service life threshold, mark the aged battery cell as a faulty battery cell; An automatic bridging task is generated in combination with the battery position information and static parameters of the faulty battery unit, and an automatic bridging device is controlled by the automatic bridging task to realize automatic bridging processing of the faulty battery unit.

3. The method for automatically bridging a battery based on battery data detection according to claim 1, characterized in that: Preprocessing battery parameters, extracting static parameters and dynamic parameters from the preprocessed battery parameters, and integrating the static parameters and dynamic parameters into battery characteristic parameters includes the following steps: The median filtering method based on the adaptive window mechanism is used to process the abnormal values ​​in the battery parameters, and the multiple interpolation method is used to fill the missing values ​​in the battery parameters to obtain the preprocessed target battery parameters; According to the preset parameter selection rules, static parameters and dynamic parameters are extracted from the target battery parameters, and both the static parameters and the dynamic parameters are time series data; After time alignment of the static parameters and the dynamic parameters, for each parameter timestamp, the static parameters and the dynamic parameters belonging to the parameter timestamp are merged into a multi-dimensional feature vector; The multi-dimensional feature vectors of all parameter timestamps are integrated into battery feature parameters.

4. The method for automatically bridging a battery based on battery data detection according to claim 3, characterized in that: Static parameters include open circuit voltage data, internal resistance data and operating temperature data; dynamic parameters include capacity attenuation rate, charging curve characteristics, discharge curve characteristics, charging efficiency, self-discharge rate, coulomb efficiency, internal resistance growth rate and cycle life attenuation rate.

5. The method for automatically bridging a battery based on battery data detection according to claim 4, characterized in that: The characteristic fuzzy relationship matrix is ​​constructed based on the battery characteristic parameters, and the battery unit is fuzzily comprehensively evaluated according to the characteristic fuzzy relationship matrix to obtain the battery status score of each battery unit, which includes the following steps: The parameter weights of static parameters and dynamic parameters are calculated by combining the analytic hierarchy process and the entropy weight method respectively; The step function is used as the membership function of the static parameter, the S-type function is used as the membership function of the dynamic parameter, and the characteristic fuzzy relationship matrix is ​​constructed based on the battery characteristic parameters and combined with the parameter weights and membership functions; The battery cells are subjected to fuzzy comprehensive evaluation according to the characteristic fuzzy relationship matrix to obtain the battery status score of each battery cell.

6. An automatic jumper device according to any one of claims 1 to 5, characterized in that: The device includes an insulating shell and a mobile chassis. The insulating shell is fixedly arranged on the top of the mobile chassis. The insulating shell is divided into a plurality of isolation chambers. A cable sealing sleeve is arranged at the cable crossing point between any two isolation chambers. Each isolation chamber is equipped with an independent grounding system. The mobile chassis is used to provide a basic support for the insulating shell and also to drive the insulating shell to move. The device also includes a main control module, a power management module, a crossover protection module and a crossover execution module. The main control module, the power management module and the crossover protection module are respectively deployed in different isolation chambers in the insulating shell according to different voltage levels. The power management module is respectively connected to the main control module, the crossover protection module and the crossover execution module, and supplies power to the main control module, the crossover protection module and the crossover execution module. The crossover execution module is arranged on the outside of the insulating shell. The main control module is connected to the mobile chassis, the jumper protection module and the jumper execution module respectively. The jumper protection module is electrically connected to the jumper execution module. The main control module is used to receive the automatic jumper task and generate an automatic jumper instruction according to the automatic jumper task. The mobile chassis responds to the automatic jumper instruction and moves to the location of the faulty battery unit. The jumper execution module responds to the automatic jumper instruction and completes the automatic jumper processing of the faulty battery unit. The jumper protection module responds to the automatic jumper instruction and protects the jumper execution module during the automatic jumper processing.

7. The automatic jumper device according to claim 6, characterized in that: The main control module includes a central processing unit, a positioning unit and a communication unit. The positioning unit and the communication unit are both connected to the central processing unit. The positioning unit is used to obtain the real-time position information of the automatic jumper device. The communication unit is used to receive the automatic jumper task. The central processing unit is used to parse the battery position information and static parameters of the faulty battery unit from the automatic jumper task. The central processing unit is used to generate an automatic jumper instruction in combination with the real-time position information, the battery position information and the static parameters. The automatic jumper instruction includes a movement instruction, a protection module configuration instruction and a jumper execution instruction that are executed in sequence. The movement instruction is generated by combining the real-time position information and the battery position information and adopting a path planning algorithm. The protection module configuration instruction is generated based on the static parameters. The communication unit is used to send the movement instruction to the mobile chassis, send the protection module configuration instruction to the jumper protection module, and send the jumper execution instruction to the jumper execution module.

8. The automatic jumper device according to claim 7, characterized in that: The jumper protection module includes an array switching controller, a relay array, a diode array and a drive circuit. The relay array includes a plurality of solid-state relays connected in parallel, and the diode array includes a plurality of Schottky diodes connected in parallel. The relay array is used to reduce the current impact of the automatic jumper processing process, and the diode array is used to prevent reverse current and voltage spikes from occurring during the automatic jumper processing process. The relay array and the diode array constitute a jumper protection circuit which is electrically connected to the jumper execution module to form a jumper loop. The jumper protection circuit is electrically connected to the array switching controller through the drive circuit, and the array switching controller is connected to the communication unit. The array switching controller is used to receive the protection module configuration instruction, and adjust the parallel number of solid-state relays in the relay array according to the protection module configuration instruction, and adjust the parallel number of Schottky diodes in the diode array.

9. The automatic jumper device according to claim 8, characterized in that: The automatic jumper device also includes a jumper monitoring module, which is deployed in an isolated chamber in the insulating shell. The power management module is connected to the jumper monitoring module and supplies power to the jumper monitoring module. The jumper monitoring module includes a voltage monitoring unit, a current monitoring unit, a temperature monitoring unit and a data processing unit. The voltage monitoring unit, the current monitoring unit and the temperature monitoring unit are all connected to the data processing unit. The data processing unit is connected to the main control module. The voltage monitoring unit and the current monitoring unit are used to monitor the jumper voltage data and the jumper current data of the jumper circuit during the automatic jumper processing. The temperature monitoring unit is used to monitor the device temperature data of the automatic jumper device. The data processing unit is used to monitor the data changes of the jumper voltage data, the jumper current data and the device temperature data. When the jumper voltage data and the jumper current data suddenly change or the device temperature data exceeds the preset temperature threshold, the data processing unit is also used to send an alarm message to the main control module. The main control module is also used to generate a jumper interrupt instruction according to the alarm information, and send the jumper interrupt instruction to the jumper execution module, so that the jumper execution module interrupts the automatic jumper processing of the faulty battery unit.

10. The automatic jumper device according to claim 7, characterized in that: The jumper execution module includes a microprocessor, a visual perception unit, a servo motor, a linear drive and a robotic arm. The robotic arm is a six-degree-of-freedom articulated structure. A battery clamp is provided at the end of the robotic arm. The visual perception unit, the servo motor and the linear drive are all connected to the microprocessor. The microprocessor is connected to the communication unit. The microprocessor is used to receive jumper execution instructions and generate image acquisition instructions according to the jumper execution instructions. The visual perception unit responds to the image acquisition instructions and collects battery image information of the faulty battery unit. The microprocessor is also used to identify and locate the jumper interface on the faulty battery unit through image recognition technology, and generate robotic arm operation instructions based on the position of the jumper interface. The servo motor and the linear drive respond to the robotic arm operation instructions at the same time, and coordinately control the robotic arm to connect the battery clamp to the jumper interface.

Images