A multi-channel signal linkage control system for a controller of an energy storage fire control centralized control cabinet

Abstract

The present application relates to the technical field of energy storage power station fire safety, and particularly relates to a multi-path signal linkage control system of a controller of a fire control centralized control cabinet of energy storage, which comprises a main controller, a PACK level fire control slave station network and a BMS cooperative communication interface; the main controller collects battery PACK temperature and gas concentration signals to form time sequence data and acquires cell parameters output by a battery management system; the time sequence data and the cell parameters are processed by a built-in time sequence prediction model to output a thermal runaway risk probability value; according to a comparison result of the risk probability value and a first risk probability threshold value, a heat dissipation device is controlled to be started by the fire control slave station, and a battery cluster power reduction control message is sent to the battery management system. In the present application, the thermal runaway risk probability value is calculated by fusing temperature, gas concentration characteristics and cell electrical parameters through the time sequence prediction model; when the risk probability exceeds the threshold value and the detector does not alarm, the heat dissipation device is started and the power reduction instruction is sent, so that the energy storage fire linkage control is realized.

Description

Technical Field

[0001] This invention relates to the field of fire safety technology for energy storage power stations, and in particular to a multi-signal linkage control system for an energy storage fire control cabinet. Background Technology

[0002] With the development of new energy storage technology, energy storage power stations are gradually moving towards larger capacity and higher integration. Energy storage containers typically integrate multiple battery packs, battery management systems, and fire detection and control equipment. Existing energy storage fire protection systems mainly employ a multi-level fire protection architecture composed of smoke detectors, temperature detectors, combustible gas detectors, fire controllers, and fire extinguishing actuators. This architecture collects environmental parameters and battery status information from inside the energy storage compartment, detects abnormal conditions, and controls alarm devices, cooling devices, and fire extinguishing equipment to perform corresponding actions based on the detection results. Simultaneously, with the development of energy storage safety monitoring technology, some fire control equipment is beginning to integrate the data acquisition capabilities of the battery management system to comprehensively analyze electrical status parameters such as battery voltage and temperature. This data exchange between fire protection equipment, the battery management system, and the upper-level monitoring platform is achieved through bus communication to meet the status monitoring and safety control needs during the operation of energy storage power stations.

[0003] Existing energy storage fire-fighting linkage control systems mostly rely on detector signals exceeding preset alarm thresholds to trigger a response. Since the system is in a silent waiting state in the early stages of thermal runaway, this results in missing the opportunity for early intervention. Summary of the Invention

[0004] To overcome the above shortcomings, this invention provides a multi-signal linkage control system for energy storage fire protection central control cabinets, aiming to improve the problem that existing energy storage fire protection linkage control mostly relies on detector signals exceeding preset alarm thresholds to trigger a response, which easily leads to missing the opportunity for early intervention.

[0005] This invention provides the following technical solution: a multi-signal linkage control system for an energy storage fire protection central control cabinet, including a main controller, a PACK-level fire control slave network, and a BMS (Battery Management System) collaborative communication interface; The main controller acquires a time-series data sequence consisting of temperature and gas concentration signals for each battery pack, and obtains cell electrical parameters from the battery management system through the BMS collaborative communication interface. The main controller has a built-in time-series prediction model that takes time-series data sequences and cell electrical parameters as inputs and calculates and outputs the thermal runaway risk probability value for each battery pack. The PACK-level fire control slave network includes multiple fire control slaves corresponding to each battery PACK. Each fire control slave is equipped with a heat dissipation drive interface and is communicatively connected to the main controller. The main controller stores a first risk probability threshold. When the thermal runaway risk probability value of a certain battery pack exceeds the first risk probability threshold and the measured signal values ​​of the detectors used to form the time-series data sequence in the battery pack do not reach the preset alarm threshold, the main controller controls the fire control slave station corresponding to the battery pack to start the heat dissipation device corresponding to the battery pack through the heat dissipation drive interface, and at the same time sends a control message to the battery management system through the BMS cooperative communication interface to request a reduction in the charging and discharging power of the battery cluster where the battery pack is located.

[0006] By adopting the above technical solution, the timing characteristics of the temperature signal and gas concentration signal of each battery pack are integrated with the electrical parameters of the battery cell into the timing prediction model of the main controller to obtain the probability value of thermal runaway risk. Then, when the probability value of thermal runaway risk exceeds the first probability threshold and the measured signal values ​​of the detectors have not reached the preset alarm threshold, the heat dissipation device is activated in advance and a power reduction command is sent to the battery management system. This improves the problem that most existing energy storage fire-fighting linkage control systems rely on the detector signal exceeding the preset alarm threshold to trigger a response. Since the system is in a silent waiting state in the early stage of thermal runaway evolution, the opportunity for early intervention is missed.

[0007] Furthermore, the main controller polls the fire control slave stations corresponding to each battery pack via the bus, collecting temperature and gas concentration signals uploaded by each fire control slave station, including: The main controller labels the collected temperature signals and gas concentration signals with corresponding battery PACK identifiers according to the address codes of each fire control slave station; The main controller arranges the temperature signal and gas concentration signal corresponding to the same battery PACK identifier into a time-series data sequence according to the order of acquisition time.

[0008] Furthermore, the main controller sending a request to the battery management system via the BMS collaborative communication interface includes: The main controller receives the cell electrical parameters returned by the battery management system through the BMS collaborative communication interface. The cell electrical parameters include the corresponding battery PACK identifier. The main controller associates and stores the received cell electrical parameters with the corresponding timing data sequence based on the battery PACK identifier.

[0009] Furthermore, the time-series prediction model uses the time-series data sequence corresponding to each battery pack and the cell electrical parameters as inputs. It calculates the temperature rise rate based on the temperature signal in the time-series data sequence, the gas concentration change rate based on the gas concentration signal in the time-series data sequence, and the voltage drop rate and internal resistance growth rate based on the cell electrical parameters, including: The time-series prediction model calculates a risk score based on the rate of temperature rise, the rate of change of gas concentration, the rate of voltage drop, and the rate of increase of internal resistance, and outputs the thermal runaway risk probability value corresponding to each battery pack. The thermal runaway risk probability value is a continuous value between 0 and 1.

[0010] Furthermore, each fire control slave station is installed inside or adjacent to its corresponding battery pack. Each fire control slave station is connected to the main controller via a bus, which simultaneously provides DC power and data communication to each fire control slave station, including: Each fire control slave station is equipped with an address code, and the main controller performs point-to-point addressing of each fire control slave station through the bus using the address code of each fire control slave station; Each fire control slave station is also equipped with a fire extinguishing drive interface, which is connected to the fire extinguishing actuator of the corresponding battery pack via a wire.

[0011] Furthermore, the control message includes a control message identifier, including: After the main controller sends a control message through the BMS collaborative communication interface, the BMS collaborative communication interface receives an acknowledgment message returned by the battery management system, and the acknowledgment message contains the identifier of the control message. If the BMS collaborative communication interface does not receive an acknowledgment message from the battery management system within a preset waiting time, the main controller will send the control message to the battery management system again through the BMS collaborative communication interface.

[0012] Furthermore, the main controller reads the thermal runaway risk probability values ​​corresponding to each battery PACK output by the time-series prediction model, including: The main controller compares the thermal runaway risk probability value corresponding to each battery PACK with the first risk probability threshold one by one. The main controller marks battery packs whose thermal runaway risk probability value exceeds a first risk probability threshold as target battery packs.

[0013] Furthermore, the main controller reads, via the bus, the measured signal values ​​of the detectors uploaded by the fire control slave station corresponding to the target battery PACK, which are used to constitute the timing data sequence, including: The main controller retrieves a preset alarm threshold corresponding to the detector used to constitute the time-series data sequence; The main controller compares the measured signal values ​​with the corresponding preset alarm thresholds one by one. When all measured signal values ​​are less than the corresponding preset alarm thresholds, the target battery PACK is marked as being in a non-alarm state.

[0014] Furthermore, the main controller generates a heat dissipation start command based on the target battery PACK being in a non-alarm state, including: The main controller sends the heat dissipation start command to the fire control slave station corresponding to the target battery PACK via the bus; After parsing the heat dissipation start command, the fire control slave station outputs a drive voltage to the heat dissipation device through the heat dissipation drive interface to start the heat dissipation device; When the thermal runaway risk probability value of the target battery PACK falls below the first risk probability threshold, the main controller controls the fire control slave station corresponding to the battery PACK to shut down the heat dissipation device through the heat dissipation drive interface, and sends a control message to the battery management system to restore charging and discharging power through the BMS cooperative communication interface.

[0015] Furthermore, the main controller determines the battery cluster identifier of the target battery pack based on the battery pack identifier of the target battery pack in a non-alarm state, including: The main controller generates a control message based on the battery cluster identifier, the control message containing the battery cluster identifier and an instruction to request a reduction in charging and discharging power; The main controller sends the control message to the battery management system via the BMS collaborative communication interface.

[0016] The present invention has the following beneficial effects: 1. In this invention, by inputting the timing characteristics of the temperature signal and gas concentration signal of each battery pack and the cell electrical parameters into the timing prediction model of the main controller, a thermal runaway risk probability value is obtained. Then, when the thermal runaway risk probability value exceeds the first risk probability threshold and the measured signal values ​​of the detectors have not reached the preset alarm threshold, the heat dissipation device is activated in advance and a power reduction command is sent to the battery management system. This improves the problem that most existing energy storage fire-fighting linkage control systems rely on the detector signal exceeding the preset alarm threshold to trigger a response. Since the system is in a silent waiting state in the early stage of thermal runaway evolution, the opportunity for early intervention is missed.

[0017] 2. In this invention, a PACK-level fire control slave network includes multiple fire control slave stations corresponding to each battery PACK. Each fire control slave station is equipped with a heat dissipation drive interface and is communicatively connected to the main controller. When the thermal runaway risk probability value of a certain battery PACK triggers an early warning, the main controller only controls the fire control slave station corresponding to that battery PACK to start the heat dissipation device corresponding to that battery PACK through the heat dissipation drive interface. This improves the problem that most existing energy storage fire protection systems use the energy storage compartment as the control unit for full-compartment linkage fire suppression. Because the control granularity is at the compartment level, it causes collateral damage to battery PACKs that have not experienced failures due to the spraying of fire extinguishing agents throughout the compartment.

[0018] 3. In this invention, by setting up a BMS collaborative communication interface through the main controller, when the thermal runaway risk probability value of a certain battery PACK exceeds the first risk probability threshold and the measured signal values ​​of the detectors in the battery PACK do not reach the preset alarm threshold, the main controller actively sends a control message to the battery management system through the BMS collaborative communication interface, requesting a reduction in the charging and discharging power of the battery cluster where the battery PACK is located. This improves the problem that most existing energy storage fire protection systems only receive information from the battery management system in one direction. Since the fire protection system cannot actively issue control commands to the battery management system, there is a lack of two-way collaboration between the fire protection system and the battery management system, and thermal runaway cannot be suppressed from the electrical source. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall structure of a multi-signal linkage control system for an energy storage fire protection central control cabinet according to the present invention; Figure 2 This is a schematic diagram of the process by which the main controller acquires temperature signals, gas concentration signals, and battery cell electrical parameters and forms associated data in a multi-channel signal linkage control system for an energy storage fire-fighting central control cabinet according to the present invention. Figure 3 This is a flowchart illustrating the process of calculating the probability value of thermal runaway risk based on time-series data sequence and cell electrical parameters in a multi-channel signal linkage control system for an energy storage fire-fighting central control cabinet according to the present invention. Figure 4 This is a schematic diagram of the process of fire linkage control based on thermal runaway risk probability value and actual detector signal value in the multi-channel signal linkage control system of the energy storage fire-fighting central control cabinet controller of the present invention; Figure 5 This is a schematic diagram of the connection relationship between the fire control slave station, the main controller, the heat dissipation device, and the fire extinguishing actuator in the multi-signal linkage control system of the energy storage fire control cabinet controller of the present invention. Detailed Implementation

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1: In the first embodiment of the present invention, the present invention provides a multi-signal linkage control system for an energy storage fire protection central control cabinet, such as... Figure 1 The aforementioned includes a main controller, a PACK-level fire control slave network, and a BMS collaborative communication interface; like Figure 2 As shown, the main controller acquires a time-series data sequence consisting of temperature and gas concentration signals for each battery pack, and obtains cell electrical parameters from the battery management system through the BMS collaborative communication interface. Furthermore, the main controller polls the fire control slave stations corresponding to each battery pack via the bus, collecting temperature and gas concentration signals uploaded by each fire control slave station, including: The main controller labels the collected temperature and gas concentration signals with the corresponding battery PACK identifiers based on the address codes of each fire control slave station; The main controller arranges the temperature signal and gas concentration signal corresponding to the same battery PACK identifier into a time-series data sequence according to the order of acquisition time.

[0022] Specifically, the main controller polls the fire control slave stations corresponding to each battery pack at a fixed period via the bus, collecting temperature and gas concentration signals uploaded by each fire control slave station. This fixed period is preferably 1 second. The main controller reads the address code of each fire control slave station and uses the address code as the battery pack identifier to mark the temperature and gas concentration signals collected in the current polling period. The main controller maintains a circular buffer with a length of 60 sampling points in the local memory for each battery pack identifier. Since the polling period is 1 second, the 60 sampling points correspond to the time-series data within a 60-second time window. The temperature and gas concentration signals corresponding to the same battery pack identifier are stored sequentially into the corresponding circular buffer according to the order of collection time. The 60 sets of data in the buffer constitute the time-series data sequence of the battery pack.

[0023] Furthermore, the main controller sends requests to the battery management system via the BMS collaborative communication interface, including: The main controller receives the cell electrical parameters returned by the battery management system via the BMS collaborative communication interface. The cell electrical parameters include the corresponding battery PACK identifier. The main controller associates and stores the received cell electrical parameters with the corresponding timing data sequence based on the battery PACK identifier.

[0024] Specifically, the main controller sends a data acquisition request to the battery management system via the BMS collaborative communication interface using the Controller Area Network (CLAN) bus protocol. This request message contains a target cell address list, which includes the cell number of each cell. The main controller also receives the cell electrical parameters returned by the battery management system via the BMS collaborative communication interface. These parameters include the battery PACK identifier mapped by the battery management system according to the cell's physical location, the voltage value of each cell, and the internal resistance value of each cell. The main controller associates the received cell voltage values ​​and internal resistance values ​​with the corresponding battery PACK timing data sequence in the same data table based on the battery PACK identifier.

[0025] like Figure 3 As shown, the main controller has a time-series prediction model embedded in it. It takes the time-series data sequence and the electrical parameters of the battery cell as inputs and calculates and outputs the thermal runaway risk probability value of each battery pack. Furthermore, the time-series prediction model uses the time-series data sequence corresponding to each battery pack and the cell electrical parameters as inputs. It calculates the temperature rise rate based on the temperature signal in the time-series data sequence, the gas concentration change rate based on the gas concentration signal in the time-series data sequence, and the voltage drop rate and internal resistance growth rate based on the cell electrical parameters, including: The time-series prediction model calculates risk scores based on the rate of temperature rise, the rate of change of gas concentration, the rate of voltage drop, and the rate of increase of internal resistance, and outputs the thermal runaway risk probability value for each battery pack. The thermal runaway risk probability value is a continuous value between 0 and 1.

[0026] Specifically, the main controller has a time-series prediction model embedded within it. This model is a lightweight time-series convolutional neural network model with fewer than 500K parameters. Before leaving the factory, it has been trained using the publicly available lithium-ion battery thermal runaway experimental dataset from NASA's Ames Research Center and then embedded in the main controller's flash memory in quantized format, automatically loading and running upon power-up. The time-series prediction model takes the time-series data sequence corresponding to each battery pack and the cell's electrical parameters as input, extracting the temperature rise rate from the temperature signal in the time-series data sequence. ;in This represents the current temperature value. for Temperature value before the time, It lasts for 10 seconds. The unit is degrees Celsius per second; the rate of change of carbon monoxide concentration is extracted from the gas concentration signal in the time-series data sequence. and changes in volatile organic compound concentration ;in and These represent the current carbon monoxide concentration and volatile organic compound concentration, respectively. and They are respectively The carbon monoxide concentration and volatile organic compound concentration values ​​before the specified time. The unit is ppm per second. The unit is ppm; the rate of voltage drop of each individual cell is extracted from the voltage values ​​of each individual cell in the cell's electrical parameters. ;in This represents the current voltage value of a single cell. for The unit voltage value before the time, The unit is volts per second; the internal resistance growth rate is extracted from the internal resistance values ​​of each individual cell in the battery cell's electrical parameters. ;in The internal resistance value at the current moment. for The internal resistance value before time, when When the internal resistance is less than the preset minimum internal resistance value, it is replaced by the preset minimum internal resistance value. Participate in calculation, The unit is per second; the time-series prediction model concatenates the temperature rise rate, carbon monoxide concentration change rate, volatile organic compound concentration change, single cell voltage drop rate, and internal resistance growth rate into a five-dimensional feature vector, which is then fused through four time-series convolutional layers and one fully connected layer to calculate the risk score. After mapping by the Sigmoid function, the thermal runaway risk probability value corresponding to each battery PACK is output, which is a continuous value between 0 and 1.

[0027] The PACK-level fire control slave network includes multiple fire control slaves corresponding to each battery PACK. Each fire control slave is equipped with a heat dissipation drive interface and is connected to the main controller. Furthermore, each fire control slave station is installed inside or adjacent to its corresponding battery pack. Each fire control slave station is connected to the main controller via a bus, which simultaneously provides DC power and data communication to each fire control slave station, including: Each fire control slave station is equipped with an address code. The main controller uses the address code of each fire control slave station to perform point-to-point addressing of each fire control slave station via the bus. Each fire control slave station is also equipped with a fire extinguishing drive interface, which is connected to the fire extinguishing actuator of the corresponding battery pack via a wire.

[0028] Specifically, the PACK-level fire control slave network includes multiple fire control slave stations corresponding to each battery PACK. Each fire control slave station has a heat dissipation drive interface and is communicatively connected to the main controller. Each fire control slave station is installed inside or adjacent to its corresponding battery PACK. Each fire control slave station is connected to the main controller via a bus, which uses a fire protection dual bus and provides 24V DC power supply and 500kbps data communication. Each fire control slave station is configured with a unique address code, which is set via a DIP switch. The main controller encapsulates the target address code in the address field of the communication frame. The system broadcasts via the bus, and each fire control slave station compares its address field with its own address code. Matching slave stations respond to commands from the main controller, thus achieving point-to-point addressing for each slave station. Each slave station, in addition to its heat dissipation drive interface, also has a fire extinguishing drive interface. The heat dissipation drive interface is connected to a miniature cooling fan within the pack via a wire, and the fire extinguishing drive interface is connected to a miniature aerosol fire extinguishing device within the pack via a wire. Both the heat dissipation drive interface and the fire extinguishing drive interface are MOSFET switching circuits. The main controller sends high-level or low-level commands via the bus to control the on / off state of the corresponding interface. Figure 5 As shown, each fire control slave station is connected to the main controller via a bus and configured with an address code; after receiving the control command sent by the main controller, the fire control slave station controls the heat dissipation device through the heat dissipation drive interface and controls the fire extinguishing actuator through the fire extinguishing drive interface.

[0029] like Figure 4 As shown, the main controller stores a first risk probability threshold. When the thermal runaway risk probability value of a certain battery PACK exceeds the first risk probability threshold and the measured signal values ​​of the detectors used to form the time-series data sequence in the battery PACK do not reach the preset alarm threshold, the main controller controls the fire control slave station corresponding to the battery PACK to start the heat dissipation device corresponding to the battery PACK through the heat dissipation drive interface. At the same time, it sends a control message to the battery management system through the BMS cooperative communication interface to request a reduction in the charging and discharging power of the battery cluster where the battery PACK is located. Furthermore, the control message includes a control message identifier, including: After the main controller sends a control message through the BMS collaborative communication interface, the BMS collaborative communication interface receives an acknowledgment message returned by the battery management system. The acknowledgment message contains the control message identifier. If the BMS collaborative communication interface does not receive an acknowledgment message from the battery management system within the preset waiting time, the main controller will send a control message to the battery management system again through the BMS collaborative communication interface.

[0030] Specifically, the main controller stores a first risk probability threshold, preferably 0.6. When the thermal runaway risk probability value of a certain battery pack exceeds the first risk probability threshold and the measured signal values ​​of the detectors used to form the time-series data sequence within the battery pack do not reach the preset alarm threshold, the main controller controls the fire control slave station corresponding to the battery pack to start the heat dissipation device corresponding to the battery pack via the heat dissipation drive interface. Simultaneously, it sends a control message requesting a reduction in the charging and discharging power of the battery cluster containing the battery pack to the battery management system via the BMS collaborative communication interface. The control message contains a control message identifier, which is determined by… The main controller generates messages sequentially according to the sending order. After sending a control message via the BMS collaborative communication interface, the main controller starts a timer with a preset waiting time, preferably 500 milliseconds. The BMS collaborative communication interface receives an acknowledgment message returned by the battery management system. The acknowledgment message contains an acknowledgment message identifier corresponding to the control message identifier. The main controller clears the timer after comparing the acknowledgment message identifier with the control message identifier. If the timer expires and the BMS collaborative communication interface still does not receive an acknowledgment message returned by the battery management system, the main controller resends the control message to the battery management system via the BMS collaborative communication interface. The maximum number of resends is preferably 3.

[0031] Furthermore, the main controller reads the thermal runaway risk probability values ​​corresponding to each battery pack output by the timing prediction model, including: The main controller compares the thermal runaway risk probability value corresponding to each battery PACK with the first risk probability threshold one by one; The main controller marks battery packs whose thermal runaway risk probability value exceeds the first risk probability threshold as target battery packs.

[0032] Specifically, the main controller reads the thermal runaway risk probability value corresponding to each battery PACK output by the timing prediction model in each calculation cycle. The calculation cycle is the same as the polling cycle, which is 1 second. The main controller compares the thermal runaway risk probability value corresponding to each battery PACK with the first risk probability threshold of 0.6 one by one. The main controller marks the battery PACK with the thermal runaway risk probability value exceeding 0.6 as the target battery PACK.

[0033] Furthermore, the main controller reads the measured signal values ​​of the detectors used to construct the timing data sequence from the fire control slave station corresponding to the target battery PACK via the bus, including: The main controller retrieves the preset alarm thresholds corresponding to the detectors used to construct the time-series data sequence; The main controller compares the measured signal values ​​with the corresponding preset alarm thresholds one by one. When all measured signal values ​​are less than the corresponding preset alarm thresholds, the target battery PACK is marked as being in a non-alarm state.

[0034] Specifically, the main controller reads the measured signal values ​​of the detectors used to construct the time-series data sequence from the fire control slave station corresponding to the target battery PACK via the bus, including the real-time temperature value of the temperature detector and the real-time concentration value of the gas concentration detector; the main controller retrieves the preset alarm thresholds corresponding to the detectors used to construct the time-series data sequence from the parameter storage area. The preset alarm threshold for the temperature detector is 65 degrees Celsius, the preset alarm threshold for the carbon monoxide concentration detector is 200 ppm, and the preset alarm threshold for the volatile organic compound concentration detector is 100 ppm; the main controller compares the measured signal values ​​with the corresponding preset alarm thresholds one by one. When the measured temperature value is less than 65 degrees Celsius, the measured carbon monoxide concentration value is less than 200 ppm, and the measured volatile organic compound concentration value is less than 100 ppm, the target battery PACK is marked as being in a non-alarm state.

[0035] Furthermore, the main controller generates a heat dissipation start command based on the target battery PACK being in a non-alarm state, including: The main controller sends the heat dissipation start command to the fire control slave station corresponding to the target battery PACK via the bus; After the fire control slave station parses the heat dissipation start command, it outputs a drive voltage to the heat dissipation device through the heat dissipation drive interface to start the heat dissipation device. When the thermal runaway risk probability value of the target battery PACK falls below the first risk probability threshold, the main controller controls the fire control slave station corresponding to the battery PACK to shut down the heat dissipation device through the heat dissipation drive interface, and sends a control message to the battery management system to restore charging and discharging power through the BMS cooperative communication interface.

[0036] Specifically, during system initialization, the main controller establishes a mapping table between battery pack identifiers and address codes based on the address codes of each fire control slave station, and stores it in the main controller's local memory. The main controller then searches the fire control slave station address mapping table for the corresponding address code of the target battery pack in a non-alarm state, generating a cooling start command frame containing that address code. The main controller broadcasts the cooling start command frame via the bus. The fire control slave station corresponding to the target battery pack with the matching address code receives and parses the command frame. After obtaining the cooling start command, the microcontroller of the fire control slave station sets the gate of the MOS transistor in the cooling drive interface to a high level, outputting 24V to the cooling device. The driving voltage is applied to start the miniature cooling fan. Subsequently, the main controller continuously acquires the thermal runaway risk probability value of the target battery pack in each calculation cycle. When the thermal runaway risk probability value falls below the first risk probability threshold of 0.6, the main controller generates a heat dissipation shutdown instruction frame containing the address code corresponding to the target battery pack and sends it to the fire control slave station corresponding to the target battery pack via the bus. The fire control slave station sets the gate of the MOS transistor of the heat dissipation drive interface to a low level to shut down the heat dissipation device. At the same time, the main controller sends a control message to the battery management system to restore charging and discharging power through the BMS cooperative communication interface. The control message to restore charging and discharging power carries the target battery cluster identifier and the instruction to restore normal charging and discharging power.

[0037] Furthermore, the main controller determines the battery cluster identifier of the target battery pack based on the battery pack identifier of the target battery pack in a non-alarm state, including: The main controller generates a control message based on the battery cluster identifier. The control message contains the battery cluster identifier and an instruction to reduce the charging and discharging power. The main controller sends control messages to the battery management system via the BMS collaborative communication interface.

[0038] Specifically, the main controller queries the pre-configured PACK-cluster mapping table of the battery management system based on the battery PACK identifier of the target battery PACK in the non-alarm state to determine the battery cluster identifier of the target battery PACK; the main controller generates a power reduction control message according to the battery management system communication protocol based on the battery cluster identifier. The power reduction control message contains the battery cluster identifier and an instruction to reduce the charging and discharging power. The target charging and discharging power value carried by the instruction is preferably 0.2C rate; the main controller sends the power reduction control message to the battery management system via the CAN bus through the BMS cooperative communication interface.

[0039] Example 2: A storage container in an energy storage power station contains 24 battery packs. Under heavy discharge conditions in winter, battery pack #18 experienced slow self-discharge due to partial damage to its internal separator. Within 10 minutes, the temperature inside the pack rose from 28°C to 36°C, the carbon monoxide concentration increased from 2 ppm to 12 ppm, and the volatile organic compound (VOC) concentration simultaneously showed a continuous slight increase. Data from the battery management system showed a slight decreasing trend in the individual cell voltages within the pack, and the internal resistance gradually increased. However, at this time, the measured signal values ​​of the temperature detector, carbon monoxide detector, and VOC concentration detector within the pack were all far below their respective alarm thresholds. Traditional fire suppression systems cannot respond to such early anomalies. To solve the above problem, a multi-signal linkage control system for an energy storage fire suppression control cabinet, provided by this invention, was adopted. The specific implementation process of this system is as follows:

[0040] First, the main controller polls the fire control slave stations corresponding to each battery pack every 1 second via the fire protection bus. Each fire control slave station is installed inside the corresponding battery pack and is set with a unique address code via a DIP switch. The main controller reads the temperature signal and gas concentration signal collected by the composite detector in this pack uploaded by each fire control slave station, and marks the address code as the battery pack identifier in the signal data. According to the acquisition time, the 60 consecutive sets of temperature signals and gas concentration signals corresponding to the same battery pack identifier are arranged into the time sequence data sequence of that pack, thereby continuously recording the dynamic evolution process of the multidimensional status signal of each pack.

[0041] Then, the main controller sends a data acquisition request to the battery management system via the CAN bus through the BMS collaborative communication interface and receives the cell electrical parameters returned by the battery management system. These cell electrical parameters include the individual cell voltage value and individual cell internal resistance value in each battery pack and are marked with the corresponding battery pack identifier. The main controller associates and stores the cell electrical parameters with the time-series data sequence of the same battery pack identifier, so that the thermal characteristics and electrical characteristics of each pack correspond to each other in the same time dimension.

[0042] Next, the main controller loads a lightweight temporal convolutional neural network model embedded in flash memory. This model has been trained before leaving the factory, has 450K parameters, and takes about 35 milliseconds for a single inference. The model takes the temporal data sequence of each battery pack and the associated cell electrical parameters as input. It calculates the temperature rise rate from the temperature signal sequence, the carbon monoxide concentration change rate and volatile organic compound concentration change from the gas concentration signal sequence, and the cell voltage drop rate and internal resistance growth rate from the cell electrical parameters. After concatenating the above five features, the model extracts the temporal coupling features through four temporal convolutional layers and one fully connected layer. After mapping by the Sigmoid function, it outputs the probability value of thermal runaway of the pack within the next 2 minutes. The five features of the 18th pack show a synchronous deterioration trend on the time axis. The thermal runaway risk probability value output by the model climbs from 0.15 to 0.72 within 10 seconds, thus identifying the abnormal pack in the very early stage of thermal runaway before the traditional threshold is triggered.

[0043] Then, the main controller compares the thermal runaway risk probability value of PACK 18 (0.72) with the first risk probability threshold (0.6). If it determines that the probability exceeds the threshold, PACK 18 is marked as the target battery PACK. The main controller then reads the measured temperature and gas concentration values ​​uploaded by the fire control slave station corresponding to PACK 18 via the fire protection bus. It retrieves the preset alarm thresholds stored in the parameter area and compares them one by one. It confirms that the measured temperature value of 36 degrees Celsius is lower than the 65 degrees Celsius threshold, the measured carbon monoxide concentration value of 12 ppm is lower than the 200 ppm threshold, and the measured volatile organic compound concentration value is lower than the 100 ppm threshold. Since none of the measured signal values ​​have triggered an alarm, PACK 18 is marked as being in a non-alarm state, thus completing double confirmation before active intervention.

[0044] Subsequently, the main controller generates a heat dissipation start command based on the battery PACK identifier of PACK 18. The communication frame of the heat dissipation start command encapsulates the address code of the fire control slave station corresponding to PACK 18, and broadcasts it via the fire control bus. After receiving and parsing the command, the fire control slave station of PACK 18 outputs a 24V drive voltage to the miniature cooling fan in this PACK through the MOSFET switching circuit, forcibly dissipating the heat and gas accumulated inside the PACK. At the same time, the main controller queries the PACK-cluster mapping table based on the battery PACK identifier of PACK 18 to determine that this PACK belongs to the second battery cluster, and generates a power reduction control message. The control message contains the cluster identifier of the second battery cluster, a command to request that the charging and discharging power be reduced to 0.2C rate, and an incrementally generated control message identifier. It is sent to the battery management system through the BMS cooperative communication interface. After the battery management system performs the power reduction operation, it returns an acknowledgment message containing the corresponding control message identifier. After receiving the acknowledgment message within a preset waiting time of 500 milliseconds, the main controller completes the closed loop. Thus, by enhancing heat dissipation and limiting electrical load, the risk is eliminated in its infancy before a fire develops.

[0045] After the intervention measures were implemented, the temperature of PACK 18 dropped from 36°C to 33°C within 15 seconds, and the carbon monoxide concentration stabilized at 12ppm and stopped rising. The main controller continuously monitored the thermal runaway risk probability value of this PACK. When the probability value dropped to 0.55, which is lower than the first risk probability threshold of 0.6, the main controller generated a heat dissipation shutdown command and sent it to the fire control slave station of PACK 18 via the fire protection bus to shut down the cooling fan. It also sent a control message to the battery management system through the BMS collaborative communication interface to restore the normal charging and discharging power of the second battery cluster. The system automatically returned to normal operation. During the entire process, no audible or visual alarms were triggered in the energy storage container, and the remaining 23 battery PACKs remained in normal operation.

[0046] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-signal linkage control system for an energy storage fire protection central control cabinet, characterized in that, This includes the main controller, a PACK-level fire control slave network, and a BMS collaborative communication interface; The main controller acquires a time-series data sequence consisting of temperature and gas concentration signals for each battery pack, and obtains cell electrical parameters from the battery management system through the BMS collaborative communication interface. The main controller has a built-in time-series prediction model that takes time-series data sequences and cell electrical parameters as inputs and calculates and outputs the thermal runaway risk probability value for each battery pack. The PACK-level fire control slave network includes multiple fire control slaves corresponding to each battery PACK. Each fire control slave is equipped with a heat dissipation drive interface and is communicatively connected to the main controller. The main controller stores a first risk probability threshold. When the thermal runaway risk probability value of a certain battery pack exceeds the first risk probability threshold and the measured signal values ​​of the detectors used to form the time-series data sequence in the battery pack do not reach the preset alarm threshold, the main controller controls the fire control slave station corresponding to the battery pack to start the heat dissipation device corresponding to the battery pack through the heat dissipation drive interface, and at the same time sends a control message to the battery management system through the BMS cooperative communication interface to request a reduction in the charging and discharging power of the battery cluster where the battery pack is located.

2. The multi-signal linkage control system for an energy storage fire protection central control cabinet according to claim 1, characterized in that, The main controller polls the fire control slave stations corresponding to each battery pack via the bus, collecting temperature and gas concentration signals uploaded by each fire control slave station, including: The main controller labels the collected temperature signals and gas concentration signals with corresponding battery PACK identifiers according to the address codes of each fire control slave station; The main controller arranges the temperature signal and gas concentration signal corresponding to the same battery PACK identifier into a time-series data sequence according to the order of acquisition time.

3. The multi-signal linkage control system for an energy storage fire protection central control cabinet according to claim 1, characterized in that, The main controller sends requests to the battery management system via the BMS collaborative communication interface, including: The main controller receives the cell electrical parameters returned by the battery management system through the BMS collaborative communication interface. The cell electrical parameters include the corresponding battery PACK identifier. The main controller associates and stores the received cell electrical parameters with the corresponding timing data sequence based on the battery PACK identifier.

4. The multi-signal linkage control system for an energy storage fire protection central control cabinet according to claim 1, characterized in that, The time-series prediction model takes the time-series data sequence corresponding to each battery pack and the cell electrical parameters as inputs. It calculates the temperature rise rate based on the temperature signal in the time-series data sequence, the gas concentration change rate based on the gas concentration signal in the time-series data sequence, and the voltage drop rate and internal resistance growth rate based on the cell electrical parameters, including: The time-series prediction model calculates a risk score based on the rate of temperature rise, the rate of change of gas concentration, the rate of voltage drop, and the rate of increase of internal resistance, and outputs the thermal runaway risk probability value corresponding to each battery pack. The thermal runaway risk probability value is a continuous value between 0 and 1.

5. The multi-signal linkage control system for an energy storage fire protection central control cabinet according to claim 1, characterized in that, Each fire control slave station is installed inside or adjacent to its corresponding battery pack. Each fire control slave station is connected to the main controller via a bus, which simultaneously provides DC power and data communication to each fire control slave station, including: Each fire control slave station is equipped with an address code, and the main controller performs point-to-point addressing of each fire control slave station through the bus using the address code of each fire control slave station; Each fire control slave station is also equipped with a fire extinguishing drive interface, which is connected to the fire extinguishing actuator of the corresponding battery pack via a wire.

6. The multi-signal linkage control system for an energy storage fire-fighting central control cabinet according to claim 1, characterized in that, The control message contains a control message identifier, including: After the main controller sends a control message through the BMS collaborative communication interface, the BMS collaborative communication interface receives an acknowledgment message returned by the battery management system, and the acknowledgment message contains the identifier of the control message. If the BMS collaborative communication interface does not receive an acknowledgment message from the battery management system within a preset waiting time, the main controller will send the control message to the battery management system again through the BMS collaborative communication interface.

7. The multi-signal linkage control system for an energy storage fire protection central control cabinet according to claim 1, characterized in that, The main controller reads the thermal runaway risk probability values ​​corresponding to each battery pack output by the time-series prediction model, including: The main controller compares the thermal runaway risk probability value corresponding to each battery PACK with the first risk probability threshold one by one. The main controller marks battery packs whose thermal runaway risk probability value exceeds a first risk probability threshold as target battery packs.

8. A multi-signal linkage control system for an energy storage fire-fighting central control cabinet according to claim 7, characterized in that, The main controller reads, via the bus, the measured signal values ​​of the detectors used to construct the timing data sequence uploaded by the fire control slave station corresponding to the target battery PACK, including: The main controller retrieves a preset alarm threshold corresponding to the detector used to constitute the time-series data sequence; The main controller compares the measured signal values ​​with the corresponding preset alarm thresholds one by one. When all measured signal values ​​are less than the corresponding preset alarm thresholds, the target battery PACK is marked as being in a non-alarm state.

9. A multi-signal linkage control system for an energy storage fire-fighting central control cabinet according to claim 8, characterized in that, The main controller generates a heat dissipation start command based on the target battery PACK being in a non-alarm state, including: The main controller sends the heat dissipation start command to the fire control slave station corresponding to the target battery PACK via the bus; After parsing the heat dissipation start command, the fire control slave station outputs a drive voltage to the heat dissipation device through the heat dissipation drive interface to start the heat dissipation device; When the thermal runaway risk probability value of the target battery PACK falls below the first risk probability threshold, the main controller controls the fire control slave station corresponding to the battery PACK to shut down the heat dissipation device through the heat dissipation drive interface, and sends a control message to the battery management system to restore charging and discharging power through the BMS cooperative communication interface.

10. A multi-signal linkage control system for an energy storage fire-fighting central control cabinet according to claim 8, characterized in that, The main controller determines the battery cluster identifier of the target battery pack based on the battery pack identifier of the target battery pack when it is in a non-alarm state, including: The main controller generates a control message based on the battery cluster identifier, the control message containing the battery cluster identifier and an instruction to request a reduction in charging and discharging power; The main controller sends the control message to the battery management system via the BMS collaborative communication interface.

Images