Intelligent fire prevention and control fire-fighting management system and method for lithium battery energy storage power station

Abstract

The invention provides an intelligent fire prevention and control fire-fighting management system and method for a lithium battery energy storage power station, and relates to the technical field of data processing, and the method comprises the steps: arranging a multi-level temperature sensor, a gas detector and a smoke detector in and outside a battery module, and collecting and forming multi-source monitoring data; performing time sequence rate analysis, and if a preset condition is met, generating early warning data; in combination with sensor matrix distribution in the battery module, spatial cross comparison is executed to obtain abnormal positioning data, and false anomalies are eliminated through multi-point consistency verification; generating a directional processing instruction corresponding to the injection unit, and setting injection data in the processing instruction; the spraying unit is controlled to release an atomized fire extinguishing medium, so that the medium forms a permeation flow field in the gaps of the battery cells, and treatment feedback data is acquired and formed; performing trend analysis, and if the trend is not enough to recover to the preset safety range, generating a secondary processing instruction; according to the invention, the autonomy and accuracy of fire control management are improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a smart fire prevention and control management system and method for lithium battery energy storage power stations. Background Technology

[0002] Currently, the common approach to fire prevention in lithium battery energy storage power stations is to deploy temperature sensors, smoke detectors, gas detectors, and other devices around the battery compartment or battery modules, combined with a monitoring system or battery management system, to monitor environmental parameters in real time. When the monitored values ​​exceed a set threshold, the system will trigger an alarm and activate corresponding fire extinguishing devices, such as inert gas extinguishing or spray cooling devices, to achieve initial fire control.

[0003] In practical applications, such as within the containerized battery compartments of energy storage power stations, battery modules are densely packed. If a single cell experiences thermal runaway, heat rapidly accumulates inside the module. External sensors, due to their limited detection range, are often only triggered after the gas or smoke has diffused to a certain concentration. By this time, the cell may have already decomposed or ejected. Even if a fire extinguishing device is activated, the extinguishing medium may not be able to fully penetrate the module, leading to persistent localized high temperatures and potentially causing secondary combustion or reignition. Summary of the Invention

[0004] The purpose of this invention is to provide a smart fire prevention and control management system and method for lithium battery energy storage power stations, aiming to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: Firstly, a smart fire prevention and fire management method for lithium battery energy storage power stations, the method comprising: Multi-level temperature sensors, gas detectors, and smoke detectors are deployed inside and outside the battery module to collect and form multi-source monitoring data. Time series rate analysis is performed on multi-source monitoring data. If the temperature rise rate of the internal monitoring signal exceeds the preset rate threshold and there is a trigger time difference with the external monitoring signal, early warning data is generated and a priority label is assigned to the corresponding cell area. Based on early warning data and combined with the sensor matrix distribution within the battery module, spatial cross-comparison is performed to obtain anomaly location data, and false anomalies are eliminated through multi-point consistency verification. Based on the abnormal location data, a directional treatment command for the corresponding injection unit is generated, and the injection data is set in the treatment command, including the injection dose, injection angle, injection pressure and injection duration; According to the directional disposal command, the spray unit is controlled to release atomized fire extinguishing medium, so that the medium forms a permeation flow field in the gap between the battery cells, and disposal feedback data is collected; The feedback data is analyzed for trends. If the temperature or gas value does not drop sufficiently to return to the preset safe range within the preset time limit, a secondary treatment instruction is generated, and the injection data is adjusted in the secondary treatment instruction. Based on the feedback data and secondary handling instructions, event log data is generated, and the event log data is transformed into optimization parameters and fed back to the monitoring platform.

[0006] Preferably, time-series rate analysis is performed on the multi-source monitoring data. If the temperature rise rate of the internal monitoring signal exceeds a preset rate threshold and there is a trigger time difference with the external monitoring signal, early warning data is generated, and a priority label is assigned to the corresponding cell area, including: Based on multi-source monitoring data, time synchronization and noise reduction processing are performed to generate aligned monitoring data; Based on the aligned monitoring data, time-by-time differential analysis is performed to generate internal temperature rise rate sequences and external temperature rise rate sequences. Based on the internal temperature rise rate sequence, threshold comparison is performed and time difference evaluation is conducted with the trigger time of the external temperature rise rate sequence to generate candidate warning segments; Based on the candidate early warning segments, the continuity and stability are checked, transient fluctuations are screened out, and early warning data is generated; Based on early warning data, and combined with the magnitude, duration and spatial location of temperature rise, priority labels are generated and linked to the early warning data.

[0007] Preferably, based on early warning data and the sensor matrix distribution within the battery module, spatial cross-comparison is performed to obtain anomaly location data, and false anomalies are eliminated through multi-point consistency verification, including: Based on the temperature sensors, gas detectors, and smoke detectors deployed within the battery module, a sensor matrix is ​​established according to their spatial coordinate relationships, generating matrix distribution data. Based on matrix distribution data and early warning data, temperature and gas signals of the target core point and its multiple neighboring points in the sensor matrix are extracted to generate neighborhood monitoring data. Based on neighborhood monitoring data, coordinate mapping and amplitude normalization are performed to generate a local spatial feature map; Based on the local spatial feature map, perform spatial cross-comparison to identify sets of adjacent points with consistent trends and generate abnormal connected regions. Based on the abnormal connectivity areas, and combined with the battery module structure layout, preliminary abnormal location data is generated. Based on the preliminary anomaly location data, multi-point consistency verification and historical comparison verification are performed to remove single-point deviations and abnormal unstable segments, thereby generating anomaly location data.

[0008] Preferably, based on the abnormal positioning data, a directional handling command for the corresponding injection unit is generated, and the injection data is set in the handling command, including: Based on the abnormal location data, the target injection unit is generated by registering it with the spatial arrangement mapping relationship of the injection unit. Based on the anomaly location data, combined with the temperature rise rate, anomaly duration, number of involved battery cells, and distance between battery cell gaps and nozzles, a penetration coverage requirement dataset is generated. Based on the penetration coverage requirement dataset, the initial injection dose, injection angle, injection pressure, and injection duration are calculated to generate a draft of targeted treatment instructions; According to the draft of the targeted disposal instruction, an interlock check and spray path accessibility check are performed before execution to generate an executable targeted disposal instruction; Based on the executable directional treatment command, control is issued to the target injection unit to initiate directional injection.

[0009] Preferably, trend analysis is performed on the treatment feedback data. If the temperature or gas value does not decrease sufficiently to recover to a preset safe range within a preset time limit, a secondary treatment command is generated, and the injection data is adjusted in the secondary treatment command, including: Based on the processing feedback data, time synchronization and noise reduction are performed to generate aligned feedback data; Based on the alignment feedback data, the rate of temperature decrease and the rate of gas concentration decrease are extracted, and the magnitude of change within a preset time limit is calculated to generate residual anomaly assessment results. Based on the residual anomaly assessment results, if the decrease is insufficient to restore the situation to a safe level, a draft secondary response instruction will be generated. According to the draft secondary treatment instruction, the spraying data is adjusted, and continuous spraying, intermittent spraying, or pulse spraying is selected based on the distribution of residual anomalies to generate a secondary treatment instruction; According to the secondary treatment command, the linkage control spray unit implements secondary spraying and stores the secondary treatment command and treatment feedback data together.

[0010] Preferably, based on the local spatial feature map, spatial cross-comparison is performed to identify sets of adjacent points with consistent trends, generating abnormal connected regions, including: Based on the local spatial feature map, the target core point signal and the adjacent point signal are synchronously compared within a time window to generate multi-point trend comparison data. Based on multi-point trend comparison data, the difference magnitude and trend similarity between adjacent points and core points are calculated to form a similarity assessment result; Based on the similarity assessment results, a set of adjacent points with consistent trends and differences in amplitude changes within a set range is selected to form an outlier set. Based on the set of anomalies, generate an anomaly connected region containing the target core point and its adjacent points.

[0011] Preferably, based on the preliminary anomaly location data, multi-point consistency verification and historical comparison verification are performed to remove single-point deviations and unstable segments, generating anomaly location data, including: Based on the preliminary anomaly location data, multiple monitoring signals within the area are extracted to form a location neighborhood signal set; Based on the location neighborhood signal set, consistency verification is performed. If the neighboring points and the core point maintain continuous similarity in trend and amplitude, they are judged as valid anomalies. Based on the location neighborhood signal set, and in conjunction with the stored historical operation data, a comparison and verification are performed. If the abnormal signal deviates significantly from the historical stable pattern, it is determined to be a valid anomaly; otherwise, it is determined to be a false anomaly. Based on the combined results of consistency verification and historical comparison verification, single-point deviations and abnormal unstable segments are removed to generate corrected anomaly location data.

[0012] Preferably, based on the anomaly location data, combined with the temperature rise rate, anomaly duration, number of involved battery cells, and distance between the battery cell gap and the nozzle, a penetration coverage requirement dataset is generated, including: Based on the anomaly location data, the temperature rise rate and duration within the anomaly area are extracted to form thermal intensity parameters; Based on the abnormal location data, the number of battery cells involved and the density of battery cell arrangement are counted to form regional impact parameters; The injection geometry parameters are formed based on the cell gap width and the distance from the nozzle to the target area; The thermal intensity parameters, regional influence parameters, and injection geometry parameters are combined to generate a permeability coverage requirement dataset.

[0013] Preferably, based on the penetration coverage requirement dataset, the initial injection dose, injection angle, injection pressure, and injection duration are calculated to generate a draft targeted treatment instruction, including: Based on the thermal intensity parameter in the penetration coverage requirement parameters, combined with the temperature rise rate and the duration of the anomaly, the cooling requirement is determined, and the initial injection dose is obtained. Based on the geometric parameters in the penetration coverage requirement parameters, and combined with the spatial angle between the nozzle position and the cell gap, the spray angle is determined to achieve spray coverage of the center of the abnormal area. Based on the geometric parameters in the penetration coverage requirement parameters, combined with the nozzle distance and cell gap width, the driving force requirement is determined, and the injection pressure is obtained; Based on the regional impact parameters in the penetration coverage requirement parameters, combined with the parameters of the number of battery cells and thermal intensity involved, the required continuous cooling time is calculated to obtain the spraying duration; Based on the initial injection dose, injection angle, injection pressure, and injection duration, a draft directional treatment instruction containing multiple parameters is generated.

[0014] Secondly, a smart fire prevention and fire management system for lithium battery energy storage power stations, the system comprising: The multi-source monitoring module is used to collect data from multi-level temperature sensors, gas detectors, and smoke detectors inside and outside the battery module, forming multi-source monitoring data. The early warning generation module is used to perform time series rate analysis on multi-source monitoring data. If the temperature rise rate of the internal monitoring signal exceeds the preset rate threshold and there is a trigger time difference with the external monitoring signal, early warning data is generated and a priority label is assigned to the corresponding cell area. The anomaly localization module is used to perform spatial cross-comparison based on early warning data and the sensor matrix distribution within the battery module to obtain anomaly localization data, and to eliminate false anomalies through multi-point consistency verification. The treatment instruction module is used to generate directional treatment instructions for the corresponding injection unit based on the abnormal location data, and to set the injection data in the treatment instructions, including the injection dose, injection angle, injection pressure and injection duration. The directional spray module is used to control the spray unit to release atomized fire extinguishing medium according to the directional disposal command, so that the medium forms a permeation flow field in the gap between the battery cells, and to collect and form disposal feedback data; The trend analysis module is used to perform trend analysis on the treatment feedback data. If the temperature or gas value does not drop enough to recover to the preset safety range within the preset time limit, a secondary treatment instruction is generated, and the injection data is adjusted in the secondary treatment instruction. The event logging module is used to generate event logging data based on the handling feedback data and secondary handling instructions, and to convert the event logging data into optimization parameters and feed them back to the monitoring platform.

[0015] The above-described solution of the present invention has at least the following beneficial effects: First, by simultaneously deploying multi-level temperature sensors, gas detectors, and smoke detectors in the cooling channels inside the battery module, between the cells, and outside the module, multi-source monitoring data is generated. Compared with deploying detectors only on the outside, the temperature rise signal inside the cell can be captured earlier, thereby solving the response lag problem caused by the limited detection location in the existing technology.

[0016] Furthermore, by performing time-series rate analysis on multi-source monitoring data and generating early warning data using the trigger time difference between internal and external signals, anomalies can be identified before external smoke or gas accumulates to the trigger concentration. Simultaneously, by combining temperature rise magnitude and duration to generate priority labels, the system can rationally prioritize based on risk severity, ensuring that the most dangerous areas are dealt with first.

[0017] Furthermore, by combining spatial cross-comparison with sensor matrix distribution and eliminating false anomalies through multi-point consistency verification, early warning positioning can be accurate down to specific cell areas. This approach not only improves the accuracy of anomaly positioning but also avoids misjudgments caused by single-point anomalies or instantaneous fluctuations, ensuring the reliability of handling commands.

[0018] Subsequently, based on the abnormal location data, a directional treatment command is generated, including the spray dosage, spray angle, spray pressure, and spray duration. This allows the extinguishing medium to be sprayed directly into the gaps between the battery cells, creating a penetrating flow field. This process effectively solves the problem of traditional fire extinguishing methods failing to penetrate into the battery cells, enabling timely cooling and suppression of hotspots and reducing the risk of reignition.

[0019] Finally, after the injection is executed, the system collects and analyzes the feedback data. If the temperature or gas values ​​fail to return to a safe range within a preset time limit, a secondary response command is automatically generated and the injection data is adjusted. This closed-loop control mechanism ensures that the fire extinguishing process can adaptively adjust according to the actual cooling effect, thereby improving the thoroughness and safety of the overall fire extinguishing. Simultaneously, the generation of event log data and the feedback of optimized parameters enable the system to continuously improve, providing data support for the long-term operation and maintenance and risk management of energy storage power stations.

[0020] In summary, this invention, through a progressive mechanism of "early warning - precise positioning - targeted handling - closed-loop verification," not only overcomes the shortcomings of existing technologies such as monitoring lag and insufficient media penetration, but also reduces the risk of secondary combustion and reignition, achieving rapid, accurate, and intelligent prevention and control of fires in lithium battery energy storage power stations. Attached Figure Description

[0021] Figure 1 This is a flowchart of a smart fire prevention and fire management method for lithium battery energy storage power stations provided in an embodiment of the present invention. Detailed Implementation

[0022] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0023] like Figure 1 As shown, embodiments of the present invention propose a smart fire prevention and fire management method for lithium battery energy storage power stations, the method comprising: Multi-level temperature sensors, gas detectors, and smoke detectors are deployed inside and outside the battery module to collect and form multi-source monitoring data. Time series rate analysis is performed on multi-source monitoring data. If the temperature rise rate of the internal monitoring signal exceeds the preset rate threshold and there is a trigger time difference with the external monitoring signal, early warning data is generated and a priority label is assigned to the corresponding cell area. Based on early warning data and combined with the sensor matrix distribution within the battery module, spatial cross-comparison is performed to obtain anomaly location data, and false anomalies are eliminated through multi-point consistency verification. Based on the abnormal location data, a directional treatment command for the corresponding injection unit is generated, and the injection data is set in the treatment command, including the injection dose, injection angle, injection pressure and injection duration; According to the directional disposal command, the spray unit is controlled to release atomized fire extinguishing medium, so that the medium forms a permeation flow field in the gap between the battery cells, and disposal feedback data is collected; The feedback data is analyzed for trends. If the temperature or gas value does not drop sufficiently to return to the preset safe range within the preset time limit, a secondary treatment instruction is generated, and the injection data is adjusted in the secondary treatment instruction. Based on the feedback data and secondary handling instructions, event log data is generated, and the event log data is transformed into optimization parameters and fed back to the monitoring platform.

[0024] In this embodiment of the invention, by deploying multi-level temperature sensors, gas detectors, and smoke detectors inside and outside the battery module, data can be collected from different locations and levels to form multi-source monitoring information, thereby achieving comprehensive coverage inside the battery cell, at the module boundary, and in the external space. This multi-level acquisition method avoids the monitoring blind spot problem caused by relying on a single-area sensor, and is conducive to timely detection of early signs of thermal runaway.

[0025] Time-series rate analysis is performed on the collected multi-source monitoring data. When the temperature rise rate of the internal signal significantly exceeds a preset threshold and there is a trigger time difference with the external signal, early warning data can be generated and assigned a priority label. This processing mechanism can identify potential risks before external detectors are triggered, improving the timeliness of fire detection. At the same time, the priority labels are used to rank the degree of risk, making subsequent handling more targeted.

[0026] By combining early warning data with sensor matrix distribution, spatial cross-comparison is performed, and multi-point consistency verification is used to obtain anomaly location results, which can effectively improve the accuracy of location. This not only narrows down the anomaly point to a specific cell area, but also avoids misjudgments caused by single-point errors or sporadic signal interference, thus improving the reliability of the system in complex operating environments.

[0027] Based on the anomaly location results, a targeted response command is generated, specifying spray data such as spray dosage, spray angle, spray pressure, and spray duration. This ensures a high degree of matching between the extinguishing medium and the target area. In this way, the extinguishing medium can act more precisely on hot spots within the cell gaps, thereby improving penetration and cooling efficiency.

[0028] After the targeted response command is issued, the spray unit releases atomized extinguishing medium, creating a permeating flow field within the gaps between the battery cells. This ensures complete coverage of the internal battery cells, effectively preventing hot spots from remaining due to insufficient extinguishing agent or uneven diffusion. Simultaneously, the system collects real-time response feedback data, dynamically monitoring the extinguishing effect and providing a basis for subsequent adjustments.

[0029] The system performs trend analysis on the feedback data from the fire suppression efforts. If the temperature or gas levels fail to return to a safe range within a preset time limit, a secondary response command is automatically generated, and the spray data is adjusted. This closed-loop control mechanism avoids the possibility of residual hazards after a single fire suppression operation, enabling the fire suppression process to adaptively adjust according to the actual situation and reduce the risk of reignition.

[0030] Ultimately, the feedback data from the initial response and secondary response instructions are used to generate event logs, which are then transformed into optimization parameters and fed back to the monitoring platform. This enables the tracing and learning of the entire fire response process. This not only provides a basis for subsequent operation and maintenance but also allows for continuous optimization of the early warning and response models, giving the system the ability to self-improve.

[0031] For example, in a containerized energy storage power station, when a row of cells experiences a rapid temperature rise due to localized thermal runaway, internal sensors first detect the rate of temperature rise and generate early warning data. The system uses sensor matrix comparison to pinpoint the abnormal area as a cluster of cells within a specific module. It then issues a directional spray command, controlling the corresponding nozzles to release atomized fire extinguishing media to cover the gaps between the abnormal cells. When feedback data shows insufficient temperature reduction, the system immediately triggers a secondary spray and adjusts the pressure and duration until the data returns to a safe range. Records generated throughout the process are transmitted to the monitoring platform, serving as the basis for subsequent operation and maintenance and risk modeling, thus achieving rapid, accurate, and closed-loop fire prevention and control.

[0032] This includes deploying multiple layers of temperature sensors, gas detectors, and smoke detectors inside and outside the battery module to collect and generate multi-source monitoring data, specifically including: Miniature temperature sensors are installed in the internal cooling channels and cell gaps of each battery module to collect real-time temperature information of the cell surface and the air in the gaps. Smoke detectors and gas detectors are installed around the module casing to detect combustible gases and smoke generated by thermal runaway. A sensor matrix is ​​established by uniformly calibrating sensors at different locations according to spatial coordinates and time sampling points. During operation, the controller collects data from each sensor according to a set cycle to form temperature sequences, gas concentration sequences, and smoke concentration sequences. These sequences are then integrated into multi-source monitoring data after time synchronization and signal denoising processing for subsequent anomaly analysis and judgment.

[0033] The preset rate threshold specifically includes: Before system deployment, experimental data from multiple lithium battery modules under different operating conditions were analyzed to determine the typical rate range of cell temperature rise under normal operating conditions. The upper limit of this range was used as a safety reference, and after adjustments based on factors such as cell capacity, cooling conditions, and ambient temperature, a critical value for the temperature rise rate was determined as a rate threshold. During operation, when the rate of internal temperature change detected by the sensors exceeds this critical value, it can be determined that the current cell temperature change exceeds the normal operating fluctuation range, providing a basis for subsequent early warning.

[0034] The injection unit specifically includes: Atomizing nozzles are installed on the outer casing of the battery module and in the internal cooling air ducts. These nozzles are connected to the extinguishing agent storage tank via pipelines. Solenoid valves and pressure regulators are installed on the pipelines to control the release volume and spray pressure of the extinguishing agent. Upon receiving a targeted disposal command, the controller drives the corresponding solenoid valve to open, allowing the extinguishing agent to be atomized into fine mist particles through the nozzles and enter the gaps between the modules. The spray unit can adjust the coverage area by adjusting the nozzle angle and opening duration, enabling the medium to target the cell gaps in abnormal areas, achieving rapid cooling and suffocation extinguishing effects.

[0035] Specifically, according to the targeted disposal command, the spray unit is controlled to release atomized fire extinguishing medium, causing the medium to form a permeation flow field within the gaps between the battery cells, and disposal feedback data is collected, including: Based on the anomaly location results, the controller selects the nozzle corresponding to the target area and drives the nozzle to release the extinguishing medium according to the spray dosage, spray angle, spray pressure, and spray duration set in the disposal command. After the extinguishing medium enters the cell gap, it forms an enveloping and penetrating flow field structure, allowing heat to be quickly conducted and carried away. During the spraying process, temperature sensors, gas detectors, and smoke detectors deployed in the cell gap and outside the module continue to collect data in real time. After synchronization and filtering, this data generates disposal feedback data to reflect the cooling effect and the distribution of the extinguishing medium, providing a basis for subsequent trend analysis and secondary disposal.

[0036] This includes generating event log data based on the handling feedback data and secondary handling instructions, converting the event log data into optimization parameters, and feeding them back to the monitoring platform. Specifically, this includes: During the primary and secondary spraying processes, the system records key parameters in real time, such as the rate of temperature drop, the magnitude of gas concentration change, the spray dosage, the spray pressure, and the spray duration. After the response is completed, these parameters are compared with the monitoring data before the incident to generate complete event log data, which describes the entire process of anomaly occurrence, response execution, and effectiveness evaluation. Subsequently, the event log data is structured, for example, by extracting the minimum spray dosage required for effective fire extinguishing, the optimal spray duration range, and the medium penetration characteristics under different cell gaps, and converting them into optimization parameters. These optimization parameters are transmitted to the monitoring platform through the communication interface and used to update subsequent response strategies, enabling the system to automatically adjust warning thresholds and spray control parameters based on historical experience in subsequent events, thereby improving the overall prevention and control effectiveness and reliability.

[0037] In a preferred embodiment of the present invention, time series rate analysis is performed on multi-source monitoring data. If the temperature rise rate of the internal monitoring signal exceeds a preset rate threshold and there is a trigger time difference with the external monitoring signal, early warning data is generated, and a priority label is assigned to the corresponding cell area, including: Based on multi-source monitoring data, time synchronization and noise reduction processing are performed to generate aligned monitoring data; Based on the aligned monitoring data, time-by-time differential analysis is performed to generate internal temperature rise rate sequences and external temperature rise rate sequences. Based on the internal temperature rise rate sequence, threshold comparison is performed and time difference evaluation is conducted with the trigger time of the external temperature rise rate sequence to generate candidate warning segments; Based on the candidate early warning segments, the continuity and stability are checked, transient fluctuations are screened out, and early warning data is generated; Based on early warning data, and combined with the magnitude, duration and spatial location of temperature rise, priority labels are generated and linked to the early warning data.

[0038] In this embodiment of the invention, by performing time synchronization and noise reduction processing on multi-source monitoring data, the sampling delay between sensors and the impact of environmental noise can be reduced, making the subsequent data analysis results more reliable. After the monitoring data is aligned, time-by-time difference can be performed to obtain the internal and external temperature rise rate sequences, thereby more intuitively reflecting the dynamic trend of temperature changes inside the battery cell.

[0039] Based on this, by comparing the internal temperature rise rate with a threshold and combining it with the trigger time of the external temperature rise rate for time difference assessment, situations where the internal temperature rise precedes the external response can be effectively identified, generating candidate warning segments. This method enables early detection of early signs of thermal runaway. Continuity and stability checks on candidate warning segments further filter out false alarms caused by transient fluctuations, ensuring the accuracy of the warning signals. Finally, priority labels are generated by combining temperature rise amplitude, duration, and spatial location. This not only indicates the degree of risk but also provides a reference for the sequence of subsequent location and spraying operations, making the fire response process more efficient and orderly.

[0040] Specifically, based on the internal temperature rise rate sequence, a threshold comparison is performed, and the time difference between the threshold comparison and the trigger time of the external temperature rise rate sequence is evaluated to generate candidate warning segments, including: The system continuously calculates the rate of temperature change inside the battery cell according to the sampling period, comparing this rate sequence with a preset rate threshold one by one. Once several consecutive sampling points exceed the rate threshold, an internal anomaly signal is generated. Simultaneously, the system compares the temperature rise trigger time of external monitoring points. If the internal signal triggers before the external signal, it indicates that the anomaly may originate from thermal runaway within the battery cell rather than environmental disturbances. By combining the duration of the internal signal exceeding the threshold with the relative delay of the external signal trigger, candidate warning segments are formed as potential early warning criteria.

[0041] This process involves verifying the continuity and stability of candidate warning segments, filtering out transient fluctuations, and generating early warning data. Specifically, this includes: The system checks the duration of candidate warning segments. If the duration of the anomaly is less than the set lower limit, it is judged as transient interference and discarded. Furthermore, the system evaluates the volatility of the temperature rise rate at each sampling point in the candidate segment. If the rate change is extremely unstable and does not form a sustained upward trend, it is also rejected. Through the above continuity and stability checks, it can be ensured that only when the abnormal signal conforms to the characteristics of thermal runaway in both time and trend is it confirmed as valid and ultimately output as early warning data.

[0042] Among these, priority labels are generated based on early warning data, combined with the magnitude, duration, and spatial location of temperature rise, and are then linked to the early warning data. Specifically, these include: After generating early warning data, the system first calculates the highest temperature rise of the anomaly within the warning segment. If the rise is significantly higher than the normal operating range, a higher risk score is assigned. Simultaneously, the duration of the anomaly is assessed; the longer the duration, the higher the risk level. Then, combined with the spatial location of the sensors, the system correlates the anomaly's location with the arrangement of the battery cell modules. For example, anomalies located in the center of the battery cluster have a higher priority than those in the edge areas. The system comprehensively scores these factors to form a priority label, which is then linked to the early warning data. This allows for priority response to high-risk areas during subsequent anomaly localization and spraying procedures, thereby improving the overall targeting and timeliness of prevention and control.

[0043] In a preferred embodiment of the present invention, based on early warning data and combined with the sensor matrix distribution within the battery module, spatial cross-comparison is performed to obtain anomaly location data, and false anomalies are eliminated through multi-point consistency verification, including: Based on the temperature sensors, gas detectors, and smoke detectors deployed within the battery module, a sensor matrix is ​​established according to their spatial coordinate relationships, generating matrix distribution data. Based on matrix distribution data and early warning data, temperature and gas signals of the target core point and its multiple neighboring points in the sensor matrix are extracted to generate neighborhood monitoring data. Based on neighborhood monitoring data, coordinate mapping and amplitude normalization are performed to generate a local spatial feature map; Based on the local spatial feature map, perform spatial cross-comparison to identify sets of adjacent points with consistent trends and generate abnormal connected regions. Based on the abnormal connectivity areas, and combined with the battery module structure layout, preliminary abnormal location data is generated. Based on the preliminary anomaly location data, multi-point consistency verification and historical comparison verification are performed to remove single-point deviations and abnormal unstable segments, thereby generating anomaly location data.

[0044] In this embodiment of the invention, a sensor matrix is ​​established by deploying temperature, gas, and smoke sensors within the battery module, and matrix distribution data is generated according to spatial coordinate relationships. This provides a clear spatial reference basis for anomaly localization. Within this matrix framework, the core target point and its neighborhood monitoring data are extracted, and coordinate mapping and amplitude normalization are performed. This eliminates the measurement scale differences between different sensors, making the signals comparable.

[0045] By cross-referencing local spatial feature maps, sets of adjacent points with consistent trends can be identified, and abnormal connected regions can be generated. This enables spatial aggregation of abnormal signals, avoiding erroneous localization caused by single-point fluctuations. Matching the connected regions with the structural layout of the battery modules can further narrow down the anomaly range, yielding preliminary localization results. Finally, multi-point consistency verification and comparison with historical data can effectively eliminate single-point deviations and unstable segments, ensuring that the final anomaly localization results are both accurate and reliable, providing a precise basis for directional spraying.

[0046] Specifically, based on temperature sensors, gas detectors, and smoke detectors deployed within the battery module, a sensor matrix is ​​established according to their spatial coordinate relationships, generating matrix distribution data, which includes: Sensors are pre-deployed at different levels and locations within the battery module, and the installation coordinates of each sensor are recorded and mapped onto the 3D spatial model of the battery module. Subsequently, the coordinates of various sensors are arranged into a matrix according to the module's layout, with each row or column corresponding to a spatial direction of the module, thus forming a full-space sensor network covering cell gaps, module edges, and cooling channels. In this way, the system generates matrix distribution data containing coordinate indices and sensor types for subsequent spatial comparison of abnormal signals.

[0047] Specifically, based on matrix distribution data and early warning data, temperature and gas signals of the target core point and its multiple neighboring points in the sensor matrix are extracted to generate neighborhood monitoring data, including: When early warning data indicates a potential anomaly in a certain area, the system determines the location of the core sensor that triggered the warning based on matrix distribution data. Building upon this, it further extracts signals from multiple spatially adjacent sensors around the core point, including temperature and gas concentration change curves. These signals are then aligned chronologically with their spatial locations to form a neighborhood monitoring dataset. This neighborhood monitoring data reflects both the anomaly at the core point and includes auxiliary information from neighboring points, helping to eliminate single-point errors and noise interference.

[0048] Specifically, based on neighborhood monitoring data, coordinate mapping and amplitude normalization are performed to generate a local spatial feature map, including: The system maps each sensor signal from the neighborhood monitoring data to its spatial coordinates in the matrix distribution data, thus forming a point distribution map in three-dimensional space. Then, it normalizes the amplitude of the temperature and gas concentration values ​​from different sensors, converting them into dimensionless values ​​of the same magnitude for lateral comparison. These values ​​are then mapped back to spatial coordinates, forming a local spatial feature map with color or intensity distribution, allowing abnormal signals to be visually presented spatially. This feature map provides a unified input data foundation for subsequent spatial cross-comparison.

[0049] Specifically, preliminary anomaly location data is generated by matching abnormal connectivity areas with the structural layout of battery modules, including: After obtaining the local spatial feature map and identifying abnormal connected regions, the system matches these regions with the physical structure model of the battery module to determine the location of the corresponding cell group or cooling channel. Furthermore, it compares the geometric boundaries of these regions with the boundaries of the module cells to determine whether the anomaly is concentrated in a single cell, spans multiple cells, or is located at a specific level within the module. Through this combination of spatial and structural comparison, the system generates preliminary anomaly location data containing the specific cell number or module number, which is used in subsequent processing steps.

[0050] In a preferred embodiment of the present invention, a directional handling instruction for the corresponding injection unit is generated based on the abnormal positioning data, and the injection data is set in the handling instruction, including: Based on the abnormal location data, the target injection unit is generated by registering it with the spatial arrangement mapping relationship of the injection unit. Based on the anomaly location data, combined with the temperature rise rate, anomaly duration, number of involved battery cells, and distance between battery cell gaps and nozzles, a penetration coverage requirement dataset is generated. Based on the penetration coverage requirement dataset, the initial injection dose, injection angle, injection pressure, and injection duration are calculated to generate a draft of targeted treatment instructions; According to the draft of the targeted disposal instruction, an interlock check and spray path accessibility check are performed before execution to generate an executable targeted disposal instruction; Based on the executable directional treatment command, control is issued to the target injection unit to initiate directional injection.

[0051] In this embodiment of the invention, registration is performed based on the anomaly location results and the spatial arrangement of the spraying units. This ensures that the anomaly area corresponds one-to-one with the actual spraying hardware, thereby avoiding situations where the location is accurate but the spray cannot reach the intended area. By combining the temperature rise rate, anomaly duration, number of involved battery cells, and distance between the battery cell gap and the nozzle to generate a penetration coverage requirement dataset, the impact of different factors on fire extinguishing requirements can be quantified. For example, when the high temperature duration is long or the number of battery cells is large, the requirements for spray coverage and cooling intensity also increase.

[0052] Based on this dataset, parameters such as spray dosage, spray angle, spray pressure, and spray duration are calculated. This enables fire extinguishing commands to not only execute actions but also include adaptively adjustable multi-dimensional parameters, significantly improving the targeting and effectiveness of the spraying. Through pre-execution interlock checks and spray path reachability checks, it can be confirmed that the spraying action will not be restricted by obstacles or equipment status before the command is issued, ensuring the command's executability. The final generated directional disposal command not only specifies the spraying unit but also enhances the accuracy and safety of the fire extinguishing process through multi-parameter control.

[0053] Specifically, based on the abnormal location data, registration is performed with the spatial arrangement mapping relationship of the injection units to generate the target injection unit, which includes: During the design phase, the system has established a spatial mapping database between battery modules and injection units, with each injection unit corresponding to a specific cell area or module channel. When abnormal location data indicates an anomaly in a certain area, the system first searches the database to find the injection unit closest to the spatial coordinates of that area, and compares its coverage area with the anomaly area to ensure that at least one injection unit can completely cover the anomaly location. If multiple candidate injection units exist, they are prioritized based on distance, coverage angle, and unobstructedness of the injection path to ultimately determine the optimal target injection unit.

[0054] According to the draft targeted disposal instructions, pre-execution interlock verification and spray path accessibility checks are performed to generate executable targeted disposal instructions, specifically including: Before issuing the command, the system performs an interlock check on the draft parameters, checking whether the current spray unit is available, whether the pressure of the extinguishing medium is within a safe range, and whether there are any conflicts with other parallel tasks. Then, it performs a spray path reachability check, simulating the spray angle and the medium diffusion path to confirm that it will not be blocked by structural components, cable bundles, or battery casings, and can effectively cover abnormal areas. If the check and path check pass, the draft is converted into a formal executable command. If there are any conditions that are not met, parameter correction or switching to a backup spray unit is triggered.

[0055] In a preferred embodiment of the present invention, trend analysis is performed on the treatment feedback data. If the temperature or gas value does not decrease sufficiently to recover to a preset safe range within a preset time limit, a secondary treatment command is generated, and the injection data is adjusted in the secondary treatment command, including: Based on the processing feedback data, time synchronization and noise reduction are performed to generate aligned feedback data; Based on the alignment feedback data, the rate of temperature decrease and the rate of gas concentration decrease are extracted, and the magnitude of change within a preset time limit is calculated to generate residual anomaly assessment results. Based on the residual anomaly assessment results, if the decrease is insufficient to restore the situation to a safe level, a draft secondary response instruction will be generated. According to the draft secondary treatment instruction, the spraying data is adjusted, and continuous spraying, intermittent spraying, or pulse spraying is selected based on the distribution of residual anomalies to generate a secondary treatment instruction; According to the secondary treatment command, the linkage control spray unit implements secondary spraying and stores the secondary treatment command and treatment feedback data together.

[0056] In this embodiment of the invention, after the injection is executed, the feedback data is time-synchronized and denoised to ensure that the analysis results are not affected by signal delay or environmental interference, thus more realistically reflecting the cooling effect. Extracting the rate of temperature decrease and the rate of gas concentration decrease and calculating the magnitude of change within a time limit allows for objective quantification of residual anomalies, avoiding erroneous judgments based solely on instantaneous values. The resulting residual anomaly assessment results can accurately determine whether the fire extinguishing effect meets the standards.

[0057] When the extinguishing effect is insufficient to restore the fire to a safe level, a draft secondary response instruction is automatically generated, in which the spray data is adjusted to make the response process adaptive. By introducing the option of continuous spray, intermittent spray, or pulse spray modes into the instruction, different modes bring effects such as enhanced cooling persistence or improved penetration capacity, thereby taking the best measures for different residual risks. This process not only improves the fire extinguishing effect but also effectively reduces the risk of reignition, achieving closed-loop control.

[0058] Specifically, based on the alignment feedback data, the rate of temperature decrease and the rate of gas concentration decrease are extracted, and the magnitude of change within a preset time limit is calculated to generate residual anomaly assessment results, including: The system first samples the temperature and gas concentration signals on the time axis, calculates the difference between adjacent moments and records the trend of change, thereby obtaining the rate of temperature decrease and the rate of gas concentration decrease. Then, within a preset time limit, these rates are accumulated and statistically analyzed to obtain the overall decrease amplitude. This amplitude is then compared with the set safe decrease threshold. If the decrease is insufficient, a residual anomaly assessment result is generated to reflect the presence of abnormal hot spots or incompletely dissipated combustible gases.

[0059] Based on the residual anomaly assessment results, if the decrease is insufficient to restore the situation to a safe level, a draft secondary response instruction will be generated, specifically including: Based on the residual anomaly assessment results, the system determines whether the abnormal parameters have returned to the safe range. If the temperature is still higher than the safety limit or the gas concentration is still higher than the flammability threshold, the risk of reignition is confirmed. In this case, the system combines the location, duration and extent of failure of the abnormal area to deduce the required additional spray volume and duration, and summarizes these parameters into a draft secondary response instruction as the basis for further fire extinguishing operations.

[0060] Specifically, according to the draft secondary treatment instruction, the injection data is adjusted, and based on the distribution of residual anomalies, continuous injection, intermittent injection, or pulse injection is selected to generate the secondary treatment instruction, which includes: The system first modifies the injection dosage, injection pressure, injection angle, and injection duration based on the additional injection requirements proposed in the draft. Then, it selects the injection mode according to the distribution characteristics of the residual anomaly. When the residual anomaly exhibits a sustained high temperature, continuous injection is selected to ensure stable cooling; when the anomaly shows periodic fluctuations, intermittent injection is selected to combine energy saving with effective suppression; if the anomaly is concentrated in a local hotspot area, pulse injection is selected to achieve high-intensity, short-term penetration. After these modifications, the system finally generates a secondary response command and issues it to the corresponding injection unit to execute the fire extinguishing.

[0061] In a preferred embodiment of the present invention, spatial cross-comparison is performed based on the local spatial feature map to identify a set of adjacent points with consistent trends and generate anomaly connected regions, including: Based on the local spatial feature map, the target core point signal and the adjacent point signal are synchronously compared within a time window to generate multi-point trend comparison data. Based on multi-point trend comparison data, the difference magnitude and trend similarity between adjacent points and core points are calculated to form a similarity assessment result; Based on the similarity assessment results, a set of adjacent points with consistent trends and differences in amplitude changes within a set range is selected to form an outlier set. Based on the set of anomalies, generate an anomaly connected region containing the target core point and its adjacent points.

[0062] In this embodiment of the invention, by performing spatial cross-comparison on the local spatial feature map, trend consistency analysis can be performed on the signals of the target core point and neighboring points within a time window, thereby avoiding erroneous judgments caused by single-point anomalies. The similarity evaluation process can distinguish the set of adjacent points that are highly consistent with the changing trend of the core point, and then filter to form an anomaly point set.

[0063] The abnormal connected regions generated based on anomaly point sets can not only indicate the specific location of the anomaly but also reveal its extent of expansion within the local space. This method achieves a transition from single-point signals to regional anomalies, making the location results more stable and reliable. For subsequent directional spraying, this regionalized information allows the extinguishing medium to more effectively cover the target area, thereby improving overall fire extinguishing efficiency.

[0064] Specifically, based on multi-point trend comparison data, the difference magnitude and trend similarity between adjacent points and the core point are calculated to form a similarity assessment result, which includes: The system first compares the temperature and gas signals of the target core point with those of multiple neighboring points within a time window, extracting the numerical change trajectory of each point within that window. Then, it calculates the numerical differences between the core point and neighboring points within the same time period to measure the magnitude of the difference, and analyzes whether their directions of change are consistent to measure trend similarity. Finally, it combines the magnitude of the difference and trend similarity to form a similarity assessment result used to determine signal consistency. Thus, when neighboring points and the core point show a high degree of similarity in both the rate of change and the direction of trend, the similarity assessment result is marked as highly consistent; otherwise, it is marked as low consistent.

[0065] Based on the similarity assessment results, a set of adjacent points with consistent trends and amplitude variations within a defined range is selected to form an outlier set, specifically including: Based on the similarity assessment results, the system prioritizes selecting adjacent points with the same trend direction as the core point to eliminate noise and reverse interference signals. Then, it further compares the amplitude differences of the selected adjacent points, retaining points with amplitude changes within the set range and removing points that exceed the range. After completing the dual screening of trend and amplitude, an abnormal point set containing the core point and several highly consistent adjacent points is finally formed. This abnormal point set serves as the basis for the subsequent generation of abnormal connected regions, effectively avoiding positioning deviations caused by false alarms of single points or individual fluctuations.

[0066] In a preferred embodiment of the present invention, based on the preliminary anomaly location data, multi-point consistency verification and historical comparison verification are performed to remove single-point deviations and abnormally unstable segments, generating anomaly location data, including: Based on the preliminary anomaly location data, multiple monitoring signals within the area are extracted to form a location neighborhood signal set; Based on the location neighborhood signal set, consistency verification is performed. If the neighboring points and the core point maintain continuous similarity in trend and amplitude, they are judged as valid anomalies. Based on the location neighborhood signal set, and in conjunction with the stored historical operation data, a comparison and verification are performed. If the abnormal signal deviates significantly from the historical stable pattern, it is determined to be a valid anomaly; otherwise, it is determined to be a false anomaly. Based on the combined results of consistency verification and historical comparison verification, single-point deviations and abnormal unstable segments are removed to generate corrected anomaly location data.

[0067] In this embodiment of the invention, by verifying the consistency of the multi-point monitoring signals obtained from the initial positioning, false alarms caused by occasional fluctuations or malfunctions of a single sensor can be effectively avoided. The continuous consistency in trend and amplitude among multiple points within the neighborhood proves that the anomaly is real and has spatial scalability, thereby enhancing the credibility of the judgment. Comparing the positioning results with historical operational data can identify abnormal changes that are significantly different from the normal operating mode, further improving the accuracy of anomaly identification. Finally, through the dual screening of consistency verification and historical comparison, single-point deviations and unstable segments are eliminated, making the obtained anomaly positioning results more robust and providing accurate spatial basis for subsequent injection commands.

[0068] Among these steps, consistency verification is performed based on the location neighborhood signal set. If neighboring points and the core point maintain continuous similarity in both trend and amplitude, they are determined to be valid anomalies. Specifically, this includes: The system first compares the core point signal with the signals of multiple neighboring points within the same time window, analyzes whether the direction of change of the values ​​of each point is consistent, and calculates the difference in the change amplitude between the neighboring points and the core point. If these differences remain within the allowable range and the trend direction is consistent within multiple consecutive sampling periods, the abnormal signal is considered to have stability and continuity and can be judged as a valid anomaly. Conversely, if the change direction of some points is opposite or the value difference is too large, it is judged as an inconsistent signal and marked as low confidence.

[0069] Specifically, based on the location neighborhood signal set and combined with stored historical operational data for verification, if the abnormal signal deviates significantly from the historical stable pattern, it is determined to be a valid anomaly; otherwise, it is determined to be a false anomaly. This includes: The system retrieves historical operating data of the battery module under normal operating conditions and establishes a baseline curve for temperature and gas signals. It then compares the current location neighborhood signal set with the baseline curve hour by hour. If there are significant differences between the current signal and the baseline curve in terms of fluctuation amplitude, trend pattern, or duration, the abnormal signal is considered to be different from normal fluctuations and can be identified as a valid anomaly. If the comparison results show that the current signal is highly consistent with the baseline curve and has only a slight deviation, the signal is identified as a false anomaly, thereby avoiding false triggering.

[0070] Based on the combined results of consistency verification and historical comparison verification, single-point deviations and abnormally unstable segments are removed to generate corrected anomaly localization data, specifically including: The system comprehensively evaluates the results of consistency verification and historical comparison verification. Only signals that satisfy both verifications are retained. If some points perform well in consistency verification but are highly consistent with the baseline curve in historical comparison, they are judged as false alarms and removed. If some signal segments fluctuate drastically in a short period of time but fail to form a sustained trend, they are considered unstable segments and filtered out. Finally, the system retains only the set of points with high confidence after dual verification, outputting corrected anomaly location data for subsequent targeted processing.

[0071] In a preferred embodiment of the present invention, based on the anomaly location data, combined with the temperature rise rate, anomaly duration, number of involved battery cells, and distance between the battery cell gap and the nozzle, a penetration coverage requirement dataset is generated, including: Based on the anomaly location data, the temperature rise rate and duration within the anomaly area are extracted to form thermal intensity parameters; Based on the abnormal location data, the number of battery cells involved and the density of battery cell arrangement are counted to form regional impact parameters; The injection geometry parameters are formed based on the cell gap width and the distance from the nozzle to the target area; The thermal intensity parameters, regional influence parameters, and injection geometry parameters are combined to generate a permeability coverage requirement dataset.

[0072] In this embodiment of the invention, by extracting the temperature rise rate and duration from the anomaly location results, parameters reflecting thermal intensity can be generated to assess the energy level of the anomaly region. Statistical analysis of the number and density of involved battery cells generates regional impact parameters, which can intuitively reflect the scale of the anomaly region and the likelihood of risk diffusion. Combining the battery cell gap width with the distance from the nozzle to the target area to generate injection geometry parameters provides spatial constraint information for the injection path and penetration effect.

[0073] By merging thermal intensity parameters, regional influence parameters, and injection geometry parameters to form a penetration coverage requirement dataset, multi-dimensional factors such as temperature, quantity, and geometry can be uniformly expressed. This synthetic data not only facilitates the subsequent calculation of injection parameters but also ensures that the injection design closely integrates with actual risk characteristics, avoiding insufficient dosage or uneven coverage, thereby significantly improving the targeting and reliability of the treatment.

[0074] Specifically, based on the anomaly location data, the temperature rise rate and duration within the anomaly area are extracted to form thermal intensity parameters, including: The system first calls up the anomaly location data to pinpoint the specific area where the abnormal battery cell or module is located; then it analyzes the continuous monitoring signals from the sensors in that area to calculate the rate of temperature increase per unit time, thus obtaining the temperature rise rate; next, it calculates the duration of the anomaly based on the time period in which it persists; finally, it combines the temperature rise rate with the duration to form a thermal intensity parameter that reflects the level of local heat accumulation, which is used to assess subsequent injection requirements.

[0075] Based on the abnormal location data, the number of battery cells involved and their density are statistically analyzed to form regional impact parameters, specifically including: The system determines the number of battery cells covered by the spatial range of the anomaly location data, and calculates the compactness between the cells based on their arrangement within the module. If the cells are densely packed within the anomaly area, it means that heat is more easily conducted between adjacent cells, resulting in a larger regional impact parameter value; conversely, if the cells are sparsely packed, the regional impact parameter value is smaller. In this way, the regional impact parameter can accurately reflect the potential risk range of anomaly spread.

[0076] The injection geometry parameters are formed based on the cell gap width and the distance from the nozzle to the target area, specifically including: The system extracts the width information of the cell gap based on the structural data of the module, and establishes a geometric model of the medium flow path by combining the spatial distance between the nozzle and the abnormal area. In this model, the width of the cell gap determines the resistance of the medium flow, and the distance from the nozzle to the target area determines the range requirement of the medium. The system combines the two to generate injection geometry parameters so as to evaluate the injection pressure and injection angle in subsequent calculations.

[0077] In a preferred embodiment of the present invention, based on the penetration coverage requirement dataset, the initial injection dose, injection angle, injection pressure, and injection duration are calculated to generate a draft targeted treatment instruction, including: Based on the thermal intensity parameter in the penetration coverage requirement parameters, combined with the temperature rise rate and the duration of the anomaly, the cooling requirement is determined, and the initial injection dose is obtained. Based on the geometric parameters in the penetration coverage requirement parameters, and combined with the spatial angle between the nozzle position and the cell gap, the spray angle is determined to achieve spray coverage of the center of the abnormal area. Based on the geometric parameters in the penetration coverage requirement parameters, combined with the nozzle distance and cell gap width, the driving force requirement is determined, and the injection pressure is obtained; Based on the regional impact parameters in the penetration coverage requirement parameters, combined with the parameters of the number of battery cells and thermal intensity involved, the required continuous cooling time is calculated to obtain the spraying duration; Based on the initial injection dose, injection angle, injection pressure, and injection duration, a draft directional treatment instruction containing multiple parameters is generated.

[0078] In this embodiment of the invention, based on the penetration coverage requirement dataset, the injection dosage, injection angle, injection pressure, and injection duration are calculated item by item, transforming abstract requirements into executable control commands. During this process, the thermal intensity parameter determines the required cooling dosage, the geometric parameters determine the injection angle and pressure, and the area influence parameter determines the continuous injection time. In this way, each injection data point directly corresponds to a specific risk characteristic, ensuring precise and controllable commands.

[0079] The final draft of the targeted response instructions includes multi-parameter linkage control information, which can guide the spraying units to carry out more precise and efficient fire extinguishing actions. In this way, spraying is no longer a single fixed pattern, but is tailored to the specific risk characteristics, thereby achieving optimal fire extinguishing coverage and cooling effects during fire fighting and reducing the possibility of reignition.

[0080] Specifically, based on the thermal intensity parameter in the penetration coverage requirement parameters, combined with the temperature rise rate and the duration of the anomaly, the cooling requirement is determined to obtain the initial injection dose, which includes: The system first analyzes the temperature rise in the abnormal area, calculating the accumulated heat based on the temperature rise rate and duration per unit time. Then, it compares this accumulated heat with the cooling capacity of the extinguishing medium to derive the required cooling demand, converting it into a corresponding extinguishing medium dosage as a reference value for the initial spray dosage. This step ensures that the spray dosage matches the actual heat load, avoiding incomplete extinguishing due to insufficient dosage or resource waste due to excessive dosage. The cooling capacity of the extinguishing medium can be experimentally determined, its value being jointly determined by the medium's specific heat capacity, latent heat of vaporization, and atomization particle size characteristics. The heat that can be removed per unit mass or volume of spray volume serves as a reference for calculating the spray dosage.

[0081] Specifically, based on the geometric parameters in the penetration coverage requirement parameters, and combined with the spatial angle between the nozzle position and the cell gap, the injection angle is determined to achieve injection coverage of the center of the abnormal area. This includes: The system establishes the angular relationship between the nozzle and the target area based on the spatial coordinates of the cell gap and the nozzle. By comparing the simulation results of the medium flow field at different angles, the system selects the angle that allows the spray path to reach the center of the abnormal area and avoids uneven coverage caused by angle deviation. Finally, this angle is set as part of the spray parameters to ensure that the spray direction coincides with the abnormal area to the greatest extent, thereby improving the fire extinguishing effect.

[0082] Specifically, based on the geometric parameters in the penetration coverage requirement parameters, combined with the nozzle distance and cell gap width, the driving force requirement is determined, and the injection pressure is obtained, including: The system first obtains the straight-line distance from the nozzle to the abnormal area, and estimates the resistance that the extinguishing medium needs to overcome during the flow process by combining the width of the cell gap; then, based on the physical properties of the extinguishing medium (such as density and particle size) and the path resistance, it derives the minimum driving force required to ensure that the medium can penetrate into the target area; finally, it converts the driving force into the pressure set value of the spray unit to ensure that the sprayed medium can reach and cover the abnormal point.

[0083] Specifically, based on the regional impact parameters in the penetration coverage requirement parameters, combined with the parameters of the number of battery cells and thermal intensity, the required continuous cooling time is calculated to obtain the spraying duration, which includes: The system first counts the number of affected battery cells in the abnormal area and estimates the area that needs to be cooled simultaneously based on the battery cell density. Then, it combines this area with the heat intensity parameter to calculate the cooling effect that needs to be maintained per unit time. Finally, based on the total cooling demand and the cooling efficiency of the extinguishing medium, it determines the duration required for spraying and sets it as the spray duration parameter to ensure that the abnormal area can be sufficiently cooled and to prevent residual heat from causing reignition.

[0084] Embodiments of the present invention also provide a smart fire prevention and fire management system for lithium battery energy storage power stations, the system comprising: The multi-source monitoring module is used to collect data from multi-level temperature sensors, gas detectors, and smoke detectors inside and outside the battery module, forming multi-source monitoring data. The early warning generation module is used to perform time series rate analysis on multi-source monitoring data. If the temperature rise rate of the internal monitoring signal exceeds the preset rate threshold and there is a trigger time difference with the external monitoring signal, early warning data is generated and a priority label is assigned to the corresponding cell area. The anomaly localization module is used to perform spatial cross-comparison based on early warning data and the sensor matrix distribution within the battery module to obtain anomaly localization data, and to eliminate false anomalies through multi-point consistency verification. The treatment instruction module is used to generate directional treatment instructions for the corresponding injection unit based on the abnormal location data, and to set the injection data in the treatment instructions, including the injection dose, injection angle, injection pressure and injection duration. The directional spray module is used to control the spray unit to release atomized fire extinguishing medium according to the directional disposal command, so that the medium forms a permeation flow field in the gap between the battery cells, and to collect and form disposal feedback data; The trend analysis module is used to perform trend analysis on the treatment feedback data. If the temperature or gas value does not drop enough to recover to the preset safety range within the preset time limit, a secondary treatment instruction is generated, and the injection data is adjusted in the secondary treatment instruction. The event logging module is used to generate event logging data based on the handling feedback data and secondary handling instructions, and to convert the event logging data into optimization parameters and feed them back to the monitoring platform.

[0085] It should be noted that this system is a system corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0086] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0087] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0088] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A smart fire prevention and fire management method for lithium battery energy storage power stations, characterized in that, The method includes: Multi-level temperature sensors, gas detectors, and smoke detectors are deployed inside and outside the battery module to collect and form multi-source monitoring data. Time series rate analysis is performed on multi-source monitoring data. If the temperature rise rate of the internal monitoring signal exceeds the preset rate threshold and there is a trigger time difference with the external monitoring signal, early warning data is generated and a priority label is assigned to the corresponding cell area. Based on early warning data and combined with the sensor matrix distribution within the battery module, spatial cross-comparison is performed to obtain anomaly location data, and false anomalies are eliminated through multi-point consistency verification. Based on the abnormal location data, a directional treatment command for the corresponding injection unit is generated, and the injection data is set in the treatment command, including the injection dose, injection angle, injection pressure and injection duration; According to the directional disposal command, the spray unit is controlled to release atomized fire extinguishing medium, so that the medium forms a permeation flow field in the gap between the battery cells, and disposal feedback data is collected; The feedback data is analyzed for trends. If the temperature or gas value does not drop sufficiently to return to the preset safe range within the preset time limit, a secondary treatment instruction is generated, and the injection data is adjusted in the secondary treatment instruction. Based on the feedback data and secondary handling instructions, event log data is generated, and the event log data is transformed into optimization parameters and fed back to the monitoring platform.

2. The intelligent fire prevention and fire management method for lithium battery energy storage power stations according to claim 1, characterized in that, Time series rate analysis is performed on multi-source monitoring data. If the temperature rise rate of the internal monitoring signal exceeds a preset rate threshold and there is a trigger time difference with the external monitoring signal, early warning data is generated, and priority tags are assigned to the corresponding cell areas, including: Based on multi-source monitoring data, time synchronization and noise reduction processing are performed to generate aligned monitoring data; Based on the aligned monitoring data, time-by-time differential analysis is performed to generate internal temperature rise rate sequences and external temperature rise rate sequences. Based on the internal temperature rise rate sequence, threshold comparison is performed and time difference evaluation is conducted with the trigger time of the external temperature rise rate sequence to generate candidate warning segments; Based on the candidate early warning segments, the continuity and stability are checked, transient fluctuations are screened out, and early warning data is generated; Based on early warning data, and combined with the magnitude, duration and spatial location of temperature rise, priority labels are generated and linked to the early warning data.

3. The intelligent fire prevention and fire management method for lithium battery energy storage power stations according to claim 1, characterized in that, Based on early warning data and the sensor matrix distribution within the battery module, spatial cross-comparison is performed to obtain anomaly location data. False anomalies are then eliminated through multi-point consistency verification, including: Based on the temperature sensors, gas detectors, and smoke detectors deployed within the battery module, a sensor matrix is ​​established according to their spatial coordinate relationships, generating matrix distribution data. Based on matrix distribution data and early warning data, temperature and gas signals of the target core point and its multiple neighboring points in the sensor matrix are extracted to generate neighborhood monitoring data. Based on neighborhood monitoring data, coordinate mapping and amplitude normalization are performed to generate a local spatial feature map; Based on the local spatial feature map, perform spatial cross-comparison to identify sets of adjacent points with consistent trends and generate abnormal connected regions. Based on the abnormal connectivity areas, and combined with the battery module structure layout, preliminary abnormal location data is generated. Based on the preliminary anomaly location data, multi-point consistency verification and historical comparison verification are performed to remove single-point deviations and abnormal unstable segments, thereby generating anomaly location data.

4. The intelligent fire prevention and fire management method for lithium battery energy storage power stations according to claim 1, characterized in that, Based on the abnormal location data, a directional handling command for the corresponding injection unit is generated, and the injection data is set in the handling command, including: Based on the abnormal location data, the target injection unit is generated by registering it with the spatial arrangement mapping relationship of the injection unit. Based on the anomaly location data, combined with the temperature rise rate, anomaly duration, number of involved battery cells, and distance between battery cell gaps and nozzles, a penetration coverage requirement dataset is generated. Based on the penetration coverage requirement dataset, the initial injection dose, injection angle, injection pressure, and injection duration are calculated to generate a draft of targeted treatment instructions; According to the draft of the targeted disposal instruction, an interlock check and spray path accessibility check are performed before execution to generate an executable targeted disposal instruction; Based on the executable directional treatment command, control is issued to the target injection unit to initiate directional injection.

5. The intelligent fire prevention and fire management method for lithium battery energy storage power stations according to claim 1, characterized in that, Trend analysis is performed on the response feedback data. If the temperature or gas value does not decrease sufficiently to return to the preset safe range within a preset time limit, a secondary response command is generated, and the injection data is adjusted in the secondary response command, including: Based on the processing feedback data, time synchronization and noise reduction are performed to generate aligned feedback data; Based on the alignment feedback data, the rate of temperature decrease and the rate of gas concentration decrease are extracted, and the magnitude of change within a preset time limit is calculated to generate residual anomaly assessment results. Based on the residual anomaly assessment results, if the decrease is insufficient to restore the situation to a safe level, a draft secondary response instruction will be generated. According to the draft secondary treatment instruction, the spraying data is adjusted, and continuous spraying, intermittent spraying, or pulse spraying is selected based on the distribution of residual anomalies to generate a secondary treatment instruction; According to the secondary treatment command, the linkage control spray unit implements secondary spraying and stores the secondary treatment command and treatment feedback data together.

6. The intelligent fire prevention and fire management method for lithium battery energy storage power stations according to claim 3, characterized in that, Based on the local spatial feature map, spatial cross-comparison is performed to identify sets of adjacent points with consistent trends, generating abnormal connected regions, including: Based on the local spatial feature map, the target core point signal and the adjacent point signal are synchronously compared within a time window to generate multi-point trend comparison data. Based on multi-point trend comparison data, the difference magnitude and trend similarity between adjacent points and core points are calculated to form a similarity assessment result; Based on the similarity assessment results, a set of adjacent points with consistent trends and differences in amplitude changes within a set range is selected to form an outlier set. Based on the set of anomalies, generate an anomaly connected region containing the target core point and its adjacent points.

7. The intelligent fire prevention and fire management method for lithium battery energy storage power stations according to claim 3, characterized in that, Based on the preliminary anomaly localization data, multi-point consistency verification and historical comparison verification are performed to remove single-point deviations and unstable segments, generating anomaly localization data, including: Based on the preliminary anomaly location data, multiple monitoring signals within the area are extracted to form a location neighborhood signal set; Based on the location neighborhood signal set, consistency verification is performed. If the neighboring points and the core point maintain continuous similarity in trend and amplitude, they are judged as valid anomalies. Based on the location neighborhood signal set, and in conjunction with the stored historical operation data, a comparison and verification are performed. If the abnormal signal deviates significantly from the historical stable pattern, it is determined to be a valid anomaly; otherwise, it is determined to be a false anomaly. Based on the combined results of consistency verification and historical comparison verification, single-point deviations and abnormal unstable segments are removed to generate corrected anomaly location data.

8. The intelligent fire prevention and fire management method for lithium battery energy storage power stations according to claim 4, characterized in that, Based on the anomaly location data, combined with the temperature rise rate, anomaly duration, number of involved battery cells, and distance between the battery cell gap and the nozzle, a penetration coverage requirement dataset is generated, including: Based on the anomaly location data, the temperature rise rate and duration within the anomaly area are extracted to form thermal intensity parameters; Based on the abnormal location data, the number of battery cells involved and the density of battery cell arrangement are counted to form regional impact parameters; The injection geometry parameters are formed based on the cell gap width and the distance from the nozzle to the target area; The thermal intensity parameters, regional influence parameters, and injection geometry parameters are combined to generate a permeability coverage requirement dataset.

9. The intelligent fire prevention and fire management method for lithium battery energy storage power stations according to claim 8, characterized in that, Based on the penetration coverage requirement dataset, the initial injection dose, injection angle, injection pressure, and injection duration are calculated to generate a draft targeted treatment instruction, including: Based on the thermal intensity parameter in the penetration coverage requirement parameters, combined with the temperature rise rate and the duration of the anomaly, the cooling requirement is determined, and the initial injection dose is obtained. Based on the geometric parameters in the penetration coverage requirement parameters, and combined with the spatial angle between the nozzle position and the cell gap, the spray angle is determined to achieve spray coverage of the center of the abnormal area. Based on the geometric parameters in the penetration coverage requirement parameters, combined with the nozzle distance and cell gap width, the driving force requirement is determined, and the injection pressure is obtained; Based on the regional impact parameters in the penetration coverage requirement parameters, combined with the parameters of the number of battery cells and thermal intensity involved, the required continuous cooling time is calculated to obtain the spraying duration; Based on the initial injection dose, injection angle, injection pressure, and injection duration, a draft directional treatment instruction containing multiple parameters is generated.

10. A smart fire prevention and fire protection management system for lithium battery energy storage power stations, characterized in that: The system, used in any one of claims 1 to 9, comprises: The multi-source monitoring module is used to collect data from multi-level temperature sensors, gas detectors, and smoke detectors inside and outside the battery module, forming multi-source monitoring data. The early warning generation module is used to perform time series rate analysis on multi-source monitoring data. If the temperature rise rate of the internal monitoring signal exceeds the preset rate threshold and there is a trigger time difference with the external monitoring signal, early warning data is generated and a priority label is assigned to the corresponding cell area. The anomaly localization module is used to perform spatial cross-comparison based on early warning data and the sensor matrix distribution within the battery module to obtain anomaly localization data, and to eliminate false anomalies through multi-point consistency verification. The treatment instruction module is used to generate directional treatment instructions for the corresponding injection unit based on the abnormal location data, and to set the injection data in the treatment instructions, including the injection dose, injection angle, injection pressure and injection duration. The directional spray module is used to control the spray unit to release atomized fire extinguishing medium according to the directional disposal command, so that the medium forms a permeation flow field in the gap between the battery cells, and to collect and form disposal feedback data. The trend analysis module is used to perform trend analysis on the treatment feedback data. If the temperature or gas value does not drop enough to recover to the preset safety range within the preset time limit, a secondary treatment instruction is generated and the injection data is adjusted in the secondary treatment instruction. The event logging module is used to generate event logging data based on the handling feedback data and secondary handling instructions, and to convert the event logging data into optimization parameters and feed them back to the monitoring platform.

Images