Gas sampling-based early fire warning method and system for lithium ion energy storage system

Abstract

This invention discloses a method and system for early fire warning of lithium-ion energy storage systems based on gas sampling. The method includes: acquiring gas samples and operating status of the energy storage device's battery; preprocessing the gas samples to obtain clean sampling gas; supplementing the sampling loop with nitrogen based on differential pressure data and recording the supplementary gas flow rate; constructing a monitoring dataset based on the characteristic gases after supplementation; performing wavelet transform on the monitoring dataset to extract data abrupt change features; estimating the long-term trend of gas concentration using unscented Kalman filtering; adjusting the weight ratio of the transform and filtering based on the operating status; fusing the characteristic gas data based on the weight ratio; comparing the fused characteristic data with a preset warning threshold to generate a warning signal and provide tiered warnings. This method utilizes characteristic gas sampling, dual-path cross-verification, and fused characteristic warnings to improve the detection accuracy and reliability of energy storage systems in fire warning scenarios, reducing missed and false alarms.

Description

Technical Field

[0001] This invention relates to the field of early fire warning, and in particular to a method and system for early fire warning of lithium-ion energy storage systems based on gas sampling. Background Technology

[0002] With the rapid development of the new energy industry, lithium-ion energy storage systems have been widely used in industrial and commercial energy storage, distributed energy storage, grid-side energy storage, power plant energy storage and other power energy storage scenarios due to their advantages such as high energy density, high charge and discharge efficiency and long cycle life. They have become important equipment to support the consumption of new energy, grid peak and frequency regulation and energy efficiency. However, the energy storage device integrates multiple lithium-ion battery modules. Whether in industrial and commercial high-power charge and discharge scenarios or in long-term stable operation scenarios of power energy storage, abnormalities such as SEI film decomposition, lithium dendrite growth and internal short circuits are prone to occur inside the battery, which can lead to thermal runaway.

[0003] In the early stages of thermal runaway, lithium-ion batteries release characteristic gases such as CO, CH4, and H2. Changes in the concentration of these gases are key warning signs of thermal runaway fires and can serve as crucial early warning indicators. Currently, fire warnings for various lithium-ion energy storage systems, including industrial and commercial energy storage and power energy storage, largely rely on monitoring conventional parameters such as temperature and voltage. This makes it difficult to capture changes in characteristic gases in the early stages of thermal runaway, resulting in delayed warnings that fail to contain the spread of fires in a timely manner and can easily lead to economic losses and casualties.

[0004] Since power plants typically have a lifespan of 20 to 50 years, existing gas detection devices mostly employ electrochemical or semiconductor designs. In lithium-ion energy storage systems, the confined and enclosed space, coupled with a lack of necessary safeguards, makes them prone to missed alarms due to zero-point drift or false alarms caused by interfering gases, thus affecting alarm accuracy. Gas monitoring devices often use a single monitoring method without cross-verification monitoring units, making them susceptible to interference from temperature and pressure fluctuations within the energy storage cabinet under different scenarios. Furthermore, they lack synchronous monitoring of characteristic gases, making it difficult to distinguish between normal gas fluctuations and early-stage thermal runaway characteristic gas changes. How to accurately monitor early-stage thermal runaway characteristic gases and provide early fire warnings has become an urgent problem to be solved in the field of new energy battery system safety. This has significant practical implications and application value for promoting the safe and sustainable development of the lithium-ion energy storage industry in various scenarios, including industrial, commercial, and power sectors. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for early fire warning of lithium-ion energy storage systems based on gas sampling.

[0006] To achieve the above objectives, the present invention is implemented according to the following technical solution: The first aspect of this invention provides an early fire warning method for lithium-ion energy storage systems based on gas sampling, comprising: Based on the early warning system, gas samples and operating status of the energy storage device battery are obtained. The gas samples are then cooled, dried, and impurities are filtered out to obtain pre-treated clean sampled gas. The differential pressure data at the sampling interface is detected, nitrogen is added to the sampling loop according to the differential pressure data, and the replenishment flow rate is recorded. A monitoring dataset is constructed based on the characteristic gases after replenishment, including CO, CH4 and H2. The concentration data of the monitoring dataset is diluted and compensated according to the gas replenishment flow rate. The cross-verification deviation after compensation is calculated based on the dual-channel monitoring data. The monitoring dataset is then updated by self-calibrating through the monitoring zero point based on the cross-verification deviation. Wavelet transform is performed on the monitoring dataset to extract data mutation features, and long-term trends of gas concentration are estimated by unscented Kalman filtering. The weight ratio of transform to filtering is adjusted according to the operating status. Based on the weight ratio, the characteristic gas data of each gas is fused, and the fused characteristic data is compared with the preset warning threshold to generate a warning signal. The warning signal is then used to issue a graded warning through the warning system.

[0007] Furthermore, the method for obtaining the clean sampling gas includes: Gas samples are collected through the gas emission interface of the energy storage device battery based on a preset flow rate. During the sampling process, the pressure difference at the sampling interface is less than or equal to 20 kPa, and the preset flow rate is 0.5 L / min to 1 L / min. The operating status is determined based on the charging and discharging current of the battery. If the charging and discharging current of the energy storage device battery is greater than or equal to 0.5C, it is marked as a high-rate operating status. If the charging and discharging current is less than 0.5C, it is marked as a normal operating status. The sampling gas is cooled to room temperature by the early warning system, and then dried to remove water vapor, silica gel volatiles and industrial volatile impurities, thus obtaining clean sampling gas.

[0008] Furthermore, the method for obtaining the monitoring dataset includes: Nitrogen is added to the sampling loop based on differential pressure data, the gas pressure is adjusted to the sampling gas pressure range, and the replenishment flow rate is recorded. CO concentration data is extracted from the clean sampling gas after replenishment using an electrochemical sensor, CH4 concentration data is extracted using the NDIR optical analysis method, H2 concentration data is extracted using the TCD thermal conductivity technology, and temperature data of the clean sampling gas is collected using an N-type thermocouple. A monitoring dataset is obtained based on the concentration data, temperature data, and air pressure data, and the monitoring dataset is a dual-path parallel dataset.

[0009] Furthermore, the method for obtaining the cross-check deviation includes: The dilution coefficient is calculated based on the preset flow rate and the make-up gas flow rate. The dilution coefficient is the ratio of the total flow rate of the sampling loop to the preset flow rate. The total flow rate of the sampling loop is the sum of the preset flow rate and the make-up gas flow rate. The original concentration values ​​of each characteristic gas are corrected based on the dilution factor to obtain diluted concentration data. The cross-verification deviation of the dual-channel monitoring data is calculated based on the diluted concentration data. The gas supply flow rate is adjusted based on the cross-verification deviation based on the monitoring zero point to complete self-calibration. The dilution factor is recalculated based on the calibration result, and the original concentration values ​​in the monitoring dataset are corrected to complete the monitoring dataset update. The monitoring zero point is the stable output value in the pure nitrogen environment in the monitoring pipeline.

[0010] Further, the method for obtaining the weight ratio includes: If the battery system is operating at high rate, the output weight of the wavelet transform is between 60% and 70%, and the output weight of the unscented Kalman filter is between 30% and 40%. If the battery system is operating normally, the output weight of the unscented Kalman filter is between 60% and 70%, and the output weight of the wavelet transform is between 30% and 40%.

[0011] Furthermore, the method for obtaining the warning signal includes: Based on the comparison between fused feature data and preset warning thresholds, if the fused feature data is greater than or equal to the preset warning threshold, the warning level is determined according to the long-term trend of gas concentration. If the feature gas concentration increases by 20% compared to the initial value, a Level 1 warning is output; if the feature gas concentration increases by 50% compared to the initial value, a Level 2 warning is output. Warning signals are generated according to the warning level.

[0012] A second aspect of the present invention provides a fire early warning system for lithium-ion energy storage systems based on gas sampling, comprising: The system includes a gas sampling module, a gas pretreatment module, a gas circulation module, a nitrogen cylinder, a gas monitoring module, a data processing module, and an early warning module. The gas sampling module includes a gas adapter interface and a DN6 insulating hose. The gas adapter interface is made of corrosion-resistant and high-temperature-resistant material. The specifications of the gas adapter interface are matched with the gas emission interface of the energy storage cabinet battery. One end of the gas adapter interface is connected to the gas emission interface, and one end of the DN6 insulating hose is connected to the gas adapter interface, while the other end is connected to the gas pretreatment module. The gas pretreatment module includes a gas cooling unit and a gas drying unit. The gas cooling unit includes a natural air-cooled cooling coil, and the gas drying unit includes a high-efficiency molecular sieve desiccant. The gas inlet of the gas pretreatment module is connected to the DN6 insulated hose, and the gas outlet of the gas pretreatment module is connected to the gas monitoring module through a dedicated pipeline. The gas circulation module includes a gas replenishment module and a gas pressure monitoring module. The gas pressure monitoring module includes a high-precision pressure sensor, and the gas replenishment module includes a flow controller. The gas pressure monitoring module is located at the gas adapter interface. The gas replenishment module is connected to the nitrogen cylinder and the sampling circuit of the gas sampling module through a pipeline. The nitrogen cylinder is a high-pressure sealed gas cylinder, and the outlet of the nitrogen cylinder is equipped with a pressure regulating valve and the flow controller. The gas monitoring module includes two mutually calibrated gas monitoring devices and a temperature monitoring module. The gas monitoring devices are connected to the outlet of the gas pretreatment module through a dedicated pipeline. The temperature monitoring module is installed in the sampling gas path of the gas monitoring module and includes an N-type thermocouple. The data processing module integrates a distributed multi-source data fusion unit and an early warning algorithm unit. The data processing module is electrically connected to the gas monitoring module, the air pressure monitoring module, the gas replenishment module, and the early warning module through circuits. The early warning module includes an audible and visual alarm unit, a warning light, a signal output unit, and a remote communication unit. The remote communication unit includes a 4G communication module and a 5G communication module.

[0013] Furthermore, including: The energy storage device includes a cabinet-type lithium-ion energy storage device, wherein the cabinet is a 20-foot standard container, the energy storage capacity of the energy storage device is 5MWh to 6MWh, the length of the DN6 insulating hose is 5.0m to 20.0m, the drying efficiency of the high-efficiency molecular sieve desiccant is greater than or equal to 95%, and the measurement range of the high-precision pressure sensor is 0 to 100kPa and the measurement accuracy is less than or equal to ±1kPa; The nitrogen cylinder has a rated pressure of 12.0 MPa, and the nitrogen purity is greater than or equal to 99.99%. The CO concentration data measurement range of the gas monitoring device is 0 ppm to 5000 ppm with a measurement accuracy of less than or equal to ±3%FS. The CH4 concentration data measurement range of the gas monitoring device is 0 ppm to 1000 ppm with a measurement accuracy of less than or equal to ±5%FS. The H2 concentration data measurement range of the gas monitoring device is 0 ppm to 1000 ppm with a measurement accuracy of less than or equal to ±4%FS. The temperature monitoring module has a temperature measurement range of -20℃ to 200℃ with a temperature measurement accuracy of less than or equal to ±0.5℃.

[0014] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: This invention collects gas samples of various fire characteristics from an energy storage device, replenishes the sampling circuit with nitrogen, and controls the interface pressure difference to maintain stability in the gas sampling process within a closed energy storage environment. This avoids the introduction of combustible gases and maintains pressure balance. By cooling and drying the gas samples, impurities such as water vapor and silica gel volatiles are filtered out to obtain clean sampling gas, ensuring the long-term monitoring accuracy of the sensor. During the monitoring process, dual-path mutual verification eliminates concentration deviations caused by gas replenishment dilution and sensor zero-point drift, improving the reliability of monitoring data. The weight ratio of wavelet transform and unscented Kalman filtering is dynamically adjusted based on the battery operating status, and graded early warning is performed based on the fusion characteristics of each characteristic gas. This approach takes into account both data mutations in high-rate scenarios and long-term trend monitoring in conventional scenarios, reducing false alarms and missed alarms caused by a single algorithm, and providing early warning of fire risks in energy storage systems. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the steps of an early fire warning method for lithium-ion energy storage systems based on gas sampling, as described in this embodiment of the invention. Detailed Implementation

[0016] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0017] Reference Figure 1 As shown, this invention provides an early fire warning method for lithium-ion energy storage systems based on gas sampling, including: Based on the early warning system, gas samples and operating status of the energy storage device battery are obtained. The gas samples are then cooled, dried, and impurities are filtered out to obtain pre-treated clean sampled gas. In the actual assessment, the monitoring object was a 5MWh lithium iron phosphate energy storage system integrated in a 20-foot standard container, which was in a high-rate charge and discharge operation state. The gas adapter interface of the gas sampling module was connected to the gas emission interface of the energy storage cabinet battery pack. The gas emission interface was made of corrosion-resistant and high-temperature resistant 316L stainless steel with a coating, and its specifications matched the DN15 gas emission interface of the battery pack. The gas adapter interface was connected to the gas pretreatment module through an 8m long DN6 insulated hose. The gas sampling module, gas pretreatment module, gas circulation module, gas monitoring module, data processing module and early warning module of the early warning system were self-tested. The gas sampling module collected gas samples from the battery pack at a flow rate of 0.8L / min. The temperature of the sampled gas was about 35℃. It was cooled to room temperature of 25℃ by the natural air cooling coil of the gas pretreatment module. Then, it was filtered by high-efficiency molecular sieve desiccant to remove water vapor, silica gel volatiles and industrial volatile impurities to obtain clean sampled gas. The differential pressure data at the sampling interface is detected, nitrogen is added to the sampling loop according to the differential pressure data, and the replenishment flow rate is recorded. A monitoring dataset is constructed based on the characteristic gases after replenishment, including CO, CH4 and H2. In the actual evaluation, the gas pressure monitoring module of the gas circulation module monitors the pressure difference at the gas adapter interface in real time. When the pressure difference rises to 22 kPa, exceeding the 20 kPa threshold, the data processing module immediately controls the gas replenishment module to start. Nitrogen is replenished to the sampling loop through the nitrogen cylinder via the pressure regulating valve and flow controller at the outlet. The replenishment flow rate is 0.2 L / min. After adjustment, the pressure difference drops to 18 kPa. The gas monitoring devices of the A and B paths in the gas monitoring module simultaneously monitor the pretreated clean sampled gas. CO data is monitored by electrochemical sensors, CH4 concentration data is monitored by NDIR optical analysis, and H2 concentration data is monitored by TCD thermal conductivity technology. The temperature monitoring module collects the sampled gas temperature of 25.2℃ through an N-type thermocouple. The concentration, temperature, and pressure data are transmitted to the data processing module in real time to obtain the monitoring dataset. The concentration data of the monitoring dataset is diluted and compensated according to the gas replenishment flow rate. The cross-verification deviation after compensation is calculated based on the dual-channel monitoring data. The monitoring dataset is then updated by self-calibrating through the monitoring zero point based on the cross-verification deviation. In the actual evaluation, the data processing module uses an industrial-grade RK3588 high-performance processing chip. Based on the preset flow rate of 0.8 L / min and the make-up gas flow rate of 0.2 L / min, the data processing module calculates the dilution factor, which is (0.8 + 0.2) / 0.8 = 1.25. Based on this dilution factor, the original concentration values ​​of each characteristic gas are corrected. The original readings of sensor A are: CO - 16.0 ppm, CH4 - 4.0 ppm, H2 - 6.4 ppm. Multiplying the original readings by 1.25, the corrected values ​​are: CO 20.0 ppm, CH4 5.0 ppm, and H2 8.0 ppm. For sensor B... After correction, CO is 20.25 ppm, CH4 is 5.125 ppm, and H2 is 8.125 ppm. The cross-verification deviation between the two channels is calculated based on the corrected concentration data, where the cross-verification deviation of CO = |20.25 - 20.0| / 20.0 = 1.25%. The data processing module calibrates the CO zero point of channel A to 1.5 ppm and channel B to 1.8 ppm under pure nitrogen environment during the system debugging phase. The data processing module compares the pre-stored monitoring zero point to confirm that the sensor has no significant drift. There is no need to adjust the gas supply flow rate. The diluted concentration data directly replaces the original concentration value in the monitoring dataset, thus completing the monitoring dataset update. In actual evaluation, another lithium iron phosphate energy storage system experienced fluctuations in battery pack gas pressure, resulting in CO levels of 20.0 ppm for channel A and 22.0 ppm for channel B after correction, with a cross-calibration deviation of 10%. At this point, the data processing module confirmed the sensor was normal based on the monitoring zero point, and then controlled the gas replenishment module to adjust the gas replenishment flow rate from 0.2 L / min to 0.15 L / min. The dilution factor was recalculated to 1.1875, and the concentration was corrected again until the deviation between the two channels no longer decreased significantly, thus completing the self-calibration. Wavelet transform is performed on the monitoring dataset to extract data mutation features, and long-term trends of gas concentration are estimated by unscented Kalman filtering. The weight ratio of transform to filtering is adjusted according to the operating status. In the actual evaluation, the battery pack of the current lithium iron phosphate energy storage system is operating at a high rate. The early warning algorithm unit of the data processing module sets the output weight of wavelet transform to 65% and the output weight of unscented Kalman filter (UKF) to 35%. Wavelet transform multi-scale decomposition is performed on the CO concentration sequence of 20.0ppm, 20.5ppm, 21.0ppm and 21.8ppm in the past 30 minutes to extract the data mutation feature of an increase of 0.8ppm in 15 minutes. At the same time, the UKF estimates the long-term trend of CO gas concentration based on the state space model. Among them, the CO gas concentration has increased slowly and linearly from 20.0ppm to 21.8ppm in the past 2 hours. Based on the weight ratio, the characteristic gas data of each gas is fused, and the fused characteristic data is compared with the preset warning threshold to generate a warning signal. The warning signal is then used to issue a graded warning through the warning system.

[0018] In the actual assessment, the data processing module integrates CO, CH4, and H2 concentration data, along with gas temperature and pressure data, through a distributed multi-source data fusion unit. The data is fused using a 65:35 weighting ratio, outputting fused feature data. The current fused feature data value is 0.72, which is greater than or equal to the preset warning threshold of 0.60, triggering a warning level determination. The data processing module determines the warning level based on the long-term trend of gas concentrations. Using initial values ​​of CO -20.0 ppm, CH4 -5.0 ppm, and H2 -8.0 ppm as a baseline, the current corrected CO concentration has risen to 24.0 ppm, an increase of 20% from the initial value. Upon reaching the first-level warning, the audible and visual alarm unit of the warning module performs intermittent alarms, sounding for 5 seconds and then stopping for 5 seconds, while the warning light flashes yellow. The remote communication unit pushes the first-level warning signal and real-time monitoring data "CO-24.0ppm, CH4-5.2ppm, H2-8.5ppm, temperature 25.2℃, air pressure 18kPa" to the background monitoring center via the 4G and 5G modules, prompting a reduction in the charging and discharging rate and providing an early fire warning 60 minutes in advance. The data processing module stores historical monitoring data with a storage capacity of ≥10,000 records, which can be cyclically overwritten, and subsequent tracing and analysis can be performed based on the historical monitoring data. In the actual assessment, the CO concentration continued to rise to 30.0 ppm, an increase of 50% from the initial value, triggering a level two warning. The audible and visual alarm unit continued to sound, the warning light remained constantly red, the signal output unit linked the PCS system to shut down, and the remote communication unit pushed an emergency warning signal. After the situation was handled, the warning device was manually reset and monitoring continued in a loop.

[0019] In this embodiment, the method for obtaining the clean sampling gas includes: Gas samples are collected through the gas emission interface of the energy storage device battery based on a preset flow rate. During the sampling process, the pressure difference at the sampling interface is less than or equal to 20 kPa, and the preset flow rate is 0.5 L / min to 1 L / min. The operating status is determined based on the charging and discharging current of the battery. If the charging and discharging current of the energy storage device battery is greater than or equal to 0.5C, it is marked as a high-rate operating status. If the charging and discharging current is less than 0.5C, it is marked as a normal operating status. The sampling gas is cooled to room temperature by the early warning system, and then dried to remove water vapor, silica gel volatiles and industrial volatile impurities, thus obtaining clean sampling gas.

[0020] In this embodiment, the method for obtaining the monitoring dataset includes: Nitrogen is added to the sampling loop based on differential pressure data, the gas pressure is adjusted to the sampling gas pressure range, and the replenishment flow rate is recorded. CO concentration data is extracted from the clean sampling gas after replenishment using an electrochemical sensor, CH4 concentration data is extracted using the NDIR optical analysis method, H2 concentration data is extracted using the TCD thermal conductivity technology, and temperature data of the clean sampling gas is collected using an N-type thermocouple. A monitoring dataset is obtained based on the concentration data, temperature data, and air pressure data, and the monitoring dataset is a dual-path parallel dataset.

[0021] In this embodiment, the method for obtaining the cross-check deviation includes: The dilution coefficient is calculated based on the preset flow rate and the make-up gas flow rate. The dilution coefficient is the ratio of the total flow rate of the sampling loop to the preset flow rate. The total flow rate of the sampling loop is the sum of the preset flow rate and the make-up gas flow rate. The original concentration values ​​of each characteristic gas are corrected based on the dilution factor to obtain diluted concentration data. The cross-verification deviation of the dual-channel monitoring data is calculated based on the diluted concentration data. The gas supply flow rate is adjusted based on the cross-verification deviation based on the monitoring zero point to complete self-calibration. The dilution factor is recalculated based on the calibration result, and the original concentration values ​​in the monitoring dataset are corrected to complete the monitoring dataset update. The monitoring zero point is the stable output value in the pure nitrogen environment in the monitoring pipeline.

[0022] In this embodiment, the method for obtaining the weight ratio includes: If the battery system is operating at high rate, the output weight of the wavelet transform is between 60% and 70%, and the output weight of the unscented Kalman filter is between 30% and 40%. If the battery system is operating normally, the output weight of the unscented Kalman filter is between 60% and 70%, and the output weight of the wavelet transform is between 30% and 40%.

[0023] In this embodiment, the method for obtaining the warning signal includes: Based on the comparison between fused feature data and preset warning thresholds, if the fused feature data is greater than or equal to the preset warning threshold, the warning level is determined according to the long-term trend of gas concentration. If the feature gas concentration increases by 20% compared to the initial value, a Level 1 warning is output; if the feature gas concentration increases by 50% compared to the initial value, a Level 2 warning is output. Warning signals are generated according to the warning level.

[0024] A second aspect of the present invention also provides an early fire warning system for lithium-ion energy storage systems based on gas sampling, comprising: The system includes a gas sampling module, a gas pretreatment module, a gas circulation module, a nitrogen cylinder, a gas monitoring module, a data processing module, and an early warning module. The gas sampling module includes a gas adapter interface and a DN6 insulating hose. The gas adapter interface is made of corrosion-resistant and high-temperature-resistant material. The specifications of the gas adapter interface are matched with the gas emission interface of the energy storage cabinet battery. One end of the gas adapter interface is connected to the gas emission interface, and one end of the DN6 insulating hose is connected to the gas adapter interface, while the other end is connected to the gas pretreatment module. The gas pretreatment module includes a gas cooling unit and a gas drying unit. The gas cooling unit includes a natural air-cooled cooling coil, and the gas drying unit includes a high-efficiency molecular sieve desiccant. The gas inlet of the gas pretreatment module is connected to the DN6 insulated hose, and the gas outlet of the gas pretreatment module is connected to the gas monitoring module through a dedicated pipeline. The gas circulation module includes a gas replenishment module and a gas pressure monitoring module. The gas pressure monitoring module includes a high-precision pressure sensor, and the gas replenishment module includes a flow controller. The gas pressure monitoring module is located at the gas adapter interface. The gas replenishment module is connected to the nitrogen cylinder and the sampling circuit of the gas sampling module through a pipeline. The nitrogen cylinder is a high-pressure sealed gas cylinder, and the outlet of the nitrogen cylinder is equipped with a pressure regulating valve and the flow controller. The gas monitoring module includes two mutually calibrated gas monitoring devices and a temperature monitoring module. The gas monitoring devices are connected to the outlet of the gas pretreatment module through a dedicated pipeline. The temperature monitoring module is installed in the sampling gas path of the gas monitoring module and includes an N-type thermocouple. The data processing module integrates a distributed multi-source data fusion unit and an early warning algorithm unit. The data processing module is electrically connected to the gas monitoring module, the air pressure monitoring module, the gas replenishment module, and the early warning module through circuits. The early warning module includes an audible and visual alarm unit, a warning light, a signal output unit, and a remote communication unit. The remote communication unit includes a 4G communication module and a 5G communication module.

[0025] In this embodiment, it includes: The energy storage device includes a cabinet-type lithium-ion energy storage device, wherein the cabinet is a 20-foot standard container, the energy storage capacity of the energy storage device is 5MWh to 6MWh, the length of the DN6 insulating hose is 5.0m to 20.0m, the drying efficiency of the high-efficiency molecular sieve desiccant is greater than or equal to 95%, and the measurement range of the high-precision pressure sensor is 0 to 100kPa and the measurement accuracy is less than or equal to ±1kPa; The nitrogen cylinder has a rated pressure of 12.0 MPa, and the nitrogen purity is greater than or equal to 99.99%. The CO concentration data measurement range of the gas monitoring device is 0 ppm to 5000 ppm with a measurement accuracy of less than or equal to ±3%FS. The CH4 concentration data measurement range of the gas monitoring device is 0 ppm to 1000 ppm with a measurement accuracy of less than or equal to ±5%FS. The H2 concentration data measurement range of the gas monitoring device is 0 ppm to 1000 ppm with a measurement accuracy of less than or equal to ±4%FS. The temperature monitoring module has a temperature measurement range of -20℃ to 200℃ with a temperature measurement accuracy of less than or equal to ±0.5℃.

[0026] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

Claims

1. A method for early fire warning of lithium-ion energy storage systems based on gas sampling, characterized in that, Includes the following steps: Based on the early warning system, gas samples and operating status of the energy storage device battery are obtained. The gas samples are then cooled, dried, and impurities are filtered out to obtain pre-treated clean sampled gas. The differential pressure data at the sampling interface is detected, nitrogen is added to the sampling loop according to the differential pressure data, and the replenishment flow rate is recorded. A monitoring dataset is constructed based on the characteristic gases after replenishment, including CO, CH4 and H2. The concentration data of the monitoring dataset is diluted and compensated according to the gas replenishment flow rate. The cross-verification deviation after compensation is calculated based on the dual-channel monitoring data. The monitoring dataset is then updated by self-calibrating through the monitoring zero point based on the cross-verification deviation. Wavelet transform is performed on the monitoring dataset to extract data mutation features, and long-term trends of gas concentration are estimated by unscented Kalman filtering. The weight ratio of transform to filtering is adjusted according to the operating status. Based on the weight ratio, the characteristic gas data of each gas is fused, and the fused characteristic data is compared with the preset warning threshold to generate a warning signal. The warning signal is then used to issue a graded warning through the warning system.

2. The gas sampling based early fire warning method for lithium ion energy storage systems according to claim 1, characterized in that, A method for obtaining the clean sampling gas includes: Gas samples are collected through the gas emission interface of the energy storage device battery based on a preset flow rate. During the sampling process, the pressure difference at the sampling interface is less than or equal to 20 kPa, and the preset flow rate is 0.5 L / min to 1 L / min. The operating status is determined based on the charging and discharging current of the battery. If the charging and discharging current of the energy storage device battery is greater than or equal to 0.5C, it is marked as a high-rate operating status. If the charging and discharging current is less than 0.5C, it is marked as a normal operating status. The sampling gas is cooled to room temperature by the early warning system, and then dried to remove water vapor, silica gel volatiles and industrial volatile impurities, thus obtaining clean sampling gas.

3. The gas sampling based early fire warning method for lithium ion energy storage systems of claim 1, wherein, The method for obtaining the monitoring dataset includes: Nitrogen is added to the sampling loop based on differential pressure data, the gas pressure is adjusted to the sampling gas pressure range, and the replenishment flow rate is recorded. CO concentration data is extracted from the clean sampling gas after replenishment using an electrochemical sensor, CH4 concentration data is extracted using the NDIR optical analysis method, H2 concentration data is extracted using the TCD thermal conductivity technology, and temperature data of the clean sampling gas is collected using an N-type thermocouple. A monitoring dataset is obtained based on the concentration data, temperature data, and air pressure data, and the monitoring dataset is a dual-path parallel dataset.

4. The gas sampling based early fire warning method for lithium ion energy storage systems of claim 1, wherein, The method for obtaining the cross-check deviation includes: The dilution coefficient is calculated based on the preset flow rate and the make-up gas flow rate. The dilution coefficient is the ratio of the total flow rate of the sampling loop to the preset flow rate. The total flow rate of the sampling loop is the sum of the preset flow rate and the make-up gas flow rate. The original concentration values ​​of each characteristic gas are corrected based on the dilution factor to obtain diluted concentration data. The cross-verification deviation of the dual-channel monitoring data is calculated based on the diluted concentration data. The gas supply flow rate is adjusted based on the cross-verification deviation based on the monitoring zero point to complete self-calibration. The dilution factor is recalculated based on the calibration result, and the original concentration values ​​in the monitoring dataset are corrected to complete the monitoring dataset update. The monitoring zero point is the stable output value in the pure nitrogen environment in the monitoring pipeline.

5. The gas sampling based early fire warning method for lithium ion energy storage systems of claim 1, wherein, The method for obtaining the weight ratio includes: If the battery system is operating at high rate, the output weight of the wavelet transform is between 60% and 70%, and the output weight of the unscented Kalman filter is between 30% and 40%. If the battery system is operating normally, the output weight of the unscented Kalman filter is between 60% and 70%, and the output weight of the wavelet transform is between 30% and 40%.

6. The gas sampling based early fire warning method for lithium ion energy storage systems of claim 1, wherein, The method for obtaining the warning signal includes: Based on the comparison between fused feature data and preset warning thresholds, if the fused feature data is greater than or equal to the preset warning threshold, the warning level is determined according to the long-term trend of gas concentration. If the feature gas concentration increases by 20% compared to the initial value, a Level 1 warning is output; if the feature gas concentration increases by 50% compared to the initial value, a Level 2 warning is output. Warning signals are generated according to the warning level.

7. A gas sampling based early fire warning system for lithium ion energy storage systems, for performing the gas sampling based early fire warning method according to any one of claims 1 to 6, characterized in that The system includes: The system includes a gas sampling module, a gas pretreatment module, a gas circulation module, a nitrogen cylinder, a gas monitoring module, a data processing module, and an early warning module. The gas sampling module includes a gas adapter interface and a DN6 insulating hose. The gas adapter interface is made of corrosion-resistant and high-temperature-resistant material. The specifications of the gas adapter interface are matched with the gas emission interface of the energy storage cabinet battery. One end of the gas adapter interface is connected to the gas emission interface, and one end of the DN6 insulating hose is connected to the gas adapter interface, while the other end is connected to the gas pretreatment module. The gas pretreatment module includes a gas cooling unit and a gas drying unit. The gas cooling unit includes a natural air-cooled cooling coil, and the gas drying unit includes a high-efficiency molecular sieve desiccant. The gas inlet of the gas pretreatment module is connected to the DN6 insulated hose, and the gas outlet of the gas pretreatment module is connected to the gas monitoring module through a dedicated pipeline. The gas circulation module includes a gas replenishment module and a gas pressure monitoring module. The gas pressure monitoring module includes a high-precision pressure sensor, and the gas replenishment module includes a flow controller. The gas pressure monitoring module is located at the gas adapter interface. The gas replenishment module is connected to the nitrogen cylinder and the sampling circuit of the gas sampling module through a pipeline. The nitrogen cylinder is a high-pressure sealed gas cylinder, and the outlet of the nitrogen cylinder is equipped with a pressure regulating valve and the flow controller. The gas monitoring module includes two mutually calibrated gas monitoring devices and a temperature monitoring module. The gas monitoring devices are connected to the outlet of the gas pretreatment module through a dedicated pipeline. The temperature monitoring module is installed in the sampling gas path of the gas monitoring module and includes an N-type thermocouple. The data processing module integrates a distributed multi-source data fusion unit and an early warning algorithm unit. The data processing module is electrically connected to the gas monitoring module, the air pressure monitoring module, the gas replenishment module, and the early warning module through circuits. The early warning module includes an audible and visual alarm unit, a warning light, a signal output unit, and a remote communication unit. The remote communication unit includes a 4G communication module and a 5G communication module.

8. The gas sampling based early fire warning system for lithium ion energy storage systems of claim 7, wherein, include: The energy storage device includes a cabinet-type lithium-ion energy storage device, wherein the cabinet is a 20-foot standard container, the energy storage capacity of the energy storage device is 5MWh to 6MWh, the length of the DN6 insulating hose is 5.0m to 20.0m, the drying efficiency of the high-efficiency molecular sieve desiccant is greater than or equal to 95%, and the measurement range of the high-precision pressure sensor is 0 to 100kPa and the measurement accuracy is less than or equal to ±1kPa; The nitrogen cylinder has a rated pressure of 12.0 MPa, and the nitrogen purity is greater than or equal to 99.99%. The CO concentration data measurement range of the gas monitoring device is 0 ppm to 5000 ppm with a measurement accuracy of less than or equal to ±3%FS. The CH4 concentration data measurement range of the gas monitoring device is 0 ppm to 1000 ppm with a measurement accuracy of less than or equal to ±5%FS. The H2 concentration data measurement range of the gas monitoring device is 0 ppm to 1000 ppm with a measurement accuracy of less than or equal to ±4%FS. The temperature monitoring module has a temperature measurement range of -20℃ to 200℃ with a temperature measurement accuracy of less than or equal to ±0.5℃.

Images