A safety warning method suitable for lithium iron phosphate battery test process

Abstract

The application relates to the technical field of battery testing, in particular to a safety early warning method suitable for a lithium iron phosphate battery testing process, which comprises the following steps: S1, performing thermal runaway testing on a battery to be tested, collecting and storing battery parameters to generate historical data in the testing process, and turning to S2 when a safety valve of the battery to be tested is broken; S2, generating a monitoring threshold according to the battery parameters, and respectively generating a first prediction result of a first time period and a second prediction result of a second time period according to the historical data; and S3, generating corresponding alarm messages according to the first prediction result and the second prediction result. The method has the beneficial effects that the historical data are generated by acquiring multiple battery parameters, the first prediction result corresponding to the first time period and the second prediction result corresponding to the second time period are respectively generated, the thermal runaway state of the battery is effectively predicted based on the multiple battery parameters, the risk of battery thermal runaway can be effectively judged by engineering personnel, and the accuracy of risk judgment is improved.

Description

Technical Field

[0001] This invention relates to the field of battery testing technology, and specifically to a safety warning method applicable to the testing process of lithium iron phosphate batteries. Background Technology

[0002] Lithium-ion batteries release various flammable and explosive gases such as H2, CO, and CH4 when they experience thermal runaway, posing significant safety hazards. This is especially true in energy storage systems, where individual battery cells are connected in series and parallel to form modules for centralized use. If a single battery experiences thermal runaway, the released heat can propagate throughout the entire module; simultaneously, the released flammable gases can accumulate within the enclosed space of the energy storage enclosure, potentially leading to a fire or explosion at the entire energy storage power station.

[0003] In the existing technology, there are corresponding monitoring methods for the thermal runaway of lithium-ion batteries. Such technical solutions usually design corresponding sensors for lithium-ion batteries, including temperature and gas concentration sensors, and monitor whether the lithium-ion battery has an abnormal temperature rise or gas leakage to monitor whether the lithium-ion battery has a thermal runaway condition.

[0004] However, in actual implementation, the inventors found that the above solution only makes simple judgments based on a single factor, such as whether the gas concentration has reached the leakage threshold or whether the temperature rise has reached the temperature rise threshold, to trigger an alarm. This results in relatively simple judgment elements, which are prone to misjudgment and cannot effectively predict the future condition of lithium-ion batteries, leading to insufficient judgment accuracy. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, a safety early warning method applicable to the testing process of lithium iron phosphate batteries is provided.

[0006] The specific technical solution is as follows:

[0007] A safety warning method applicable to the testing process of lithium iron phosphate batteries includes:

[0008] Step S1: Perform thermal runaway test on the battery under test, collect and store battery parameters to generate historical data during the test, and proceed to step S2 when the safety valve of the battery under test breaks.

[0009] Step S2: Generate a monitoring threshold based on the battery parameters, and simultaneously generate a first prediction result for the first time period and a second prediction result for the second time period based on the historical data.

[0010] Step S3: Generate corresponding alarm messages based on the first prediction result and the second prediction result.

[0011] On the other hand, step S1 includes:

[0012] Step S11: Set the battery under test in the test system and perform a thermal runaway test;

[0013] Step S12: Collect and store the historical data of the battery under test, and acquire the image data of the battery under test;

[0014] Step S13: Determine whether the safety valve of the battery under test is damaged based on the image data;

[0015] If so, proceed to step S2;

[0016] If not, return to step S12.

[0017] On the other hand, step S2 includes:

[0018] Step S21: Generate a safety threshold, an insulation threshold, and a pressure threshold as the monitoring thresholds based on the battery parameters;

[0019] Step S22: Generate a first trend of change corresponding to the first time period and a second trend of change corresponding to the second time period based on the historical data;

[0020] Step S23: Compare the first trend and the second trend with each of the monitoring thresholds to obtain the first prediction result and the second prediction result.

[0021] On the other hand, step S21 includes:

[0022] Step S211: Record the first temperature value, the pressure in the first cavity and the initial gas concentration when the safety valve of the battery under test breaks as the safety threshold, and generate the pressure threshold based on the pressure in the first cavity and the battery volume of the battery under test.

[0023] Step S212: Obtain the temperature rise rate of the battery under test. When the temperature rise rate reaches the preset temperature rise threshold, proceed to step S213.

[0024] Step S213: Use the current second temperature value, the pressure inside the second cavity, and the current gas concentration of the battery under test as the insulation threshold.

[0025] On the other hand, in step S22, the first trend of change is generated by a pre-trained first neural network model, and the second trend of change is generated by a pre-trained second neural network model.

[0026] The first neural network model and the second neural network model are either BP neural network models, long short-term memory neural network models, fuzzy neural network models, or convolutional neural network models.

[0027] On the other hand, step S3 includes:

[0028] Based on the first prediction result and the second prediction result, it is determined whether the battery under test will trigger the monitoring threshold within the first time period or the second time period;

[0029] When the battery under test triggers the monitoring threshold in the first time period, a level 1 alarm message is generated and the operation and maintenance personnel are notified to be on standby remotely.

[0030] When the battery under test triggers the monitoring threshold in the second time period, a level 2 alarm message is generated and the maintenance personnel are notified to be on-site.

[0031] When the battery under test does not trigger the monitoring threshold, return to step S2 to update the first prediction result and the second prediction result.

[0032] On the other hand, during step S2, a real-time prediction process is also performed, which includes:

[0033] Step A1: Obtain the real-time monitoring temperature, real-time monitoring chamber pressure, and real-time gas concentration of the battery under test, and generate a level 3 alarm message when the real-time monitoring temperature, real-time monitoring chamber pressure, or real-time gas concentration triggers the safety threshold, then proceed to step A2.

[0034] Step A2: Based on the historical data, a data-driven model is used to generate a third prediction result for the third time period. Based on the third prediction result, it is determined whether the battery under test will trigger the insulation threshold during the third time period.

[0035] If so, generate a level 4 alarm message and power off the battery management unit;

[0036] If not, return to step A2.

[0037] On the other hand, during step S2, a real-time monitoring process is also performed, which includes:

[0038] Step B1: Obtain the real-time monitoring temperature, real-time monitoring chamber pressure, and real-time gas concentration of the battery under test;

[0039] Step B2: Determine whether the real-time monitored temperature, the real-time monitored pressure value in the chamber, or the real-time gas concentration triggers a pre-calibrated absolute threshold.

[0040] If so, generate a level 5 alarm message and power off the battery management unit and automatically activate the fire suppression system;

[0041] If not, return to step B2.

[0042] On the other hand, in step S11, the thermal runaway test includes an adiabatic thermal runaway test and an overheating thermal runaway test;

[0043] When the thermal runaway test is the overheating thermal runaway test, if the battery under test triggers the test termination condition, the execution of step S11 shall be stopped.

[0044] The test termination conditions include: the voltage drop of the battery under test exceeds 25% of the initial voltage; or the temperature rise rate of the battery under test reaches the test temperature rise rate threshold; or, a puncture appears on the surface of the battery under test.

[0045] A security early warning system, characterized in that the security early warning method described above includes:

[0046] The test module performs thermal runaway tests on the battery under test, and collects and stores battery parameters to generate historical data during the test.

[0047] A prediction module is connected to the test module. The prediction module generates a monitoring threshold based on the battery parameters, and simultaneously generates a first prediction result for a first time period and a second prediction result for a second time period based on the historical data.

[0048] The processing module is connected to the prediction module, and the processing module generates a corresponding alarm message based on the first prediction result and the second prediction result.

[0049] The above technical solution has the following advantages or beneficial effects:

[0050] To address the issue that existing technologies relying solely on single-factor assessments are ineffective, this solution generates historical data by acquiring multiple battery parameters. It then generates a first prediction result corresponding to a first time period and a second prediction result corresponding to a second time period. This enables effective prediction of battery thermal runaway states based on multiple battery parameters, helping engineers effectively assess the risk of battery thermal runaway and improving the accuracy of risk assessment. Attached Figure Description

[0051] Embodiments of the invention will be described more fully with reference to the accompanying drawings. However, the drawings are for illustration and explanation only and do not constitute a limitation on the scope of the invention.

[0052] Figure 1This is a schematic diagram of a security early warning method according to an embodiment of the present invention;

[0053] Figure 2 This is a schematic diagram of sub-step S1 in an embodiment of the present invention;

[0054] Figure 3 This is a schematic diagram of sub-step S2 in an embodiment of the present invention;

[0055] Figure 4 This is a schematic diagram of sub-step S21 in an embodiment of the present invention;

[0056] Figure 5 This is a schematic diagram of sub-step S3 in an embodiment of the present invention;

[0057] Figure 6 This is a schematic diagram of the real-time prediction process in an embodiment of the present invention;

[0058] Figure 7 This is a schematic diagram of the real-time monitoring process in an embodiment of the present invention;

[0059] Figure 8 This is a schematic diagram of a safety early warning system in an embodiment of the present invention. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0062] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0063] This invention includes:

[0064] A safety early warning method applicable to the testing process of lithium iron phosphate batteries, such as Figure 1 As shown, it includes:

[0065] Step S1: Perform thermal runaway test on the battery under test, collect and store battery parameters to generate historical data during the test, and proceed to step S2 when the safety valve of the battery under test breaks.

[0066] Step S2: Generate monitoring thresholds based on battery parameters, and simultaneously generate a first prediction result for the first time period and a second prediction result for the second time period based on historical data.

[0067] Step S3: Generate corresponding alarm messages based on the first prediction result and the second prediction result.

[0068] Specifically, in response to the problem that existing monitoring schemes rely solely on a single factor for simple judgment and thus fail to achieve good results, this embodiment collects and stores multiple sets of battery parameters as historical data during the thermal runaway test, generates monitoring thresholds based on the battery parameters, and generates a first prediction result and a second prediction result based on the historical data, thereby achieving a better prediction effect on battery condition and improving the accuracy of the judgment.

[0069] Furthermore, addressing the problem that existing technologies, which rely solely on sensor thresholds for judgment, struggle to predict battery thermal runaway, this embodiment collects historical data of the battery under test during the thermal runaway test. Subsequently, when the safety valve of the battery under test ruptures, the progress of the thermal runaway state of the battery under test is predicted based on the historical data within a first time period and a second time period, yielding a first prediction result and a second prediction result. This generates corresponding alarm messages, thereby achieving a better prediction effect for the battery under test.

[0070] In implementation, the above scheme is set up as a software embodiment in the corresponding testing equipment. It is used to further predict the condition of the battery under test when thermal runaway testing occurs and a safety valve rupture occurs, thereby determining whether the thermal runaway state of the battery under test will further develop and whether intervention is necessary. To achieve better intervention results, the first time period is generally longer than the second time period. For example, in one embodiment, the first time period is 24 hours and the second time period is 6 hours, thus informing the test personnel of the battery's progress. Battery parameters are parameters that can be collected by deploying corresponding sensors on the battery under test, including temperature, air pressure, gas concentration, and voltage.

[0071] In one embodiment, such as Figure 2 As shown, step S1 includes:

[0072] Step S11: Set up the battery under test in the test system and perform a thermal runaway test;

[0073] Step S12: Collect and store historical data of the battery under test, and acquire image data of the battery under test;

[0074] Step S13: Determine whether the safety valve of the battery under test is damaged based on the image data;

[0075] If so, proceed to step S2;

[0076] If not, return to step S12.

[0077] Specifically, in order to achieve better prediction results, in this embodiment, during the thermal runaway test of the battery under test, the test system is set up to collect battery parameters of the battery under test and store them as historical data, which provides corresponding data for the subsequent prediction process. Furthermore, by collecting image data of the battery under test, a better monitoring effect of the battery under test is achieved.

[0078] During implementation, the setup process of the test system includes: placing the battery under test in an accelerated calorimeter (ARC), setting temperature and voltage sensors at the tabs of the battery under test, winding heating wires around both sides of the battery under test and setting temperature sensors, arranging pressure sensors and gas detectors above the cavity of the accelerated calorimeter, and setting a camera in the internal viewing window of the accelerated calorimeter to collect and record the battery's appearance as image data in real time, and to measure the battery temperature and voltage, as well as the pressure and gas concentration inside the cavity of the accelerated calorimeter.

[0079] In one embodiment, in step S11, the thermal runaway test includes an adiabatic thermal runaway test and an overheating thermal runaway test;

[0080] When the thermal runaway test is an overheat thermal runaway test, if the battery under test triggers the test termination condition, step S11 is stopped.

[0081] Test termination conditions include: the voltage drop of the battery under test exceeds 25% of the initial voltage; or the temperature rise rate of the battery under test reaches the test temperature rise rate threshold; or, a puncture appears on the surface of the battery under test.

[0082] Specifically, in order to achieve better adaptability to various types of thermal runaway tests, this embodiment further divides thermal runaway tests into adiabatic thermal runaway tests and overheating thermal runaway tests. When the thermal runaway test is an overheating thermal runaway test, a test termination condition is further set to achieve better test results.

[0083] In one embodiment, such as Figure 3 As shown, step S2 includes:

[0084] Step S21: Generate safety threshold, insulation threshold and pressure threshold as monitoring thresholds based on battery parameters;

[0085] Step S22: Generate a first trend of change corresponding to the first time period and a second trend of change corresponding to the second time period based on historical data;

[0086] Step S23: Compare the first trend and the second trend with each monitoring threshold to obtain the first prediction result and the second prediction result.

[0087] Specifically, to achieve better monitoring results, the battery parameters are pre-processed to obtain safety thresholds, insulation thresholds, and pressure thresholds as monitoring thresholds during the battery prediction process. Subsequently, based on historical data, the battery status in two time periods is predicted to obtain a first trend and a second trend, which are then compared with the monitoring thresholds to obtain a first prediction result and a second prediction result. That is, whether the battery under test will further experience thermal runaway in the next 24 hours or 6 hours, thereby achieving better prediction results.

[0088] As an optional implementation, during step S23, the voltage value of the battery under test is further read to determine whether it is overcharged. If overcharged, the overcharge thermal runaway safety warning threshold is used; otherwise, the adiabatic thermal runaway safety warning threshold is used. If overcharge thermal runaway occurs, the safety warning threshold for the lithium iron phosphate battery and its module is set to: voltage safety threshold U. G Voltage absolute threshold U G1 Temperature safety threshold T PG Temperature absolute threshold T 2G Temperature change rate ΔT G Gas concentration safety threshold (H2) G and CO G ), gas concentration insulation threshold (H2) G1 and CO G1 ), gas concentration change rate (ΔH) 2G and △CO G );

[0089] In one embodiment, such as Figure 4 As shown, step S21 includes:

[0090] Step S211: Record the first temperature value, the pressure in the first chamber and the initial gas concentration when the safety valve of the battery under test breaks as a safety threshold, and generate a pressure threshold based on the pressure in the first chamber and the battery volume of the battery under test.

[0091] Step S212: Obtain the temperature rise rate of the battery under test. When the temperature rise rate reaches the preset temperature rise threshold, proceed to step S213.

[0092] Step S213: Use the current second temperature value of the battery under test, the pressure inside the second chamber, and the current gas concentration as the insulation threshold.

[0093] Specifically, in order to achieve better testing results, in this embodiment, by constructing the above-mentioned acquisition process, the battery parameters when the safety valve of the battery under test breaks are first used as the safety threshold, and then the thermal runaway state is judged by the temperature rise rate, so as to achieve better identification results.

[0094] During implementation, when conducting adiabatic thermal runaway tests on lithium iron phosphate batteries, the temperature rise rate on the battery surface exceeding 0.02℃ / min is first defined as T. 0J (Unit: °C) Starting from this temperature point, record the battery temperature and voltage U. J (Unit: V) Trend of change over time;

[0095] Then, starting from the time the safety valve ruptured as recorded in the video, the highest battery temperature T was recorded. PJ Pressure P within the ARC cavity J (Unit: Bar / L) and gas concentration (H2) J (Unit: ppm or %) and CO J (Unit: ppm or %)), and use the temperature, pressure and gas concentration values ​​at this moment as the safety threshold;

[0096] Next, the temperature point T at which the temperature rise reaches 1℃ / min. 2J As the endpoint, the temperature and pressure values ​​P at this moment are... J1 and gas concentration value H2 J1 CO J1 As an insulation threshold, the rate of temperature change ΔT was calculated. J (Unit: ℃ / min), Pressure change rate ΔP J (Unit: (Bar / L) / min) and gas concentration change rate (ΔH2) J (Units: ppm / min or % / min) and ΔCO J (Unit: ppm / min or % / min)

[0097] Furthermore, P J It is the ratio of the measured pressure value P (in Bar) inside the ARC cavity to the volume of the ARC cavity (in L), used to characterize the pressure threshold per unit volume in a relatively sealed space (such as a battery module) when a lithium iron phosphate battery experiences thermal runaway.

[0098] Furthermore, when conducting overheating and thermal runaway tests on lithium iron phosphate batteries, firstly, the lithium iron phosphate battery is overcharged at a certain charging rate C (unit: A); the overcharge test is stopped when two or more of the following phenomena occur: a) voltage drop exceeds 25%; b) significant temperature rise occurs; c) battery rupture.

[0099] Secondly, when overcharging causes the voltage to drop but the drop does not exceed 25%, this voltage is set to U. G (Unit: V), and use it as the voltage safety threshold, and record the battery temperature and voltage change trends over time;

[0100] Secondly, starting from the time the safety valve ruptures as recorded in the video, the highest battery temperature T is recorded. PG Pressure P within the ARC cavity G Gas concentration (H2) G and CO G ), and use this value as a safety threshold;

[0101] Finally, the temperature point T at which the temperature rise reaches 1℃ / min is taken as an example. 2G As the endpoint, the temperature and pressure values ​​P at this moment are... G1 and gas concentration value H2 G1 and CO G1 As the insulation threshold, and the highest voltage value during this process is set as the insulation threshold U. G1 And calculate the rate of temperature change ΔT G Pressure change rate ΔP G and the rate of change of gas concentration (ΔH) 2G and △CO G ).

[0102] Furthermore, considering that the gas pressure per unit volume and its rate of change after thermal runaway in a lithium iron phosphate battery are measured at high temperatures, it is necessary to convert them to volumes under normal temperature and pressure for measurement. The conversion process is as follows:

[0103] According to the ideal gas law:

[0104] PV=nRT (1)

[0105] In the formula, P is the pressure (Pa), and V is the gas volume (m³). 3 ), n is the amount of substance of the gas (mol), R is the molar gas constant (J / (mol K)), and T is the temperature (K).

[0106] Assuming the gas temperature is the same in any region within the cavity, then at any given time...

[0107] nR=(PV) / T (2)

[0108] Converted to gas volume at room temperature and pressure:

[0109] V0=(nR*298.15) / P0 (3)

[0110] In the formula, P0 is the pressure (Pa) of P at any temperature converted to room temperature, and V0 is the gas volume (m³) at room temperature and pressure. 3 The increase in volume compared to the previous moment is the difference in gas production ΔV(L) between two adjacent moments ΔTime.

[0111] ΔV=V n+1 -V n=(nR*298.15) / P 0 n+1 -(nR*298.15) / P 0 n (4)

[0112] The gas production rate ΔV / ΔT (L / s) is

[0113] ΔV / ΔT=(V n+1 -V n ) / (t n+1 -t n )=((nR*298.15) / P 0 n+1 -(nR*298.15) / P 0 n ) / (t n+1 -t n (5)

[0114] In one embodiment, in step S22, a first trend of change is generated using a pre-trained first neural network model, and a second trend of change is generated using a pre-trained second neural network model.

[0115] The first and second neural network models are either BP neural network models, long short-term memory neural network models, fuzzy neural network models, or convolutional neural network models.

[0116] Specifically, to achieve better prediction of battery thermal runaway, this embodiment selects a corresponding neural network model and trains it on data from different time periods to achieve better prediction results. Generally, to achieve better prediction results, for the first time period of 24 hours, historical data from the previous 6 hours can be used as input data, and the model runs for no more than 30 minutes to predict the first trend within the first time period; for the second time period of 6 hours, historical data from the previous hour can be used as input data, and the model runs for no more than 10 minutes to predict the second trend within the second time period.

[0117] In one embodiment, such as Figure 5 As shown, step S3 includes:

[0118] Based on the first and second prediction results, determine whether the battery under test will trigger the monitoring threshold in the first or second time period.

[0119] When the battery under test will trigger the monitoring threshold in the first time period, a level 1 alarm message will be generated and the operation and maintenance personnel will be notified to be on standby remotely.

[0120] When the battery under test will trigger the monitoring threshold in the second time period, a level 2 alarm message will be generated and the maintenance personnel will be notified to be on duty on site.

[0121] If the battery under test does not trigger the monitoring threshold, return to step S2 to update the first prediction result and the second prediction result.

[0122] Specifically, in order to achieve a better early warning effect, in this embodiment, the appropriate emergency response method is selected by obtaining the first prediction result and the second prediction result, thereby achieving a better prediction effect.

[0123] In one embodiment, during step S2, a real-time prediction process is also performed, such as... Figure 6 As shown, the real-time prediction process includes:

[0124] Step A1: Obtain the real-time monitoring temperature, real-time monitoring chamber pressure, and real-time gas concentration of the battery under test, and generate a level 3 alarm message when the real-time monitoring temperature, real-time monitoring chamber pressure, or real-time gas concentration triggers the safety threshold, then proceed to step A2.

[0125] Step A2: Based on historical data, a data-driven model is used to generate a third prediction result for the third time period. Based on the third prediction result, it is determined whether the battery under test will trigger the insulation threshold during the third time period.

[0126] If so, generate a level 4 alarm message and power off the battery management unit, then manually activate the fire suppression system;

[0127] If not, return to step A2.

[0128] Specifically, in order to achieve a better early warning effect, this embodiment further constructs a real-time prediction process. This real-time prediction process is triggered after the execution of step S2. When the real-time monitoring temperature of the battery under test, the real-time monitoring pressure value in the chamber, or the real-time gas concentration triggers the safety threshold, a data-driven model is further used to predict the battery status in a shorter third time period, thereby achieving a better real-time judgment effect.

[0129] During implementation, data-driven models include, but are not limited to, Partial Least Squares Regression (PLS model), Autoregressive (AR model), Moving Average (MA model), Autoregressive Moving Average (ARMA model), Autoregressive Integrated Moving Average (ARIMA) model, and Elastic Network Regression (EN model). The third time period is 5 minutes, and its input data consists of historical data from the previous minute.

[0130] In one embodiment, during step S2, a real-time monitoring process is also performed, such as... Figure 7 As shown, the real-time monitoring process includes:

[0131] Step B1: Acquire the real-time monitoring temperature, real-time monitoring chamber pressure, and real-time gas concentration of the battery under test;

[0132] Step B2: Determine whether the real-time monitored temperature, the real-time monitored pressure value in the chamber, or the real-time gas concentration triggers the pre-calibrated absolute threshold.

[0133] If so, generate a level 5 alarm message and power off the battery management unit and automatically activate the fire suppression system;

[0134] If not, return to step B2.

[0135] A security early warning system, characterized in that it is used to implement the above-mentioned security early warning method, such as Figure 8 As shown, it includes:

[0136] Test Module 1 performs thermal runaway tests on the battery under test, and collects and stores battery parameters to generate historical data during the test.

[0137] Prediction module 2 is connected to test module 1. Prediction module 2 generates monitoring thresholds based on battery parameters, and generates a first prediction result for the first time period and a second prediction result for the second time period based on historical data.

[0138] The processing module 3 is connected to the prediction module 2. The processing module 3 generates corresponding alarm messages based on the first prediction result and the second prediction result.

[0139] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made based on the description and illustrations of the present invention should be included within the protection scope of the present invention.

Claims

1. A safety early warning method applicable to the testing process of lithium iron phosphate batteries, characterized in that, include: Step S1: Perform thermal runaway test on the battery under test, collect and store battery parameters to generate historical data during the test, and proceed to step S2 when the safety valve of the battery under test breaks. Step S2: Generate a monitoring threshold based on the battery parameters, and simultaneously generate a first prediction result for the first time period and a second prediction result for the second time period based on the historical data. Step S3: Generate corresponding alarm messages based on the first prediction result and the second prediction result; Step S2 includes: Step S21: Generate a safety threshold, an insulation threshold, and a pressure threshold as the monitoring thresholds based on the battery parameters; Step S22: Generate a first trend of change corresponding to the first time period and a second trend of change corresponding to the second time period based on the historical data; Step S23: Compare the first trend and the second trend with each of the monitoring thresholds to obtain the first prediction result and the second prediction result; Step S21 includes: Step S211: Record the first temperature value, the pressure in the first cavity and the initial gas concentration when the safety valve of the battery under test breaks as the safety threshold, and generate the pressure threshold based on the pressure in the first cavity and the battery volume of the battery under test. Step S212: Obtain the temperature rise rate of the battery under test. When the temperature rise rate reaches the preset temperature rise threshold, proceed to step S213. Step S213: Use the current second temperature value, the pressure inside the second cavity, and the current gas concentration of the battery under test as the insulation threshold.

2. The safety early warning method according to claim 1, characterized in that, Step S1 includes: Step S11: Set the battery under test in the test system and perform a thermal runaway test; Step S12: Collect and store the historical data of the battery under test, and acquire the image data of the battery under test; Step S13: Determine whether the safety valve of the battery under test is damaged based on the image data; If so, proceed to step S2; If not, return to step S12.

3. The safety early warning method according to claim 1, characterized in that, In step S22, the first trend of change is generated using a pre-trained first neural network model, and the second trend of change is generated using a pre-trained second neural network model. The first neural network model and the second neural network model are either BP neural network models, long short-term memory neural network models, fuzzy neural network models, or convolutional neural network models.

4. The safety early warning method according to claim 1, characterized in that, Step S3 includes: Based on the first prediction result and the second prediction result, it is determined whether the battery under test will trigger the monitoring threshold within the first time period or the second time period; When the battery under test triggers the monitoring threshold in the first time period, a level 1 alarm message is generated and the operation and maintenance personnel are notified to be on standby remotely. When the battery under test triggers the monitoring threshold in the second time period, a level 2 alarm message is generated and the maintenance personnel are notified to be on-site. When the battery under test does not trigger the monitoring threshold, return to step S2 to update the first prediction result and the second prediction result.

5. The safety early warning method according to claim 1, characterized in that, During step S2, a real-time prediction process is also performed, which includes: Step A1: Obtain the real-time monitoring temperature, real-time monitoring chamber pressure, and real-time gas concentration of the battery under test, and generate a level 3 alarm message when the real-time monitoring temperature, real-time monitoring chamber pressure, or real-time gas concentration triggers the safety threshold, then proceed to step A2. Step A2: Based on the historical data, a data-driven model is used to generate a third prediction result for the third time period. Based on the third prediction result, it is determined whether the battery under test will trigger the insulation threshold during the third time period. If so, generate a level 4 alarm message and power off the battery management unit, then manually activate the fire suppression system; If not, return to step A2.

6. The safety early warning method according to claim 1, wherein when performing step S2, a real-time monitoring process is further performed, the real-time monitoring process comprising: Step B1: Obtain the real-time monitoring temperature, real-time monitoring chamber pressure, and real-time gas concentration of the battery under test; Step B2: Determine whether the real-time monitored temperature, the real-time monitored pressure value in the chamber, or the real-time gas concentration triggers a pre-calibrated absolute threshold. If so, generate a level 5 alarm message and power off the battery management unit and automatically activate the fire suppression system; If not, return to step B2.

7. The safety early warning method according to claim 2, characterized in that, In step S11, the thermal runaway test includes an adiabatic thermal runaway test and an overheating thermal runaway test. When the thermal runaway test is the overheating thermal runaway test, if the battery under test triggers the test termination condition, the execution of step S11 shall be stopped. The test termination conditions include: the voltage drop of the battery under test exceeds 25% of the initial voltage; or the temperature rise rate of the battery under test reaches the test temperature rise rate threshold; or, a puncture appears on the surface of the battery under test.

8. A safety early warning system, characterized in that, For implementing the security early warning method as described in any one of claims 1-7, comprising: The test module performs thermal runaway tests on the battery under test, and collects and stores battery parameters to generate historical data during the test. A prediction module is connected to the test module. The prediction module generates a monitoring threshold based on the battery parameters, and simultaneously generates a first prediction result for a first time period and a second prediction result for a second time period based on the historical data. The processing module is connected to the prediction module, and the processing module generates a corresponding alarm message based on the first prediction result and the second prediction result.

Images