A power battery thermal runaway detection method and system based on multi-source correlation analysis

Abstract

The application discloses a kind of based on multi-source association analysis's power battery thermal runaway detection method and system, belong to power battery field.The method includes: according to vehicle state, charging state, single number judges whether vehicle triggers high charging low discharging or low charging high discharging early warning;Under charging condition, according to the signal collected by vehicle judges whether vehicle triggers power battery water inlet or liquid leakage insulation risk early warning;According to vehicle state and collected historical data judges whether vehicle triggers battery overuse early warning;Monitoring the voltage difference of each single battery, judges whether triggers pressure difference early warning;After vehicle triggers high charging low discharging or low charging low discharging risk early warning, according to the single number of early warning output judges whether triggers temperature anomaly early warning;Based on preset early warning strategy judges whether vehicle triggers battery thermal runaway risk early warning.The application realizes multiple power battery risk early warning by integrating multi-source data, and realizes early warning of thermal runaway based on preset comprehensive early warning strategy.

Description

Technical Field

[0001] This invention belongs to the field of power batteries. Specifically, this invention relates to a method and system for detecting thermal runaway of power batteries based on multi-source correlation analysis. Background Technology

[0002] With the rapid development of the global new energy vehicle industry, its ownership continues to climb. The massive market base means that even low-probability safety accidents have an absolute number that cannot be ignored. As the "heart" of new energy vehicles, the thermal runaway of power batteries, leading to fires, explosions, and other safety incidents, has become a focus of industry and public attention, and a key issue that must be addressed for the industry's large-scale development.

[0003] Current technologies for detecting thermal runaway in power batteries rely on a single dimension, which presents several challenges: single voltage / current monitoring struggles to capture early, weak signals and is susceptible to noise interference; single gas sensors suffer from response delays, insufficient stability, and poor environmental adaptability; and single pressure monitoring is affected by battery pack sealing and daily temperature fluctuations. These issues mean that the timeliness of early warning systems in extreme and rapid thermal runaway scenarios remains a challenge. Summary of the Invention

[0004] This invention aims to overcome the shortcomings of the prior art and proposes a method and system for detecting thermal runaway of power batteries based on multi-source correlation analysis, so as to achieve the following objectives: by integrating multi-source data such as voltage, temperature, and current, multiple power battery risk warnings can be realized, and early warning of thermal runaway can be realized based on a preset comprehensive warning strategy.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for detecting thermal runaway of power batteries based on multi-source correlation analysis, the method comprising:

[0007] Step S1: Determine whether the vehicle has triggered a high-charge-low-release or low-charge-high-release warning based on the vehicle status, charging status, and the collected unit number.

[0008] Step S2: Under charging conditions, determine whether the vehicle has triggered a warning for water ingress or leakage insulation risk in the power battery based on the signals collected by the vehicle.

[0009] Step S3: Determine whether the vehicle has triggered a battery overuse warning based on the vehicle status and collected historical data;

[0010] Step S4: Monitor the voltage difference of each individual battery cell and determine whether a differential voltage warning is triggered.

[0011] Step S5: After the vehicle triggers a high charge / low charge risk warning or a low charge / low charge risk warning, determine whether a temperature abnormality warning is triggered based on the individual unit number output by the warning.

[0012] Step S6: Based on the vehicle warning status in steps S1 to S5, determine whether the vehicle has triggered a battery thermal runaway risk warning based on a preset warning strategy.

[0013] Preferably, step S1 includes: under discharge conditions, periodically identifying and recording cells with continuously low voltage; under charging conditions, periodically identifying and recording cells with continuously high voltage and cells with continuously low voltage; based on the correlation and overlap of the above cells in the same charge-discharge cycle, determining whether the cell's internal resistance is abnormal or its capacity decay is abnormal, generating a corresponding tag and incrementing the count value by 1 for each abnormality determined; within a preset sliding time window, triggering a high-charge-low-discharge or low-charge-low-discharge risk warning based on whether the count value reaches a threshold.

[0014] Preferably, step S1 includes:

[0015] In the same charge-discharge cycle, if there is the same cell number in the cell number set Idx_ConstDchMin and the cell number set Idx_ConstChgMax that is continuously low during discharge, the cell number corresponding to that cell number is determined to have an abnormal internal resistance. A charge-high discharge tag is generated and displayed on the user interface. At the same time, the charge-high discharge count value of that cell is incremented by one. Within a time window of the first time length, the charge-high discharge determination for the same cell number is counted only once.

[0016] If the same cell number exists in the set of cell numbers with continuous low discharge (Idx_ConstDchMin) and the set of cell numbers with continuous low charge (Idx_ConstChgMin), then the cell number corresponding to that cell number is determined to have abnormal capacity decay. A low charge / low discharge tag is generated and displayed on the user interface. At the same time, the low charge / low discharge count value of that cell is incremented by one. Within a time window of the first time length, the low charge / low discharge determination for the same cell number is counted only once.

[0017] Preferably, step S1 includes:

[0018] Establish a sliding time window of a second time length, and count the charge-high discharge count and charge-low discharge count of all individual cells within the sliding time window of the second time length. The second time length is longer than the first time length.

[0019] If, within the sliding time window of the second time length, the count value of the same unit number being charged up and released down is greater than or equal to the preset value, and the unit number is displayed as a charge-up-release-down label on the user interface, then a charge-up-release-down risk warning is triggered, and the repeatedly appearing unit number is output.

[0020] If, within the sliding time window of the second time length, the charge-low-release count value of the same unit number is greater than or equal to the preset value, and the unit number is displayed as a charge-low-release label on the user interface, a charge-low-release risk warning is triggered, and the recurring unit number is output.

[0021] Preferably, step S2 includes:

[0022] Under charging conditions:

[0023] (1) The insulation resistance value is less than the preset insulation resistance value threshold.

[0024] (2) The vehicle is in the off state and the charging state is not charging;

[0025] If both of the above conditions are met simultaneously and continue for a preset period of time, a warning of water ingress or leakage insulation risk in the power battery will be triggered.

[0026] Preferably, step S3 includes: when the total mileage of the vehicle or the data upload time difference reaches the corresponding threshold, counting the number of fast charging times and the number of slow charging times within a preset period; if the number of fast charging times exceeds the first preset threshold, or the number of fast charging times exceeds the second preset threshold and the ratio of the number of fast charging times to the total number of charging times exceeds the preset ratio threshold, then triggering a battery overuse warning.

[0027] Preferably, step S4 includes:

[0028] Acquire and preprocess the current operating data of the power battery. The data includes at least the highest single cell voltage, the lowest single cell voltage, the current, the state of charging, the current, and the time.

[0029] Based on the preprocessed data, early warning judgments are made according to instantaneous pressure difference, average pressure difference over a time period, and pressure difference at the end of discharge or charging, in the discharge, charging, and resting states, respectively, combined with the battery type.

[0030] Preferably, step S5 includes:

[0031] After triggering a high-charge-low-discharge or low-charge-low-discharge risk warning, all individual cell numbers from the warning output are acquired. A sampling period is defined as each change in the total SOC of the power battery within a preset value. Within each sampling period, the temperature and individual cell number of all individual cells are collected in real time. The individual cell numbers of the n cells with the highest temperatures at that time are acquired. If any of the individual cell numbers among the n cells with the highest temperatures is the same as any of the individual cell numbers in the warning output, the count value X of that same individual cell number is incremented by one. The initial value of X is 0. When X ≥ a preset threshold, the corresponding individual cell number is recorded as a continuously high-temperature individual cell number, and the continuously high-temperature event count value of the corresponding individual cell number is incremented by 1. Within the first time period, the same individual cell number is counted only once for a continuously high-temperature event. If, within the second time period, the continuously high-temperature event count value of any individual cell number is greater than or equal to the preset value, a temperature anomaly warning is triggered.

[0032] Preferably, step S6 includes:

[0033] When the vehicle triggers a high charge / low discharge risk warning, a differential pressure warning, or a temperature abnormality warning, a battery thermal runaway risk warning will be triggered.

[0034] If the vehicle triggers a warning for water ingress or leakage insulation risk in the power battery, it will directly trigger a warning for battery thermal runaway risk.

[0035] If the vehicle triggers the battery overuse warning and temperature abnormality warning, it will trigger the battery thermal runaway risk warning.

[0036] This invention also provides a power battery thermal runaway detection system based on multi-source correlation analysis. Using the aforementioned power battery thermal runaway detection method based on multi-source correlation analysis, the system includes:

[0037] The high charge / low discharge or low charge / high discharge warning detection module is used to determine whether the vehicle has triggered a high charge / low discharge or low charge / high discharge warning based on the vehicle status, charging status, and the collected individual unit number.

[0038] The power battery water ingress or leakage insulation risk warning detection module is used to determine whether the vehicle has triggered a power battery water ingress or leakage insulation risk warning based on the signals collected by the vehicle during charging.

[0039] The battery overuse warning detection module is used to determine whether the vehicle has triggered a battery overuse warning based on the vehicle status and collected historical data.

[0040] The differential pressure warning detection module is used to monitor the voltage difference of each individual battery cell and determine whether a differential pressure warning is triggered.

[0041] The temperature anomaly warning detection module is used to determine whether a temperature anomaly warning has been triggered based on the individual unit number output by the vehicle after the vehicle triggers a high charge / low charge risk warning or a low charge / low charge risk warning.

[0042] The battery thermal runaway risk warning and detection module is used to determine whether the vehicle has triggered a battery thermal runaway risk warning based on the current vehicle warning status and a preset warning strategy.

[0043] The technical effects of this invention are as follows:

[0044] This invention integrates multi-source data such as voltage, temperature, and current to achieve multiple risk warning outputs for power batteries and realizes early warning of thermal runaway. Attached Figure Description

[0045] Figure 1 A flowchart of a power battery thermal runaway detection method based on multi-source correlation analysis is provided in an embodiment of the present invention.

[0046] Figure 2 This is a schematic diagram of a battery thermal runaway risk warning strategy provided in an embodiment of the present invention. Detailed Implementation

[0047] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. This is to help those skilled in the art to have a more complete, accurate, and in-depth understanding of the inventive concept and technical solution of the present invention, and to facilitate its implementation. It should be noted that the terms "first," "second," etc., used in this application are only for the convenience of describing the technical solution and to distinguish components. The corresponding component configurations may be the same or different, and are not intended to limit the present application. Furthermore, the various preset values, thresholds, etc., described in this application are preferred values ​​set in the embodiments of this application unless otherwise specified, and are for reference only. In specific implementation, they can be flexibly set according to the actual situation. To make the technical solution of the present invention clearer, the present invention will be explained and illustrated through the following embodiments.

[0048] This invention provides a method for detecting thermal runaway in power batteries based on multi-source correlation analysis. It constructs a multi-layered protection system through multi-dimensional early warning detection, and ultimately determines the risk of thermal runaway based on a preset strategy. Figure 1 As shown, the method includes:

[0049] Step S1: Determine whether the vehicle has triggered a high-charge-low-release or low-charge-high-release warning based on the vehicle status, charging status, and the collected unit number.

[0050] Step S2: Under charging conditions, determine whether the vehicle has triggered a warning for water ingress or leakage insulation risk in the power battery based on the signals collected by the vehicle.

[0051] Step S3: Determine whether the vehicle has triggered a battery overuse warning based on the vehicle status and collected historical data;

[0052] Step S4: Monitor the voltage difference of each individual battery cell and determine whether a differential voltage warning is triggered.

[0053] Step S5: After the vehicle triggers a high charge / low charge risk warning or a low charge / low charge risk warning, determine whether a temperature abnormality warning is triggered based on the individual unit number output by the warning.

[0054] Step S6: Based on the vehicle warning status in steps S1 to S5, determine whether the vehicle has triggered a battery thermal runaway risk warning based on a preset warning strategy.

[0055] Thermal runaway in power batteries often originates from abnormalities in individual cells (increased internal resistance, capacity decay). These abnormalities manifest as low voltage during discharge (high internal resistance leads to a large voltage drop), and as high voltage (high internal resistance leads to large polarization) or low voltage (capacity decay prevents incomplete charging) during charging. While voltage abnormalities under a single operating condition may be caused by temporary factors, persistent abnormalities across operating conditions strongly suggest a battery malfunction. This embodiment, by correlating low discharge voltage with high / low charging voltage, can achieve early warning of high charge / low discharge or low charge / low discharge risks. Refer to step S1; the specific operation of this embodiment is as follows.

[0056] Step S11: When the vehicle is in the state of starting discharge and moving, and the charging state is not charging:

[0057] A sampling period is defined as each time the SOC of the entire power battery decreases by a preset value (in this embodiment, it is set to each sampling period as a 5% decrease in the SOC of the entire power battery). In each sampling period, the individual cell voltage and cell number of all individual cells are collected in real time. The cell numbers of the n individual cells with the lowest cell voltage at this time are added to the set Idx_DchMinZ, where Idx_DchMinZ represents the set of cell numbers created in the Zth sampling period, Z=1, 2, 3, ...

[0058] It then determines whether the n individual battery cell numbers contained in the Idx_DchMinZ set in the current sampling period are the same as the n individual battery cell numbers contained in the Idx_DchMin(Z-1) set in the previous sampling period. If they are the same, the count value N of the corresponding individual cell number is incremented by 1, where the initial value of N is 0. When N ≥ a preset threshold (set to 3 in this embodiment), the corresponding individual cell number is recorded as a low discharge cell number and added to the low discharge cell number set Idx_ConstDchMin.

[0059] Step S12: When the vehicle is in a state of being charged while the engine is off and stationary, and the charging state is "charging while parked":

[0060] A sampling period is defined as each increment of the total SOC of the power battery by a preset value (in this embodiment, it is defined as each increment of the total SOC of the power battery by 5%). Within each sampling period, the individual cell voltage and cell number of all individual cells are collected in real time. The cell numbers of the n individual cells with the lowest individual cell voltage at this time are added to the set Idx_ChgMinZ, where Idx_ChgMinZ represents the set of cell numbers created in the Zth sampling period, Z=1, 2, 3, ...

[0061] Add the cell numbers of the n cells with the highest cell voltage at this time to the set Idx_ChgMaxZ, where Idx_ChgMaxZ represents the set of cell numbers created in the Zth sampling period, Z=1, 2, 3, ...

[0062] Subsequently, it is determined whether the n individual battery cell numbers contained in the Idx_ChgMinZ set in the current sampling period are the same as the n individual battery cell numbers contained in the Idx_ChgMin(Z-1) set in the previous sampling period. If they are the same, the count value M of the corresponding individual cell number is incremented by 1, where the initial value of M is 0. When M ≥ a preset threshold (set to 3 in this embodiment), the corresponding individual cell number is recorded as a continuously low charging cell number and added to the continuously low charging cell number set Idx_ConstChgMin.

[0063] Determine whether the n individual battery cell numbers contained in the Idx_ChgMaxZ set in the current sampling period are the same as the n individual battery cell numbers contained in the Idx_ChgMax(Z-1) set in the previous sampling period. If they are the same, increment the count value P of the corresponding individual cell number by 1, where the initial value of P is 0. When P ≥ a preset threshold (set to 3 in this embodiment), record the corresponding individual cell number as a continuously high or low charging cell number and add it to the continuously high charging cell number set Idx_ConstChgMax.

[0064] Step S13: Each time it is confirmed that the vehicle is in a state of being charged while the engine is off and stationary, and the charging state is being charged while the vehicle is parked, the set of cell numbers with the longest continuous discharge (Idx_ConstDchMin), the set of cell numbers with the longest continuous charge (Idx_ConstChgMax), and the set of cell numbers with the longest continuous charge (Idx_ConstChgMin) recorded during the previous complete charge and discharge cycle are retrieved, and the following association determination is performed:

[0065] If the same cell number exists in the Idx_ConstDchMin set and the Idx_ConstChgMax set, the cell number corresponding to that cell number is determined to have an abnormal internal resistance. A charge-high discharge tag is generated and displayed on the user interface. At the same time, the charge-high discharge count value of that cell number is incremented by one. Within a 24-hour time window, the charge-high discharge determination for the same cell number is counted only once.

[0066] If the same cell number exists in the Idx_ConstDchMin set and the Idx_ConstChgMin set, the cell number corresponding to that cell number is determined to have abnormal capacity decay. A low charge / low discharge tag is generated and displayed on the user interface. At the same time, the low charge / low discharge count value of that cell is incremented by one. Within a 24-hour time window, the low charge / low discharge determination for the same cell number is counted only once.

[0067] Step S14: Establish a 120-hour sliding time window, and count the charge-high discharge count and charge-low discharge count of all individual batteries within the 120-hour sliding time window:

[0068] If, within the 120-hour sliding time window, the count value of the same unit number being charged high and released low is greater than or equal to the preset value, and the unit number is displayed as a charge-high-release-low label on the user interface, a charge-high-release-low risk warning is triggered, and the repeatedly appearing unit number is output.

[0069] If, within the 120-hour sliding time window, the charge-low-release count value of the same unit number is greater than or equal to the preset value, and the unit number is displayed as a charge-low-release label on the user interface, a charge-low-release risk warning is triggered, and the recurring unit number is output.

[0070] Step S1 reduces the false alarm rate of single-cell anomaly identification through correlation verification under discharge and charging conditions, providing reliable support for risk warning of high charge / low discharge or low charge / low discharge. Specifically, when the same cell is simultaneously abnormal under two opposing conditions of low discharge voltage and high charge voltage, it physically points only to a continuous increase in internal resistance; similarly, low discharge voltage and low charge voltage point only to capacity decay. Through this cross-condition correlation, the fault type can be accurately identified, significantly reducing the probability of false alarms.

[0071] Meanwhile, step S1 of this embodiment establishes a 120-hour sliding time window and a 24-hour deduplication counting mechanism. The 24-hour deduplication counting mechanism avoids continuous alarms from a single fault. The 120-hour sliding time window eliminates occasional interference. By accumulating the high-low count and low-low count over time, it only triggers a risk warning when the anomaly persists over a longer time scale, further ensuring the reliability of the warning results.

[0072] Water ingress or electrolyte leakage into the battery pack can lead to a decrease in insulation resistance, causing short circuits and arcing, which are important precursors to thermal runaway. Furthermore, during charging, the high-voltage system is active, making insulation defects more easily exposed and with more severe consequences. Therefore, referring to step S2, this embodiment introduces a judgment to determine whether the vehicle has triggered a warning about water ingress or electrolyte leakage in the power battery insulation risk. Specifically, the operation of step S2 is as follows.

[0073] Under charging conditions (not charging and not in a high-voltage state):

[0074] (1) The insulation resistance value is detected to be less than the preset insulation resistance value threshold (set to 200KΩ in this embodiment); the insulation resistance value can be obtained through insulation testing. The national standard stipulates that insulation testing should be performed by the charging pile before charging and by the vehicle during charging; the European standard stipulates that insulation testing should be performed by the charging pile and the vehicle can be selected freely.

[0075] (2) The vehicle is off and the charging status is not charging (the charging gun is plugged in but charging is not successful or charging has just ended, at which time the high voltage port of the battery pack is exposed).

[0076] If both of the above conditions are met simultaneously and last for a preset period of time (set to 60 seconds in this embodiment), a warning of water ingress or leakage insulation risk in the power battery will be triggered.

[0077] Step S2 implements proactive risk warning for insulation safety before charging. The vehicle being off ensures that insulation testing is performed in a static state, eliminating interference from dynamic loads such as the motor and inverter. The charging status being "not charging" locks in the high-risk window of the charging preparation phase. If an insulation abnormality is detected during this phase, charging can be prevented from starting in time, achieving proactive risk prevention rather than passive accident alarm. The duration setting avoids misjudgments caused by instantaneous insulation fluctuations such as rain splashes.

[0078] Excessive battery use, especially frequent fast charging, is a significant factor accelerating battery aging and increasing the risk of thermal runaway. Existing technologies typically assess battery usage intensity based on simple charging frequency statistics or cumulative charging amounts over fixed time windows, which has the following shortcomings: First, it fails to differentiate between fast and slow charging, treating these two charging modes with vastly different impacts on battery aging as equivalent; second, it does not consider the actual usage stage of the vehicle, applying the same judgment criteria to new and old vehicles, lacking specificity. Therefore, referring to step S3, this embodiment introduces a battery overuse warning system. By statistically analyzing the number of fast and slow charges and the amount of charge within 30 days, combined with a preset threshold strategy, it identifies the risk of accelerated battery aging caused by frequent fast charging and introduces total vehicle mileage or usage time as trigger conditions to achieve differentiated overuse warnings. Specifically, the judgment process in step S3 is as follows.

[0079] If any of the following conditions are met:

[0080] (1) The total mileage of the vehicle is greater than or equal to the preset mileage threshold (in this embodiment, the mileage threshold is set to 120,000 kilometers for passenger cars and 200,000 kilometers for commercial vehicles); the new car battery is in good health and has a certain fast charging tolerance, so the risk of frequent fast charging in the short term is relatively controllable; however, when the vehicle reaches a certain mileage, the battery has aged to a certain extent, and frequent fast charging will significantly accelerate the aging process, so it is necessary to strengthen monitoring.

[0081] (2) The time difference between the initial data upload time and the current data upload time of the vehicle background data is greater than the preset difference threshold (in this embodiment, it is set to 8 years for passenger cars and 5 years for commercial vehicles). Even if the vehicle mileage is low, if the service life is long, the battery may still experience performance degradation due to calendar aging. At this time, frequent fast charging also has a high risk.

[0082] The next step is to determine if the vehicle is off and charging is in a parked charging state.

[0083] The cumulative fast charging charge amount ∑△SOC during each fast charging session over 30 days is calculated. Therefore, the number of fast charging sessions over 30 days is calculated as (∑△SOC) / 100%.

[0084] The cumulative slow charging charge amount ∑△SOC is calculated for each slow charging session over 30 days. Therefore, the number of slow charging sessions over 30 days is calculated as (∑△SOC) / 100%.

[0085] The above calculation method normalizes the charging amount to the equivalent number of full charges, avoiding statistical bias caused by a single fast charge not being fully charged. The determination of fast charging and slow charging states is as follows: if the maximum charging current Imax is detected to be greater than 20A, it is fast charging state; if the maximum charging current Imax is detected to be less than or equal to 20A, it is slow charging state.

[0086] Then, a warning for excessive battery usage can be issued.

[0087] If the number of fast charges within 30 days exceeds the preset first fast charge threshold (set to 20 times in this embodiment), a battery overuse warning will be triggered.

[0088] Alternatively, if the number of fast charges within 30 days exceeds the preset second fast charge threshold (set to 15 times in this embodiment), and satisfies:

[0089] (In this embodiment, it is set to 50%)

[0090] This triggers a battery overuse warning. This indicates that even if the total number of charging cycles is not high, if fast charging accounts for the vast majority, it means that the user is extremely dependent on fast charging, and the battery is in a high-stress charging state for a long time, resulting in a significant risk of aging.

[0091] The voltage difference between individual cells (voltage differential) is a core indicator reflecting battery consistency and an important dimension for thermal runaway early warning. Excessive voltage differential can be dangerous because: during charging and discharging, the cell with the highest voltage may reach the charging cutoff voltage first, causing premature charging termination while other cells are not fully charged; or the cell with the lowest voltage may reach the discharging cutoff voltage first, causing premature discharging termination while other cells are not fully discharged. More seriously, at the end of charging, high-voltage cells may overcharge, triggering side reactions such as lithium plating and gas production; at the end of discharging, low-voltage cells may over-discharge, leading to copper foil dissolution, internal short circuits, and other malfunctions. These extreme situations are all significant causes of thermal runaway. Therefore, this embodiment includes step S4 for voltage differential early warning judgment. The specific operation of step S4 is as follows.

[0092] First, the current operating data of the power battery is acquired and preprocessed. This data includes at least the highest single-cell voltage, lowest single-cell voltage, current, state of charge, and time. The preprocessing includes data cleaning to improve the reliability of the data source. For example, abnormal data meeting one of the following conditions is removed: highest single-cell voltage less than 0.5V or greater than 5V; lowest single-cell voltage less than 0.5V or greater than 5V; highest temperature greater than 125℃ or less than -40℃; lowest temperature greater than 125℃ or less than -40℃; or the data type is resubmitted data reporting.

[0093] Subsequently, based on the preprocessed data, differential pressure warning judgments are made in the discharge state, charging state, and static state, respectively.

[0094] A. The differential pressure warning judgment process is as follows when the device is in a discharge state.

[0095] Step S411: Determine the instantaneous pressure difference:

[0096] For lithium iron phosphate batteries:

[0097] If the discharge current is less than 1C and the highest single-cell voltage is greater than or equal to 3.5V, then SOC is not determined. When the absolute value of the difference between the highest single-cell voltage and the lowest single-cell voltage is greater than 350mV and this condition is met for 5 consecutive frames, an excessive voltage difference warning is issued.

[0098] If the lowest single-cell voltage is less than or equal to 3.0V, no warning judgment will be made;

[0099] If the discharge current is less than 1C, the highest single-cell voltage is less than 3.5V, and the lowest single-cell voltage is greater than 3.0V, then when the absolute value of the difference between the highest single-cell voltage and the lowest single-cell voltage is greater than the discharge voltage difference threshold corresponding to the current SOC and this condition is met for 5 consecutive frames, an excessive voltage difference warning will be issued.

[0100] For ternary lithium batteries, if the discharge current is less than 1C, and the absolute value of the difference between the highest and lowest single cell voltages is greater than the current SOC corresponding to the discharge voltage difference threshold, and this condition is met for 5 consecutive frames, an excessive voltage difference warning will be issued.

[0101] Step S412: Determine the average pressure difference over the time period:

[0102] Select a time period (set to 100s in this embodiment) that is prior to the current data. If there is no discharge data within this time period, no warning is issued; otherwise, calculate the average voltage of each cell within this time period and select the maximum and minimum values ​​of the average voltage of the cells.

[0103] For lithium iron phosphate batteries:

[0104] If the discharge current is less than 1C and the average voltage of the highest single cell is greater than or equal to 3.5V, then the SOC is not judged. When the absolute value of the difference between the average voltage of the highest single cell and the average voltage of the lowest single cell is greater than 350mV, an excessive voltage difference warning is issued.

[0105] If the lowest single-cell voltage is less than or equal to 3.0V, no warning judgment will be made;

[0106] If the discharge current is <1C and the lowest average single-cell voltage is ≤3.0V, SOC is not determined and no warning is issued;

[0107] If the maximum average voltage of a single cell is less than 3.5V and the minimum average voltage of a single cell is greater than 3.0V, when the absolute value of the difference between the maximum and minimum average voltages of a single cell is greater than the voltage difference threshold corresponding to the current SOC, an excessive voltage difference warning will be issued directly.

[0108] For ternary lithium batteries, if the discharge current is less than 1C, and the absolute value of the difference between the highest average voltage of a single cell and the lowest average voltage of a single cell is greater than the discharge voltage difference threshold corresponding to the current SOC, and this condition is met for 5 consecutive frames, an excessive voltage difference warning will be issued.

[0109] Compared to instantaneous data-based judgment, judgment based on average data further smooths out temporary fluctuations and is more suitable for identifying persistent pressure difference anomalies. When both judgment methods are triggered, the system can upgrade the warning level or shorten the warning response time.

[0110] Step S413, Discharge End Judgment: During the discharge end period, if the average voltage of all cells is detected to be greater than the set threshold 1 (3400mV for ternary lithium, 3200mV for lithium iron phosphate) and the lowest cell voltage is less than the set threshold 2 (3200mV for ternary lithium, 3000mV for lithium iron phosphate), it is considered that the battery has excessive voltage difference, an excessive voltage difference warning is issued, and the frequency of warning occurrence is statistically analyzed. The significance of this judgment is that: at the end of the discharge period, the overall voltage is still acceptable, but the voltage of individual cells is extremely low, indicating that there are cells with severe capacity decay. These cells are prematurely depleted in each discharge, which will cause over-discharge damage or even internal short circuit in the long run.

[0111] B. The differential pressure warning judgment process is as follows when charging.

[0112] Step S421: Determine the instantaneous pressure difference:

[0113] For lithium iron phosphate batteries:

[0114] If the highest single-cell voltage is ≥3.5V, SOC is not judged. When the absolute value of the difference between the highest single-cell voltage and the lowest single-cell voltage is >350mV and the condition is met for 5 consecutive frames, an excessive voltage difference warning is issued.

[0115] If the highest single-cell voltage is <3.5V and the lowest single-cell voltage is ≥3.1V, and the absolute value of the difference between the highest single-cell voltage and the lowest single-cell voltage is > the charging voltage difference threshold corresponding to the current SOC and the condition is met for 5 consecutive frames, a voltage difference warning will be issued.

[0116] No differential voltage warning is made if the lowest single-cell voltage is <3.1V;

[0117] For ternary lithium batteries, if the absolute value of the difference between the highest and lowest single cell voltages is greater than the current SOC's corresponding discharge voltage difference threshold and this condition is met for 5 consecutive frames, an excessive voltage difference warning will be issued.

[0118] Step S422: Determine the average pressure difference over the time period:

[0119] Select a time period prior to the current data. If there is no discharge data within this time period, no warning is issued; otherwise, calculate the average voltage of each cell within this time period and select the maximum and minimum values ​​of the average voltage of each cell.

[0120] For lithium iron phosphate batteries:

[0121] If the average voltage of the highest single cell is ≥3.5V, the SOC is not judged. When the absolute value of the difference between the average voltage of the highest single cell and the average voltage of the lowest single cell is >350mV, an excessive voltage difference warning is issued.

[0122] If the highest average voltage of a single cell is less than 3.5V and the lowest average voltage of a single cell is greater than 3.1V, when the absolute value of the difference between the highest and lowest average voltages of a single cell is greater than the charging voltage difference threshold corresponding to the current SOC, an excessive voltage difference warning will be issued directly.

[0123] No warning is issued if the lowest average cell voltage is less than 3.1V.

[0124] For ternary lithium batteries, if the absolute value of the difference between the highest average voltage of a single cell and the lowest average voltage of a single cell is greater than the voltage difference threshold corresponding to the current SOC and this condition is met for 5 consecutive frames, an excessive voltage difference warning will be issued.

[0125] Step S423, End of charging determination: If it is detected during the end of charging period;

[0126] (1) The highest single cell voltage is greater than or equal to the charging cutoff voltage threshold 1 (4200mV for ternary lithium, 3650mV for lithium iron phosphate).

[0127] (2) The average voltage of all cells is less than threshold 2 (4170mV for ternary lithium, 3370mV for lithium iron phosphate) or the lowest cell voltage is less than threshold 3 (4140mV for ternary lithium, 3330mV for lithium iron phosphate).

[0128] When the above conditions are met simultaneously for 5 consecutive frames, an excessive voltage difference warning is issued and the frequency of warning occurrence is statistically analyzed. The significance of this judgment is that: at the end of the charging period, some cells are fully charged while the whole cell or some cells are far from being fully charged, indicating that there are cells with serious capacity decay or increased internal resistance. These cells reach the cutoff condition prematurely in each charge, causing the charging to terminate prematurely. Other cells are undercharged for a long time, and high-voltage cells may be at risk of overcharging.

[0129] C. Under static conditions, the differential pressure warning judgment process is as follows.

[0130] The definition of a static state is: the absolute value of the current is less than or equal to 3A, the duration is greater than 1 hour, and the minimum battery temperature is higher than 15 degrees Celsius.

[0131] When the condition is determined to be static, the voltage difference after static is calculated as the difference between the highest and lowest individual cell voltages, that is, the voltage difference after static = highest individual cell voltage - lowest individual cell voltage.

[0132] Subsequently, based on the current minimum and maximum voltages, combined with the average temperature and the OCV-SOC table (the open-circuit voltage-state-of-charge comparison table, which was obtained through battery calibration experiments, and the OCV-SOC curves differ at different temperatures), the highest and lowest SOCs after static calibration are obtained respectively. If the highest and lowest SOCs fall within the same SOC range, a warning of excessive pressure difference is issued when the pressure difference after static calibration exceeds the static pressure difference threshold corresponding to the current SOC range. If the highest and lowest SOCs fall within different SOC ranges, no warning is issued this time, because the OCVs of different SOC ranges differ significantly, and the pressure difference across ranges is mainly caused by SOC differences, rather than inconsistencies in the battery itself.

[0133] In addition, in step S4 of this embodiment, to facilitate subsequent anomaly tracing, the fault cell number can be preliminarily determined and recorded, including: in the above differential voltage warning determination process, calculating the standard deviation σ and the average μ of the average voltage of the cells in the power battery, and recording all cell numbers that meet the condition that the cell voltage minus the average is greater than 3 times the standard deviation; if there is no cell number that meets the condition, then recording the cell number corresponding to the lowest voltage.

[0134] Step S4 of this embodiment establishes differentiated differential pressure judgment strategies based on the different voltage characteristics of different batteries, avoiding missed or false alarms caused by applying a one-size-fits-all threshold. Simultaneously, three complete differential pressure judgment systems are established for discharge, charging, and stationary states. Effective differential pressure monitoring can be performed regardless of whether the vehicle is in motion, charging, or parked, achieving full-scenario coverage. The method based on average voltage over a time period effectively filters out instantaneous differential pressure fluctuations caused by temporary factors such as sudden current changes and temperature gradients, ensuring consistent warnings for the specific battery. Furthermore, through special judgments at the discharge and charging ends, dangerous scenarios of "overall acceptable, but individual extremes" are identified. These scenarios often correspond to cells with severe capacity decay or increased internal resistance, representing high-risk precursors to thermal runaway.

[0135] Temperature is the most direct and sensitive characteristic parameter of thermal runaway. When a single battery cell experiences faults such as increased internal resistance or micro-short circuits, its heat generation rate increases, manifesting as a temperature rise during charging and discharging. In particular, for cells with increased internal resistance, Joule heating increases significantly under the influence of charging and discharging current, resulting in a higher cell temperature than normal cells. If not identified and intervened in time, the continued temperature rise will trigger a chain reaction of thermal runaway. Therefore, this embodiment introduces step S5 for temperature anomaly warning. After the vehicle has triggered a high-charge-low-discharge or low-charge-low-discharge risk warning, it monitors the temperature of a specific cell number in the warning output, achieving precise identification of the risk cell and verification of the temperature dimension. Specifically, the judgment process of step S5 is as follows.

[0136] After triggering a high-charge-low-discharge or low-charge-low-discharge risk warning, all individual cell numbers from the warning output are acquired. A sampling period is defined as each change in the total SOC of the power battery within a preset value. Within each sampling period, the temperature and individual cell number of all individual cells are collected in real time. The individual cell numbers of the n cells with the highest temperatures at that time are obtained. If any of the individual cell numbers among the n cells with the highest temperatures is the same as any of the individual cell numbers in the warning output, the count value X of that same individual cell number is incremented by one. The initial value of X is 0. When X ≥ a preset threshold, the corresponding individual cell number is recorded as a continuously high-temperature individual cell number, and the continuously high-temperature event count value of the corresponding individual cell number is incremented by 1. Within 24 hours, the same individual cell number is counted only once for a continuously high-temperature event. If within 120 hours, the continuously high-temperature event count value of any individual cell number is greater than or equal to a preset value, a temperature anomaly warning is triggered.

[0137] Thermal runaway is a complex process involving multiple coupled factors. A single-dimensional anomaly may not be sufficient to trigger true thermal runaway, but the superposition of multiple-dimensional anomalies significantly increases the probability of thermal runaway. In existing technologies, thermal runaway early warning is typically based on a threshold judgment of a single dimension (such as temperature or voltage), or simply by logically superimposing multiple dimensions. This strategy has the following problems: first, the false negative rate of single-dimensional warnings is high because precursors to thermal runaway may alternate across different dimensions; second, the false positive rate of simple OR logic is high because anomalies in different dimensions may be caused by independent factors, not synergistically pointing to thermal runaway. Therefore, this embodiment introduces step S6, which sets differentiated thermal runaway risk warning triggering rules based on the combination relationship of different warning dimensions. For example... Figure 2 As shown, step S6 includes:

[0138] When the vehicle triggers a high charge / low discharge risk warning, a differential pressure warning, or a temperature abnormality warning, a battery thermal runaway risk warning will be triggered.

[0139] If the vehicle triggers a warning for water ingress or leakage insulation risk in the power battery, it will directly trigger a warning for battery thermal runaway risk.

[0140] If the vehicle triggers the battery overuse warning and temperature abnormality warning, it will trigger the battery thermal runaway risk warning.

[0141] Based on step S6, this embodiment achieves a comprehensive assessment of multi-dimensional risks. By logically combining the warning outputs from multiple dimensions, it identifies coupled risk scenarios that cannot be identified by a single dimension, thereby improving the accuracy and completeness of thermal runaway warnings.

[0142] This embodiment also provides a power battery thermal runaway detection system based on multi-source correlation analysis. Using the above-mentioned power battery thermal runaway detection method based on multi-source correlation analysis, the system includes:

[0143] The high charge / low discharge or low charge / high discharge warning detection module is used to determine whether the vehicle has triggered a high charge / low discharge or low charge / high discharge warning based on the vehicle status, charging status, and the collected individual unit number.

[0144] The power battery water ingress or leakage insulation risk warning detection module is used to determine whether the vehicle has triggered a power battery water ingress or leakage insulation risk warning based on the signals collected by the vehicle during charging.

[0145] The battery overuse warning detection module is used to determine whether the vehicle has triggered a battery overuse warning based on the vehicle status and collected historical data.

[0146] The differential pressure warning detection module is used to monitor the voltage difference of each individual battery cell and determine whether a differential pressure warning is triggered.

[0147] The temperature anomaly warning detection module is used to determine whether a temperature anomaly warning has been triggered based on the individual unit number output by the vehicle after the vehicle triggers a high charge / low charge risk warning or a low charge / low charge risk warning.

[0148] The battery thermal runaway risk warning and detection module is used to determine whether the vehicle has triggered a battery thermal runaway risk warning based on the current vehicle warning status and a preset warning strategy.

[0149] This invention provides a method and system for detecting thermal runaway in power batteries based on multi-source correlation analysis. By integrating multi-source data, a multi-dimensional early warning and detection system is constructed, enabling early identification and accurate warning of thermal runaway risks in power batteries. Specifically, this invention establishes an early warning mechanism from five dimensions: battery cell anomaly, insulation safety, operating conditions, differential pressure, and temperature anomaly. Finally, based on a preset comprehensive early warning strategy, the thermal runaway risk is determined, forming a comprehensive five-dimensional power battery thermal runaway safety protection system.

[0150] The present invention has been described above by way of example with reference to the accompanying drawings. Obviously, the specific implementation of the present invention is not limited to the above-described manner. Any non-substantial improvements made using the inventive concept and technical solution; or the direct application of the inventive concept and technical solution to other situations without modification, are all within the protection scope of the present invention.

Claims

1. A method for detecting thermal runaway of a power battery based on multi-source correlation analysis, characterized in that: The method includes: Step S1: Determine whether the vehicle has triggered a high-charge-low-release or low-charge-high-release warning based on the vehicle status, charging status, and the collected unit number. Step S2: Under charging conditions, determine whether the vehicle has triggered a warning for water ingress or leakage insulation risk in the power battery based on the signals collected by the vehicle. Step S3: Determine whether the vehicle has triggered a battery overuse warning based on the vehicle status and collected historical data; Step S4: Monitor the voltage difference of each individual battery cell and determine whether a differential voltage warning is triggered. Step S5: After the vehicle triggers a high charge / low charge risk warning or a low charge / low charge risk warning, determine whether a temperature abnormality warning is triggered based on the individual unit number output by the warning. Step S6: Based on the vehicle warning status in steps S1 to S5, determine whether the vehicle has triggered a battery thermal runaway risk warning based on a preset warning strategy.

2. The method for detecting thermal runaway of a power battery based on multi-source correlation analysis according to claim 1, characterized in that: Step S1 includes: under discharge conditions, periodically identifying and recording cells with continuously low voltage; under charging conditions, periodically identifying and recording cells with continuously high voltage and cells with continuously low voltage; based on the correlation and overlap of the above cells in the same charge-discharge cycle, determining whether the cell's internal resistance is abnormal or its capacity decay is abnormal, generating a corresponding tag and incrementing the count value by 1 for each abnormality determined; within a preset sliding time window, triggering a high-charge-low-discharge or low-charge-low-discharge risk warning based on whether the count value reaches a threshold.

3. The method for detecting thermal runaway of a power battery based on multi-source correlation analysis according to claim 2, characterized in that: Step S1 includes: In the same charge-discharge cycle, if there is the same cell number in the cell number set Idx_ConstDchMin and the cell number set Idx_ConstChgMax that is continuously low during discharge, the cell number corresponding to that cell number is determined to have an abnormal internal resistance. A charge-high discharge tag is generated and displayed on the user interface. At the same time, the charge-high discharge count value of that cell is incremented by one. Within a time window of the first time length, the charge-high discharge determination for the same cell number is counted only once. If the same cell number exists in the set of cell numbers with continuous low discharge (Idx_ConstDchMin) and the set of cell numbers with continuous low charge (Idx_ConstChgMin), then the cell number corresponding to that cell number is determined to have abnormal capacity decay. A low charge / low discharge tag is generated and displayed on the user interface. At the same time, the low charge / low discharge count value of that cell is incremented by one. Within a time window of the first time length, the low charge / low discharge determination for the same cell number is counted only once.

4. The method for detecting thermal runaway of a power battery based on multi-source correlation analysis according to claim 3, characterized in that: Step S1 includes: Establish a sliding time window of a second time length, and count the charge-high discharge count and charge-low discharge count of all individual cells within the sliding time window of the second time length. The second time length is longer than the first time length. If, within the sliding time window of the second time length, the count value of the same unit number being charged up and released down is greater than or equal to the preset value, and the unit number is displayed as a charge-up-release-down label on the user interface, then a charge-up-release-down risk warning is triggered, and the repeatedly appearing unit number is output. If, within the sliding time window of the second time length, the charge-low-release count value of the same unit number is greater than or equal to the preset value, and the unit number is displayed as a charge-low-release label on the user interface, a charge-low-release risk warning is triggered, and the recurring unit number is output.

5. The method for detecting thermal runaway of a power battery based on multi-source correlation analysis according to claim 1, characterized in that: Step S2 includes: Under charging conditions: (1) The insulation resistance value is less than the preset insulation resistance value threshold. (2) The vehicle is in the off state and the charging state is not charging; If both of the above conditions are met simultaneously and continue for a preset period of time, a warning of water ingress or leakage insulation risk in the power battery will be triggered.

6. The method for detecting thermal runaway of a power battery based on multi-source correlation analysis according to claim 1, characterized in that: Step S3 includes: when the total mileage of the vehicle or the data upload time difference reaches the corresponding threshold, counting the number of fast charging and slow charging within a preset period; if the number of fast charging exceeds the first preset threshold, or the number of fast charging exceeds the second preset threshold and the ratio of the number of fast charging to the total number of charging exceeds the preset ratio threshold, then triggering a battery overuse warning.

7. The method for detecting thermal runaway of a power battery based on multi-source correlation analysis according to claim 1, characterized in that: Step S4 includes: Acquire and preprocess the current operating data of the power battery. The data includes at least the highest single cell voltage, the lowest single cell voltage, the current, the state of charging, the current, and the time. Based on the preprocessed data, early warning judgments are made according to instantaneous pressure difference, average pressure difference over a time period, and pressure difference at the end of discharge or charging, in the discharge, charging, and resting states, respectively, combined with the battery type.

8. The method for detecting thermal runaway of a power battery based on multi-source correlation analysis according to claim 1, characterized in that: Step S5 includes: After triggering a high-charge-low-discharge or low-charge-low-discharge risk warning, all individual cell numbers from the warning output are acquired. A sampling period is defined as each change in the total SOC of the power battery within a preset value. Within each sampling period, the temperature and individual cell number of all individual cells are collected in real time. The individual cell numbers of the n cells with the highest temperatures at that time are acquired. If any of the individual cell numbers among the n cells with the highest temperatures is the same as any of the individual cell numbers in the warning output, the count value X of that same individual cell number is incremented by one. The initial value of X is 0. When X ≥ a preset threshold, the corresponding individual cell number is recorded as a continuously high-temperature individual cell number, and the continuously high-temperature event count value of the corresponding individual cell number is incremented by 1. Within the first time period, the same individual cell number is counted only once for a continuously high-temperature event. If, within the second time period, the continuously high-temperature event count value of any individual cell number is greater than or equal to the preset value, a temperature anomaly warning is triggered.

9. The method for detecting thermal runaway of a power battery based on multi-source correlation analysis according to claim 1, characterized in that: Step S6 includes: When the vehicle triggers a high charge / low discharge risk warning, a differential pressure warning, or a temperature abnormality warning, a battery thermal runaway risk warning will be triggered. If the vehicle triggers a warning for water ingress or leakage insulation risk in the power battery, it will directly trigger a warning for battery thermal runaway risk. If the vehicle triggers the battery overuse warning and temperature abnormality warning, it will trigger the battery thermal runaway risk warning.

10. A power battery thermal runaway detection system based on multi-source correlation analysis, using the power battery thermal runaway detection method based on multi-source correlation analysis according to any one of claims 1-9, characterized in that: The system includes: The high charge / low discharge or low charge / high discharge warning detection module is used to determine whether the vehicle has triggered a high charge / low discharge or low charge / high discharge warning based on the vehicle status, charging status, and the collected individual unit number. The power battery water ingress or leakage insulation risk warning detection module is used to determine whether the vehicle has triggered a power battery water ingress or leakage insulation risk warning based on the signals collected by the vehicle during charging. The battery overuse warning detection module is used to determine whether the vehicle has triggered a battery overuse warning based on the vehicle status and collected historical data. The differential pressure warning detection module is used to monitor the voltage difference of each individual battery cell and determine whether a differential pressure warning is triggered. The temperature anomaly warning detection module is used to determine whether a temperature anomaly warning has been triggered based on the individual unit number output by the vehicle after the vehicle triggers a high charge / low charge risk warning or a low charge / low charge risk warning. The battery thermal runaway risk warning and detection module is used to determine whether the vehicle has triggered a battery thermal runaway risk warning based on the current vehicle warning status and a preset warning strategy.

Images