Electric vehicle battery management system for thermal runaway monitoring

Abstract

The application relates to the technical field of electric vehicle battery management and safety monitoring, in particular to an electric vehicle battery management system for thermal runaway monitoring, which comprises the following modules: a data acquisition module, which is used for acquiring cell temperature, cell voltage, battery pack current, module gas concentration, gas sensor humidity, gas sensor temperature, cooling liquid temperature and cooling liquid flow; a data validity judgment module; an internal resistance reference and gas correction module, which is used for correcting the gas concentration; a monitoring parameter calculation module, which is used for calculating the temperature rise rate, the temperature outlier, the internal resistance relative change and the corrected gas concentration; a state judgment module, which is used for determining the abnormal source candidate cell and determining the normal, abnormal early warning, high-risk or sensor failure state; a thermal propagation time calculation module, which is used for calculating the target thermal propagation remaining time; a control module, which is used for outputting corresponding control instructions; and a motion feedback and state latching module, which is used for maintaining or releasing the state latching; and the application improves the system reliability.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle battery management and safety monitoring technology, specifically to an electric vehicle battery management system for thermal runaway monitoring. Background Technology

[0002] With the continuous development of electric vehicle technology, the complexity of thermal safety management of power battery systems has increased significantly, leading to a significant increase in the difficulty of comprehensive monitoring of multi-dimensional physical quantities and identification of thermal anomalies.

[0003] Currently, the status of battery cells is generally assessed by collecting absolute values ​​from a single sensor through the battery management system, using a fixed threshold for safety warnings, and recording the occurrence of each abnormal trigger. However, traditional methods of monitoring a single physical quantity rely on fixed references and isolated signals, making it difficult to distinguish between overall temperature rise caused by normal charging and discharging of the battery pack and abnormal temperature rise of local cells. Although these methods can provide basic temperature or voltage status information, they are usually less accurate and reliable, easily affected by sensor noise and fluctuations in operating conditions, leading to false alarms. Furthermore, relying solely on absolute temperature can result in a lag in anomaly detection.

[0004] Therefore, how to detect the risk of thermal runaway of power batteries in a timely and accurate manner and predict the evolution of thermal propagation has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] The purpose of this invention is to provide an electric vehicle battery management system for thermal runaway monitoring, avoiding false alarms and delays caused by noise from a single sensor and fluctuations in operating conditions, and more easily distinguishing between overall temperature rise and local anomalies, thereby enabling timely and accurate detection of thermal runaway risks and prediction of heat propagation evolution; the purpose of this invention can be achieved through the following technical solutions:

[0006] The electric vehicle battery management system for thermal runaway monitoring includes: a data acquisition module for acquiring cell temperature, cell voltage, battery pack current, module gas concentration, gas sensor humidity, gas sensor temperature, coolant temperature, and coolant flow rate;

[0007] The data validity assessment module is used to generate data validity indicators;

[0008] The internal resistance reference and gas correction module is used to determine the internal resistance reference value and correct the gas concentration.

[0009] The monitoring parameter calculation module is used to calculate the temperature rise rate, temperature outlier, relative change in internal resistance, and correction gas concentration.

[0010] The status judgment module is used to identify candidate cells for abnormal sources and determine the status as normal, abnormal warning, high risk, or sensor failure.

[0011] The heat propagation time calculation module is used to calculate the remaining heat propagation time of the target based on the abnormal source candidate cells determined by the state judgment module and the adjacent cells that have a physical adjacency relationship with the abnormal source candidate cells.

[0012] The control module includes a sampling communication control module, a thermal management control module, and an electrical isolation control module, which are used to output corresponding control commands;

[0013] The action feedback and state latch module is used to obtain the execution feedback of the control command and maintain or release the state latch based on the execution feedback.

[0014] Preferably, the data acquisition module sends raw data with sampling time and channel identifier to the data validity judgment module;

[0015] The data validity judgment module marks data as invalid for the same channel or the same cell if the range is out of bounds, open circuit, short circuit, absolute value of the difference with the previous valid sampling point is greater than the channel's preset jump threshold, the rate of change of adjacent sampling points is greater than the preset rate of change threshold, timestamp is discontinuous, communication verification error, or gas sensor has not completed preheating.

[0016] When the current step amplitude is less than the preset amplitude, a steady-state time window is formed before and after the step and the steady-state time window is less than the preset duration, or the current change rate within the steady-state time window is greater than the preset stability threshold, the data required for this internal resistance calculation will be marked as invalid.

[0017] Freeze the gas zero-point baseline under abnormal warning or high-risk conditions.

[0018] As a preferred embodiment, the monitoring parameter calculation module obtains valid temperature samples with valid data indicators, and filters the valid temperature samples to obtain valid filtered temperature samples. Three to seven valid filtered temperature samples covering a time window of 0.5s to 2s are selected, and linear fitting is performed according to the sampling time. The fitting slope is used as the temperature rise rate.

[0019] For any given cell as the cell to be evaluated, cells that have been latched as high-risk are excluded from the other cells in the module where the cell to be evaluated is located. The median of the effective filtered temperature samples of the remaining effective cells is used as the reference temperature, and the difference between the effective filtered temperature sample of the cell to be evaluated and the reference temperature is used as the temperature outlier.

[0020] When the number of remaining valid cells is less than 50% of the total number of cells in the module and less than 3, the temperature outlier will be marked as invalid.

[0021] Preferably, the monitoring parameter calculation module calculates the average cell voltage and average battery pack current within the steady-state time window before and after the current step, and according to... , Determine the relative change in internal resistance;

[0022] in, This represents the difference in average cell voltage within the steady-state time window. This represents the difference in average battery pack current within the steady-state time window. The values ​​are obtained by looking up a calibration table or by piecewise linear interpolation based on the state of charge, state of health, and cell temperature.

[0023] according to Determine the concentration of the correction gas, where, For the original converted gas concentration, As the zero-point baseline for the gas, This refers to the humidity correction coefficient obtained by looking up a table or by piecewise linear interpolation within the calibrated humidity and temperature ranges of the gas sensor. This refers to the temperature correction coefficient obtained by looking up a table or piecewise linear interpolation within the calibrated humidity and temperature ranges of the gas sensor.

[0024] Preferably, the state determination module uses one or more pre-selected thermal runaway characteristic gases as target thermal runaway characteristic gases. The target thermal runaway characteristic gases include at least one of carbon monoxide, carbon dioxide, hydrogen or hydrocarbon gases, and the gas threshold is determined from the corresponding gas calibration threshold table according to the type and concentration unit of the target thermal runaway characteristic gas.

[0025] When the rate of temperature rise is not less than 0.5℃ / s and the temperature outlier is not less than 5℃, evidence of strong thermal anomaly is confirmed.

[0026] When the rate of temperature rise is not less than 0.2℃ / s or the temperature outlier is not less than 3℃, evidence of weak thermal anomaly is confirmed.

[0027] When the relative change in internal resistance is no greater than -20% and the effective filtered temperature sample is no lower than 45℃, evidence of abnormal internal resistance is confirmed.

[0028] When the concentration of the corrected gas for the target thermal runaway characteristic gas is not lower than the corresponding gas threshold, evidence of module gas anomaly is determined.

[0029] As a preferred option, the data acquisition module also collects module pressure and pressure relief switch status;

[0030] The status judgment module obtains the target coolant flow rate from the thermal management control module and the measured coolant flow rate; the status judgment module determines an abnormal warning state when any of the following conditions are met: weak thermal abnormal evidence lasts for at least 0.5s;

[0031] The difference between the first occurrence of evidence of weak thermal abnormality and evidence of internal resistance abnormality shall not be greater than the preset correlation time, which is a fixed value within the range of 1s to 3s.

[0032] The evidence of abnormal module gas occurs simultaneously with the condition that the temperature rise rate is not less than 0.1℃ / s or the effective filtered temperature sample is not less than 40℃.

[0033] The temperature outliers of at least two cells with direct physical adjacency within the same module have a fitting slope greater than 0.1℃ / s within a trend time window of 1s to 10s.

[0034] Alternatively, the measured coolant flow rate is less than 80% of the target coolant flow rate, and the fitting slope of the average effective filtered temperature sample of the module calculated by the monitoring parameter calculation module based on the effective filtered temperature sample within the same module is greater than zero within 1 second.

[0035] A state is identified as high-risk when any of the following conditions are met: at least two valid filtered temperature samples are obtained consecutively as evidence of strong thermal anomalies.

[0036] The difference between the first occurrence of evidence of abnormal internal resistance and evidence of abnormal weak heat should not exceed the preset correlation time.

[0037] The effective filtered temperature sample reaches the emergency temperature threshold; the module pressure change rate is not lower than the preset pressure change rate threshold; the pressure relief switch is activated; or the remaining heat propagation time of the target is lower than the emergency time threshold.

[0038] When key temperature, internal resistance, gas, pressure, or communication data do not meet the validity conditions, the status judgment module will mark the corresponding feature as invalid and determine whether to enter the sensor failure state based on the number of invalid sensors and their location.

[0039] High-risk status takes precedence over sensor failure status, sensor failure status takes precedence over abnormal warning status, and a normal status is determined when the above conditions are not met.

[0040] Preferably, the state judgment module identifies cells that meet the criteria for strong or weak thermal anomaly evidence as candidate cells for anomaly sources. Based on the pre-stored physical adjacency table and the thermal coupling coefficient between each physical adjacency cell, combined with the difference between the effective filtered temperature samples of the candidate cells for anomaly sources and the physical adjacency cells, and the temperature rise rate of the adjacency cells, it determines the remaining heat propagation time of each physical adjacency cell. Specifically, when the temperature rise rate of the physical adjacency cell is less than zero, it is treated as zero. When the effective filtered temperature sample of the candidate cells for anomaly sources is not higher than the effective filtered temperature sample of the physical adjacency cells, the difference between the effective filtered temperature samples is treated as zero.

[0041] The heat propagation time calculation module determines the minimum value among the remaining heat propagation times of each physically adjacent cell as the target remaining heat propagation time.

[0042] Preferably, the sampling communication control module allocates a sampling frequency of 0.5Hz to 1Hz for the temperature and voltage of cells in normal state, a sampling frequency of 2Hz to 5Hz for cells in abnormal warning state, and a sampling frequency of 10Hz to 20Hz for cells in high-risk state and their first-order physical adjacent cells. The first-order physical adjacent cells are those cells that are 1 distance away from cells in high-risk state in the physical adjacency table.

[0043] Cells with a distance of 2 or greater in the diagram are considered distantly adjacent cells.

[0044] When communication capacity is insufficient, the reporting frequency of cells in normal state is reduced first, and then the reporting frequency of cells in abnormal warning state in the distant adjacent cells corresponding to high-risk cells is reduced. At the same time, the sampling frequency of high-risk cells and adjacent cells with the shortest remaining heat propagation time is kept not lower than 10Hz.

[0045] During the switching interval between the first sampling period before the data acquisition module switches to the second sampling period after the switch, the last valid filtered temperature sample, the corresponding sampling time, and the temperature rise rate are saved, and then... Calculate the predicted temperature, where, This is the last valid filtered temperature sample. This refers to the sampling time corresponding to the last valid filtered temperature sample. For the rate of temperature rise, This is the current forecast time; the forecast temperature is not used to lift abnormal warning or high-risk status.

[0046] Preferably, the thermal management control module is connected to the control terminal of the cooling pump and the cooling branch switching valve. In the abnormal warning state, the cooling pump is operated at 20% to 40% of the rated speed, and in the high-risk state, the cooling pump is operated at 80% to 100% of the rated speed, and the coolant is switched to the cooling branch corresponding to the module where the battery cell is located in the high-risk state.

[0047] The electrical isolation control module stops the balancing operation of the high-risk cells and the cells in the module where the high-risk cells are located, sends a current reduction request to the vehicle controller and stops charging. When the high-risk cells are located in an independently isolable branch, the corresponding branch contactor is disconnected first; otherwise, the main contactor is disconnected when the battery pack current is not higher than the contactor's allowable breaking current.

[0048] Within 1 to 5 seconds after the cooling control command is issued, if the action feedback and status latching module detects that the measured coolant flow rate is lower than 80% of the target coolant flow rate, the contactor auxiliary contact does not change accordingly, or the bus current does not drop below the preset safe current, it maintains the high-risk state and triggers electrical isolation, expands the high-frequency sampling range to the module where the high-risk cell is located, or issues a stop warning.

[0049] In a high-risk state, the state latch is released only after receiving an authorized reset command from the vehicle controller after it has completed power-down and confirmed that the bus current is lower than the preset safe current and the insulation resistance is not lower than the insulation threshold, and after all effective battery cell temperatures are lower than the recovery temperature of 35°C to 45°C and the correction gas concentration is lower than 50% of the corresponding gas threshold.

[0050] Compared with the prior art, the present invention has the following advantages:

[0051] 1. This invention collects multi-dimensional data and uses the median temperature of effective battery cells to calculate temperature outliers and temperature rise rate. The system can accurately distinguish between overall temperature rise caused by normal charging and discharging and local abnormal heating of battery cells, overcoming the serious lag in anomaly detection caused by the traditional solution that relies solely on absolute temperature.

[0052] 2. This invention combines temperature, relative changes in internal resistance, and correction gas concentration to form a hierarchical, multi-dimensional chain of abnormal evidence for comprehensive analysis, effectively avoiding false alarms caused by fluctuations in a single sensor and improving system reliability.

[0053] 3. Based on real physical adjacency relationships, current temperature difference, and thermal coupling coefficient, the system can dynamically estimate the remaining time of heat propagation of surrounding adjacent cells and lock the most urgent target; this mechanism transforms the spatial dimension of heat diffusion risk into a time indicator, enabling the prediction of the evolution of heat propagation. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort:

[0055] Figure 1 This is a schematic diagram of a battery management system for electric vehicle thermal runaway monitoring provided in an embodiment of this application. Detailed Implementation

[0056] To make the above features and advantages of the present invention more apparent and understandable, exemplary embodiments are described below in detail with reference to the accompanying drawings.

[0057] This embodiment provides a battery management system for electric vehicles used for thermal runaway monitoring, such as... Figure 1As shown, the system first asynchronously and concurrently collects data on cell temperature, cell voltage, battery pack current, module gas concentration, gas sensor humidity, gas sensor temperature, coolant temperature, and coolant flow rate. Based on the system master clock, it allocates multi-modal aligned sampling times and channel identifiers to eliminate the physical response hysteresis of different sensors.

[0058] Further, data validity indicators are formed by judging data validity, and the valid temperature data is filtered and feature calculated to obtain the temperature rise rate and temperature outlier.

[0059] Based on this, the status judgment module outputs normal, abnormal warning, high risk or sensor failure status, and the control module and action feedback and status latching module perform corresponding sampling, thermal management, electrical isolation and status holding operations.

[0060] Through the above processing, the system can distinguish between overall battery pack temperature rise and local cell abnormal temperature rise, and provide time-continuous data input for subsequent internal resistance, gas and heat propagation analysis;

[0061] Electric vehicle battery management systems for thermal runaway monitoring include:

[0062] The data acquisition module is used to collect data such as cell temperature, cell voltage, battery pack current, module gas concentration, gas sensor humidity, gas sensor temperature, coolant temperature, and coolant flow rate.

[0063] The data validity assessment module is used to generate data validity indicators;

[0064] The internal resistance reference and gas correction module is used to determine the internal resistance reference value and correct the gas concentration.

[0065] The monitoring parameter calculation module is used to calculate the temperature rise rate, temperature outlier, relative change in internal resistance, and correction gas concentration.

[0066] The status judgment module is used to identify candidate cells for abnormal sources and determine the status as normal, abnormal warning, high risk, or sensor failure.

[0067] The heat propagation time calculation module is used to calculate the remaining heat propagation time of the target based on the abnormal source candidate cells determined by the state judgment module and the adjacent cells that have a physical adjacency relationship with the abnormal source candidate cells.

[0068] The control module includes a sampling communication control module, a thermal management control module, and an electrical isolation control module, which are used to output corresponding control commands;

[0069] The action feedback and state latch module is used to obtain the execution feedback of the control command and maintain or release the state latch based on the execution feedback.

[0070] In a practical implementation, the data acquisition module can consist of a battery monitoring simulation front end, a current acquisition unit, a gas sensor, a humidity sensor, a temperature sensor, and a coolant flow sensor; the cell temperature acquisition channel is used to obtain the temperature of each cell or the corresponding casing; the cell voltage acquisition channel is used to obtain the terminal voltage of each cell; and the battery pack current acquisition channel is used to obtain the charging or discharging current of the battery pack; the module gas sensor is located near the module pressure relief path and is used to detect one or more of carbon monoxide, carbon dioxide, hydrogen, or hydrocarbon gases.

[0071] Among them, the data validity flag is used to indicate whether the corresponding sampled data can be used in subsequent calculations. For example, when the temperature channel is open, only the temperature data is marked as invalid, and the invalid data is not used as low temperature data. The internal resistance reference value is used to characterize the normal equivalent internal resistance of the cell under the current state of charge, health state and cell temperature conditions.

[0072] The calibration gas concentration is used to eliminate the influence of zero point, humidity and gas sensor temperature changes on gas detection results; the temperature rise rate is used to reflect how fast the cell temperature changes; and the temperature outlier is used to reflect the degree of local temperature deviation of the cell under evaluation relative to other effective cells in the same module.

[0073] When filtering the effective temperature data, the monitoring parameter calculation module processes the temperature samples of the same cell according to the channel identifier, and only sends the samples with valid identifiers and whose sampling time is continuously determinable to the filtering process; the filtering adopts a fixed-length median filtering window, and the window length is 3 or 5 valid samples;

[0074] When the number of valid samples in the window is less than half the window length, no new valid filtered temperature samples are output; when the number of valid samples reaches half the window length, the valid samples in the window are sorted by temperature value and the sample in the middle position is taken as the valid filtered temperature sample. When the window length is even, the arithmetic mean of the two middle samples is taken; the filtering window slides forward according to the sampling time. Invalid samples in the window are not replaced with zero values, nor are they included in the calculation of temperature rise rate and temperature outlier.

[0075] Jump threshold and rate of change threshold Based on the sensor's measurement resolution Maximum permissible rate of change Sampling frequency Confirmed, the specific calculation formula is as follows:

[0076]

[0077]

[0078] in, and This is a preset safety margin factor;

[0079] The stability threshold is set as the fundamental variance of the corresponding physical quantity under normal charge-discharge steady state. The fundamental variance is calculated by recording data within a normal charge-discharge steady-state time window of no less than a preset period at the calibration temperature.

[0080] Specifically, the effective variation range of each channel is first obtained under normal charging and discharging, static, low temperature and high temperature operating conditions, and then the boundary that exceeds the effective variation range and has abnormal line or communication characteristics is set as invalid judgment boundary.

[0081] For the temperature channel, the jump threshold and the rate of change threshold should be less than the maximum single sample change and the maximum continuous rate of change of the calibrated cell under normal operating conditions, and allowance should be reserved for sensor noise and sampling jitter; different channels should save corresponding thresholds, and the same threshold should not be directly used for temperature, voltage, gas concentration and flow channels.

[0082] This embodiment uses the coordinated changes of multiple channels such as temperature, internal resistance, and gas to make judgments, preventing the overall temperature rise caused by normal charging and discharging from being mistaken for thermal runaway; at the same time, it calculates the remaining heat propagation time through physical adjacency relationships, expanding the monitoring and control objects to adjacent battery cells; this system achieves local thermal anomaly identification based on multi-physical quantity acquisition, data validity screening, and state classification.

[0083] Furthermore, based on the above embodiments, the data acquisition module sends raw data with sampling time and channel identifier to the data validity judgment module; the sampling time is used to determine the sequential relationship and time interval between different data, and the channel identifier is used to determine the cell, module or sensor channel to which the data belongs. For example, different cell temperature channels in the same module have different channel identifiers.

[0084] In some possible implementations, the data acquisition module sends raw data with sampling time and channel identifier to the data validity judgment module;

[0085] The data validity judgment module marks data as invalid for the same channel or the same cell if the range is out of bounds, open circuit, short circuit, absolute value of the difference with the previous valid sampling point is greater than the channel's preset jump threshold, the rate of change of adjacent sampling points is greater than the preset rate of change threshold, timestamp is discontinuous, communication verification error, or gas sensor has not completed preheating.

[0086] When the current step amplitude is less than the preset amplitude, a steady-state time window is formed before and after the step and the steady-state time window is less than the preset duration, or the current change rate within the steady-state time window is greater than the preset stability threshold, the data required for this internal resistance calculation will be marked as invalid.

[0087] Freeze the gas zero-point baseline under abnormal warning or high-risk conditions;

[0088] The monitoring parameter calculation module obtains valid temperature samples with valid data indicators, and filters the valid temperature samples to obtain valid filtered temperature samples. It selects 3 to 7 valid filtered temperature samples covering a time window of 0.5s to 2s, performs linear fitting according to the sampling time, and uses the fitting slope as the temperature rise rate.

[0089] For any given cell as the cell to be evaluated, cells that have been latched as high-risk are excluded from the other cells in the module where the cell to be evaluated is located. The median of the effective filtered temperature samples of the remaining effective cells is used as the reference temperature, and the difference between the effective filtered temperature sample of the cell to be evaluated and the reference temperature is used as the temperature outlier.

[0090] When the number of remaining valid cells is less than 50% of the total number of cells in the module and less than 3, the temperature outlier will be marked as invalid.

[0091] In practical implementation, the data validity judgment module establishes a record of the previous valid sampling point for each channel; range out of bounds refers to the sampling value exceeding the measurement range specified by the sensor or analog front end; open circuit and short circuit are judged by channel diagnostic value, port level or acquisition result respectively; jump threshold is used to suppress abrupt data that does not conform to the physical change law of the measured object;

[0092] The rate of change between adjacent sampling points is determined based on the change in temperature, voltage, or gas concentration between adjacent valid sampling times and the sampling interval. The sampling interval is calculated from timestamped data, rather than being determined according to the message arrival time.

[0093] For internal resistance calculation, the data validity judgment module first confirms that there is a step change in the battery pack current that meets the requirements, and then confirms that the steady-state time window before and after the step change has a sufficient length, and that the rate of change of current within the steady-state time window does not exceed the stability threshold.

[0094] The steady-state time window is a continuous sampling interval used to reduce the impact of current fluctuations on voltage difference calculation. For example, when the current is still changing, the average voltage and average current of this interval are not used for internal resistance calculation. When any condition is not met, the current internal resistance result is simply marked as invalid, and no evidence of internal resistance anomaly is generated based on it.

[0095] The aforementioned preset amplitude is comprehensively calibrated based on the rated capacity of the battery pack, the typical current fluctuation characteristics under normal charging and discharging conditions, and the minimum resolvable accuracy of the current sensor, to ensure that the step current signal is sufficient to cause an accurately measurable voltage response and can be significantly distinguished from conventional electrical noise interference.

[0096] When the gas sensor has not completed preheating, the gas concentration data will not be used for the module's gas anomaly judgment. After the gas sensor enters an abnormal warning state or a high-risk state, the zero-point baseline will stop updating, thereby preventing the gas sensor from slowly absorbing the actual gas evolution process as a new zero point. The gas zero-point baseline can be slowly updated after the vehicle is powered on, when the battery pack is not under high current charging or discharging and the temperature is within a stable range. The specific update conditions are determined by the calibration parameters of the vehicle and the battery pack.

[0097] This embodiment uses multiple conditions for filtering, including range, line status, jump amplitude, rate of change, timestamp, communication verification, and sensor preheating status, to clearly distinguish between no data and normal data. For internal resistance calculation, it further confirms the results by using current step amplitude and steady-state condition to avoid internal resistance anomalies caused by current disturbances. This embodiment uses data validity flags to isolate abnormal sampling values ​​and internal resistance results that do not meet the calculation conditions.

[0098] In this embodiment, gas concentration correction can be achieved using zero-point, humidity, and temperature compensation parameters obtained from gas sensor calibration; for any module, the original converted gas concentration is assumed to be... The gas zero-point baseline is The humidity of the gas sensor is The gas sensor temperature is The correction result Determined according to the following formula:

[0099]

[0100] The above correction first deducts the zero-point baseline, and then compensates according to the current humidity and sensor temperature; the compensation coefficient is obtained within its respective calibration range by looking up a table or piecewise linear interpolation, and is matched with the gas type and sensor calibration parameters;

[0101] Maximum value calculation is used to avoid negative values ​​in the compensated gas concentration. For example, when the original concentration is below the zero baseline, the correction result is limited to zero. In abnormal warning or high-risk states, Maintain the baseline value before entering the corresponding state, and do not update the zero point using gas data during the abnormal period;

[0102] Furthermore, the monitoring parameter calculation module selects effective filtered temperature samples covering a preset time window from the effective temperature sampling data, and performs linear fitting based on the sampling time; the time window coverage ranges from 0.5s to 2s, and the number of effective filtered temperature samples participating in the fitting is 3 to 7;

[0103] Let the sampling time of the selected effective filtered temperature sample be... The effective filtered temperature sample is The sample size is Then the rate of temperature rise Determined by dividing the covariance of each sample's sampling time and the effective filtered temperature sample by the variance of the sampling time:

[0104]

[0105] in, and Arithmetic averages were calculated for the selected samples based on their sampling time and the effective filtered temperature samples. The temperature rise rate was calculated only when the number of effective samples was between 3 and 7 and the variance of the sampling time was greater than zero; otherwise, the temperature rise rate was marked as invalid. The actual sampling time was used as the time reference for fitting to reduce the impact of sampling jitter and noise from individual temperature sampling on the results.

[0106] Compared with the method of simply subtracting two adjacent temperature points, linear fitting can reduce the impact of single temperature sampling noise on the temperature rise rate and can adapt to the situation where there is slight jitter in the sampling time; when switching between different sampling frequencies, the fitting is still based on the actual sampling time, so there will be no obvious rate deviation due to the assumption that the fixed sampling period is not valid.

[0107] Temperature outliers are used to reflect the degree of localized temperature rise of the cell under evaluation relative to the effective cell group in the same module; for the cell under evaluation Cells that have been latched as high-risk are excluded from the other cells in the same module, and the median of the effective filtered temperature samples of the remaining valid cells is used as the reference temperature. The temperature outlier was determined by subtracting the reference temperature from the effective filtering temperature sample of the cell to be evaluated.

[0108]

[0109] in, Indicates the battery cell to be evaluated Temperature outlier Indicates the battery cell to be evaluated Effective filtered temperature samples, Indicates reference temperature;

[0110] During the calculation, effective filtered temperature samples are first screened from other cells in the same module, and cells that have been latched as high-risk are excluded. The remaining samples are arranged in ascending order of temperature and the median value is taken to determine the reference level. When the number of remaining samples is even, the arithmetic mean of the two middle temperatures is taken. When the outlier is positive, it indicates that the temperature of the cell to be evaluated is higher than the reference level.

[0111] Using the median as the reference temperature can reduce the effect of one or more abnormally high-temperature cells on the reference value. For example, if there is already a high-temperature cell in the same module, that cell will not significantly change the reference level composed of the remaining cells. When the number of remaining valid cells is less than 50% of the total number of cells in the module and less than 3, no valid temperature outlier is generated. Instead, the temperature outlier is marked as invalid.

[0112] This embodiment obtains the local temperature deviation of a single cell relative to its surrounding cells within the same module by excluding cells in high-risk states that have been latched and using the median; it obtains the continuous temperature rise trend through linear fitting with sampling time; this embodiment characterizes the cell temperature rise trend and local temperature difference based on the temperature rise rate and temperature outlier, respectively, providing objective temperature characteristics for state classification.

[0113] Based on the aforementioned data validity judgment and temperature characteristic calculation, this embodiment performs internal resistance calculation on the effective current step data, corrects the gas sensor output, and combines and judges the abnormal evidence formed by temperature, internal resistance, and gas concentration.

[0114] After an abnormal source candidate cell is found, the remaining heat propagation time of adjacent cells is further estimated according to the physical adjacency relationship and thermal coupling coefficient between cells, so as to determine the most urgent adjacent cell and the target remaining heat propagation time.

[0115] In some possible implementations, the monitoring parameter calculation module calculates the average cell voltage and average battery pack current within the steady-state time window before and after the current step, and according to... , Determine the relative change in internal resistance;

[0116] in, This represents the difference in average cell voltage within the steady-state time window. This represents the difference in average battery pack current within the steady-state time window. The values ​​are obtained by looking up a calibration table or by piecewise linear interpolation based on the state of charge, state of health, and cell temperature.

[0117] according to Determine the concentration of the correction gas, where, For the original converted gas concentration, As the zero-point baseline for the gas, This refers to the humidity correction coefficient obtained by looking up a table or by piecewise linear interpolation within the calibrated humidity and temperature ranges of the gas sensor. This refers to the temperature correction coefficient obtained by looking up a table or piecewise linear interpolation within the calibrated humidity and temperature ranges of the gas sensor.

[0118] In practical implementation, the monitoring parameter calculation module extracts cell voltage and battery pack current samples from the steady-state time window before the current step and the steady-state time window after the current step, and calculates the corresponding average values.

[0119] Let the average cell voltage before the step be The average cell voltage after the step jump is The average battery pack current before the step jump is The average battery pack current after the step jump is Then determine the voltage difference. Current difference and equivalent DC internal resistance ;

[0120] During calculation, the average values ​​before and after the step are obtained from the effective samples within the corresponding steady-state time window, and the voltage sampling position and current definition are kept consistent. When the absolute value of the current difference before and after the step does not reach the preset amplitude, or the number of effective samples is insufficient, the internal resistance result is marked as invalid.

[0121] When the battery pack adopts a series cell structure and the cell branch current is consistent with the battery pack current, the battery pack current can be directly used for the above internal resistance calculation; when the battery pack includes parallel branches, the battery pack current is converted into the corresponding branch current according to the current distribution relationship of the parallel branch where the tested cell is located to form the average battery pack current. The current distribution relationship is determined by the battery pack topology or calibration parameters.

[0122] Internal resistance reference value With equivalent DC internal resistance By using the same current definition, measurement location and voltage sampling standard, the relative change in internal resistance reflects the change in operating conditions under the same current caliber, and does not misjudge the difference in the number of parallel branches or the current distribution as an abnormality in the cell internal resistance.

[0123] The absolute value is used to eliminate the influence of different current direction definitions during charging and discharging on the sign of the internal resistance result; the internal resistance calculation is only valid when the current step amplitude, the steady-state time window before and after the step, and the current stability within the steady-state time window all meet the requirements; if any condition is not met, and All results were marked as invalid and will not be used as evidence of abnormal internal resistance.

[0124] The relative change in internal resistance is as described above. The relationship is determined; to avoid ratio distortion caused by abnormal reference values, when the internal resistance reference value is... When calibration is incomplete, data is invalid, or the internal resistance is less than the preset minimum reference internal resistance, the relative change in internal resistance during this test is as follows: Marked as invalid; It can be obtained by looking up the calibration table, or by using piecewise linear interpolation between adjacent calibration points;

[0125] Preset minimum reference internal resistance The average DC internal resistance of the new battery cells measured at the optimal operating temperature. and its standard deviation Confirmed, the formula is:

[0126]

[0127] in, This is the error compensation amount for the sensor system, used to avoid the risk of division by zero or distortion of calculation results caused by abnormal data in the calibration table;

[0128] This embodiment determines the internal resistance reference value based on the state of charge, health status, and cell temperature, so that the change in internal resistance matches the current cell operating condition; at the same time, the current and voltage are averaged through a steady-state time window to reduce the impact of current fluctuations and sampling noise on the internal resistance calculation; this embodiment obtains a comparable relative change in internal resistance based on the internal resistance benchmark related to the operating condition, avoiding insufficient sensitivity or false alarms caused by a fixed benchmark.

[0129] Gas concentration as described above The relationship is corrected; during correction, the original concentration is first converted based on the gas sensor output and the zero baseline is subtracted, and then the corresponding compensation coefficient is obtained by referring to a table or interpolating based on the current humidity and gas sensor temperature.

[0130] When the humidity or gas sensor temperature exceeds the calibration range, the gas concentration data is marked as having limited validity and is not used as an unconditional basis for high-risk judgment; when the gas sensor has not completed preheating, communication verification is incorrect, or the zero-point baseline is invalid, the calibrated gas concentration is marked as invalid; the calibration result is only used in the module's gas anomaly judgment when the original concentration, humidity, gas sensor temperature, and zero-point baseline all have corresponding valid marks.

[0131] Furthermore, the status judgment module generates different types of abnormal evidence based on the target thermal runaway characteristic gas, temperature rise rate, temperature outlier, and relative change in internal resistance.

[0132] In some possible implementations, the state determination module uses one or more pre-selected thermal runaway characteristic gases as target thermal runaway characteristic gases. The target thermal runaway characteristic gases include at least one of carbon monoxide, carbon dioxide, hydrogen or hydrocarbon gases, and determines the gas threshold from the corresponding gas calibration threshold table according to the type and concentration unit of the target thermal runaway characteristic gas.

[0133] When the rate of temperature rise is not less than 0.5℃ / s and the temperature outlier is not less than 5℃, evidence of strong thermal anomaly is confirmed.

[0134] When the rate of temperature rise is not less than 0.2℃ / s or the temperature outlier is not less than 3℃, evidence of weak thermal anomaly is confirmed.

[0135] When the relative change in internal resistance is no greater than -20% and the effective filtered temperature sample is no lower than 45℃, evidence of abnormal internal resistance is confirmed.

[0136] When the concentration of the corrected gas for the target thermal runaway characteristic gas is not lower than the corresponding gas threshold, evidence of module gas anomaly is determined.

[0137] In practical implementation, the gas calibration threshold table is distinguished at least according to gas type and concentration unit; gas type is used to determine the correspondence between the output of different gas sensors and thermal runaway gas evolution, and concentration unit is used to avoid directly comparing concentration values ​​under different dimensions or different calibration ranges.

[0138] The gas threshold is determined based on the gas sensor selectivity, module volume, ventilation conditions, and gas evolution conditions. The maximum gas reading under normal charging and discharging and changes in ambient temperature and humidity is used as the upper limit of the normal range. A safety margin is set between this upper limit and the concentration range that can characterize abnormal gas evolution.

[0139] For different gases, the gas type, concentration unit, sensor range, applicable temperature range and corresponding threshold are stored separately. When the gas concentration units are inconsistent, they are first converted according to the corresponding calibration relationship before the threshold comparison is performed.

[0140] Strong thermal anomaly evidence is used to indicate that the cell simultaneously meets the preset temperature rise rate threshold and temperature outlier threshold; weak thermal anomaly evidence is used to indicate that the cell has shown a temperature rise trend or local temperature deviation, but has not yet met both conditions of strong thermal anomaly evidence.

[0141] Internal resistance anomaly evidence is used to indicate that when the effective filtered temperature sample reaches the corresponding condition, the decrease in the equivalent DC internal resistance of the cell relative to the current operating condition reference value reaches the preset threshold for relative change in internal resistance; module gas anomaly evidence is used to indicate that the concentration of the calibration gas within the module range reaches the calibration threshold of the corresponding gas; the above evidence is only generated when its corresponding input data is valid; when the corresponding input is invalid, the corresponding anomaly evidence will not be generated with the invalid input.

[0142] Among them, the temperature rise rate threshold and the temperature outlier threshold have different physical meanings: the temperature rise rate threshold is used to determine whether the abnormal temperature rise process has a sufficiently fast change trend, and the temperature outlier threshold is used to determine whether the cell is significantly different from other cells in the module; combining the two for strong heat abnormal evidence can reduce false alarms caused by the synchronous temperature rise of the entire module during fast charging or high power discharge.

[0143] The threshold values ​​for temperature rise rate, temperature outlier, relative change in internal resistance, and effective filtered temperature sample are determined based on the cell material system, sensor installation location, sampling error, and normal operation data. First, the upper limit of normal fluctuations for the corresponding characteristics is determined using normal operation data. Then, the abnormal threshold value is determined by combining the response margin required for thermal safety determination. The threshold values ​​are then written into the calibration parameter table corresponding to the cell model and sensor arrangement.

[0144] This embodiment integrates strong thermal anomalies, weak thermal anomalies, internal resistance anomalies, and module gas anomalies across channels, ensuring that short-term fluctuations from a single sensor are not directly equivalent to a high-risk state, and uses gas signals to supplement temperature signals. This embodiment forms layered anomaly evidence based on the synergistic changes between different physical quantities, avoiding the risk of a single signal being overwhelmed by noise or detection lag caused by relying solely on absolute temperature.

[0145] Furthermore, the state judgment module determines the cell state by combining the duration of abnormal evidence, the time of its first appearance, and the module-level signal; for the difference in the time of the first appearance of different abnormal evidence, the absolute time difference between the two times of the first appearance is compared to avoid ambiguity caused by different order of appearance of evidence.

[0146] In some possible implementations, the data acquisition module also acquires module pressure and pressure relief switch status;

[0147] The status judgment module obtains the target coolant flow rate from the thermal management control module and the measured coolant flow rate; the status judgment module determines an abnormal warning state when any of the following conditions are met: weak thermal abnormal evidence lasts for at least 0.5s;

[0148] The difference between the first occurrence of evidence of weak thermal abnormality and evidence of internal resistance abnormality shall not be greater than the preset correlation time, which is a fixed value within the range of 1s to 3s.

[0149] The evidence of abnormal module gas occurs simultaneously with the condition that the temperature rise rate is not less than 0.1℃ / s or the effective filtered temperature sample is not less than 40℃.

[0150] The temperature outliers of at least two cells with direct physical adjacency within the same module have a fitting slope greater than 0.1℃ / s within a trend time window of 1s to 10s.

[0151] Alternatively, the measured coolant flow rate is less than 80% of the target coolant flow rate, and the fitting slope of the average effective filtered temperature sample of the module calculated by the monitoring parameter calculation module based on the effective filtered temperature sample within the same module is greater than zero within 1 second.

[0152] A state is considered high-risk if any of the following conditions are met: at least two consecutive valid filtered temperature samples are obtained as evidence of strong thermal anomalies.

[0153] The difference between the first occurrence of evidence of abnormal internal resistance and evidence of abnormal weak heat should not exceed the preset correlation time.

[0154] The effective filtered temperature sample reaches the emergency temperature threshold; the module pressure change rate is not lower than the preset pressure change rate threshold; the pressure relief switch is activated; or the remaining heat propagation time of the target is lower than the emergency time threshold.

[0155] When key temperature, internal resistance, gas, pressure, or communication data do not meet the validity conditions, the status judgment module will mark the corresponding feature as invalid and determine whether to enter the sensor failure state based on the number of invalid sensors and their location.

[0156] High-risk status takes precedence over sensor failure status, sensor failure status takes precedence over abnormal warning status, and a normal status is determined when the above conditions are not met.

[0157] In the specific implementation, the state judgment module saves the first occurrence time, current duration and valid sample flag for each abnormal evidence; the preset correlation time is used to confirm whether different physical quantities are in the same abnormal development interval, with a value range of 1s to 3s; the trend time window is used to determine whether multiple adjacent cells simultaneously show an upward trend in temperature outliers, with a value range of 1s to 10s.

[0158] The fitting slope within the trend time window is calculated using the outlier of each cell temperature and the corresponding sampling time; the difference between the first occurrence times of different abnormal evidence is calculated according to the absolute time difference, and the two are considered to be correlated only when both pieces of evidence are valid and their absolute time difference is not greater than the preset correlation time.

[0159] The simultaneous occurrence of evidence of abnormal module gas and the conditions that the effective temperature rise rate is not less than 0.1℃ / s or the effective filtered temperature sample is not less than 40℃ means that the evidence of abnormal module gas is established, and at least one of the following conditions is met within the same associated time range: the effective temperature rise rate reaches 0.1℃ / s or the effective filtered temperature sample reaches 40℃. When the temperature rise rate or the effective filtered temperature sample data is invalid, the corresponding conditions are not met by the invalid data.

[0160] The abnormal warning status is mainly used to handle situations where abnormal signs have appeared but have not yet met the high-risk conditions. For example, if the temperature rise rate of a single cell reaches the weak thermal abnormality evidence threshold and continues to meet the specified conditions, or if the temperature outliers of two directly adjacent cells in the same module rise simultaneously, the abnormal warning status can be entered. When insufficient coolant flow and continued rise in the average module temperature occur simultaneously, the abnormal warning status can also be entered to reflect the situation where the cooling system has not achieved the expected effect.

[0161] High-risk status is used to handle situations that meet the preset thermal runaway trigger conditions or where a thermal event has already occurred; when at least two valid samples are obtained consecutively as evidence of strong thermal anomalies, it indicates that the rapid temperature rise and local temperature deviation are not single sampling noise; when the effective filtered temperature sample reaches the emergency temperature threshold, the module pressure change rate reaches the pressure change rate threshold, or the pressure relief switch is activated, the high-risk status can be directly determined without relying on whether the temperature outlier is valid.

[0162] The emergency temperature threshold is determined based on the cell material system, sensor installation location, and thermal inertia calibration. It should be set below the verified self-heating acceleration temperature and allow for a safety margin for sensor response hysteresis.

[0163] The duration, correlation time, trend time window, temperature rise rate, temperature and flow thresholds used for abnormal early warning, as well as the pressure change rate, emergency temperature and heat propagation time thresholds used for high-risk conditions, are all calibrated according to the cell model, module structure, cooling circuit capacity and sensor response characteristics.

[0164] During calibration, the false trigger boundary is determined by normal operating data, the safety margin is determined by the shortest response time required for thermal safety judgment, and the thresholds corresponding to different module structures are stored separately, without replacing the thresholds of all modules with the thresholds of a single module.

[0165] When key temperature, internal resistance, gas, pressure, or communication data do not meet the validity conditions, the status judgment module marks the corresponding feature as invalid and determines whether to enter the sensor failure state based on the number of invalid sensors and their location. The sensor failure state is not the same as the normal state. The status priority is arranged as high-risk state, sensor failure state, abnormal warning state, and normal state. When both the high-risk state and the sensor failure state are met, the high-risk state is output.

[0166] This embodiment uses a combination of evidence duration, time difference of first occurrence, and adjacent cell trends for confirmation, which suppresses single-event noise and reduces waiting time for rapid thermal events. At the same time, sensor failure is set as a separate state to avoid erroneous output of normal state when data is missing. This embodiment uses multi-condition priority judgment to realize the graded output of abnormal warning and high-risk state, so as to avoid sensor failure from masking the real risk.

[0167] Furthermore, the heat propagation time calculation module estimates the remaining time required for the adjacent battery cell to reach the predetermined propagation treatment temperature based on the temperature difference between the candidate battery cell and its adjacent battery cell, the current temperature rise rate of the adjacent battery cell, and the thermal coupling coefficient.

[0168] In the specific calculation, candidate cells for abnormal sources are first read from the physical adjacency list. Adjacent cells The adjacency relationship and thermal coupling coefficient; then from the adjacent cells The effective filtered temperature sample is used to obtain its own temperature rise rate, and the self-temperature rise rate less than zero is treated as zero; only abnormal source candidate cells are processed. The effective filtering temperature sample is higher than that of the adjacent cell. Calculate the temperature difference heat transfer term in the case of anomaly source candidate cells. The effective filtering temperature sample is not higher than that of the adjacent cell. This item is not included in the time limit;

[0169] The temperature difference heat transfer term is based on the thermal coupling coefficient and the candidate cells of the abnormal source. and adjacent cells The positive temperature difference between them is jointly determined, and the unit of the thermal coupling coefficient matches the temperature difference and time unit; when different physical adjacency relationships, coolant flow rates or temperature ranges correspond to different thermal coupling coefficients, they are selected from the thermal coupling coefficient calibration table according to the relationship category in the physical adjacency table and the current cooling condition, or piecewise linear interpolation is performed between adjacent calibration points.

[0170] Adjacent cells The predicted temperature rise rate is obtained by adding its own temperature rise rate to the temperature difference heat transfer term; when any required data is invalid, the candidate cell for the abnormal source is... Adjacent cells The propagation time is marked as unreliable and no deterministic time value is output.

[0171] In some possible implementations, the state judgment module identifies cells that meet the strong thermal anomaly evidence or weak thermal anomaly evidence as candidate cells for anomaly sources, and determines the remaining heat propagation time of each physical adjacent cell based on the pre-stored physical adjacency table and the thermal coupling coefficient between each physical adjacent cell, combined with the difference between the effective filtered temperature samples of the candidate cells for anomaly sources and the physical adjacent cells and the temperature rise rate of the physical adjacent cells.

[0172] Among them, when the temperature rise rate of the physically adjacent cells is less than zero, it is treated as zero; when the effective filtered temperature sample of the candidate cell of the abnormal source is not higher than the effective filtered temperature sample of the physically adjacent cells, the difference of the effective filtered temperature samples is treated as zero.

[0173] The heat propagation time calculation module determines the minimum value among the remaining heat propagation times of each physically adjacent cell as the target remaining heat propagation time.

[0174] In practical implementation, the candidate cells for anomaly sources can be cells that satisfy strong thermal anomaly evidence or weak thermal anomaly evidence; the physical adjacency table is used to describe the adjacency relationships formed by the actual arrangement of cells, thermal insulation structure, cold plate structure and module boundary, rather than determining the adjacency relationship solely based on the electrical series sequence of cells;

[0175] The thermal coupling coefficient is used to reflect the degree of influence of the temperature change of the candidate cell of the abnormal source on the temperature rise of the adjacent cell. It can be determined by the temperature response data obtained under controlled thermal pulse conditions and classified and stored according to horizontal adjacent, vertical adjacent, adjacent across the thermal insulation structure, or edge position.

[0176] When calibrating the thermal coupling coefficient, for each type of physical adjacency relationship, the temperature difference between the candidate cell of the abnormal source and the adjacent cell, the initial temperature rise rate of the adjacent cell and the coolant flow rate are recorded, and the corresponding temperature response relationship is obtained under the same temperature range and cooling conditions. During operation, coefficients that match the current adjacency relationship, temperature range and cooling conditions are selected first. If the matching coefficient is missing, the deterministic propagation time is not generated.

[0177] To estimate the temperature rise rate of adjacent cells after being affected by a candidate cell from an abnormal source, let the candidate cell from the abnormal source be... The adjacent battery cell is ;

[0178] Adjacent cells The predicted temperature rise rate is determined by the adjacent cells. The temperature rise rate obtained by fitting its own effective temperature sample and the candidate cells for the anomaly source and adjacent cells The positive temperature difference heat transfer term between them is jointly determined; among them, the value is zero when the self-temperature rise rate is less than zero, and when the abnormal source candidate cell The effective filtering temperature sample is not higher than that of the adjacent cell. At that time, the value of the heat transfer term due to temperature difference is zero;

[0179] Forward temperature difference heat transfer term according to the candidate cells of the abnormal source Effective filtering temperature sample and adjacent cell The effective filtered temperature difference is used to determine the adjacent cells, and the thermal coupling coefficient under the corresponding adjacency relationship, current temperature range, and cooling conditions is converted into the adjacent cell value. The rate of temperature rise;

[0180] in, This represents the rate of temperature rise of the battery cell relative to the sampling time. Indicates candidate cells for abnormal sources For adjacent cells Predicted temperature rise rate when it has an effect; Predicted temperature rise rate The calculation formula is:

[0181]

[0182] in, Indicates adjacent battery cells The self-temperature rise rate is obtained by fitting the effective temperature sample. Indicates battery cell With battery cells The thermal coupling coefficient between them Indicates candidate cells for abnormal sources Effective filtered temperature samples, Indicates adjacent battery cells Effective filtered temperature samples;

[0183] Only in adjacent cells Temperature samples are used when they meet data validity requirements and cover the specified calculation time window; when , or If any one of these conditions is invalid, the candidate cell for the anomaly source will not be processed. Adjacent cells The propagation time generates a deterministic result, and the combination is marked as having a propagation time that cannot be reliably estimated;

[0184] Instead of using fixed dimensionless weights, the positive temperature difference is converted into a calibration parameter for the temperature rise rate of adjacent cells. The value of this parameter corresponds to the cell spacing, insulation structure, cold plate position, coolant flow rate, and temperature range. When the above influencing conditions change, the coefficients under the corresponding conditions are called up, or piecewise linear interpolation is performed between adjacent calibration points.

[0185] The thermal coupling coefficient is determined by temperature response data under controlled heating conditions: for the battery cell Apply controlled heating power and record adjacent cells The initial temperature rise rate and the temperature difference between the two cells were obtained, and then the results were obtained by fitting multiple sets of temperature response data. ;

[0186] Different cell spacing, insulation structure, cold plate position, and coolant flow conditions can correspond to different thermal coupling coefficients; multiple sets of temperature response data should at least cover adjacent cells. Low temperature, medium temperature and high temperature ranges, as well as low flow rate and rated flow rate conditions of the cooling circuit;

[0187] For each set of data, data with invalid sensors, non-positive temperature differences, or unstable temperature rise rates of adjacent cells are removed. Then, a corresponding calibration table is established based on the retained temperature differences and temperature rise responses. If the calibration table does not cover the current temperature range or cooling conditions, the heat propagation time is marked as unreliable and is not directly replaced by coefficients from other conditions.

[0188] Let the predetermined transmission and disposal temperature be Then the adjacent battery cell Compared to candidate cells with abnormal sources The remaining heat propagation time depends on the adjacent cells. Temperature difference from the predetermined propagation and treatment temperature, adjacent battery cells The predicted temperature rise rate and the preset minimum calculation speed are determined;

[0189] When the predicted temperature rise rate is zero or lower than the preset minimum calculation speed, the preset minimum calculation speed is used for time estimation to prevent the calculation results from being unbounded; when the current temperature of the adjacent cell has reached or exceeded the predetermined heat propagation treatment temperature, the remaining heat propagation time of the adjacent cell is regarded as reaching the emergency treatment condition.

[0190] Specifically, first calculate the remaining temperature difference between the predetermined propagation processing temperature and the current effective filtering temperature sample of the adjacent cell; then divide the remaining temperature difference by the larger of the predicted temperature rise rate of the adjacent cell and the preset minimum calculation speed to obtain the remaining propagation time of the adjacent cell.

[0191] The calculation is only applicable when the residual temperature difference is positive and the relevant temperature, temperature rise rate, and thermal coupling coefficient are all valid; the preset minimum calculation speed is determined based on the temperature sampling resolution, temperature rise rate calculation cycle, and the maximum allowable prediction time, and the upper limit prediction time is set in the controller; when the time calculated according to the preset minimum calculation speed exceeds the upper limit prediction time, the upper limit prediction time is output and the prediction time limitation flag is retained.

[0192] in, This is the predetermined propagation handling temperature, which is lower than the cell's thermal runaway initiation temperature, used to trigger protection before thermal propagation reaches a higher risk; a preset minimum calculation speed. This is the minimum speed set to prevent the denominator of the propagation time calculation from being zero or too small. The example value is 0.01℃ / s to 0.05℃ / s.

[0193] The emergency temperature threshold in the status judgment represents the propagation handling temperature and the cell emergency temperature judgment benchmark, respectively. The two are calibrated independently according to different control purposes, and one is not directly replaced by the other.

[0194] The emergency time threshold is determined based on the cell material system, module thermal diffusion characteristics, cooling system response time, and electrical isolation time, and is lower than the thermal runaway initiation temperature; the emergency time threshold is determined based on the shortest response time of cooling control, sampling resource adjustment, and electrical isolation actions; both are written into the propagation and handling parameter table and the state judgment parameter table, respectively.

[0195] When the effective temperature or temperature rise rate of an adjacent cell is invalid, the adjacent cell is not used to generate a deterministic minimum remaining time. Instead, it is marked as a cell whose propagation time cannot be reliably estimated, and high-frequency sampling and sensor failure monitoring of the adjacent cell are maintained.

[0196] For multiple physically adjacent cells of the same candidate cell for anomaly source, calculate the remaining heat propagation time for each cell and take the minimum value as the target remaining heat propagation time for that candidate cell. For invalid adjacent cells, only retain the flag that the adjacent cells cannot be reliably estimated, and do not use them as the minimum value for comparison.

[0197] When multiple candidate cells for abnormal sources exist simultaneously, the remaining time is calculated for each candidate cell and its physical adjacent cells, and the minimum value among all valid combinations of candidate cells and adjacent cells is taken as the remaining time for the target heat propagation of the system.

[0198] When no valid adjacent cells are available, a deterministic propagation remainder time is not generated, and corresponding protective actions are still performed according to the current state of the candidate cells of the abnormal source; this result is used to subsequently determine the sampling hold range, cooling intensity, and electrical isolation timing.

[0199] This embodiment combines the spatial adjacency relationship of battery cells, thermal coupling strength, current temperature difference, and the temperature rise trend of adjacent battery cells to capture the risk of abnormal battery cells spreading to surrounding battery cells; by taking the minimum value among the remaining time of multiple adjacent battery cells, the most urgent propagation path is transformed into a quantitative result that can be used for control decisions; this embodiment determines the remaining time of target heat propagation based on thermal coupling calculation, avoiding fixed monitoring range, coarse cooling objects, and lack of time basis for isolation actions.

[0200] In this embodiment, based on the determined cell status and the remaining time of target heat propagation, the sampling and communication control module allocates sampling and communication resources according to the status level, physical adjacency distance, and propagation urgency; in the abnormal warning state, the cooling system enters the pre-circulation state, and in the high-risk state, the cooling capacity is increased and electrical isolation is performed;

[0201] At the same time, the action feedback and status latching module checks the cooling flow, contactor auxiliary contacts and bus current to prevent the high-risk state from being mistakenly released after the control command is issued because the actuator has not reached the target.

[0202] In some possible implementations, the sampling communication control module allocates a sampling frequency of 0.5Hz to 1Hz for the temperature and voltage of cells in normal state, a sampling frequency of 2Hz to 5Hz for cells in abnormal warning state, and a sampling frequency of 10Hz to 20Hz for cells in high-risk state and their first-order physical adjacent cells. The first-order physical adjacent cells are those cells that are 1 distance away from cells in high-risk state in the physical adjacency table.

[0203] Cells with a distance of 2 or greater in the diagram are considered distantly adjacent cells.

[0204] When communication capacity is insufficient, the reporting frequency of cells in normal state is reduced first, and then the reporting frequency of cells in abnormal warning state in the distant adjacent cells corresponding to high-risk cells is reduced. At the same time, the sampling frequency of high-risk cells and adjacent cells with the shortest remaining heat propagation time is kept not lower than 10Hz.

[0205] During the switching interval between the first sampling period before the data acquisition module switches to the second sampling period after the switch, the last valid filtered temperature sample, the corresponding sampling time, and the temperature rise rate are saved, and then... Calculate the predicted temperature, where, This is the last valid filtered temperature sample. This refers to the sampling time corresponding to the last valid filtered temperature sample. For the rate of temperature rise, This is the current forecast time; the forecast temperature is not used to lift abnormal warning or high-risk status.

[0206] In practical implementation, the sampling communication control module establishes state level and graph distance information for each cell; a first-order physical adjacency cell refers to a cell that can be reached from a high-risk cell in the physical adjacency table through only one adjacency edge;

[0207] A cell with a graph distance greater than 2 refers to a cell that requires more than two adjacent edges to reach from a cell in a high-risk state. The graph distance is calculated based on the physical adjacency list, and therefore reflects the proximity in terms of heat propagation, rather than the numbering order of the voltage acquisition channels.

[0208] The sampling frequency of temperature and voltage of the battery cell under normal conditions can be configured in the range of 0.5Hz to 1Hz; the sampling frequency of the battery cell under abnormal warning conditions can be configured in the range of 2Hz to 5Hz; the sampling frequency of the battery cell under high risk conditions and its first-order physical adjacent battery cells can be configured in the range of 10Hz to 20Hz.

[0209] Even if the adjacent cell with the shortest remaining heat propagation time has not yet entered a high-risk state, a sampling frequency of no less than 10Hz is maintained to ensure that temperature changes on the most urgent propagation path can be acquired in a timely manner.

[0210] When communication capacity is insufficient, let the available capacity of the communication bus be... The minimum bandwidth required for high-risk and urgent battery cells is The required frequency reduction battery cells are assembled as follows Then the latest reported frequency of the battery cell needs to be reduced. Determined according to the following formula:

[0211]

[0212] in, For the corresponding data frame length, To maintain the preset minimum frequency limit for communication handshake;

[0213] Resource adjustments are prioritized for cells in normal condition, and for cells in abnormal warning state that are more than 2 units away from cells in high-risk condition. Cells in high-risk condition and adjacent cells with the shortest remaining heat propagation time will not be reduced to frequencies below the specified frequency due to insufficient communication capacity. Data from cells in normal condition can be merged into data frames when the conditions for safe data transmission are met, and the uploading of historical valid sampling point data that is not related to safety can be suspended.

[0214] When the analog front end switches between the first and second sampling periods, the controller saves the effective filtered temperature, the corresponding sampling time, and the temperature rise rate at that time before the switch, and calculates the temporary predicted temperature based on the actual duration of the switching interval to maintain control continuity during the sampling switch.

[0215] The predicted temperature is only used to maintain the already triggered abnormal warning state or high-risk state, and to continue to execute the corresponding control during the sampling switch, and is not used to cancel the abnormal warning state or high-risk state; after the new sampling period takes effect, the sampling communication control module will prioritize the new effective filtered temperature to recalculate the temperature rise rate and temperature outlier.

[0216] This embodiment performs targeted sampling based on state level and physical map distance. When communication capacity is limited, data of high-risk cells and urgently adjacent cells are prioritized. By saving the effective temperature, its sampling time, and temperature rise trend, data interruption caused by analog front-end switching intervals is prevented. This embodiment allocates sampling and communication resources based on state and the urgency of heat propagation.

[0217] Furthermore, the thermal management control module and the electrical isolation control module perform different levels of protective actions based on abnormal warning states or high-risk states;

[0218] In some possible implementations, the thermal management control module is connected to the control terminal of the cooling pump and the cooling branch switching valve. In the abnormal warning state, the cooling pump is operated at 20% to 40% of the rated speed, and in the high-risk state, the cooling pump is operated at 80% to 100% of the rated speed, and the coolant is switched to the cooling branch corresponding to the module where the battery cell is located in the high-risk state.

[0219] The electrical isolation control module stops the balancing operation of the high-risk cells and the cells in the module where the high-risk cells are located, sends a current reduction request to the vehicle controller and stops charging. When the high-risk cells are located in an independently isolable branch, the corresponding branch contactor is disconnected first; otherwise, the main contactor is disconnected when the battery pack current is not higher than the contactor's allowable breaking current.

[0220] Within 1 to 5 seconds after the cooling control command is issued, if the action feedback and status latching module detects that the measured coolant flow rate is lower than 80% of the target coolant flow rate, the contactor auxiliary contact does not change accordingly, or the bus current does not drop below the preset safe current, it maintains the high-risk state and triggers electrical isolation, expands the high-frequency sampling range to the module where the high-risk cell is located, or issues a stop warning.

[0221] In a high-risk state, the state latch is released only after the vehicle controller has completed power-down and confirmed that the bus current is lower than the preset safe current and the insulation resistance is not lower than the insulation threshold, and the temperature of all effective cells is lower than the recovery temperature of 35°C to 45°C and the concentration of the correction gas is lower than 50% of the corresponding gas threshold.

[0222] In practical implementation, under abnormal warning conditions, the thermal management control module outputs a corresponding pre-circulation control command to the cooling pump, causing the cooling pump to run at 20% to 40% of its rated speed and controlling the coolant to pass through the battery cold plate.

[0223] At this time, if the current temperature of the battery pack has not yet reached the condition that requires forced heat dissipation, the actual heat dissipation intensity can be limited by bypassing the radiator, reducing the cooling capacity of the refrigerant side, or adjusting the mixing valve; the purpose of pre-circulation is to keep the pump body, coolant and cooling branch in a responsive state and reduce the response delay caused by the static start-up of the cooling pump and valve switching.

[0224] In high-risk conditions, the thermal management control module increases the cooling pump speed to 80% to 100% of the rated speed and controls the cooling branch switching valve to ensure that the coolant flows preferentially through the cooling branch of the module where the battery cell is located in high-risk conditions.

[0225] The target coolant flow rate is determined by the thermal management control module based on the high-risk module and the current cooling mode, while the actual coolant flow rate is obtained by the flow sensor; the coolant temperature is used to determine the current thermal state of the cooling medium, and the coolant flow rate is used to determine whether the cooling branch has actually established flow.

[0226] In a high-risk state, the electrical isolation control module stops the balancing operation of the high-risk cells and the cells in the module to prevent the balancing process from continuing to apply power to the abnormal area; at the same time, it sends a current reduction request to the vehicle controller and a charging stop request to the charging device.

[0227] When a high-risk cell is located in a branch that can be independently disconnected, the contactor of that branch is controlled to disconnect first; when there is no independently disconnectable branch, the main contactor is controlled to disconnect under the condition that the battery pack current is not higher than the contactor's allowable breaking current.

[0228] The action feedback and status latching module verifies the cooling and electrical isolation actions; after the cooling control command is issued, the measured coolant flow rate and the target coolant flow rate are compared within the action feedback verification time of 1s to 5s; when the measured flow rate is less than 80% of the target flow rate, it is determined that the cooling flow rate has not met the control requirements.

[0229] The contactor action is confirmed through feedback from the auxiliary contacts, and the bus current is confirmed through the current acquisition channel. If the auxiliary contacts do not change accordingly, or the bus current does not drop below the preset safe current, the electrical isolation action is not considered complete.

[0230] If the cooling flow is insufficient, the contactor auxiliary contacts do not change, or the bus current is still higher than the safe current, the action feedback and status latching module maintains a high-risk state and performs at least one of the following measures: expanding the high-frequency sampling range, performing electrical isolation in advance, or issuing a shutdown warning; the high-frequency sampling range is expanded to at least the module where the high-risk cell is located so that the temperature and voltage changes of the cells in that module can continue to be acquired when the cooling operation is abnormal.

[0231] Once a high-risk state is established, it is immediately latched and will not be automatically released due to a single temperature drop, a decrease in gas concentration, or a brief recovery of a sampling channel; the following conditions must be met simultaneously to release the latch:

[0232] Receives an authorized reset command sent by the vehicle controller after power-down is completed, the bus current is confirmed to be lower than the safe current and the insulation resistance is not lower than the insulation threshold.

[0233] The recovery temperature is between 35°C and 45°C for all effective battery cells; the calibration gas concentration is below 50% of the corresponding gas threshold. The recovery temperature is calibrated based on the battery cell material system, the installation location of the temperature sensor, and the cooling system capacity. The gas concentration recovery condition is used to confirm that the risk of gas evolution in the module is below the corresponding gas anomaly judgment level.

[0234] The preset safe current refers to the low current threshold or zero current threshold that can ensure the main contactor can safely disconnect without arcing and without causing the battery pack to continuously overheat abnormally; the insulation threshold is set according to the vehicle's high-voltage electrical safety standards and the battery pack's rated operating voltage, and is used to quantitatively characterize that the high-voltage network's insulation to ground has returned to normal, confirming that the system has no risk of high-voltage leakage or short circuit.

[0235] This embodiment uses coolant flow rate, contactor auxiliary contacts, and bus current as conditions for action completion, enabling it to identify situations such as cooling branch not being established, valves or pumps not reaching their targets, and contactors refusing to operate. Through high-risk state latching and authorized reset conditions, it avoids automatic resumption of normal operation when risks have not been eliminated. This embodiment achieves graded protection based on cooling pre-circulation, emergency cooling, electrical isolation, and execution feedback verification, avoiding inconsistencies between control commands and actual execution states.

[0236] In an electric vehicle power battery pack, the battery pack consists of multiple modules, each module including multiple cells; each cell is equipped with a temperature acquisition channel and a voltage acquisition channel, each module is equipped with at least one target thermal runaway characteristic gas sensor, and a cooling pump, a cooling branch switching valve and a coolant flow sensor are installed in the cooling circuit; the controller pre-stores cell physical adjacency table, thermal coupling coefficient, internal resistance reference value calibration table, gas humidity correction coefficient, gas temperature correction coefficient and various state judgment thresholds;

[0237] After the vehicle is powered on, the data acquisition module collects cell temperature, cell voltage, battery pack current, module gas concentration, gas sensor humidity, gas sensor temperature, coolant temperature, and coolant flow rate at the normal sampling frequency, and adds sampling time and channel identifier to each sampled data. The data validity judgment module judges the range, line status, sampling jump, rate of change, timestamp, communication verification, and gas sensor preheating status, and generates corresponding data validity flags.

[0238] When the effective filter temperature of a certain battery cell deviates significantly from the reference temperature of the same module, and the temperature rise rate reaches the weak thermal anomaly condition, the status judgment module determines that the battery cell is in an abnormal warning state; the sampling communication control module increases the sampling frequency of the battery cell and related adjacent battery cells, the thermal management control module makes the cooling pump run at 20% to 40% of the rated speed, and makes the coolant enter the corresponding cooling branch; at the same time, the gas zero-point baseline is frozen to preserve subsequent gas concentration changes;

[0239] When the cell further meets the evidence of strong thermal anomaly, or the evidence of abnormal internal resistance, abnormal module gas, pressure change rate, or pressure relief switch status meets the high-risk judgment conditions, the status judgment module determines it as a high-risk cell; the heat propagation time calculation module calculates the remaining heat propagation time of each adjacent cell based on the temperature difference between the high-risk cell and adjacent cells, the temperature rise rate of adjacent cells, and the thermal coupling coefficient, and determines the minimum value as the target remaining heat propagation time;

[0240] The sampling communication control module determines the high-frequency sampling target based on the status of the high-risk battery cell and the adjacent battery cell with the shortest remaining heat propagation time. When the communication capacity is insufficient, it reduces the reporting frequency of the normal battery cell or the distant adjacent battery cell, and keeps the sampling frequency of the high-risk battery cell and the most urgent adjacent battery cell not lower than the specified value.

[0241] When the front end switches the sampling period, the controller saves an effective filtered temperature, the corresponding sampling time and the temperature rise rate, and calculates the predicted temperature within the switching interval; after the new sampled data becomes valid, the temperature rise rate is updated using the new data, and the predicted temperature is not used to remove the high-risk state.

[0242] In high-risk conditions, the thermal management control module increases the cooling pump speed to 80% to 100% of its rated speed and switches the coolant to the cooling branch of the module where the high-risk battery cell is located; the electrical isolation control module stops the relevant equalization operation, sends a current reduction request to the vehicle controller and stops charging, and prioritizes disconnecting the corresponding branch contactor when an independent isolation branch is available; otherwise, it disconnects the main contactor after the battery pack current drops to the contactor's allowable breaking range.

[0243] During the action feedback and status latching time, the action feedback and status latching module continuously acquires the measured coolant flow rate, contactor auxiliary contact status, bus current and insulation resistance. When the measured coolant flow rate does not reach 80% of the target coolant flow rate, the contactor auxiliary contact does not change accordingly or the bus current does not drop below the safe current, it maintains a high-risk state and expands the high-frequency sampling range, performs further electrical isolation or issues a shutdown warning.

[0244] The high-risk state latch is released only after the vehicle is powered off, the bus current and insulation resistance meet safety conditions, the temperature of all effective cells is below the recovery temperature, the concentration of the calibration gas is below 50% of the corresponding threshold, and an authorized reset command is received.

[0245] Through the above application process, new temperature, voltage, current, gas, and cooling system data can be continuously fed into the data validity judgment module, and sequentially trigger the calculation of monitoring parameters, status judgment, heat propagation time calculation, sampling communication adjustment, thermal management control, electrical isolation control, and execution feedback verification. This results in the formation of a sampling frequency, communication priority, cooling capacity, and electrical isolation measures that match the risk level during the development of local thermal anomalies in the battery cell.

[0246] For any part not mentioned in this invention, existing technologies can be used or referenced. Of course, the above description is not a limitation of this invention, and this invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of this invention should also be within the protection scope of this invention.

Claims

1. A battery management system for electric vehicles used for thermal runaway monitoring, characterized in that, include: The data acquisition module is used to collect data such as cell temperature, cell voltage, battery pack current, module gas concentration, gas sensor humidity, gas sensor temperature, coolant temperature, and coolant flow rate. The data validity assessment module is used to generate data validity indicators; The internal resistance reference and gas correction module is used to determine the internal resistance reference value and correct the gas concentration. The monitoring parameter calculation module is used to calculate the temperature rise rate, temperature outlier, relative change in internal resistance, and correction gas concentration. The status judgment module is used to identify candidate cells for abnormal sources and determine the status as normal, abnormal warning, high risk, or sensor failure. The heat propagation time calculation module is used to calculate the remaining heat propagation time of the target based on the abnormal source candidate cells determined by the state judgment module and the adjacent cells that have a physical adjacency relationship with the abnormal source candidate cells. The control module includes a sampling communication control module, a thermal management control module, and an electrical isolation control module, which are used to output corresponding control commands; The action feedback and state latching module is used to obtain the execution feedback of the control command and maintain or release the state latching according to the execution feedback.

2. The electric vehicle battery management system for thermal runaway monitoring according to claim 1, characterized in that, The data acquisition module sends raw data with sampling time and channel identifier to the data validity judgment module; The data validity judgment module marks data as invalid for the same channel or the same cell if the range is out of bounds, open circuit, short circuit, absolute value of the difference with the previous valid sampling point is greater than the channel preset jump threshold, the rate of change of adjacent sampling points is greater than the preset rate of change threshold, timestamp is discontinuous, communication verification error, or gas sensor has not completed preheating. When the current step amplitude is less than the preset amplitude, a steady-state time window is formed before and after the step and the steady-state time window is less than the preset duration, or the current change rate within the steady-state time window is greater than the preset stability threshold, the data required for this internal resistance calculation will be marked as invalid. Freeze the gas zero-point baseline under abnormal warning or high-risk conditions.

3. The electric vehicle battery management system for thermal runaway monitoring according to claim 2, characterized in that, The monitoring parameter calculation module obtains valid temperature samples with the data validity flag, and filters the valid temperature samples to obtain valid filtered temperature samples. It selects 3 to 7 valid filtered temperature samples covering a time window of 0.5s to 2s, performs linear fitting according to the sampling time, and uses the fitting slope as the temperature rise rate. For any given cell as the cell to be evaluated, cells that have been latched as high-risk are excluded from the other cells in the module where the cell to be evaluated is located. The median of the effective filtered temperature samples of the remaining effective cells is used as the reference temperature, and the difference between the effective filtered temperature sample of the cell to be evaluated and the reference temperature is used as the temperature outlier. When the number of remaining valid cells is less than 50% of the total number of cells in the module and less than 3, the temperature outlier is marked as invalid.

4. The electric vehicle battery management system for thermal runaway monitoring according to claim 3, characterized in that, The monitoring parameter calculation module calculates the average cell voltage and average battery pack current within the steady-state time window before and after the current step, and according to... , Determine the relative change in internal resistance; in, The difference in average cell voltage within the steady-state time window. The difference in average battery pack current within the steady-state time window. The values ​​are obtained by looking up a calibration table or by piecewise linear interpolation based on the state of charge, state of health, and cell temperature. according to Determine the concentration of the correction gas, where, For the original converted gas concentration, As the zero-point baseline for the gas, This refers to the humidity correction coefficient obtained by looking up a table or by piecewise linear interpolation within the calibrated humidity and temperature ranges of the gas sensor. This refers to the temperature correction coefficient obtained by looking up a table or piecewise linear interpolation within the calibrated humidity and temperature ranges of the gas sensor.

5. The electric vehicle battery management system for thermal runaway monitoring according to claim 4, characterized in that, The state determination module uses one or more pre-selected thermal runaway characteristic gases as target thermal runaway characteristic gases. The target thermal runaway characteristic gases include at least one of carbon monoxide, carbon dioxide, hydrogen or hydrocarbon gases, and determines the gas threshold from the corresponding gas calibration threshold table according to the type and concentration unit of the target thermal runaway characteristic gas. When the rate of temperature rise is not less than 0.5℃ / s and the temperature outlier is not less than 5℃, evidence of strong thermal anomaly is confirmed. When the rate of temperature rise is not less than 0.2℃ / s or the temperature outlier is not less than 3℃, evidence of weak thermal anomaly is confirmed. When the relative change in internal resistance is no greater than -20% and the effective filtered temperature sample is no lower than 45℃, evidence of abnormal internal resistance is confirmed. When the concentration of the corrected gas for the target thermal runaway characteristic gas is not lower than the corresponding gas threshold, evidence of module gas anomaly is determined.

6. The electric vehicle battery management system for thermal runaway monitoring according to claim 5, characterized in that, The data acquisition module also collects module pressure and pressure relief switch status; The status judgment module obtains the target coolant flow rate from the thermal management control module and obtains the measured coolant flow rate; the status judgment module determines an abnormal warning state when any of the following conditions are met: weak thermal abnormal evidence lasts for at least 0.5s; The difference between the first occurrence of evidence of weak thermal abnormality and evidence of internal resistance abnormality shall not be greater than a preset correlation time, which is a fixed value within the range of 1s to 3s. The evidence of abnormal module gas occurs simultaneously with the condition that the temperature rise rate is not less than 0.1℃ / s or the effective filtered temperature sample is not less than 40℃. The temperature outliers of at least two cells with direct physical adjacency within the same module have a fitting slope greater than 0.1℃ / s within a trend time window of 1s to 10s. Alternatively, the measured coolant flow rate is less than 80% of the target coolant flow rate, and the fitting slope of the average effective filtered temperature sample of the module calculated by the monitoring parameter calculation module based on the effective filtered temperature sample within the same module is greater than zero within 1 second. A state is identified as high-risk when any of the following conditions are met: at least two valid filtered temperature samples are obtained consecutively as evidence of strong thermal anomalies. The difference between the first occurrence of evidence of abnormal internal resistance and evidence of abnormal weak heat is no greater than the preset correlation time; The effective filtered temperature sample reaches the emergency temperature threshold; the module pressure change rate is not lower than the preset pressure change rate threshold; the pressure relief switch is activated; or the remaining heat propagation time of the target is lower than the emergency time threshold. When key temperature, internal resistance, gas, pressure, or communication data do not meet the validity conditions, the status judgment module will mark the corresponding feature as invalid and determine whether to enter the sensor failure state based on the number of invalid sensors and their location. High-risk status takes precedence over sensor failure status, sensor failure status takes precedence over abnormal warning status, and a normal status is determined when the above conditions are not met.

7. The electric vehicle battery management system for thermal runaway monitoring according to claim 6, characterized in that, The state judgment module identifies cells that meet the strong or weak thermal anomaly evidence as candidate cells for anomaly sources. Based on the pre-stored physical adjacency table and the thermal coupling coefficient between each physical adjacency cell, combined with the difference between the effective filtered temperature samples of the candidate cells for anomaly sources and the physical adjacency cells and the temperature rise rate of the adjacency cells, it determines the remaining heat propagation time of each physical adjacency cell. Specifically, when the temperature rise rate of physically adjacent cells is less than zero, it is treated as zero; when the effective filtered temperature sample of the candidate cell for the abnormal source is not higher than the effective filtered temperature sample of the physically adjacent cell, the difference between the effective filtered temperature samples is treated as zero. The heat propagation time calculation module determines the minimum value among the remaining heat propagation times of each physically adjacent cell as the target remaining heat propagation time.

8. The electric vehicle battery management system for thermal runaway monitoring according to claim 7, characterized in that, The sampling communication control module allocates a sampling frequency of 0.5Hz to 1Hz for the temperature and voltage of cells in normal state, a sampling frequency of 2Hz to 5Hz for cells in abnormal warning state, and a sampling frequency of 10Hz to 20Hz for cells in high-risk state and their first-order physical adjacent cells. The first-order physical adjacent cells are those cells that are 1 distance away from cells in high-risk state in the physical adjacency table. Cells with a distance of 2 or greater in the diagram are considered distantly adjacent cells. When communication capacity is insufficient, the reporting frequency of cells in normal state is reduced first, and then the reporting frequency of cells in abnormal warning state in the distant adjacent cells corresponding to the high-risk cells is reduced. At the same time, the sampling frequency of the high-risk cells and the adjacent cells with the shortest remaining target heat propagation time is kept not lower than 10Hz. During the switching interval between the first sampling period before the switch and the second sampling period after the switch in the data acquisition module, the last valid filtered temperature sample, the corresponding sampling time, and the temperature rise rate are saved, and then... Calculate the predicted temperature, where, This is the last valid filtered temperature sample. This refers to the sampling time corresponding to the last valid filtered temperature sample. The temperature rise rate, The predicted time is the current time; the predicted temperature is not used to lift the abnormal warning state or high-risk state.

9. The electric vehicle battery management system for thermal runaway monitoring according to claim 8, characterized in that, The thermal management control module is connected to the control terminal of the cooling pump and the cooling branch switching valve. In the abnormal warning state, the cooling pump is operated at 20% to 40% of the rated speed. In the high-risk state, the cooling pump is operated at 80% to 100% of the rated speed, and the coolant is switched to the cooling branch corresponding to the module where the battery cell is located in the high-risk state. The electrical isolation control module stops the balancing operation of the high-risk battery cell and the battery cells in the module where the high-risk battery cell is located, sends a current reduction request to the vehicle controller and stops charging. When the high-risk battery cell is located in an independently isolated branch, the corresponding branch contactor is disconnected first; otherwise, the main contactor is disconnected when the battery pack current is not higher than the contactor's allowable breaking current. Within 1 to 5 seconds after the cooling control command is issued, if the action feedback and status latching module detects that the measured coolant flow rate is lower than 80% of the target coolant flow rate, the contactor auxiliary contact does not change accordingly, or the bus current does not drop below the preset safe current, it maintains the high-risk state and triggers electrical isolation, expands the high-frequency sampling range to the module where the high-risk state cell is located, or issues a stop warning. The high-risk state is only released after receiving an authorized reset command from the vehicle controller after it has completed power-down and confirmed that the bus current is lower than the preset safe current and the insulation resistance is not lower than the insulation threshold, and after all effective battery cell temperatures are lower than the recovery temperature of 35°C to 45°C and the correction gas concentration is lower than 50% of the corresponding gas threshold.

Images