Lithium iron phosphate battery thermal runaway early warning method based on ultrasonic gas detection

Abstract

The application provides a lithium iron phosphate battery thermal runaway early warning method based on ultrasonic gas detection, and relates to the technical field of battery management, and comprises the following steps: collecting ultrasonic signals in a battery thermal runaway process by using an ultrasonic gas detector; filtering the ultrasonic signals to obtain a detrended signal; extracting a slope feature and an auxiliary judgment feature from the detrended signal; the auxiliary judgment feature comprises an energy envelope and a Page-Hinkley cumulative test; calculating the smoothness of the detrended signal; calculating a segmented threshold based on the smoothness and the auxiliary judgment feature; and completing thermal runaway early warning by using the segmented threshold and the slope feature. The application can detect abnormalities in the early stage of battery thermal runaway, that is, when the gas is just released, realize earlier early warning than traditional temperature monitoring, and the ultrasonic signal is not sensitive to the temperature and humidity of the environment, electromagnetic interference and the like, and can work stably in a complex battery operating environment, thereby reducing the false alarm rate.

Description

Technical Field

[0001] This invention relates to the field of battery thermal runaway early warning technology, and in particular to a method for early warning of thermal runaway of lithium iron phosphate batteries based on ultrasonic gas detection. Background Technology

[0002] With the rapid development of battery technology, batteries have been widely used in various fields. However, batteries are prone to thermal runaway under conditions such as overcharging, over-discharging, short circuits, and mechanical damage, leading to a rapid increase in temperature, smoke, fire, or even explosion, seriously threatening personal and property safety. Currently, early warning methods for battery thermal runaway mainly include temperature monitoring, voltage monitoring, and gas monitoring. Among them, gas monitoring determines whether thermal runaway has occurred by detecting gases (such as hydrogen, carbon monoxide, carbon dioxide, hydrocarbons, etc.) released during battery thermal runaway. Traditional gas monitoring methods mostly use electrochemical sensors and semiconductor sensors. These sensors have problems such as slow response speed, low sensitivity, susceptibility to environmental interference, and short lifespan, making it difficult to achieve early, rapid, and accurate early warning of battery thermal runaway.

[0003] In recent years, ultrasonic gas detection technology has gradually attracted attention due to its advantages such as high sensitivity, fast response speed, strong anti-interference ability, and long lifespan. However, there is currently no effective method for applying ultrasonic gas detectors to battery thermal runaway early warning. Therefore, there is an urgent need for a new battery thermal runaway early warning method that can overcome the shortcomings of existing technologies and achieve early, rapid, and accurate early warning of battery thermal runaway. Summary of the Invention

[0004] To address the aforementioned problems, the purpose of this invention is to provide a method for early warning of thermal runaway in lithium iron phosphate batteries based on ultrasonic gas detection.

[0005] A method for early warning of thermal runaway in lithium iron phosphate batteries based on ultrasonic gas detection includes:

[0006] Step 1: Use an ultrasonic gas detector to collect ultrasonic signals during the battery thermal runaway process;

[0007] Step 2: Filter the ultrasonic signal to obtain a detrended signal;

[0008] Step 3: Extract slope features and auxiliary judgment features from the detrended signal; the auxiliary judgment features include energy envelope time-domain features and Page-Hinkley cumulative test feature values;

[0009] Step 4: Calculate the smoothness of the detrending signal;

[0010] Step 5: Calculate the segmentation threshold based on smoothness and auxiliary judgment features;

[0011] Step 6: Use the segmented threshold and slope features to complete the thermal runaway early warning.

[0012] Preferably, in step 2, a rolling median high-pass filter is performed within a set window to obtain the detrended signal; wherein the rolling median high-pass filter formula is:

[0013]

[0014] in, Calculate the frequency for ADC sampling. This is a detrending signal. Here is the ADC value of the ultrasonic signal, and median indicates that the median is being calculated. For the set time window, This refers to the current moment.

[0015] Preferably, in step 3, slope features are extracted from the absolute values ​​of the linear regression slopes of the long window and the short window; wherein, the slope feature calculation formula is:

[0016]

[0017]

[0018]

[0019] in, This represents the initial calculated value of the slope feature. Indicates the length of the sliding window. This represents the i-th time within the window. This represents the time average of the data points within the window. , This represents a dimensionless metric (robust z-score) for slope feature files. This represents the window median. This indicates the absolute deviation of the median within the window. This indicates data points within the window that are at the 95th percentile. This indicates the slope characteristic.

[0020] Preferably, in step 3, the energy envelope is calculated based on the moving average of the absolute values ​​of the detrended signal; wherein, the energy envelope calculation formula is:

[0021]

[0022]

[0023] in, Represents the time-domain characteristics of the energy envelope. This indicates calculating the average value. This indicates the initial point of the energy envelope calculation window. This indicates the end point of the energy envelope calculation window. This represents the time-domain characteristics of the normalized energy envelope.

[0024] Preferably, in step 3, a Page-Hinkley cumulative test is performed based on the absolute difference of the detrended signal; wherein, the Page-Hinkley cumulative test formula is:

[0025]

[0026]

[0027]

[0028] in, This represents the eigenvalues ​​of the Page-Hinkley cumulative test. This represents the absolute value of the difference in the de-energized signal. Indicates the smoothed mean. Indicates the pH tolerance zone. Represents the smoothing coefficient. Represents the normalized pH characteristic value. This represents the pH characteristic value data at the 95th percentile within the window.

[0029] Preferably, in step 4, a long window is used to calculate the smoothness; wherein, the smoothness calculation formula is:

[0030]

[0031]

[0032] in, This indicates the calculation of the standard deviation. Indicates the length of the long window. Indicates smoothness, Indicates normalized smoothness. This represents the smoothness data at the 95th percentile within the window. This represents the slope sample of the BASE window.

[0033] Preferably, in step 5, the time period is divided based on smoothness and energy envelope:

[0034]

[0035]

[0036] otherwise

[0037] in, Indicates the remote valve section. Indicates the section near the valve. Indicates the valve opening warning section. Indicates the smoothness criterion threshold. This represents the threshold for the envelope criterion.

[0038] Preferably, step 6: Utilizing the segmented threshold and slope features to complete thermal runaway early warning, including:

[0039] Step 6.1: Based on the segmented gating results, determine whether the slope exceeds the threshold and the number of continuous samples are met;

[0040] Step 6.2: Within a few verification windows after triggering, check whether the energy and change point evidence meet the verification conditions. If both stages pass, output the warning time Alarm(t).

[0041] The formula for the continuous sample constraint is expressed as:

[0042]

[0043] The verification formula within the verification window is expressed as follows:

[0044]

[0045] Final judgment:

[0046] Alarm(t) = C(t) V(t)

[0047] The gas detector issues an alarm when the value of Alarm(t) is non-zero; where, The flag indicates that the sample persistence condition is met. Indicates the starting sample index that triggers the test. This indicates a true / false judgment, with 1 for true and 0 for false. This indicates the threshold for the slope evidence criterion. Indicates the number of continuous samples. This indicates that the review has been approved. Indicates any, Indicates the length of the verification window. This indicates that the energy gate has multiplied. This indicates that the pH gate is multiplied.

[0048] The present invention also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are connected via the bus. The computer program, when executed by the processor, implements the steps in the above-described method for early warning of thermal runaway of lithium iron phosphate batteries based on ultrasonic gas detection.

[0049] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps in the above-described method for early warning of thermal runaway of lithium iron phosphate batteries based on ultrasonic gas detection.

[0050] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0051] This invention relates to a method for early warning of thermal runaway in lithium iron phosphate batteries based on ultrasonic gas detection. Compared with the prior art, this invention can detect anomalies in the early stage of battery thermal runaway, that is, when the gas is just released, achieving an earlier warning than traditional temperature monitoring. Furthermore, ultrasonic signals are not sensitive to environmental temperature, humidity, electromagnetic interference, etc., and can work stably in complex battery operating environments, reducing the false alarm rate.

[0052] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a flowchart of the lithium iron phosphate battery thermal runaway early warning method based on ultrasonic gas detection involved in this invention. Detailed Implementation

[0055] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0056] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0057] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0058] like Figure 1 As shown, this embodiment provides a method for early warning of thermal runaway in lithium iron phosphate batteries based on ultrasonic gas detection, including:

[0059] Step 1: Use an ultrasonic gas detector to collect ultrasonic signals during the battery thermal runaway process;

[0060] In step 1, ultrasonic gas detectors are deployed to acquire signal characteristics for real-time monitoring: at least one ultrasonic gas detector is deployed inside or near the battery module. The ultrasonic gas detector is used to detect changes in the ultrasonic wave propagation speed caused by variations in gas temperature, average molecular weight, etc., during battery thermal runaway.

[0061] Step 2: Filter the ultrasonic signal to obtain a detrended signal;

[0062] In step 2, the phase difference value of the acquired ultrasound signal is converted into an ADC value x[t] that can be used for calculation. This is then subjected to rolling median high-pass filtering within a certain window T to remove low-frequency interference and eliminate some baseline drift, resulting in a detrended signal. :

[0063]

[0064] Where f is the ADC sampling frequency, and median represents the calculation of the median.

[0065] Step 3: Extract slope features and auxiliary judgment features from the detrended signal; the auxiliary judgment features include energy envelope and Page-Hinkley cumulative test;

[0066] In step 3, to better identify the data pattern of the response data after the detector monitors the pressure relief valve opening, this step calculates the absolute value of the linear regression slope within both a long window and a short window. The short window is used to quickly respond to sudden changes in the slope, while the long window is used to discern slope changes. Using this structure for training with actual data can ensure that false alarms are reduced when the valve is opened. The specific steps are as follows:

[0067] For long windows With short window The formula for calculating the absolute value of the slope in linear regression is:

[0068]

[0069] Take the absolute values ​​of both and then calculate the robust z-score, as shown in the formula:

[0070]

[0071] For z(t), after normalization to the 95th percentile:

[0072]

[0073] This indicates data points within the window that are at the 95th percentile.

[0074] Features for auxiliary judgment are extracted from the detrended signal. These features include the energy envelope and the Page-Hinkley cumulative test. The energy envelope E reflects changes in signal amplitude, while Page-Hinkley (PH) is an online change-point detection statistic used in this algorithm to determine whether a sustained, irreversible upward shift trend has occurred locally in the signal. Both features can be used for subsequent gate judgments and for verifying the judgment results.

[0075] Among them, the energy envelope is the moving average of the absolute value of the signal, and then the quantile normalization is performed;

[0076]

[0077] In the formula, and These represent the start and end points of the energy envelope calculation, respectively.

[0078] Similarly, after normalization to the 95th percentile, the results are as follows:

[0079]

[0080] The Page-Hinkley cumulative test is performed on the absolute value of the signal difference, as shown in the following formula:

[0081]

[0082]

[0083] In the formula and The coefficients need to be adjusted based on the specific experimental data.

[0084] Similarly, after normalization to the 95th percentile, the results are as follows:

[0085]

[0086] Step 4: Calculate the smoothness of the detrending signal;

[0087] In step 4, the ratio of the slope rolling standard deviation to the local P95 is calculated as a smoothness index. The threshold is dynamically adjusted according to the smoothness level, i.e., smoothing gating is performed. Setting this step can facilitate the adjustment of the threshold of each segment based on experimental data, thereby achieving the effect of suppressing false alarms in the smoothing segment before the valve and timely alarm when the valve is opened.

[0088] Because a larger window provides a more accurate perception of smoothness, a long window can be used to calculate the smoothness r(t), as shown in the following formula:

[0089]

[0090] In the formula This indicates the calculation of the standard deviation. Indicates the length of the long window;

[0091] After normalization, it becomes:

[0092]

[0093] Step 5: Calculate the segmentation threshold based on smoothness and auxiliary judgment features;

[0094] In step 5, based on smoothness and energy envelope The time period is divided into:

[0095] Far-pre section: smooth and low energy, raises the valve opening judgment threshold and uses AND logic for the judgment parameters;

[0096] Near-pre section: moderate smoothing or energy rise, using moderate threshold and AND logic;

[0097] Valve opening warning section ( ): Uses OR logic, with a low threshold.

[0098] The formula for the segmentation criterion is as follows:

[0099]

[0100]

[0101] otherwise

[0102] Among them, smoothness and energy envelope The judgment threshold and the judgment thresholds of various characteristic parameters used for valve opening judgment both need to be adjusted using specific experimental data.

[0103] Step 6: Use the segmented threshold and slope features to complete the thermal runaway early warning.

[0104] In step 6, based on the segmented threshold given by the agent segmented valve, the valve opening is checked through two steps. The first step is to determine whether the slope exceeds the threshold and the continuous sample number requirements are met based on the segmented gating results. The second step is to check whether the energy and change point evidence meet the verification conditions within a certain number of verification windows after triggering. If both stages pass, the warning time Alarm(t) is output.

[0105] The formula for the continuous sample constraint is expressed as:

[0106]

[0107] The verification formula within the verification window is expressed as follows:

[0108]

[0109] Final judgment:

[0110] Alarm(t) = C(t) V(t)

[0111] When the value of Alarm(t) is non-zero, the gas detector issues a valve opening alarm.

[0112] The gas detector transmits the valve opening alarm information to the host computer or BMS via the communication module so that the battery that has experienced early thermal runaway can be dealt with in a timely manner.

[0113] The thermal runaway early warning method of the present invention will be further described below with reference to specific embodiments:

[0114] In an experiment conducted on a provincial-level research project, overcharge thermal runaway experiments of lithium iron phosphate batteries at 0.25C, 0.5C, and 0.75C, and overheating thermal runaway experiments at 50%, 75%, and 100% SOC were carried out on a simulated 280Ah LFP module with 2 parallel and 7 series connections. An ultrasonic gas detector was installed above the battery strings inside the module cover. The ultrasonic gas detector operates at a frequency of approximately 340kHz. Nineteen sets of data were obtained through RS-485 communication with an external host computer. Data was collected once per second. After processing these data, this method was used for offline evaluation and parameter training. Some parameters are specified as follows:

[0115] Slope window: The long window has a length of 9 data points, and the short window has a length of 5 data points;

[0116] Energy window and pH window: 7 data points in length.

[0117] Agent gating judgment window: The E and PH judgment conditions of the Far segment are satisfied for ≥6 consecutive points, and the Near segment is satisfied for ≥3 consecutive points.

[0118] The parameters obtained after calculation are shown in Table 1 below:

[0119] Table 1

[0120]

[0121] The alarm results for offline training of 19 sets of data are shown in Table 2 below:

[0122] Table 2

[0123]

[0124] The results show that using the battery thermal runaway early warning model provided by this invention, each group can have at least one effective alarm while the false alarm rate can reach 0, and the longest alarm delay is only 4 seconds.

[0125] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0126] 1. Strong early warning capability: Ultrasonic gas detectors are very sensitive to minute changes in ambient gas and can detect anomalies in the early stages of battery thermal runaway, that is, when the gas has just been released, thus providing an earlier warning than traditional temperature monitoring.

[0127] 2. Fast response speed: Ultrasonic waves propagate at extremely high speeds, and the sensor's response time can reach the millisecond level, which is much faster than the response speed of traditional electrochemical gas sensors.

[0128] 3. Strong anti-interference capability: Ultrasonic gas detectors are not sensitive to environmental temperature, humidity, electromagnetic interference, etc., and can work stably in complex battery operating environments, reducing false alarm rate.

[0129] 4. Long lifespan and low maintenance cost: Ultrasonic gas detectors have no chemical reaction, no consumables, long service life, and do not require frequent replacement, thus reducing maintenance costs.

[0130] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for early warning of thermal runaway in lithium iron phosphate batteries based on ultrasonic gas detection, characterized in that, include: Step 1: Use an ultrasonic gas detector to collect ultrasonic signals during the battery thermal runaway process; Step 2: Filter the ultrasonic signal to obtain a detrended signal; Step 3: Extract slope features and auxiliary judgment features from the detrended signal; the auxiliary judgment features include energy envelope time-domain features and Page-Hinkley cumulative test feature values; Step 4: Calculate the smoothness of the detrending signal; Step 5: Calculate the segmentation threshold based on smoothness and auxiliary judgment features; Step 6: Utilize the segmented threshold and slope features to complete the thermal runaway early warning; In step 3, slope features are extracted from the absolute values ​​of the linear regression slopes of the long and short windows; the formula for calculating the slope features is as follows: in, This represents the initial calculated value of the slope feature. Indicates the length of the sliding window. This represents the i-th time within the window. This represents the time average of the data points within the window. , The slope feature file represents a dimensionless metric, i.e., a robust z-score. This represents the window median. This indicates the absolute deviation of the median within the window. This indicates data points within the window that are at the 95th percentile. Indicates slope characteristics; In step 3, the energy envelope is calculated based on the moving average of the absolute values ​​of the detrended signal; wherein, the formula for calculating the energy envelope is: in, Represents the time-domain characteristics of the energy envelope. This indicates calculating the average value. This indicates the initial point of the energy envelope calculation window. This indicates the end point of the energy envelope calculation window. This represents the time-domain characteristics of the normalized energy envelope; In step 3, a Page-Hinkley cumulative test is performed based on the absolute difference of the detrended signal; the Page-Hinkley cumulative test formula is: in, This represents the eigenvalues ​​of the Page-Hinkley cumulative test. This represents the absolute value of the difference in the de-energized signal. Indicates the smoothed mean. Indicates the pH tolerance zone. Represents the smoothing coefficient. Represents the normalized pH characteristic value. This represents the pH characteristic value data at the 95th percentile within the window.

2. The method for early warning of thermal runaway of lithium iron phosphate batteries based on ultrasonic gas detection according to claim 1, characterized in that, In step 2, a rolling median high-pass filter is performed within a set window to obtain the detrended signal; the formula for the rolling median high-pass filter is: in, Calculate the frequency for ADC sampling. This is a detrending signal. Here is the ADC value of the ultrasonic signal, and median indicates that the median is being calculated. For the set time window, This refers to the current moment.

3. The method for early warning of thermal runaway of lithium iron phosphate batteries based on ultrasonic gas detection according to claim 2, characterized in that, In step 4, a long window is used to calculate the smoothness; the smoothness calculation formula is as follows: in, This indicates the calculation of the standard deviation. Indicates the length of the long window. Indicates smoothness, Indicates normalized smoothness. This represents the smoothness data at the 95th percentile within the window. This represents the slope sample of the BASE window.

4. The method for early warning of thermal runaway of lithium iron phosphate batteries based on ultrasonic gas detection according to claim 3, characterized in that, In step 5, the time period is divided based on smoothness and energy envelope: otherwise in, Indicates the remote valve section. Indicates the section near the valve. Indicates the valve opening warning section. This represents the threshold for the smoothness criterion. This represents the threshold for the envelope criterion.

5. The method for early warning of thermal runaway of lithium iron phosphate batteries based on ultrasonic gas detection according to claim 4, characterized in that, Step 6: Utilize the segmented threshold and slope features to complete the thermal runaway early warning, including: Step 6.1: Based on the segmented gating results, determine whether the slope exceeds the threshold and the number of continuous samples are met; Step 6.2: Within a few verification windows after triggering, check whether the energy and change point evidence meet the verification conditions. If both stages pass, output the warning time Alarm(t). The formula for the continuous sample constraint is expressed as: The verification formula within the verification window is expressed as follows: Final judgment: Alarm(t)=C(t) V(t) The gas detector issues an alarm when the value of Alarm(t) is non-zero; where, The flag indicates that the sample persistence condition is met. Indicates the starting sample index that triggers the test. This indicates a true / false judgment, with 1 for true and 0 for false. This indicates the threshold for the slope evidence criterion. Indicates the number of continuous samples. This indicates that the review has been approved. Indicates any, Indicates the length of the verification window. This indicates that the energy gate has multiplied. This indicates that the pH gate is multiplied.

6. An electronic device comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, characterized in that, When the computer program is executed by the processor, it implements the steps in the method for early warning of thermal runaway of lithium iron phosphate battery based on ultrasonic gas detection as described in any one of claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps in the method for early warning of thermal runaway of lithium iron phosphate battery based on ultrasonic gas detection as described in any one of claims 1-5.

Images