Thermal runaway detection method and warning system

Abstract

A method for detecting cell thermal runaway includes positioning a battery pack having a plurality of cells in a motor vehicle, measuring cell voltages of the plurality of cells at a predetermined sampling rate, and determining whether the cell voltages decrease and transition as cell surface temperatures increase to indicate a beginning cell short.

Description

Technical Field

[0001] This disclosure relates to detecting and mitigating thermal runaway in vehicle battery pack cells. Background Technology

[0002] Short circuits are the most common cause of thermal runaway in battery pack cells used in motor vehicle systems. A short circuit typically begins with a short circuit in the primary cell, followed by a gradual increase in cell temperature over 5-10 seconds. A short circuit can lead to thermal runaway, defined as an exothermic reaction that generates enough heat for the cell to potentially ignite the battery pack.

[0003] Battery cell short circuits can be caused by vehicle collisions, overcharging, road debris, vehicles parked in extreme temperatures, or manufacturing defects in the battery cells. Systems for detecting and identifying short circuits and preventing thermal runaway within 5 to 10 seconds after the short circuit begins are not yet known.

[0004] Therefore, while current vehicle battery pack designs have achieved their intended purpose, a new and improved system and approach are still needed to detect and mitigate short circuits and thermal runaway in vehicle battery pack cells. Summary of the Invention

[0005] According to several aspects, a method for detecting cell thermal runaway includes: positioning a battery pack having multiple cells in a motor vehicle; measuring the cell voltages of the multiple cells at a predetermined sampling rate; and determining whether the cell voltages decrease and change when or before the cell surface temperature rises to indicate the initiation of a cell short circuit, thereby indicating a thermal runaway event.

[0006] In another aspect of this disclosure, the method further includes confirming that the vehicle has stopped before measuring the unit voltage.

[0007] In another aspect of this disclosure, the method further includes: determining whether the cell temperature reaches about 70°C; and determining whether, after the cell temperature reaches about 70°C, the cell temperature rapidly rises to about 500°C within about 5 seconds, thereby determining that a thermal runaway event of the battery cell has begun.

[0008] In another aspect of this disclosure, the method further includes determining the average value of all cell voltages of a plurality of cells by subtracting the minimum cell voltage from the sum of the cell voltages.

[0009] In another aspect of this disclosure, the method further includes: calculating the derivative of the cell voltage of each cell in the plurality of cells with respect to time; and passing the derivative of the cell voltage to a buffer.

[0010] In another aspect of this disclosure, the method further includes calculating the fast Fourier transform power spectrum by applying the derivatives of the cell voltages of the multiple cells with respect to time.

[0011] In another aspect of this disclosure, the method also includes applying a power spectrum to calculate the released energy.

[0012] In another aspect of this disclosure, the method further includes: determining whether the power spectrum exceeds a predetermined spectrum threshold after the cell short circuit begins; and initiating an alarm after the predetermined threshold is exceeded.

[0013] In another aspect of this disclosure, the method also includes, upon activating the alarm, initiating at least one of the following actions: stopping the charging operation of the battery pack; releasing the pressure on the battery pack; allowing coolant to begin flowing into the battery pack; conveying a warning to the vehicle operator via a smartphone; transmitting the status of the battery pack to a cloud-based remote security service; and contacting emergency services.

[0014] In another aspect of this disclosure, the method further includes continuing to analyze the cell voltage and cell surface temperature after the cell short circuit begins and during the period of cell voltage decrease and rapid cell voltage transition, as well as as the cell surface temperature increases.

[0015] According to several aspects, a method for detecting cell thermal runaway includes: positioning a battery pack having multiple cells in a motor vehicle; measuring the cell voltages of the multiple cells at a predetermined sampling rate; and determining whether at least one of the multiple cells has experienced a cell short circuit.

[0016] In another aspect of this disclosure, the method further includes decomposing the unit voltage signal into multiple signal time-domain scales by applying a scalable wavelet filter of discrete wavelet transformation (DWT).

[0017] In another aspect of this disclosure, the method further includes: passing the unit voltage signal through a wavelet filter; and applying a scaling function to distinguish and remove frequencies of signals pre-specified as noise.

[0018] In another aspect of this disclosure, the method further includes applying a scaling function to divide the first high-frequency extraction portion into a first high-frequency component and the first low-frequency extraction portion into a first low-frequency component.

[0019] In another aspect of this disclosure, the method further includes: halving the low-frequency components from the first low-frequency extraction portion; and generating first-order coefficients.

[0020] In another aspect of this disclosure, the method further includes: calculating the derivative of the cell voltage of each cell in the plurality of cells with respect to time; applying the derivatives of the cell voltages of the plurality of cells with respect to time to calculate the fast Fourier transform power spectrum; applying the power spectrum to calculate the released energy; and activating an alarm if the power spectrum exceeds a predetermined threshold after a cell short circuit begins.

[0021] In another aspect of this disclosure, the method further includes determining whether the cell voltage decreases and changes as the cell surface temperature rises to indicate the initiation of a cell short circuit.

[0022] According to several aspects, a method for detecting cell thermal runaway includes: confirming that a motor vehicle has stopped; measuring the cell voltages of multiple cells in a battery pack of the motor vehicle at a predetermined sampling rate; determining whether the cell voltages have begun to decrease and change to indicate that a cell short circuit has occurred; and issuing a signal if the cell temperature reaches approximately 70°C to indicate that a battery cell thermal runaway event has occurred.

[0023] In another aspect of this disclosure, the method further includes: calculating the derivative of the cell voltage of each cell in the plurality of cells with respect to time; applying the derivatives of the cell voltages of the plurality of cells with respect to time to calculate the fast Fourier transform power spectrum; applying the power spectrum to calculate the released energy; and activating an alarm if the power spectrum exceeds a predetermined threshold after a cell short circuit begins.

[0024] In another aspect of this disclosure, the method further includes: passing the unit voltage signal through a wavelet filter; applying a scaling function to divide a first high-frequency extracted portion of the unit voltage signal into a first high-frequency component and a first low-frequency extracted portion into a first low-frequency component; and distinguishing and removing frequencies of signals pre-specified as noise.

[0025] Further applicability will become apparent from the description provided herein. It should be understood that the descriptions and specific examples are for illustrative purposes only and are not intended to limit the scope of this disclosure. Attached Figure Description

[0026] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of this disclosure in any way.

[0027] Figure 1 It is a schematic diagram of a battery pack having multiple battery cell modules comprising multiple individual units, based on typical aspects;

[0028] Figure 2 It presents a graph showing the relationship between unit voltage and time and unit surface temperature;

[0029] Figure 3 This is a flowchart of the method steps used to determine a unit thermal runaway event;

[0030] Figure 4 It is a graph showing the range of the derivative of the unit voltage over time (in seconds);

[0031] Figure 5 It presents a graph showing the relationship between the calculated power spectrum and time (in seconds);

[0032] Figure 6 It presents a graph showing the unit voltage relative to time (in seconds);

[0033] Figure 7 It presents a graph showing the power spectrum versus time (in seconds) when a sampling rate of 50ms is applied.

[0034] Figure 8 It provides a flowchart of the method steps for determining whether a unit thermal runaway event has occurred using Discrete Wavelet Transform (DWT);

[0035] Figure 9 It is a graph of cell voltage versus time (in seconds) used to determine thermal runaway conditions during battery pack charging operations; and

[0036] Figure 10 It is a graph that uses DWT to determine the voltage noise level as a function of time (in seconds), and is used to determine the TRA cell voltage level 1 decomposition signal during charging operation. Detailed Implementation

[0037] The following description is merely exemplary in nature and is not intended to limit this disclosure, its application, or its uses.

[0038] refer to Figure 1 A thermal runaway detection method and warning system 10 are applied to analyze the cells of one or more battery cell modules (e.g., a first module 12), each battery cell module having multiple cells connected in series, typically including a first cell 14, a second cell 16, and in one typical aspect approximately up to a sixteenth cell 18. The multiple cell modules are then connected in series to form a typical module group 20 having a first module 12, a second module 22, a third module 24, up to an Nth module 26. The multiple module groups (e.g., module group 20) are then connected in series to form a battery pack 28 having multiple modules and multiple cells.

[0039] Battery pack 28 is positioned in a motor vehicle 30 that operates using electricity generated by battery pack 28. The cells of battery pack 28 are charged as needed, monitored during operation, and controlled using controller 32. Controller 32 may be installed in the motor vehicle 30 either together with battery pack 28 or separately. According to several aspects, controller 32 may include one or more computers, each having one or more processors, at least one memory, and instructions stored in the memory. The memory is a non-transitory computer-readable medium.

[0040] refer to Figure 2 And refer again Figure 1 Graph 34 illustrates the relationship between cell voltage and time and cell surface temperature. Because the battery pack 28 consists of many cells, any one, two or more cells, or two or more modules could potentially catch fire. Therefore, the algorithm of this disclosure is applied to the battery cells individually. An exemplary thermal runaway process is depicted in graph 34, which identifies the cell voltage 36 for each time period 38 (in seconds) and provides the cell surface temperature 39. The cell voltage 40 is generally substantially stable, and the cell surface temperature 42 rises very slowly over time during normal cell operation until a cell short circuit 44 occurs at approximately 407 seconds in the example shown. Immediately following the cell short circuit 44, the cell voltage 46 decreases and rapidly changes, and the cell surface temperature rises, causing gas to be generated and released from the battery cell. After a time period 50 of approximately 7.9 seconds following the cell short circuit 44, the cell voltage 48 drops rapidly, as shown at a typical time of approximately 415 seconds.

[0041] Simultaneously, at approximately 415 seconds, a threshold cell temperature of approximately 70°C is reached, after which the cell temperature rapidly rises 52, which is identified as the initiation of a thermal runaway (TRA) event 54. During TRA event 54, within approximately 5 to 10 seconds after the cell temperature reaches and exceeds the threshold temperature of 70°C (defined as the initiation of TRA event 54), the cell temperature reaches 500°C or higher. After the cell short circuit 44 occurs, and within approximately 2 to 3 seconds after the cell voltage 48 rapidly decreases at approximately 415 seconds, the cell voltage 56 reaches approximately zero volts. The thermal runaway detection method and warning system 10 performs voltage and thermal data analysis of TRA event 54 at the initiation of cell short circuit 44, and during the period 50 of cell voltage 46 decreasing and rapidly transitioning, as the cell surface temperature rises. For a more detailed description, refer to [reference needed]. Figure 3 .

[0042] refer to Figure 3 And refer again Figure 2First, it should be noted that the thermal runaway detection method and warning system 10 are operable when the vehicle 30 is confirmed to be stopped (including when the vehicle 30 is in parking mode, and when the vehicle 30 is, for example, at a traffic light or stop sign). When the vehicle is not running, the cells (e.g., first cell 14, second cell 16, third cell 60 up to sixteenth cell 18) of each different module (e.g., first module 12) are queried at predetermined time intervals or sampling rates. For example, each battery cell is queried at a sampling rate of 0.05 seconds. During each query of each cell, the open-circuit cell voltage V is measured. oc (t). In the presented example, the open-circuit cell voltage V of the first cell 14 is obtained. oc1 (t)62.

[0043] After obtaining the open-circuit cell voltages of all cells (e.g., cell 14, cell 16, cell 60 up to cell 18), in the average value determination step 64, the average value of all open-circuit cell voltages is calculated, and the minimum open-circuit cell voltage is subtracted from the average value. Determining the average open-circuit cell voltage and subtracting the minimum open-circuit cell voltage allows the elimination of the influence of measured cell voltages that differ significantly from the remaining cell voltages. To obtain the average open-circuit cell voltage, the following equation 1 is used:

[0044] Equation 1:

[0045] Vmean(t)=1 / (n-1){[∑1^nVoc_i(t)]-min(Voc_i(t))}

[0046] Where: n = number of units.

[0047] The battery cell voltage needs to be removed using Equation 1 above only during initial parking or parking, and for approximately 2 to 3 minutes after initial parking or parking. After the removal period for the remaining battery, the minimum value V... oc (t) is essentially equal to the minimum unit voltage, and the average voltage can be set to be essentially zero.

[0048] After step 64 of determining the average value is completed, or if the average voltage is substantially equal to zero, the derivative 66 for each voltage is determined in the derivative step 68, where the derivative of the measured cell voltage with respect to time is calculated for each cell using Equation 2:

[0049] Equation 2:

[0050] dV oci (t) / dt.

[0051] Then, in buffering step 70, the results of Equation 2 for each cell are buffered. Buffering step 70 may apply a moving buffer window with a point buffer 72 having, for example, a cell voltage of 100. The cell voltage sampling rate may be, for example, t = 0.05 seconds. At each increment, the first point of buffer 72 is discarded, and a new point is added. Buffer 72 is a calibration of the trade-off between detection time and robustness, because a smaller buffer size allows for faster detection of problems, while a larger buffer size provides more accurate judgment.

[0052] After applying each buffer, a fast Fourier transform is performed in power spectrum calculation step 74 to calculate the power spectrum using Equation 3:

[0053] Equation 3:

[0054]

[0055] in In frequency The power spectral density at that location.

[0056] Then, using the power spectrum values ​​calculated above, Equation 4 is used to calculate the energy ΔE released at sampling time k:

[0057] Equation 4:

[0058] Where n < 100

[0059] It is the noise term removed from the energy calculation, where i is defined as a predetermined frequency threshold, and frequencies above the predetermined frequency threshold are considered noise values ​​that need to be removed from the energy calculation.

[0060] refer to Figure 4 And refer again Figure 2 and Figure 3 Graph 78 shows the typical range of the derivative of the cell voltage 80 as a function of time 82 (in seconds). The perturbation in the derivative of cell voltage 84 indicates the cell short circuit occurring at approximately 407 seconds, and the peak of derivative 86 occurring at approximately 415 seconds.

[0061] refer to Figure 5 And refer again Figures 1 to 4 Graph 88 shows the relationship between the calculated power spectrum 90 and time 92 (in seconds). The power spectrum 90 can be used to set alarm conditions after a thermal runaway event is detected. For example, at a power spectrum value 96 approximately 0.6 seconds after the cell short circuit begins at 407 seconds, the alarm threshold 94 is exceeded.

[0062] refer to Figure 6 And refer again Figures 1 to 5The graph 98 illustrates the runaway state caused by sudden changes in unit voltage during vehicle parking and charging. Graph 98 shows the relationship between unit voltage 100 and time 102 (in seconds). The normal or good unit voltage 104 is essentially flat over time. Compared to the good unit voltage 104, the runaway unit voltage 106 begins with an ISC (internal short circuit) 108 detected at 2601.5 seconds, followed by a rapid voltage disturbance, and then a TRA 110 occurs at 2607.9 seconds.

[0063] refer to Figure 7 And refer again Figures 1 to 6 During charging operation Figure 6 The situation is presented in graph 112, which shows the relationship between the power spectrum 114 and time 116 (in seconds) when a sampling rate of 50 ms is applied. Similar to... Figure 5 The power spectrum 114 can be used to set alarm conditions after detecting ISC and thermal runaway events. For example, approximately 2 seconds after an ISC (internal short circuit) is detected at 2601.5 seconds, the power spectrum value 120 exceeds the alarm setting threshold 118, and a TRA occurs at approximately 2607.9 seconds.

[0064] When the alarm threshold is exceeded, several actions may be taken or may occur. These actions may include: 1) sending a signal to stop charging operation; 2) opening the battery pack pressure relief valve to release battery pack pressure; 3) sending a signal to initiate mitigation, which, depending on several aspects, includes allowing cooling coolant to begin flowing into the battery pack; 4) sending a signal to convey a warning via the operator's smartphone; 5) sending a signal to a cloud-based vehicle communication system, which may, for example, automatically alert emergency services such as 911; and 6) calling fire services.

[0065] The power spectrum of battery voltage can be calculated using the Discrete Wavelet Transform (DWT) of a time-series signal instead of the Functional Fiber Transform (FFT). Compared to FFT, DWT decomposes the signal into multiple time-domain scales using scalable wavelet filters. These wavelet filters can be scaled to separate specific signal frequencies that are pre-specified as noise. DWT provides a tool for detecting signal abrupt changes that may occur during TRA events. Multiple wavelets could be potential candidates for this application.

[0066] refer to Figure 8 And refer again Figures 1 to 7DWT can be applied as follows: The unit voltage signal passes through a wavelet filter 122, which is then scaled by a scaling function 124. In the scaling function 124, the input signal 126 received from the wavelet filter 122 is segmented and scaled into a first high-frequency component in a first high-frequency extraction section 128, and scaled into a first low-frequency component in a first low-frequency extraction section 130. The low-frequency component signal from the first low-frequency extraction section 130 is halved and used to generate first-level coefficients 132. The high-frequency component generated by the first high-frequency extraction section 128 is halved and forwarded to a second high-frequency extraction section 134, where the signal is scaled into a second high-frequency component, and then scaled into a second low-frequency component in a second low-frequency extraction section 136. The low-frequency component signal from the second low-frequency extraction section 136 is halved and used to generate second-level coefficients. The high-frequency component generated by the second high-frequency extraction section 134 is halved and forwarded, if necessary, to additional high-frequency and low-frequency extraction sections to generate third-level coefficients if needed. Therefore, DWT is used to extract high-frequency noise components from unit voltage signals.

[0067] The DWT coefficients can be calculated using the following equations 5 and 6:

[0068] Equation 5:

[0069]

[0070] Equation 6:

[0071]

[0072] refer to Figure 9 And refer again Figures 1 to 8 Graph 138 illustrates the thermal runaway conditions during battery pack charging operations. Graph 138 shows the relationship between cell voltage 140 and time 142 (in seconds). Cell voltage curve 144 indicates the onset of a short circuit 146 at 2600 seconds. After the short circuit 146 occurs, cell voltage 144 oscillates or changes, and cell venting 148 begins approximately 0.2 seconds after the short circuit 146. After the voltage oscillation ends, cell voltage 144 drops significantly, and at voltage point 150 at approximately 2607 seconds, the battery cell fully enters the combustion phase.

[0073] refer to Figure 10 And refer again Figures 1 to 9Graph 152 presents the decomposed signal of the TRA cell voltage level 1 identified using DWT during charging operation. Graph 152 shows the change of voltage noise level 154 over time 156 (in seconds). Noise level threshold 158 identifies the boundary of normal cell noise level. A cell short circuit 160 is detected at approximately 2601 seconds, indicated by the noise level exceeding threshold 158. TRA 162 begins at approximately 2607 seconds.

[0074] The thermal runaway detection method and warning system 10 disclosed herein can be applied as follows: a mutation detection method for unit voltage derivative time series; using FFT to estimate the power spectrum of the voltage derivative in a moving window; using RMS to estimate the power spectrum of the voltage derivative in a moving window; detecting thermal runaway using an array of multiple diagnostic thresholds; applying a mutation detection method for unit voltage derivatives; and estimating the energy of the discrete wavelet transform (DWT) spectrum of the voltage derivative.

[0075] The thermal runaway detection method and warning system 10 disclosed herein offer several advantages. These include an early detection method that can detect potential failure modes in less than 1-2 seconds. Early detection can be used to provide immediate warnings and can activate thermal runaway mitigation systems.

[0076] The description in this disclosure is merely exemplary in nature, and variations thereof that do not depart from the spirit and scope of this disclosure are intended to fall within its scope. Such variations should not be considered as departing from the spirit and scope of this disclosure.

Claims

1. A method for detecting thermal runaway in a unit, comprising: Positioning a battery pack with multiple cells in a motor vehicle; The cell voltage of the plurality of cells is measured at a predetermined sampling rate; A scalable wavelet filter based on discrete wavelet transform is used to decompose the unit voltage signal into multiple signal time-domain scales and generate coefficients characterizing voltage changes. Based on the analysis of the discrete wavelet transform, abrupt changes in the cell voltage signal are identified to determine whether the cell voltage decreases and changes to indicate the start of a cell short circuit when or before the cell surface temperature rises, thereby indicating the start of a thermal runaway event.

2. The method of claim 1, further comprising confirming that the motor vehicle has stopped before measuring the unit voltage.

3. The method of claim 2, further comprising: Determine whether the unit threshold temperature, which is limited to approximately 70°C, has been reached; as well as Determine whether the cell temperature rapidly rises to approximately 500°C within approximately 5 seconds after the cell temperature reaches the threshold temperature when the thermal runaway event of the battery cell occurs.

4. The method of claim 3, further comprising determining the average value of all cell voltages of the plurality of cells by subtracting the minimum cell voltage from the sum of the cell voltages.

5. The method of claim 4, further comprising: Calculate the derivative of the unit voltage of each of the plurality of units with respect to time; as well as The derivative of the unit voltage is passed to the buffer.

6. The method of claim 5 further includes calculating the fast Fourier transform power spectrum by applying the derivatives of the unit voltages of the plurality of units with respect to time.

7. The method of claim 6, further comprising using the power spectrum to calculate the released energy.

8. The method of claim 7, further comprising: After a unit short circuit begins, determine whether the power spectrum exceeds a predetermined threshold. as well as An alarm is triggered after the predetermined threshold is exceeded.

9. The method of claim 8, further comprising, after activating the alarm, initiating at least one of the following actions: Stop the charging operation of the battery pack; Release pressure on the battery pack; Allow coolant to flow into the battery pack; Warnings are relayed to vehicle operators via smartphones; Transmit the status of the battery pack to a cloud-based remote security service; and Contact emergency services.

10. The method of claim 1, further comprising continuing to analyze the cell voltage and cell surface temperature after the cell short circuit begins and during the period of cell voltage decrease and rapid cell voltage transition, and as the cell surface temperature increases.

Images