A method for screening and monitoring abnormal batteries in a vehicle battery system

Abstract

The present invention provides a method for screening and monitoring abnormal batteries in a vehicle battery system: according to the voltage at the extraction time, the lowest voltage, and the lowest voltage label, the average voltage of all battery modules and the characteristic parameters of each battery module at the current moment are calculated. The battery modules with the lowest voltage are grouped, and then the average value of the characteristic parameters of each battery module between the two moments is calculated. The average value result is compared with the threshold value, and if it exceeds, the battery module is abnormal; or, the average value of the characteristic parameter and the operating and charging total energy ratio of the battery module are projected into a two-dimensional coordinate system, and the number of the abnormal battery module is obtained according to the abnormal coordinates. The advantage is that the module average and the lowest voltage difference under each battery module number based on the charging data are used as the characteristic value, and the characteristic value comparison on the time scale is combined with the number of charging times as the data segment, so that abnormal battery modules with hidden dangers can be identified, the vehicle battery module status can be effectively and timely analyzed, and enhanced early warning measures can be adopted to avoid accidents.

Description

Technical Field

[0001] The present invention relates to the field of vehicle batteries, and in particular to a method for screening and monitoring abnormal batteries in a vehicle battery system. Background Art

[0002] With the rapid increase in the penetration rate of new energy vehicles, the safety monitoring and detection of new energy vehicles is becoming more and more important. According to current statistics, spontaneous combustion accounts for a high proportion of new energy vehicle safety accidents. In spontaneous combustion accidents, 41% of vehicles are in driving state, 40% of vehicles are stationary, and 19% of vehicles are in charging state. Therefore, it is very important to accurately detect and warn the batteries of vehicles in use.

[0003] There are many reasons for vehicle spontaneous combustion, such as spontaneous combustion in a stationary state, in a driving state, and in a charging state. The spontaneous combustion conditions are different, and the triggering conditions are different. It is impossible to reproduce the accident after spontaneous combustion, and the cause analysis is very difficult. The existing monitoring methods are divided into two categories: real-time monitoring and remote data analysis monitoring. Real-time monitoring mainly performs threshold detection by analyzing the direct information collected by vehicle data.

[0004] The prior art provides some methods for battery testing, including:

[0005] The Chinese patent "CN 112622690A A monitoring method and system for safe charging of power batteries" provides a monitoring method and system for safe charging of power batteries. The charging parameters of the rechargeable battery are collected at regular intervals; the changing trend of the charging parameters is monitored; when an abnormality is detected in the changing trend of the charging parameters, abnormal processing is performed. The charging parameters of the rechargeable battery are actively obtained in an active manner, and the accident risk of the battery is evaluated through the parameter change process. The charging parameters include one or more of the maximum temperature value, the maximum single cell voltage value, the current voltage value, and the current current value of the rechargeable battery.

[0006] The Chinese patent "CN 113291200A A vehicle battery assembly monitoring method and device" provides a vehicle battery assembly monitoring and device. After the vehicle has traveled a certain mileage, the battery cells are charged and discharged with high-frequency pulses, and the temperature rise data of each battery cell is collected. Based on the temperature rise data, the current temperature characteristic value corresponding to each battery cell is generated. When the current temperature characteristic value is detected to be abnormal, it is determined that the battery assembly has a thermal runaway risk.

[0007] The Chinese patent "CN 111474488 A A method and system for detecting micro-internal short circuits in lithium-ion batteries" provides a method and system for detecting micro-internal short circuits in lithium-ion batteries. Perform cyclic constant current charge and discharge on lithium-ion batteries that have not experienced micro-internal short circuits, perform dynamic impedance testing on lithium-ion batteries, and obtain test dynamic impedance data. Test the dynamic impedance of lithium-ion batteries that have experienced micro-internal short circuits, and obtain dynamic impedance data. Determine the slope standard of the lithium-ion battery micro-internal short circuit straight line within a preset test frequency range, obtain the dynamic impedance of the lithium-ion battery to be tested as detection data, and determine the slope of the detection data. When the slope is less than or equal to the slope standard, it is determined that a micro-internal short circuit has occurred in the lithium-ion battery to be tested.

[0008] The Chinese patent "CN 113609791 A Lithium-ion Battery Energy Storage Power Station Active Safety Monitoring and Early Warning Method and System" provides an active safety monitoring and early warning method and system involving lithium-ion battery energy storage power stations. Simulate the normal charging and discharging process and the process of potential faults in the energy storage power station battery, obtain the operating environment data and the data collected by the BMS as characteristic parameters, and establish an original sample database. The BP neural network based on the LM algorithm establishes a local decision model for each characteristic parameter, normalizes the dimensions of each characteristic parameter in the original sample database, obtains a sample database of the local decision model of each parameter, and trains the local decision model of each parameter based on this. The local decision models of each parameter are integrated to establish a comprehensive early warning model for energy storage battery hidden dangers based on the LSTM algorithm. The model is trained using the sample database, firstly establishing a local decision model for the parameters of each characteristic quantity, and then integrating each local decision model to establish a comprehensive early warning model for battery hidden dangers in the energy storage power station.

[0009] However, the above technical solution is to analyze the historical data of vehicle operation through remote data analysis and monitoring to extract abnormal features. Due to hardware limitations, real-time detection can only monitor a small amount of data for abnormalities. Although battery spontaneous combustion appears to be a sudden abnormal situation, its essence is often that the quantitative change of the battery itself causes a qualitative change. Although remote data analysis and monitoring can realize the analysis of historical data, the existing methods basically monitor abnormalities by comparing the voltage and temperature of different batteries, or monitor through complex big data analysis and model prediction. How to provide a method that can monitor and detect battery abnormalities based on the data changes during battery charging has become an urgent problem to be solved. Summary of the invention

[0010] The present invention provides a method for screening and monitoring abnormal batteries in a vehicle battery system, which is used to solve the problem that the prior art cannot monitor and detect battery abnormalities based on data changes during battery charging, which has become an urgent problem to be solved. The purpose of the present invention is to screen abnormal batteries in a battery system based on historical vehicle operation data by extracting derived variables as characteristic parameters, especially for screening in the case of early internal short circuits, thereby extracting batteries with hidden dangers and then issuing early warnings.

[0011] In order to achieve the above-mentioned purpose, the technical solution of the present invention provides a method for screening and monitoring abnormal batteries in a vehicle battery system, including: extracting the charging segment data of the vehicle: voltage at a certain moment, minimum voltage, and minimum voltage label. Calculate the average voltage of all battery modules in the vehicle. Calculate the characteristic parameters of each battery module at the current moment. Group the battery modules with the lowest voltage, calculate the average value of the characteristic parameters of each battery module between two moments after grouping, and project the average value of the characteristic parameters discretely into a two-dimensional coordinate system. After calculating the operating and charging total energy ratio of the battery module, project the calculation result into a two-dimensional coordinate system. Obtain the number of the abnormal battery module according to the abnormal coordinates in the two-dimensional coordinate system.

[0012] As a preferred embodiment of the above technical solution, preferably, calculating the average voltage of all battery modules in the vehicle includes:

[0013]

[0014] V i-t : The voltage of the i-th battery module at time t, V avg-t is the average voltage.

[0015] As a preferred embodiment of the above technical solution, preferably, the characteristic parameters of all the battery modules at the current moment are calculated, including:

[0016] Record the battery module number i with the lowest voltage, and extract the lowest voltage V min-t , and the average voltage V avg-t , and obtain the characteristic parameters α of each battery module i-t :

[0017] α i-t =V avg-t -V min-t .

[0018] As a preferred embodiment of the above technical solution, the battery modules with the lowest voltage are grouped, and after grouping, the average value of the characteristic parameters of each battery module between two moments is calculated, including: after counting the number of times the lowest voltage appears in each battery module in each time period, each battery module is grouped; and the average value of the characteristic parameters of the battery modules in each group in each time period is calculated.

[0019]

[0020] Among them, j is the discrete representation of the time period, and the battery module that does not have the lowest voltage is considered to have the average value of its characteristic parameters

[0021] As a preferred embodiment of the above technical solution, preferably, calculating the total energy ratio of the operation and charging of the battery module includes:

[0022]

[0023] Among them, ratio_c / d i is the total energy ratio of the operation and charging of the i-th battery module, Energy i-c_total : Total charging energy of battery module No. i; Energy i-d_total : Total operating energy of battery module No. i.

[0024] As a preferred embodiment of the above technical solution, preferably, the serial number of the abnormal battery module is obtained according to the abnormal coordinates in the two-dimensional coordinate system, including: projecting the average values ​​of the characteristic parameters of all battery modules and the total energy ratio of operation and charging into the two-dimensional coordinate system as vertical and horizontal coordinates; scaling and clustering the two-dimensional data summarized in the two-dimensional coordinate system to obtain the classification results, and obtaining the serial number of the abnormal battery module according to the classification results.

[0025] The technical solution of the present invention provides a method for extracting the charging segment data of a vehicle: the voltage at the moment, the lowest voltage, the lowest voltage label, and calculating the average voltage of all battery modules in the vehicle. Calculate the characteristic parameters of all battery modules at the current moment. Group the battery modules with the lowest voltage, calculate the average characteristic parameter of each battery module between two moments after grouping, and project the average characteristic parameter and the operating and charging total energy ratio of the battery module into a two-dimensional coordinate system. Obtain the number of the abnormal battery module according to the abnormal coordinates in the two-dimensional coordinate system.

[0026] The advantage of the present invention is that it uses the average and minimum voltage differences of all modules under each battery number in the charging data as characteristic values, and compares the characteristic values ​​on a time scale using the number of charging times as the data segment, so as to identify abnormal batteries with hidden dangers, analyze the battery status of the vehicle more effectively and timely, and adopt enhanced early warning measures to avoid accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0028] Figure 1a Flowchart 1 of the technical solution of the present invention.

[0029] Figure 1b The process of the technical solution of the present invention Figure 2 .

[0030] Figure 2 It is a two-dimensional schematic diagram of the average value data of the total energy ratio of operation and charging and characteristic parameters in the technical solution of the present invention.

[0031] Figure 3 for Figure 2 Schematic diagram of the data results after scaling and clustering. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] Most of the natural occurrences in electric vehicles are caused by abnormal cells in the battery system. A large part of the causes of natural occurrences in batteries are caused by short circuits in the battery, including the generation of foreign matter and lithium dendrites. This state often occurs when the battery is in a high SOC state (state of charge), and the volume expansion of the high SOC battery makes the battery more susceptible to internal short circuits. Since the vehicle battery system is used under a single and consistent condition during the charging process, the present invention uses the charging segment data in the vehicle history data to extract characteristic parameters.

[0034] The technical solution of the present invention is now described in detail, referring to the process Figure 1a and Figure 1b :

[0035] Step 101: extract the charging segment data of the vehicle, including the voltage at the moment, the lowest voltage, and the lowest voltage number.

[0036] Specifically, the voltage of the battery module at time t is V i-t ; The lowest voltage at time t is Vmin-t ; The lowest voltage label is the label of the battery module when the lowest voltage appears at time t, and the specific label is N min-t .

[0037] Step 102: Calculate and record the average voltage of all battery modules in the vehicle.

[0038]

[0039] Among them, V avg-t is the average voltage of all battery modules, and N is the battery module number.

[0040] Step 103: extract the lowest voltage at time t and record the corresponding battery module number N.

[0041] Specifically, extract the minimum voltage V min-t , and record the number i of the battery module with the lowest voltage.

[0042] Step 104: Calculate the characteristic parameter α of each battery module at time t i-t .

[0043] α i-t =V avg-t -V min-t

[0044] α i-t It is the difference between the lowest voltage and the average voltage of battery i at time t.

[0045] Wherein, i represents the number of each battery in the N battery modules.

[0046] Step 105 , grouping the battery modules according to the number of times the lowest voltage appears.

[0047] For example, the first battery module has low voltage at time t2, the second battery module has low voltage at time t1, t3, t4, and the third battery module has low voltage at time t5 and t6. Then the data of the first battery module at time t2 is taken as group 1, the data of the second battery module at time t1, t3, t4 is taken as group 2, and the data of the third battery module at time t5 and t6 is taken as group 3.

[0048] Step 106: Calculate the average value of the characteristic parameters of the battery modules in each group in each time period.

[0049]

[0050] Continuing with step 105, for group 1, the average value is the low voltage at time t2; for group 2, the average value is the average value of the low voltages at times t1, t3, and t4; for group 3, the average value is the average value of the two low voltages at times t5 and t6.

[0051] Since the time in data collection is a discrete point, j is used to represent the discrete time periods between t1-t2, t2-t3...tn-1-tn, j = 0, 1, 2, 3...n.

[0052] If a battery module has never reached the minimum voltage, it is considered that the average value of its characteristic parameters

[0053] After step 106, step 1071 or step 1072 may be performed:

[0054] Step 1071, determine whether the average value of the characteristic parameter is higher than the threshold value, if so, the battery module is abnormal.

[0055] Step 108: Obtain the abnormal battery module number.

[0056] like Figure 1b As shown, step 1072, calculate the total energy ratio of operation and charging of each battery module.

[0057]

[0058] Among them, ratio_c / d i is the total energy ratio of the operation and charging of the i-th battery module, Energy i-c_total : Total charging energy of battery module No. i; Energy i-d_total : Total operating energy of battery module No. i.

[0059] Step 109: Project the operating and charging total energy ratio of each battery module and the average value of the characteristic parameters into a two-dimensional coordinate system.

[0060] The ratio of total energy of operation to charging is the horizontal axis, and the average value of characteristic parameters is the vertical axis.

[0061] Step 110: scaling and clustering the two-dimensional data to obtain classification results.

[0062] Normalizer scaling preprocessing was used for the two-dimensional data, and the KMeans 2-class clustering method was used to classify the data. The classification results were 0 and 1, with 0 representing normal and 1 representing abnormal.

[0063] Step 111, obtaining the abnormal battery module number.

[0064] It should be noted that a battery module may have one battery or multiple batteries. When a battery module has multiple batteries, the number of the abnormal battery module is finally obtained; when a battery module has only one battery, the number of the abnormal battery is finally obtained.

[0065] Now, a practical embodiment is used to illustrate the technical solution of the present invention:

[0066] Take the historical operation data of a new energy accident vehicle as an example, and take all the historical data as a whole. According to steps 101 to 107, calculate the average value of the difference between the average voltage of all modules and the lowest voltage each time the lowest voltage appears under the charging state of each numbered battery, and the charge-discharge energy ratio of each numbered battery, and record the corresponding battery number. The calculation data is as follows:

[0067]

[0068]

[0069]

[0070]

[0071] Will The average value of the difference between the average voltage of all batteries and the lowest voltage is compared with the threshold value of 20mV. Figure 2 It can be seen that the module numbers that exceed the threshold are 13, 45, 52, 54, and 71. This includes module No. 45, which has thermal runaway.

[0072] Extract two-dimensional data y: x:ratio_c / d i , Normalizer scaling preprocessing is used for the two-dimensional data, and KMeans 2-class clustering method is used to classify the data and obtain the abnormal battery number. Figure 3 The results show that the abnormal battery numbers are 13, 45, 52, 54, and 71.

[0073] In summary, the present invention provides a method for screening and monitoring abnormal batteries in a vehicle battery system to achieve:

[0074] 1. Based on the vehicle charging segment data, the data segment includes the constant voltage charging segment, the high SOC interval, and the shutdown charging segment. The use environment is single and the data reliability is strong;

[0075] 2. Based on the average and minimum voltage difference of all battery modules as the characteristic variable, it can not only detect voltage abnormalities caused by abnormal micro-short circuits inside the battery, but also avoid over-identification caused by abnormal data interference through the averaging method.

[0076] 3. By obtaining the average and minimum voltage difference of all modules under each battery number for all charging data, combined with the operation and total charging energy ratio of each battery (module), Normalizer scaling preprocessing is used, and the KMeans 2-class clustering method is used to classify the data to obtain the abnormal battery number.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for screening and monitoring abnormal batteries in a vehicle battery system, characterized in that: include: Extract characteristic parameters from the charging segment data in the vehicle history data, including the voltage at the moment, the lowest voltage, and the lowest voltage number; Calculating the average voltage of all battery modules in the vehicle, including, V i-t : The voltage of the i-th battery module at time t, V avg-t is the average voltage; Calculate the characteristic parameters of all battery modules at the current moment, including recording the battery module number i with the lowest voltage, and according to the extracted lowest voltage V min-t , and the average voltage V avg-t , and obtain the characteristic parameters α of each battery module i-t : α i-t =V avg-t -V min-t ; The battery modules with the lowest voltage are grouped, and after grouping, the average value of the characteristic parameters of each battery module between two moments is calculated, including: after counting the number of times the lowest voltage appears in each battery module in each time period, each battery module is grouped; and the average value of the characteristic parameters of the battery modules in each group in each time period is calculated. Among them, j is the discrete representation of the moment, and the battery module that does not have the lowest voltage is considered to have the average value of its characteristic parameters After discretely projecting the average value of the characteristic parameter into a two-dimensional coordinate system, calculating the total energy ratio of the operation and charging of the battery module, and projecting the calculation result into the two-dimensional coordinate system; The number of the abnormal battery module is obtained according to the abnormal coordinates in the two-dimensional coordinate system, so as to obtain the battery with the hidden danger of internal short circuit; specifically, the average value of the characteristic parameters of all battery modules and the total energy ratio of operation and charging are projected into the two-dimensional coordinate system as the vertical and horizontal coordinates, including: The calculation of the total operating and charging energy ratio of the battery module includes: Among them, ratio_c / d i is the total energy ratio of discharge and charge of battery No. i, Energy i-c_total : Total charging energy of battery No. i; Energy i-d_total : Total operating energy of battery No. i; Scaling and clustering the two-dimensional data summarized in the two-dimensional coordinate system to obtain a classification result, and obtaining the number of the abnormal battery module according to the classification result; including extracting the two-dimensional data y: α i , x: ratio_c / d i ,The Normalizer scaling preprocessing was used on the two-dimensional data, and the KMeans 2-class clustering method was used to classify the data and obtain the abnormal battery numbers.

Images