Power battery multi-fault risk assessment and early warning method

Abstract

The invention discloses a multi-fault risk assessment and early warning method for a power battery, and the method comprises the steps: requesting the recent life cycle data of an in-service electric vehicle power battery from a cloud server, carrying out the data preprocessing, and obtaining a temperature value of any single battery in the power battery through an interpolation method; differentiating the voltage value and the temperature value of any battery monomer to obtain a fault discrimination signal; establishing a multi-fault risk discrimination model based on the fault discrimination signal to calculate absolute fault degree scores corresponding to different faults of the battery; performing statistics on data distribution of various battery faults and fault degree scores, determining a fault grading range according to a distribution result, and drawing a multi-fault risk early warning graph; and carrying out early warning on results in a specific fault grading range according to severity. According to the method, various faults of the power battery can be evaluated and separated, battery maintenance and overhaul reference can be provided for a vehicle owner in time according to the multi-fault risk early warning graph, calculation is simple, deployment is convenient, and the method has practical application value.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-fault assessment and early warning of power battery systems, and in particular to a multi-fault risk assessment and early warning method for power batteries of electric vehicles. Background Art

[0002] As the sales volume of new energy vehicles increases, their safety issues are becoming increasingly apparent, especially the problems of smoke, spontaneous combustion, and explosion of on-board power batteries, which not only increase the travel anxiety of car owners and passengers, but also seriously threaten the safety of public life and property, and restrict the sustainable development of the new energy vehicle industry. Therefore, power battery fault assessment and early warning are particularly important for car owners and car companies.

[0003] At present, the methods for power battery fault diagnosis are mainly based on two categories: model-driven and data-driven. Model-driven methods can stably track fault signals under idealized approximate conditions, but they are difficult to adapt to the complex conditions in the actual use environment. On the other hand, although data-driven methods are the current mainstream solution, they are still difficult to handle the multiple types of concurrent faults that exist in actual situations. Therefore, it is urgent to propose a power battery multi-fault risk assessment and early warning method. Summary of the invention

[0004] The purpose of the present invention is to solve the deficiencies in the above-mentioned technology and provide a method for multi-fault risk assessment and early warning of power batteries. Based on the recent life cycle data of power batteries in the cloud data center, the median difference of the voltage and temperature values ​​of each battery cell in the power battery is calculated to reveal different fault characteristics. In the median difference data space, the statistical characteristics of battery cell data with different fault degrees are significantly different. Therefore, a multi-fault risk discrimination model for power batteries can be constructed based on statistical characteristics to realize multi-fault risk assessment and early warning of any battery cell in the power battery.

[0005] The purpose of the present invention can be achieved by adopting the following technical solutions: A method for risk assessment and early warning of multiple faults of electric vehicle power batteries, the method comprising: S1. Request the recent life cycle data of the power battery of the in-service electric vehicle from the cloud server, perform data preprocessing, and obtain the temperature value of any battery cell in the power battery system by interpolation method; S2. Differentiate the voltage value and temperature value of any battery cell in the power battery system to obtain a fault discrimination signal; S3. Construct a multi-fault risk discrimination model based on the fault discrimination signal to comprehensively evaluate the battery faults, including abnormal inter-battery connection of the power battery system, voltage sampling distortion, abnormal battery self-discharge, sudden internal short circuit fault, voltage inconsistency, and temperature inconsistency, and calculate the absolute fault degree scores corresponding to the above battery faults; S4. Count the data distribution of different battery failures and their absolute failure degree scores, determine the warning level and failure level based on the distribution results, and draw a multi-failure risk warning map; S5. In accordance with the preset method, the vehicle owner is prompted with the situations and treatment methods at different warning levels and fault levels in the multi-fault risk warning map, so as to achieve the multi-fault warning goal of the power battery.

[0006] Furthermore, the process of step S1 is as follows: S101. Requesting power battery life cycle data from the cloud big data center, where the data fields include: collection time, vehicle status, charging status, operating mode, vehicle speed, total voltage, total current, cumulative mileage, state of charge, battery cell voltage value, battery cell voltage maximum value, battery cell code with the highest voltage, battery cell voltage minimum value, battery cell code with the lowest voltage, maximum temperature value, maximum temperature subsystem number, maximum temperature probe serial number, maximum alarm level, probe temperature, and general alarm flag; S102, sorting the data in order according to the acquisition time, separating the data of different data fields, filtering the impulse noise in the data by using mean filtering and median filtering, and then using the least squares method to interpolate the missing data; S103. According to the layout of the limited temperature probes in the power battery system, the known temperature probe data are weightedly integrated by interpolation method to obtain the temperature value of any battery cell in the power battery system, so as to compensate for the sparsity of the temperature data in the power battery system.

[0007] Furthermore, the process of step S2 is as follows: S201, load all battery cell voltage values ​​in the power battery system and calculate the median differential voltage, calculate the median differential temperature based on the battery cell temperature value, and generate a fault discrimination signal. The fault discrimination signal highlights the data differentiation of the battery cells, forms a high-resolution data mode, and helps to extract the data characteristics of different battery faults. Calculate by the following method: exist At this moment, the power battery system Voltage value of each battery cell and temperature value Forming a battery cell voltage sequence and battery cell temperature series , and the The median differential voltage is obtained by subtracting the voltage value of each battery cell from the median of the battery cell voltage sequence. , let the The temperature value of each battery cell is subtracted from the median of the battery cell temperature sequence to obtain the median differential temperature. When the data of the tested battery cell is normal, and The closer the value is to the 0 value line; otherwise, there is data anomaly. and The value is far away from the 0 value line; S202, analyzing the data performance of the fault discrimination signal in the charging and discharging intervals, and extracting data features of different battery faults in the median difference data space.

[0008] Furthermore, the process of step S3 is as follows: S301, for battery voltage sampling distortion and battery connection abnormality, perform logical mutual exclusion judgment: divide the charging interval and the discharging interval in a charge and discharge segment, according to Slice out the first The median differential voltage of each battery cell and the discharge interval The median differential voltage of each battery cell , and calculate the arithmetic mean of the extreme values and and the variance of the time series difference statistics and , the arithmetic mean without extreme values and The calculation process is as follows: In the above formula and Respectively represent The maximum median differential voltage and the minimum median differential voltage within the charging interval of each battery cell, and Respectively represent The maximum median differential voltage and the minimum median differential voltage within the discharge interval of each battery cell, and They are the time when charging starts and ends during a single continuous charge and discharge process. and are the time when the discharge starts and ends respectively. The calculation process of the statistical variance of the time series difference of the extreme value is as follows: In the above formula and Respectively represent The time dimension of each battery cell and The sequence after difference, operator Indicates that and After sorting, remove the first and last two extreme value samples to obtain the extreme value time difference sequence and , and finally calculate the statistical variance of the two sequences and The above statistical descriptive analysis results are used to characterize the central tendency and discrete degree of different battery faults in the charging and discharging range, extract and quantify the data modes in the range, and enhance the transparency of fault analysis.

[0009] If voltage sampling distortion occurs, most battery cells in the same module will have data distortion, while abnormal connection between batteries will cause data distortion of a few battery cells in the module and there will be differences in the charging process. In view of this feature, a mutually exclusive logic judgment process is designed. Greater than the charging distortion judgment threshold ,or Exceeding the discharge distortion judgment threshold The number of battery cells is greater than 1, and Below the overcharge threshold , voltage sampling distortion occurs. Exceed If the number of is equal to 1, the battery connection is abnormal, otherwise no fault occurs. If voltage sampling distortion occurs, the mutually exclusive symbol If the battery connection is abnormal, the mutual exclusion symbol is set to -1. is +1; In order to expand the distance between voltage sampling anomalies and battery connection anomalies in the data space, and thus more effectively achieve fault separation, the first Amplification factor of each battery cell , calculated by the following formula: Definition The data distortion fraction of each battery cell is , calculated by the following formula: In the above formula, Representation function , since the value range is [-1, +1], the data distortion score is normalized to the interval [-1, +1] by using this function; Data distortion score Characterizes the fault degree of voltage sampling distortion and abnormal connection between batteries. If the battery cell is healthy, the data distortion score is close to 0. If voltage sampling distortion occurs, it will diverge to the edge of -1. If abnormal connection occurs between batteries, it will diverge to the edge of +1. S302: For abnormal self-discharge of batteries and sudden internal short circuit faults, feature confrontation discrimination is performed: the abnormal self-discharge of battery cells is manifested in the data as a larger voltage drop, which exists in the entire charging and discharging process, and the median differential voltage expands to the negative domain, gradually outliers, and the consistency difference is obvious. and Calculate the Difference score of each battery cell Characterize and design attenuation factors To reduce the interference of healthy battery data, the calculation formula is as follows: Use median filter to The median differential temperature of each battery cell Filter to get , and in the time dimension Perform the difference to obtain the temperature time series difference sequence ; Definition The sudden internal short circuit fraction of each battery cell is , the specific calculation process is as follows: In the above formula For the The voltage variation factor of each battery cell, For the Temperature variation factor of each battery cell, correction factor Used to adjust the impact of data distortion on sudden internal short circuits; It means to find the maximum value of the sequence. Express request The battery with a sudden internal short circuit has a voltage drop during the charging stage, and is accompanied by data distortion due to the coupling of different battery failures. Introducing distortion in the representation data , adjust the correction factor To control data distortion Under the temperature control system of the battery system, the temperature of a normal battery is controllable. When an internal short circuit occurs in the battery, the temperature rises suddenly, which in turn induces spontaneous combustion and explosion. Therefore, the maximum value of the temperature time series difference is used to characterize the most significant abnormal temperature change of the battery. The role of the function is to strengthen the impact of short circuits within the burst and weaken the interference caused by other data features.

[0010] right and Difference The mutation score of each battery cell is The calculation process is as follows: Mutation score Indicates the degree of battery self-discharge abnormality and sudden internal short circuit failure. Used to normalize the mutation score to the range of [-1, +1]. If the battery cell is healthy, the mutation score is close to 0. If a self-discharge abnormality occurs, it will diverge to the edge of -1. If a sudden internal short circuit occurs, it will diverge to the edge of +1. S303: For the voltage inconsistency and temperature inconsistency of the battery, define The voltage inconsistency score of each battery cell is , No. The temperature inconsistency fraction of each battery cell is , the calculation process is as follows: In the above formula The sensitivity adjustment factor representing the voltage inconsistency fraction, The sensitivity adjustment factor represents the temperature inconsistency fraction. The voltage inconsistency of the battery is The temperature consistency of the battery is proportional to Inversely proportional to the value of the battery cell inconsistency.

[0011] S304, taking the absolute value of the fault degree score calculated above as the basis for fault severity assessment. The voltage value and temperature value of the battery cell are fully utilized to construct a universal fault discrimination signal reflecting multiple types of battery faults, and the data characteristics of multiple types of faults are discovered based on the signal, thus avoiding complex data transformation and improving the diagnostic efficiency.

[0012] Furthermore, the process of step S4 is as follows: S401, counting the frequencies of all battery cells under different fault types and absolute fault severity scores, and calculating the relative density to obtain the data distribution of each fault and its fault severity; S402, according to the data distribution result, and in accordance with the general principle that the number of normal battery cells is much greater than the number of faulty battery cells, the fault degree interval is divided to obtain the warning level and the fault level; S403, determine the fault classification range based on the warning level and the fault level, and draw a multi-fault risk warning map according to the fault type. Quantifying and visualizing battery faults is helpful for car owners to obtain clear and intuitive fault information.

[0013] Furthermore, the process of step S5 is as follows: S501, according to the multi-fault risk warning diagram, if the absolute fault degree score of the battery cell is within the corresponding fault classification range, then the battery cell is warned or judged as a fault, otherwise the battery cell is judged as safe; S502. Preset corresponding prompting methods for different warning levels and fault levels, and remind and warn the vehicle owner according to the preset prompting methods to achieve multiple fault warnings for the power battery, thereby avoiding potential safety accidents and expensive maintenance costs.

[0014] Compared with the prior art, the present invention has the following advantages and beneficial effects: 1. The present invention adopts an interpolation weighted fusion method to achieve the expansion from low-density temperature sampling to full battery cell temperature monitoring, thereby improving the recognition accuracy and reliability of multiple types of faults of battery cells under abnormal temperature conditions; 2. The present invention uses the voltage and temperature values ​​of the battery cells to construct a universal fault discrimination signal that can reflect multiple types of battery faults. Based on the signal, the data characteristics of multiple types of faults are discovered, which avoids complex data transformation and improves the diagnostic efficiency. 3. In the evaluation of multiple types of faults, a multi-fault risk discrimination model is constructed based on mathematical statistics. This model can quantify the fault degree of multiple types of faults of any battery cell in the power battery system. There is no need to analyze the complex internal mechanism of the physical battery. The calculation process is fast and efficient, and is suitable for online applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 is a flow chart of a method for multi-fault risk assessment and early warning of a power battery in an embodiment of the present invention; Figure 2 is an initial data graph in an embodiment of the present invention; Figure 3 is a data graph after preprocessing in an embodiment of the present invention; Figure 4 is a layout diagram of a power battery system in an embodiment of the present invention; Figure 5 The temperature calculation result of some battery cells in the embodiment of the present invention is shown in FIG. Figure 6 is a schematic diagram of a median differential voltage of an abnormal connection between batteries in an embodiment of the present invention; Figure 7 is a schematic diagram of a median differential voltage of voltage sampling distortion in an embodiment of the present invention; Figure 8 Schematic diagram of median differential voltage of abnormal battery self-discharge and inconsistent voltage in an embodiment of the present invention; Fig. 9 is a schematic diagram of the median differential voltage of a sudden internal short circuit fault in an embodiment of the present invention; Fig.10 is a schematic diagram of a median differential voltage with inconsistent temperatures in an embodiment of the present invention; Fig.11 It is a multi-fault risk assessment and early warning diagram in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.

[0018] Example This embodiment discloses a power battery multi-fault risk assessment and early warning method, the implementation process is as follows Figure 1 As shown, the method specifically comprises the following steps: S1. Request the recent life cycle data of the power battery of the in-service electric vehicle from the cloud server, perform data preprocessing, and obtain the temperature value of any battery cell in the power battery system by interpolation method; S2. Differentiate the voltage value and temperature value of any battery cell in the power battery system to obtain a fault discrimination signal; S3. Construct a multi-fault risk discrimination model based on the fault discrimination signal to comprehensively evaluate the battery faults, including abnormal inter-battery connection of the power battery system, voltage sampling distortion, abnormal battery self-discharge, sudden internal short circuit fault, voltage inconsistency, and temperature inconsistency, and calculate the absolute fault degree scores corresponding to the above battery faults; S4. Count the data distribution of different battery failures and their absolute failure degree scores, determine the warning level and failure level based on the distribution results, and draw a multi-failure risk warning map; S5. In accordance with the preset method, the vehicle owner is prompted with the situations and treatment methods at different warning levels and fault levels in the multi-fault risk warning map, so as to achieve the multi-fault warning goal of the power battery.

[0019] In this embodiment, the in-service data of 20 passenger electric vehicles is taken as an example to describe the implementation process of the power battery multi-fault risk assessment and early warning method. The specific implementation steps are as follows: S1. Request the recent life cycle data of the power batteries of in-service electric vehicles from the cloud server and perform data preprocessing.

[0020] In this step, the original format of the collected data can selectively comply with GB / T 32960.3-2016 "Technical Specifications for Electric Vehicle Remote Service and Management System Part 3: Communication Protocol and Data Format", including: collection time, vehicle status, charging status, operating mode, vehicle speed, total voltage, total current, cumulative mileage, state of charge, battery cell voltage value, battery cell voltage maximum value, highest voltage battery cell code, battery cell voltage minimum value, lowest voltage battery cell code, maximum temperature value, maximum temperature subsystem number, maximum temperature probe serial number, maximum alarm level, probe temperature, general alarm mark. As a preferred embodiment of the present invention, the electric vehicle power battery system contains a total of 92 battery cells and 34 temperature probes, and further preferably selects the collection time, charging status, total voltage, total current, battery cell voltage value and probe temperature value.

[0021] It should be noted that the data collected in step S1 is not limited to the above-mentioned data types, and other types of data may also be collected, such as battery cell impedance and battery cell current of the power battery, etc., which may be optimized according to actual needs.

[0022] Data disturbances may occur due to vehicle power failure, abnormal operation of the communication module, or interference with information collection, which may cause data coding loss or disorder. Figure 2 As shown, the data needs to be preprocessed. As a preferred embodiment of the present invention, the data preprocessing further selectively includes: sorting the above data according to the sampling time; separating the data according to the battery cell number and the probe temperature number; filtering and interpolating the screened data.

[0023] The screened data is filtered by using mean filtering and median filtering to filter the impulse noise in the data, and the missing values ​​are interpolated using the least squares method, which can remove the noise and burrs in the data, reconstruct the missing data, and improve the data quality. Figure 3 shown.

[0024] Since the number of temperature probes in the power battery does not match the number of battery cells in the module, it is difficult to distinguish the temperature abnormality of a single battery using only the probe temperature value. Therefore, a full battery temperature generation equation is designed to construct the full battery temperature data. As a preferred embodiment of the present invention, the power battery system layout diagram is as follows Figure 4 As shown, "1P5S" represents a mixed connection of battery modules of "1 in parallel and 5 in series", and "1P6S" and so on, where all battery modules are connected in parallel to form a power battery.

[0025] At the time of collection Under this condition, the calculation method of the battery cell temperature value in the battery system is as follows: For edge modules of parallel type 1, the calculation process includes: For the middle module of parallel type 1, the calculation process includes: Finally, the temperature data of the battery cells far away from the temperature probe in the module are jointly corrected (such as , wait): In the above calculation process, The temperature of the temperature probe is , in this case ;No. The temperature of each battery cell is Indicates the generated battery cell temperature value, ; A temporary variable representing the temperature value of the battery cell during the calculation process. and Respectively represent the probe point temperature weight and the inter-group temperature transfer weight. For the temperature of a battery, a larger weight is given to the adjacent temperature probe data, and other temperature probe data are combined; since the temperature transfer efficiency within a module is generally higher than that between modules, a lower weight is given to the temperature probe data group between modules. Preferably, here let and When the value is 0.8, the calculation results that are unbiased with the temperature probe data can be obtained, which makes up for the sparsity of the battery temperature values ​​in the power battery system and fixes the abnormal values ​​after preprocessing. Some results are shown in Figure 5 The temperature values ​​of the remaining battery cells can be calculated in the same way and will not be described in detail here.

[0026] It should be noted that the power battery of the above-mentioned passenger electric vehicle can be selectively composed of multiple battery cells connected in series, in parallel or in hybrid connection, and the hybrid connection refers to a connection form in which both series and parallel connections exist at the same time. In addition, the method for generating the battery temperature value is not limited to the above-mentioned weighted fusion method, and weighted least squares method, Kalman filtering, etc. can also be used.

[0027] S2. Differentiate the preprocessed data to obtain a fault discrimination signal, and extract data features of different faults in the median difference data space.

[0028] By calculating the median differential voltage and median differential temperature of the 92 battery cells in the power battery system, the fault identification signal is obtained. Taking the fault type of abnormal connection between batteries as a reference case, the analysis of the median differential voltage is given. Figure 6 As shown, the legend labels are The curve indicates that there are battery cells with abnormal inter-battery connection. The corresponding median differential voltage, the curve with the darkest grayscale corresponds to a fault-free battery cell for reference. The battery cells with abnormal connection in the figure are obviously scattered in different modules, and the battery cells with abnormal connection not only have data abnormalities during the discharge process, but also have an increase in the median differential voltage during the charging process. The reason for the former is that when the connection is loose, the vibration of the vehicle causes the tightly connected electrodes to shake, resulting in unstable electrode voltage and data distortion; the latter is that as the degree of connection loosening deepens, the contact resistance increases. Under high charging current, Ohm's law shows that large resistance components have higher voltages under large currents.

[0029] It should be noted that the analysis process for other fault types is similar to the above, so it will not be repeated here. Only the corresponding data features are given as shown in Table 1. The specific manifestations are as follows Figure 7 , Figure 8 , Fig. 9 , Fig.10 shown.

[0030] Table 1. Data feature description table of various faults In addition, the power battery failures targeted are not limited to the above-mentioned failure types, but may also selectively include battery cell internal resistance inconsistency, battery cell life inconsistency, etc.

[0031] S3. Build a multi-fault risk discrimination model based on the fault discrimination signal, conduct a comprehensive assessment of the battery fault, and calculate the absolute fault degree score corresponding to the battery fault.

[0032] The preprocessed data is divided into multiple charge and discharge cycles according to adjacent charge and discharge behaviors, and the charge interval and discharge interval are cut out. The median differential voltage and median differential temperature of all batteries in the charge interval and discharge interval are counted, and the data distortion score, mutation score, voltage inconsistency score, and temperature inconsistency score of all batteries are input into the multi-fault risk discrimination model. The absolute value of the calculated score is taken to characterize the degree of faults such as abnormal connection between batteries, voltage sampling distortion, abnormal battery self-discharge, sudden internal short circuit fault, voltage inconsistency, and temperature inconsistency. The relevant parameters involved in the multi-fault risk discrimination model are determined by expert experience combined with the grid method, as shown in Table 2.

[0033] Table 2. Related parameters and value selection methods involved in the multi-fault risk identification model S4. Count the data distribution of different battery faults and their absolute fault degree scores, determine the warning level and fault level based on the distribution results, and draw a multi-fault risk warning diagram. As a preferred embodiment of the present invention, the frequencies of different faults and different fault degrees of all batteries in the 20 passenger electric vehicles are counted, and the frequency density is calculated to obtain the data distribution of each fault and its fault degree. According to the general principle that the number of normal batteries is far greater than the number of faulty batteries, the intervals of normal data and faulty data in terms of fault degree are divided, and the fault classification range of power batteries is shown in Table 3: Table 3. Power battery fault classification range As a preferred embodiment of the present invention, the multi-fault risk warning diagram is represented by a multi-dimensional radar diagram. The corresponding labels in the multi-fault risk warning diagram are the abnormal degree of battery connection, the voltage sampling distortion degree, the abnormal level of battery self-discharge, the risk of sudden internal short circuit failure, and the voltage inconsistency degree and temperature inconsistency degree, such as Fig.11 The fault severity of different faults can be shown to the car owner through a multi-fault risk warning diagram.

[0034] It should be noted that the multi-fault risk warning map is not limited to the expression form of the multi-dimensional radar map, and can also be presented using any other two-dimensional or three-dimensional map.

[0035] S5. In accordance with the preset method, the vehicle owner is prompted with the situations and treatment methods at different warning levels and fault levels in the multi-fault risk warning map, so as to achieve the multi-fault warning goal of the power battery.

[0036] Preferably, according to the classification range of step S4, the preset prompt is: (1) If the absolute failure degree scores of all cells in the power battery are in the safe zone, it indicates that the vehicle's main power battery system is safe; (2) If the absolute failure degree score of a battery in the power battery is in the warning level I zone, the owner will be prompted that the battery has a corresponding failure risk, and the owner can choose to have the power battery inspected regularly; (3) If the absolute fault degree score of a battery in the power battery is in the warning level II zone, the owner is reminded that the battery has a high risk of corresponding faults and the owner needs to conduct a routine inspection of the power battery; (4) If the absolute fault degree score of a battery in the power battery is in the fault level I zone, the owner will be prompted that the battery has a corresponding fault and the owner needs to inspect the power battery and optionally replace the power battery system components; (5) If the absolute fault severity score of a battery in the power battery is in the fault level II zone, the vehicle owner will be prompted that the battery has a serious fault and the vehicle owner needs to replace the power battery system components based on the power battery inspection results.

[0037] In summary, the above embodiments present an implementation process of a multi-fault risk assessment and early warning method for a power battery. The data preprocessing method reduces data analysis and modeling errors and improves data reliability. The interpolation method makes full use of low-density temperature sampling data to generate the temperature value of the entire battery cell, thereby improving the accuracy of identifying multiple types of faults in the battery cell under abnormal temperature conditions. Based on the battery cell voltage value and the battery cell temperature value, a fault discrimination signal is constructed by a differential method, and data features of different faults are extracted from it without complex data transformation, thereby improving calculation efficiency. The use of mathematical and statistical methods to construct a multi-fault risk discrimination model realizes the detection and separation of multiple faults in the power battery. Based on the multi-fault risk warning map, the owner can be provided with a reference for battery maintenance and repair in a timely manner to prevent potential dangers caused by battery system failures.

[0038] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments.

[0039] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A power battery multi-fault risk assessment and early warning method, characterized in that: The risk assessment and early warning method includes: S1. Request the recent life cycle data of the power battery of the in-service electric vehicle from the cloud server, perform data preprocessing, and obtain the temperature value of any battery cell in the power battery system by interpolation method; S2. Differentiate the voltage value and temperature value of any battery cell in the power battery system to obtain a fault discrimination signal; S3. Construct a multi-fault risk discrimination model based on the fault discrimination signal to comprehensively evaluate the battery faults, including abnormal inter-battery connection of the power battery system, voltage sampling distortion, abnormal battery self-discharge, sudden internal short circuit fault, voltage inconsistency, and temperature inconsistency, and calculate the absolute fault degree scores corresponding to the above battery faults; S4. Count the data distribution of different battery failures and their absolute failure degree scores, determine the warning level and failure level based on the distribution results, and draw a multi-failure risk warning map; S5. In accordance with the preset method, the vehicle owner is prompted with the situations and treatment methods at different warning levels and fault levels in the multi-fault risk warning map, so as to achieve the multi-fault warning goal of the power battery.

2. A power battery multi-fault risk assessment and early warning method according to claim 1, characterized in that: The process of step S1 is as follows: S101. Requesting power battery life cycle data from the cloud big data center, where the data fields include: collection time, vehicle status, charging status, operating mode, vehicle speed, total voltage, total current, cumulative mileage, state of charge, battery cell voltage value, battery cell voltage maximum value, battery cell code with the highest voltage, battery cell voltage minimum value, battery cell code with the lowest voltage, maximum temperature value, maximum temperature subsystem number, maximum temperature probe serial number, maximum alarm level, probe temperature, and general alarm flag; S102, sorting the data in order according to the acquisition time, separating the data of different data fields, filtering the impulse noise in the data by using mean filtering and median filtering, and then using the least squares method to interpolate the missing data; S103 . According to the layout of the limited temperature probes in the power battery system, the known temperature probe data are weightedly integrated by interpolation method to calculate the temperature value of any battery cell in the power battery system.

3. A power battery multi-fault risk assessment and early warning method according to claim 2, characterized in that: The process of step S2 is as follows: S201, load all battery cell voltage values ​​in the power battery system and calculate the median differential voltage, calculate the median differential temperature based on the battery cell temperature value, and generate a fault discrimination signal, which is calculated by the following method: exist At this moment, the power battery system Voltage value of each battery cell and temperature value Forming a battery cell voltage sequence and battery cell temperature series , and the The median differential voltage is obtained by subtracting the voltage value of each battery cell from the median of the battery cell voltage sequence. , let the The temperature value of each battery cell is subtracted from the median of the battery cell temperature sequence to obtain the median differential temperature. , and It is the fault identification signal; S202, analyzing the fault discrimination signal, and extracting data features of different faults in the median difference data space.

4. A power battery multi-fault risk assessment and early warning method according to claim 3, characterized in that: The process of step S3 is as follows: S301, for battery voltage sampling distortion and battery connection abnormality, perform logical mutual exclusion judgment: divide the charging interval and the discharging interval in a charge and discharge segment, according to Slice out the first The median differential voltage of each battery cell and the discharge interval The median differential voltage of each battery cell , and calculate the arithmetic mean of the extreme values and and the variance of the time series difference statistics and , the arithmetic mean without extreme values and The calculation process is as follows: In the above formula and Respectively represent The maximum median differential voltage and the minimum median differential voltage within the charging interval of each battery cell, and Respectively represent The maximum median differential voltage and the minimum median differential voltage within the discharge interval of each battery cell, and They are the time when charging starts and ends during a single continuous charge and discharge process. and are the moments when the discharge starts and ends, respectively; The calculation process of the statistical variance of the extreme value time series difference is as follows: In the above formula and Respectively represent The time dimension of each battery cell and The sequence after difference, operator Indicates that and After sorting, remove the first and last two extreme value samples to obtain the extreme value time difference sequence and , and finally calculate the statistical variance of the two sequences and ; In the same battery module, the battery cells Greater than the charging distortion judgment threshold ,or Exceeding the discharge distortion judgment threshold The number of battery cells is greater than 1, and Below the overcharge threshold , voltage sampling distortion occurs; Battery Cell Exceed If the number of is equal to 1, the battery connection is abnormal, otherwise no fault occurs; If voltage sampling distortion occurs, the mutually exclusive symbol If the battery connection is abnormal, the mutual exclusion symbol is set to -1. is +1; Definition Amplification factor of each battery cell It is used to enhance the diagnosis of abnormal voltage sampling and abnormal connection between batteries, and is calculated by the following formula: Definition The data distortion fraction of each battery cell is , calculated by the following formula: In the above formula, Representation function , since the value range is [-1, +1], this function is used to normalize the data distortion score to the interval [-1, +1]; Data distortion score Characterizes the fault degree of voltage sampling distortion and abnormal connection between batteries. If the battery cell is healthy, the data distortion score is close to 0. If voltage sampling distortion occurs, it will diverge to the edge of -1. If abnormal connection occurs between batteries, it will diverge to the edge of +1. S302: Perform feature confrontation identification for abnormal battery self-discharge and sudden internal short circuit failure: and Calculate the The voltage difference fraction of each battery cell , and design the attenuation factor To reduce the interference of healthy battery data, the specific calculation formula is as follows: Use median filter to The median differential temperature of each battery cell Filter to get , and in the time dimension Perform the difference to obtain the temperature time series difference sequence ; Definition The sudden internal short circuit fraction of each battery cell is , the specific calculation process is as follows: In the above formula For the The voltage variation factor of each battery cell, For the Temperature variation factor of each battery cell, correction factor Used to adjust the impact of data distortion on sudden internal short circuits; It means to find the maximum value of a sequence. Express request The power of right and Difference The mutation score of each battery cell is To express; The specific calculation process is as follows: Mutation score Indicates the degree of battery self-discharge abnormality and sudden internal short circuit failure. Used to normalize the mutation score to the interval [-1, +1]; If the battery cell is healthy, the mutation score approaches 0. If a self-discharge abnormality occurs, it will diverge to the -1 edge. If a sudden internal short circuit occurs, it will diverge to the +1 edge. S303: For the voltage inconsistency and temperature inconsistency of the battery, define The voltage inconsistency score of each battery cell is , No. The temperature inconsistency fraction of each battery cell is , the calculation process is as follows: In the above formula The sensitivity adjustment factor representing the voltage inconsistency fraction, represents the sensitivity adjustment factor of the temperature inconsistency fraction; S304: Take the absolute value of the fault severity score obtained by the above calculation as a basis for fault severity assessment.

5. A power battery multi-fault risk assessment and early warning method according to claim 4, characterized in that: The process of step S4 is as follows: S401, counting the frequencies of all battery cells with different faults and absolute fault severity scores, and calculating the relative density to obtain the data distribution of each fault and its fault severity; S402, dividing the fault degree interval according to the data distribution result, and obtaining the warning level and the fault level; S403: Determine the fault classification range based on the warning level and the fault level, and draw a multi-fault risk warning map according to the fault type.

6. A power battery multi-fault risk assessment and early warning method according to claim 5, characterized in that: The process of step S5 is as follows: S501, according to the multi-fault risk warning diagram, if the absolute fault degree score of the battery cell is within the corresponding fault classification range, then the battery cell is warned or judged as a fault, otherwise the battery cell is judged as safe; S502: Preset corresponding prompting methods for different warning levels and fault levels, and provide the vehicle owner with a multi-fault warning for the power battery according to the preset prompting methods.

Images