Vehicle battery health degree detection method and system based on multi-parameter fusion analysis

Abstract

This invention discloses a method and system for detecting vehicle battery health based on multi-parameter fusion analysis, relating to the field of vehicle battery detection technology. The detection method includes the following steps: collecting multi-dimensional data from the vehicle battery to generate vehicle driving records; identifying abnormal intervals of the vehicle battery in any vehicle driving record; analyzing the changes in dimensional data of each dimension and identifying the scene mode; extracting abnormal features based on the dimensional data presented in each abnormal interval; matching corresponding influence weights to the dimensions of each abnormal feature based on the scene mode, evaluating the health of the vehicle battery, and confirming the feature evaluation value of the abnormal identification; identifying the current scene mode based on the real-time collected dimensional data whenever the vehicle is in motion; analyzing the real-time health of the vehicle battery and determining whether the vehicle battery is abnormal.

Description

Technical Field

[0001] This invention relates to the field of vehicle battery testing technology, specifically to a method and system for testing the health of vehicle batteries based on multi-parameter fusion analysis. Background Technology

[0002] With the popularization of new energy vehicles, accurate assessment of vehicle battery health has become a key technology to ensure vehicle safety and range. Reliable vehicle battery health testing can significantly improve the reliability and safety of vehicle operation and ensure the user's travel experience. Although existing testing methods can perform comprehensive testing through multi-dimensional parameters, they still have significant limitations.

[0003] Traditional testing methods typically assign fixed weights to different parameters for evaluation. This fixed-rule approach struggles to effectively address the complexity and variability of vehicle battery status changes across various driving scenarios. For instance, in congested traffic, vehicle batteries may have a higher cycle count but a lower depth of discharge, while at high speeds, they may have a higher depth of discharge but a relatively lower cycle count. Furthermore, the weights of different parameters on battery degradation vary significantly across different scenarios. Applying fixed weights to health checks can easily lead to misjudgments, making it difficult to accurately assess vehicle battery health and effectively identify battery malfunctions. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for detecting the health of vehicle batteries based on multi-parameter fusion analysis, so as to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a vehicle battery health detection method based on multi-parameter fusion analysis, the detection method comprising the following steps:

[0006] Step S100: A monitoring device is installed inside the vehicle to collect multi-dimensional data on the vehicle battery during each vehicle trip, generating corresponding vehicle driving records; anomaly analysis is performed on the dimensional data of each dimension to identify abnormal intervals of the vehicle battery in any vehicle driving record.

[0007] Step S200: Analyze the changes in dimensional data of each dimension in any vehicle driving record between different intervals, and identify the scene mode of the vehicle battery; based on the dimensional data presented in each abnormal interval, extract the abnormal features that affect the abnormal condition of the vehicle battery.

[0008] Step S300: Analyze the changes of various abnormal features in different abnormal intervals in any vehicle driving record, and match the corresponding influence weights of each abnormal feature in the dimension based on the scene mode; evaluate the health of the vehicle battery according to the influence weights of each dimension and confirm the feature evaluation values ​​of the anomaly identification.

[0009] Step S400: Whenever the vehicle is in motion, the current scene mode is identified based on the real-time collected dimensional data; the influence weights of each dimension are extracted based on the current scene mode, the real-time health of the vehicle battery is analyzed, and it is determined whether there is any abnormality in the vehicle battery.

[0010] Furthermore, step S100 includes the following steps:

[0011] Step S101: Several monitoring indicators are preset for the vehicle battery. During vehicle operation, monitoring data of each monitoring indicator is collected at each unit time. Each monitoring indicator is set as a dimension of the vehicle battery, and the monitoring data of the monitoring indicator is set as the dimension data of the corresponding dimension. The dimension data of each dimension are summarized to generate a vehicle driving record. The monitoring indicators set for the vehicle battery include voltage, current, temperature, charging efficiency, and discharging efficiency.

[0012] Step S102: Preset an expected value range for each dimension, wherein the expected value range for the i-th dimension is set to [a i ,b i ]; Randomly select the i-th dimension from a vehicle driving record, and obtain the dimension data of the i-th dimension at the j-th time point as d. (i,j) If d (i,j) <a i Or d (i,j) >b i If so, then the j-th time point is set as the abnormal time point of the i-th dimension;

[0013] Step S103: Obtain all abnormal time points in the i-th dimension. If two abnormal time points are adjacent to each other, merge the two abnormal time points into one abnormal time interval. Continue to merge adjacent abnormal time points to obtain several abnormal intervals in the i-th dimension.

[0014] Furthermore, step S200 includes the following steps:

[0015] Step S201: Select any vehicle driving record, extract dimensional data of each dimension from the selected vehicle driving record, and preset several indicator features for any dimension to obtain the indicator data of each indicator feature at each time point; taking voltage as an example, the set indicator features may include the maximum voltage fluctuation amplitude, average voltage and fluctuation frequency, etc.

[0016] Step S202: A scenario database is pre-built, which stores several scenario modes. Each scenario mode is matched with several scenario features and corresponding expected data ranges. A scenario mode is arbitrarily selected, and the indicator features of each dimension of the selected vehicle driving record at the j-th time point are compared with the same scenario features in the selected scenario mode. If the indicator data of each indicator feature is within the expected data range of the compared scenario features, then the scenario mode in which the vehicle battery is located at the j-th time point is the selected scenario mode.

[0017] Step S203: Extract the scene modes of the vehicle battery at each time point in the selected vehicle driving record, and divide the selected vehicle driving record into several target time intervals according to the different scene modes; arbitrarily select a target time interval, compare the abnormal intervals of each dimension with the selected target time interval, extract the abnormal intervals with common intervals, and obtain the target abnormal intervals of each dimension in the selected target time interval.

[0018] Step S204: Arbitrarily select the target anomaly interval of the i-th dimension, and set the interval length of the target anomaly interval of the i-th dimension to t. i And the selected target time interval has an interval length of T, if T > t i Then, the monitoring indicator of the i-th dimension is extracted and set as the first abnormal feature. If T = t i Then, the adjacent time intervals that are adjacent to the selected target time interval are extracted, and the target abnormal interval of the i-th dimension in the adjacent time interval is obtained. If the target abnormal interval has the same interval length as the adjacent time interval, the monitoring indicator of the i-th dimension is set as the first abnormal feature. If the interval lengths are different, the monitoring indicator is set as the second abnormal feature, thus obtaining the first feature set and the second feature set in each scenario mode of the selected vehicle driving.

[0019] The first anomaly is the presence of abnormal monitoring indicators across various dimensions during actual driving. If the monitoring indicators fluctuate between normal and abnormal values ​​under the same scenario mode, it indicates that the abnormality reflects an abnormal condition of the battery. The second anomaly is that the values ​​are abnormal under the same scenario mode, but return to normal when the scenario mode is switched. This indicates that the abnormal values ​​generated by the second anomaly are inevitable under a specific scenario mode. Therefore, the first anomaly increases the weight in the subsequent weight allocation, while the second anomaly decreases the weight.

[0020] Furthermore, step S300 includes the following steps:

[0021] Step S301: Randomly select a scene mode, extract all vehicle driving records containing the selected scene mode, arbitrarily select the monitoring indicator of the i-th dimension, and count the number of records in each vehicle driving record that contain the monitoring indicator of the i-th dimension in the first feature set, which is m1. i The number of records in the second feature set is m2 i ;

[0022] Step S302: Randomly select a vehicle driving record containing a selected scene mode, obtain the total time length of the abnormal interval of the i-th dimension in the selected scene mode in the selected vehicle driving record, and set the total time length of the abnormal interval of the i-th dimension in the selected scene mode in the k-th vehicle driving record as L. (i,k) The total time length of the abnormal interval when the monitoring indicator of the i-th dimension is in the first feature set is L1. (i,k) The total time length of the abnormal interval when the monitoring indicator of the i-th dimension is in the second feature set is L2. (i,k) ;

[0023] Step S303: In the selected scene mode, set a first feature value η1 for the i-th dimension of the k-th vehicle driving record. (i,k) And a second eigenvalue η2 (i,k) η1 was calculated. (i,k) =L1 (i,k) / L (i,k) and η2 (i,k) =L2 (i,k) / L (i,k) According to the formula:

[0024] ;

[0025] Where k1 and k2 are both positive integers, and k1∈[1,m1] i ],k2∈[1,m2 i ], η1 (i,k1)Let η2 be the first feature value of the i-th dimension in the k1-th vehicle driving record. (i,k2) Let P be the second feature value of the i-th dimension in the k2-th vehicle driving record; calculate the comprehensive feature value P of the i-th dimension under the selected scene mode. i The comprehensive feature value is obtained by using each dimension as the first and second abnormal features respectively. The first abnormal feature can identify abnormal situations and is used as a feature to increase the influence weight. The second abnormal feature can reflect the actual fluctuation range of the corresponding dimension in a specific scenario mode. If the occurrence of abnormal values ​​is allowed outside the expected value range, the influence weight of abnormal identification will be reduced accordingly.

[0026] Step S304: If a monitoring indicator of a certain dimension is not in the second feature set in any of the vehicle driving records, it is set as a normal dimension; in the selected scenario mode, the number of all dimensions is num, and the influence weight of any normal dimension is Q=1 / num; all dimensions other than the normal dimensions are set as abnormal dimensions, and the comprehensive feature value of each abnormal dimension in the selected scenario is obtained according to the formula:

[0027] ;

[0028] Where c is the number of normal dimensions, d1 is a positive integer and d1∈[1,e], e is the number of abnormal dimensions, and P d Let P be the comprehensive feature value of the d-th anomaly dimension. d1 Let be the comprehensive feature value of the d1th anomaly dimension; calculate the influence weight Q of the dth anomaly dimension in the selected scene mode. d ;

[0029] Step S305: Randomly select the j-th time point from any vehicle driving record to obtain the scene mode at the j-th time point; obtain the dimension data of the i-th dimension at the j-th time point as d. (i,j) Let the expected value range of the i-th dimension be [a i ,b i If d (i,j) <a i Then the deviation magnitude of the i-th dimension is f. (i,j) =(a i -d (i,j) ) / a i If d (i,j) >b i The deviation amplitude is then obtained as f. (i,j) =(d (i,j) -b i ) / b i If d (i,j) ∈[a i ,bi Then the deviation amplitude f is obtained. (i,j) =0; According to the formula:

[0030] ;

[0031] Where g is the number of dimensions contained in any vehicle's driving record, and Q i The influence weight of the i-th dimension is used; the feature evaluation value Z of the vehicle battery at the j-th time point is calculated. j ;

[0032] Step S306: For all time points in each vehicle driving record where the monitoring indicators of each dimension are at the first feature, obtain the feature evaluation value of the vehicle battery at each time point, and select the feature evaluation value with the smallest value as the evaluation threshold Z for identifying abnormal vehicle battery health. th .

[0033] Furthermore, step S400 includes the following steps:

[0034] Step S401: Collect dimensional data of the vehicle in each dimension at the current moment, obtain the indicator data corresponding to each indicator feature in any dimension, and compare it with each scene mode in the scene database to obtain the real-time scene mode of the vehicle at the current moment.

[0035] Step S402: Obtain the expected value range of each dimension and the deviation magnitude of each dimension; extract the influence weight of each dimension for the real-time scene mode, and calculate the real-time evaluation value Z of the vehicle battery at the current moment. now Set the evaluation threshold to Z. th If Z now ≥Z th If so, an abnormal warning will be sent to the vehicle battery.

[0036] To better implement the above methods, a vehicle battery health detection system is also proposed. The detection system includes a historical record analysis module, a driving scenario analysis module, a weight adjustment evaluation module, and a real-time anomaly identification module.

[0037] The historical record analysis module is used to collect multi-dimensional data from the vehicle battery during each vehicle trip by monitoring equipment installed inside the vehicle, generating corresponding vehicle driving records; it performs anomaly analysis on the data in each dimension, and identifies abnormal intervals of the vehicle battery in any vehicle driving record.

[0038] The driving scenario analysis module is used to analyze the changes of dimensional data of various dimensions in different intervals in the driving records of any vehicle, and to identify the scene mode of the vehicle battery; based on the dimensional data presented in each abnormal interval, it extracts the abnormal features that affect the abnormal condition of the vehicle battery.

[0039] The weight adjustment and evaluation module is used to analyze the changes of various abnormal features in different abnormal intervals in the driving record of any vehicle, and to match the corresponding influence weights of each abnormal feature in the dimension based on the scene mode; based on the influence weights of each dimension, the health of the vehicle battery is evaluated and the feature evaluation values ​​of the anomaly identification are confirmed.

[0040] The real-time anomaly detection module is used to identify the current scene mode based on real-time collected dimensional data whenever the vehicle is in motion; it extracts the influence weights of each dimension based on the current scene mode, analyzes the real-time health of the vehicle battery, and determines whether there are any anomalies in the vehicle battery.

[0041] Furthermore, the historical record analysis module includes a historical record acquisition unit and an abnormal driving identification unit;

[0042] The historical data acquisition unit is used to install monitoring equipment inside the vehicle to collect multi-dimensional data on the vehicle battery during each vehicle trip and generate corresponding vehicle driving records; the abnormal driving identification unit is used to perform anomaly analysis on the dimensional data of each dimension and identify abnormal intervals of the vehicle battery in any vehicle driving record.

[0043] Furthermore, the driving scenario analysis module includes a scenario pattern recognition unit and an anomaly feature extraction unit;

[0044] The scene pattern recognition unit is used to analyze the changes in dimensional data of each dimension in any vehicle driving record between different intervals and to identify the scene pattern of the vehicle battery; the abnormal feature extraction unit is used to extract abnormal features that affect the abnormal condition of the vehicle battery based on the dimensional data presented in each abnormal interval.

[0045] Furthermore, the weight adjustment assessment module includes a mode weight matching unit and a battery health assessment unit;

[0046] The pattern weight matching unit is used to analyze the changes of various abnormal features in different abnormal intervals in any vehicle driving record, and to match the corresponding influence weights of each abnormal feature in the dimension based on the scene pattern; the battery health assessment unit is used to assess the health of the vehicle battery according to the influence weights of each dimension and to confirm the feature assessment values ​​of the anomaly identification.

[0047] Furthermore, the real-time anomaly identification module includes a real-time scene identification unit and a battery anomaly judgment unit;

[0048] The real-time scene recognition unit is used to identify the current scene pattern based on real-time collected dimensional data whenever the vehicle is in motion; the battery anomaly judgment unit is used to extract the influence weights of each dimension based on the current scene pattern, analyze the real-time health of the vehicle battery, and judge whether the vehicle battery has any anomalies.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] 1. This invention identifies the scene patterns of vehicle driving by constructing a scene database, and adaptively matches the corresponding influence weights to different dimensions according to different scene patterns, helping users to make a more accurate judgment on the impact on battery health, and solving the problem of poor adaptability of traditional fixed weight methods in different scenarios.

[0051] 2. This invention enables comprehensive and accurate monitoring of battery status by adaptively adjusting the influence weights of different dimensions. By distinguishing between the first abnormal feature to reflect the actual battery abnormality and the second abnormal feature to reflect scene-related fluctuations, the weight allocation is further optimized to improve the accuracy of detection.

[0052] 3. By setting the influence weights of various dimensions under different scenarios, this invention can achieve real-time monitoring of vehicle batteries, effectively prevent safety hazards caused by battery failures, and optimize the user's driving experience and vehicle maintenance efficiency. Attached Figure Description

[0053] Figure 1 This is a schematic diagram illustrating the steps of a vehicle battery health detection method based on multi-parameter fusion analysis.

[0054] Figure 2 This is a schematic diagram of a vehicle battery health detection system based on multi-parameter fusion analysis. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] Example: Figures 1 to 2 As shown, this invention provides a vehicle battery health detection method based on multi-parameter fusion analysis. The detection method includes the following steps:

[0057] Step S100: A monitoring device is installed inside the vehicle to collect multi-dimensional data on the vehicle battery during each vehicle trip, generating corresponding vehicle driving records; anomaly analysis is performed on the dimensional data of each dimension to identify abnormal intervals of the vehicle battery in any vehicle driving record.

[0058] Step S100 includes the following steps:

[0059] Step S101: Preset several monitoring indicators for the vehicle battery. Collect monitoring data of each monitoring indicator at every unit time during vehicle operation. Set one monitoring indicator as a dimension of the vehicle battery, set the monitoring data of the monitoring indicator as the dimension data of the corresponding dimension, and summarize the dimension data of each dimension to generate a vehicle driving record.

[0060] Step S102: Preset an expected value range for each dimension, wherein the expected value range for the i-th dimension is set to [a i ,b i ]; Randomly select the i-th dimension from a vehicle driving record, and obtain the dimension data of the i-th dimension at the j-th time point as d. (i,j) If d (i,j) <a i Or d (i,j) >b i If so, then the j-th time point is set as the abnormal time point of the i-th dimension;

[0061] Step S103: Obtain all abnormal time points in the i-th dimension. If two abnormal time points are adjacent to each other, merge the two abnormal time points into one abnormal time interval. Continue to merge adjacent abnormal time points to obtain several abnormal intervals in the i-th dimension.

[0062] Step S200: Analyze the changes in dimensional data of each dimension in any vehicle driving record between different intervals, and identify the scene mode of the vehicle battery; based on the dimensional data presented in each abnormal interval, extract the abnormal features that affect the abnormal condition of the vehicle battery.

[0063] Step S200 includes the following steps:

[0064] Step S201: Select any vehicle driving record, extract the dimensional data of each dimension from the selected vehicle driving record, and preset several indicator features for any dimension to obtain the indicator data of each indicator feature at each time point.

[0065] Step S202: A scenario database is pre-built, which stores several scenario modes. Each scenario mode is matched with several scenario features and corresponding expected data ranges. A scenario mode is arbitrarily selected, and the indicator features of each dimension of the selected vehicle driving record at the j-th time point are compared with the same scenario features in the selected scenario mode. If the indicator data of each indicator feature is within the expected data range of the compared scenario features, then the scenario mode in which the vehicle battery is located at the j-th time point is the selected scenario mode.

[0066] Step S203: Extract the scene modes of the vehicle battery at each time point in the selected vehicle driving record, and divide the selected vehicle driving record into several target time intervals according to the different scene modes; arbitrarily select a target time interval, compare the abnormal intervals of each dimension with the selected target time interval, extract the abnormal intervals with common intervals, and obtain the target abnormal intervals of each dimension in the selected target time interval.

[0067] Step S204: Arbitrarily select the target anomaly interval of the i-th dimension, and set the interval length of the target anomaly interval of the i-th dimension to t. i And the selected target time interval has an interval length of T, if T > t i Then, the monitoring indicator of the i-th dimension is extracted and set as the first abnormal feature. If T = t i Then, the adjacent time intervals that are adjacent to the selected target time interval are extracted, and the target abnormal interval of the i-th dimension in the adjacent time interval is obtained. If the target abnormal interval has the same interval length as the adjacent time interval, the monitoring indicator of the i-th dimension is set as the first abnormal feature. If the interval lengths are different, the monitoring indicator is set as the second abnormal feature, thus obtaining the first feature set and the second feature set in each scenario mode of the selected vehicle driving.

[0068] Step S300: Analyze the changes of various abnormal features in different abnormal intervals in any vehicle driving record, and match the corresponding influence weights of each abnormal feature in the dimension based on the scene mode; evaluate the health of the vehicle battery according to the influence weights of each dimension and confirm the feature evaluation values ​​of the anomaly identification.

[0069] Step S300 includes the following steps:

[0070] Step S301: Randomly select a scene mode, extract all vehicle driving records containing the selected scene mode, arbitrarily select the monitoring indicator of the i-th dimension, and count the number of records in each vehicle driving record that contain the monitoring indicator of the i-th dimension in the first feature set, which is m1. iThe number of records in the second feature set is m2 i ;

[0071] Step S302: Randomly select a vehicle driving record containing a selected scene mode, obtain the total time length of the abnormal interval of the i-th dimension in the selected scene mode in the selected vehicle driving record, and set the total time length of the abnormal interval of the i-th dimension in the selected scene mode in the k-th vehicle driving record as L. (i,k) The total time length of the abnormal interval when the monitoring indicator of the i-th dimension is in the first feature set is L1. (i,k) The total time length of the abnormal interval when the monitoring indicator of the i-th dimension is in the second feature set is L2. (i,k) ;

[0072] Step S303: In the selected scene mode, set a first feature value η1 for the i-th dimension of the k-th vehicle driving record. (i,k) And a second eigenvalue η2 (i,k) η1 was calculated. (i,k) =L1 (i,k) / L (i,k) and η2 (i,k) =L2 (i,k) / L (i,k) According to the formula:

[0073] ;

[0074] Where k1 and k2 are both positive integers, and k1∈[1,m1] i ],k2∈[1,m2 i ], η1 (i,k1) Let η2 be the first feature value of the i-th dimension in the k1-th vehicle driving record. (i,k2) Let P be the second feature value of the i-th dimension in the k2-th vehicle driving record; calculate the comprehensive feature value P of the i-th dimension under the selected scene mode. i ;

[0075] Example 1: A scenario mode of urban congestion is selected. Within this scenario, the number of charge / discharge cycles is chosen as a monitoring indicator. The number of charge / discharge cycles in each vehicle driving record is calculated as follows: 5 cycles in the first feature set and 20 cycles in the second feature set. A vehicle driving record containing urban congestion is randomly selected. The total length of the abnormal interval in the selected record is 50 minutes, with the abnormal interval for the first abnormal feature being 5 minutes and the abnormal interval for the second abnormal feature being 15 minutes. The first feature value is calculated as η1 = 0.1, and the second feature value as η2 = 0.3. Simultaneously, other vehicle driving records containing urban congestion are obtained, and the first and second feature values ​​are summed to obtain 0.8 and 5, respectively. The comprehensive feature value of the charge / discharge cycles in the selected scenario is calculated as P = (5 × 0.8) / (15 × 5) = 5.33%.

[0076] Step S304: If a monitoring indicator of a certain dimension is not in the second feature set in any of the vehicle driving records, it is set as a normal dimension; in the selected scenario mode, the number of all dimensions is num, and the influence weight of any normal dimension is Q=1 / num; all dimensions other than the normal dimensions are set as abnormal dimensions, and the comprehensive feature value of each abnormal dimension in the selected scenario is obtained according to the formula:

[0077] ;

[0078] Where c is the number of normal dimensions, d1 is a positive integer and d1∈[1,e], e is the number of abnormal dimensions, and P d Let P be the comprehensive feature value of the d-th anomaly dimension. d1 Let be the comprehensive feature value of the d1th anomaly dimension; calculate the influence weight Q of the dth anomaly dimension in the selected scene mode. d ;

[0079] Example 2: In the selected scene mode, there are 4 normal dimensions and 4 abnormal dimensions. The influence weight of each normal dimension is 1 / 8 = 12.5%. The comprehensive feature values ​​of each abnormal dimension are 5%, 4%, 6% and 5%, respectively. The influence weight of the first abnormal dimension is calculated as Q = (1 - 12.5% ​​× 4) × (5% / 20%) = 50% × 25% = 12.5%.

[0080] Step S305: Randomly select the j-th time point from any vehicle driving record to obtain the scene mode at the j-th time point; obtain the dimension data of the i-th dimension at the j-th time point as d. (i,j) Let the expected value range of the i-th dimension be [a i ,bi If d (i,j) <a i Then the deviation magnitude of the i-th dimension is f. (i,j) =(a i -d (i,j) ) / a i If d (i,j) >b i The deviation amplitude is then obtained as f. (i,j) =(d (i,j) -b i ) / b i If d (i,j) ∈[a i ,b i Then the deviation amplitude f is obtained. (i,j) =0; According to the formula:

[0081] ;

[0082] Where g is the number of dimensions contained in any vehicle's driving record, and Q i The influence weight of the i-th dimension is used; the feature evaluation value Z of the vehicle battery at the j-th time point is calculated. j ;

[0083] Step S306: For all time points in each vehicle driving record where the monitoring indicators of each dimension are at the first feature, obtain the feature evaluation value of the vehicle battery at each time point, and select the feature evaluation value with the smallest value as the evaluation threshold Z for identifying abnormal vehicle battery health. th .

[0084] Step S400: Whenever the vehicle is in motion, the current scene mode is identified based on the real-time collected dimensional data; the influence weights of each dimension are extracted based on the current scene mode, the real-time health of the vehicle battery is analyzed, and it is determined whether there is any abnormality in the vehicle battery.

[0085] Step S400 includes the following steps:

[0086] Step S401: Collect dimensional data of the vehicle in each dimension at the current moment, obtain the indicator data corresponding to each indicator feature in any dimension, and compare it with each scene mode in the scene database to obtain the real-time scene mode of the vehicle at the current moment.

[0087] Step S402: Obtain the expected value range of each dimension and the deviation magnitude of each dimension; extract the influence weight of each dimension for the real-time scene mode, and calculate the real-time evaluation value Z of the vehicle battery at the current moment. now Set the evaluation threshold to Z. th If Znow ≥Z th If so, an abnormal warning will be sent to the vehicle battery.

[0088] The vehicle battery health detection system includes a historical record analysis module, a driving scenario analysis module, a weight adjustment evaluation module, and a real-time anomaly identification module.

[0089] The historical record analysis module is used to collect multi-dimensional data from the vehicle battery during each vehicle trip by monitoring equipment installed inside the vehicle, generating corresponding vehicle driving records; it performs anomaly analysis on the data in each dimension, and identifies abnormal intervals of the vehicle battery in any vehicle driving record.

[0090] The driving scenario analysis module is used to analyze the changes of dimensional data of various dimensions in different intervals in the driving records of any vehicle, and to identify the scene mode of the vehicle battery; based on the dimensional data presented in each abnormal interval, it extracts the abnormal features that affect the abnormal condition of the vehicle battery.

[0091] The weight adjustment and evaluation module is used to analyze the changes of various abnormal features in different abnormal intervals in the driving record of any vehicle, and to match the corresponding influence weights of each abnormal feature in the dimension based on the scene mode; based on the influence weights of each dimension, the health of the vehicle battery is evaluated and the feature evaluation values ​​of the anomaly identification are confirmed.

[0092] The real-time anomaly detection module is used to identify the current scene mode based on real-time collected dimensional data whenever the vehicle is in motion; it extracts the influence weights of each dimension based on the current scene mode, analyzes the real-time health of the vehicle battery, and determines whether there are any anomalies in the vehicle battery.

[0093] The historical record analysis module includes a historical record acquisition unit and an abnormal driving identification unit.

[0094] The historical data acquisition unit is used to install monitoring equipment inside the vehicle to collect multi-dimensional data on the vehicle battery during each vehicle trip and generate corresponding vehicle driving records; the abnormal driving identification unit is used to perform anomaly analysis on the dimensional data of each dimension and identify abnormal intervals of the vehicle battery in any vehicle driving record.

[0095] The driving scenario analysis module includes a scenario pattern recognition unit and an anomaly feature extraction unit.

[0096] The scene pattern recognition unit is used to analyze the changes in dimensional data of each dimension in any vehicle driving record between different intervals and to identify the scene pattern of the vehicle battery; the abnormal feature extraction unit is used to extract abnormal features that affect the abnormal condition of the vehicle battery based on the dimensional data presented in each abnormal interval.

[0097] The weight adjustment evaluation module includes a mode weight matching unit and a battery health evaluation unit.

[0098] The pattern weight matching unit is used to analyze the changes of various abnormal features in different abnormal intervals in any vehicle driving record, and to match the corresponding influence weights of each abnormal feature in the dimension based on the scene pattern; the battery health assessment unit is used to assess the health of the vehicle battery according to the influence weights of each dimension and to confirm the feature assessment values ​​of the anomaly identification.

[0099] The real-time anomaly identification module includes a real-time scene identification unit and a battery anomaly judgment unit.

[0100] The real-time scene recognition unit is used to identify the current scene pattern based on real-time collected dimensional data whenever the vehicle is in motion; the battery anomaly judgment unit is used to extract the influence weights of each dimension based on the current scene pattern, analyze the real-time health of the vehicle battery, and judge whether the vehicle battery has any anomalies.

[0101] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for detecting vehicle battery health based on multi-parameter fusion analysis, characterized in that: The detection method includes the following steps: Step S100: A monitoring device is installed inside the vehicle to collect multi-dimensional data on the vehicle battery during each vehicle trip, generating corresponding vehicle driving records; anomaly analysis is performed on the dimensional data of each dimension to identify abnormal intervals of the vehicle battery in any vehicle driving record. Step S200: Analyze the changes in dimensional data of each dimension in any vehicle driving record between different intervals, and identify the scene mode of the vehicle battery; based on the dimensional data presented in each abnormal interval, extract the abnormal features that affect the abnormal condition of the vehicle battery. Step S300: Analyze the changes of various abnormal features in different abnormal intervals in any vehicle driving record, and match the corresponding influence weights of each abnormal feature in the dimension based on the scene mode; evaluate the health of the vehicle battery according to the influence weights of each dimension and confirm the feature evaluation values ​​of the anomaly identification. Step S400: Whenever the vehicle is in motion, the current scene mode is identified based on the real-time collected dimensional data; the influence weights of each dimension are extracted based on the current scene mode, the real-time health of the vehicle battery is analyzed, and it is determined whether there is any abnormality in the vehicle battery. Step S200 includes the following steps: Step S201: Select any vehicle driving record, extract the dimensional data of each dimension from the selected vehicle driving record, and preset several indicator features for any dimension to obtain the indicator data of each indicator feature at each time point. Step S202: A scenario database is pre-built, which stores several scenario modes. Each scenario mode is matched with several scenario features and corresponding expected data ranges. A scenario mode is arbitrarily selected, and the indicator features of each dimension of the selected vehicle driving record at the j-th time point are compared with the same scenario features in the selected scenario mode. If the indicator data of each indicator feature is within the expected data range of the compared scenario features, then the scenario mode in which the vehicle battery is located at the j-th time point is the selected scenario mode. Step S203: Extract the scene modes of the vehicle battery at each time point in the selected vehicle driving record, and divide the selected vehicle driving record into several target time intervals according to the different scene modes; arbitrarily select a target time interval, compare the abnormal intervals of each dimension with the selected target time interval, extract the abnormal intervals with common intervals, and obtain the target abnormal intervals of each dimension in the selected target time interval. Step S204: Arbitrarily select the target anomaly interval of the i-th dimension, and set the interval length of the target anomaly interval of the i-th dimension to t. i And the selected target time interval has an interval length of T, if T > t i Then, the monitoring indicator of the i-th dimension is extracted and set as the first abnormal feature. If T = t i Then, the adjacent time intervals that are adjacent to the selected target time interval are extracted, and the target abnormal interval of the i-th dimension in the adjacent time interval is obtained. If the target abnormal interval has the same interval length as the adjacent time interval, the monitoring indicator of the i-th dimension is set as the first abnormal feature. If the interval lengths are different, the monitoring indicator is set as the second abnormal feature, thus obtaining the first feature set and the second feature set in each scenario mode of the selected vehicle driving.

2. The vehicle battery health detection method based on multi-parameter fusion analysis according to claim 1, characterized in that: Step S100 includes the following steps: Step S101: Preset several monitoring indicators for the vehicle battery. Collect monitoring data of each monitoring indicator at every unit time during vehicle operation. Set one monitoring indicator as a dimension of the vehicle battery, set the monitoring data of the monitoring indicator as the dimension data of the corresponding dimension, and summarize the dimension data of each dimension to generate a vehicle driving record. Step S102: Preset an expected value range for each dimension, wherein the expected value range for the i-th dimension is set to [a i ,b i ]; Randomly select the i-th dimension from a vehicle driving record, and obtain the dimension data of the i-th dimension at the j-th time point as d. (i,j) If d (i,j) <a i Or d (i,j) >b i If so, then the j-th time point is set as the abnormal time point of the i-th dimension; Step S103: Obtain all abnormal time points in the i-th dimension. If two abnormal time points are adjacent to each other, merge the two abnormal time points into one abnormal time interval. Continue to merge adjacent abnormal time points to obtain several abnormal intervals in the i-th dimension.

3. The vehicle battery health detection method based on multi-parameter fusion analysis according to claim 2, characterized in that: Step S300 includes the following steps: Step S301: Randomly select a scene mode, extract all vehicle driving records containing the selected scene mode, arbitrarily select the monitoring indicator of the i-th dimension, and count the number of records in each vehicle driving record that contain the monitoring indicator of the i-th dimension in the first feature set, which is m1. i The number of records in the second feature set is m2 i ; Step S302: Randomly select a vehicle driving record containing a selected scene mode, obtain the total time length of the abnormal interval of the i-th dimension in the selected scene mode in the selected vehicle driving record, and set the total time length of the abnormal interval of the i-th dimension in the selected scene mode in the k-th vehicle driving record as L. (i,k) The total time length of the abnormal interval when the monitoring indicator of the i-th dimension is in the first feature set is L1. (i,k) The total time length of the abnormal interval when the monitoring indicator of the i-th dimension is in the second feature set is L2. (i,k) ; Step S303: In the selected scene mode, set a first feature value η1 for the i-th dimension of the k-th vehicle driving record. (i,k) And a second eigenvalue η2 (i,k) η1 was calculated. (i,k) =L1 (i,k) / L (i,k) and η2 (i,k) =L2 (i,k) / L (i,k) According to the formula: ； Where k1 and k2 are both positive integers, and k1∈[1,m1] i ],k2∈[1,m2 i ], η1 (i,k1) Let η2 be the first feature value of the i-th dimension in the k1-th vehicle driving record. (i,k2) Let P be the second feature value of the i-th dimension in the k2-th vehicle driving record; calculate the comprehensive feature value P of the i-th dimension under the selected scene mode. i ; Step S304: If a monitoring indicator of a certain dimension is not in the second feature set in any of the vehicle driving records, it is set as a normal dimension; in the selected scenario mode, the number of all dimensions is num, and the influence weight of any normal dimension is Q=1 / num; all dimensions other than the normal dimensions are set as abnormal dimensions, and the comprehensive feature value of each abnormal dimension in the selected scenario is obtained according to the formula: ； Where c is the number of normal dimensions, d1 is a positive integer and d1∈[1,e], e is the number of abnormal dimensions, and P d Let P be the comprehensive feature value of the d-th anomaly dimension. d1 Let be the comprehensive feature value of the d1th anomaly dimension; calculate the influence weight Q of the dth anomaly dimension in the selected scene mode. d ; Step S305: Randomly select the j-th time point from any vehicle driving record to obtain the scene mode at the j-th time point; obtain the dimension data of the i-th dimension at the j-th time point as d. (i,j) Let the expected value range of the i-th dimension be [a i ,b i If d (i,j) <a i Then the deviation magnitude of the i-th dimension is obtained as f. (i,j) =(a i -d (i,j) ) / a i If d (i,j) >b i The deviation amplitude is then obtained as f. (i,j) =(d (i,j) -b i ) / b i If d (i,j) ∈[a i ,b i Then the deviation amplitude f is obtained. (i,j) =0; According to the formula: ； Where g is the number of dimensions contained in any vehicle's driving record, and Q i The influence weight of the i-th dimension is used; the feature evaluation value Z of the vehicle battery at the j-th time point is calculated. j ; Step S306: For all time points in each vehicle driving record where the monitoring indicators of each dimension are at the first feature, obtain the feature evaluation value of the vehicle battery at each time point, and select the feature evaluation value with the smallest value as the evaluation threshold Z for identifying abnormal vehicle battery health. th .

4. The vehicle battery health detection method based on multi-parameter fusion analysis according to claim 3, characterized in that: Step S400 includes the following steps: Step S401: Collect dimensional data of the vehicle in each dimension at the current moment, obtain the indicator data corresponding to each indicator feature in any dimension, and compare it with each scene mode in the scene database to obtain the real-time scene mode of the vehicle at the current moment. Step S402: Obtain the expected value range of each dimension and the deviation magnitude of each dimension; extract the influence weight of each dimension for the real-time scene mode, and calculate the real-time evaluation value Z of the vehicle battery at the current moment. now Set the evaluation threshold to Z. th If Z now ≥Z th If so, an abnormal warning will be sent to the vehicle battery.

5. A vehicle battery health detection system, used to execute the vehicle battery health detection method based on multi-parameter fusion analysis as described in any one of claims 1-4, characterized in that: The detection system includes a historical record analysis module, a driving scenario analysis module, a weight adjustment evaluation module, and a real-time anomaly identification module. The historical record analysis module is used to install monitoring equipment inside the vehicle to collect multi-dimensional data on the vehicle battery during each vehicle trip, generate corresponding vehicle driving records, perform anomaly analysis on the dimensional data of each dimension, and identify abnormal intervals of the vehicle battery in any vehicle driving record. The driving scenario analysis module is used to analyze the changes in dimensional data of each dimension in any vehicle driving record between different intervals, identify the scenario mode of the vehicle battery, and extract abnormal features that affect the abnormal condition of the vehicle battery based on the dimensional data presented in each abnormal interval. The weight adjustment and evaluation module is used to analyze the changes of various abnormal features in different abnormal intervals in any vehicle driving record, match the corresponding influence weights of each abnormal feature in the dimension based on the scene mode, and evaluate the health of the vehicle battery and confirm the feature evaluation value of the anomaly identification based on the influence weights of each dimension. The real-time anomaly identification module is used to identify the current scene pattern based on real-time collected dimensional data whenever the vehicle is in motion. Based on the current scenario, the influence weights of each dimension are extracted to analyze the real-time health of the vehicle battery and determine whether there are any abnormalities in the vehicle battery.

6. The vehicle battery health detection system according to claim 5, characterized in that: The historical record analysis module includes a historical record acquisition unit and an abnormal driving identification unit; The historical record acquisition unit is used to install monitoring equipment inside the vehicle to collect multi-dimensional data on the vehicle battery during each vehicle trip and generate corresponding vehicle driving records; the abnormal driving identification unit is used to perform anomaly analysis on the dimensional data of each dimension and identify abnormal intervals of the vehicle battery in any vehicle driving record.

7. The vehicle battery health detection system according to claim 5, characterized in that: The driving scenario analysis module includes a scenario pattern recognition unit and an anomaly feature extraction unit; The scene pattern recognition unit is used to analyze the changes in dimensional data of each dimension in any vehicle driving record between different intervals, and to identify the scene pattern of the vehicle battery; the abnormal feature extraction unit is used to extract abnormal features that affect the abnormal condition of the vehicle battery based on the dimensional data presented in each abnormal interval.

8. The vehicle battery health detection system according to claim 5, characterized in that: The weight adjustment evaluation module includes a mode weight matching unit and a battery health evaluation unit; The pattern weight matching unit is used to analyze the changes of various abnormal features in different abnormal intervals in any vehicle driving record, and match the corresponding influence weights of each abnormal feature in the dimension based on the scene pattern; the battery health assessment unit is used to assess the health of the vehicle battery according to the influence weights of each dimension and confirm the feature assessment values ​​of the anomaly identification.

9. The vehicle battery health detection system according to claim 5, characterized in that: The real-time anomaly identification module includes a real-time scene identification unit and a battery anomaly judgment unit; The real-time scene recognition unit is used to identify the current scene mode based on real-time collected dimensional data whenever the vehicle is in motion; the battery anomaly judgment unit is used to extract the influence weights of each dimension based on the current scene mode, analyze the real-time health of the vehicle battery, and judge whether the vehicle battery has any anomalies.

Images