A new energy vehicle health evaluation method and system

By combining online and offline detection methods, multiple evaluation index parameters are obtained and battery health status prediction is predicted using MLR models, the problem of low accuracy in detection of new energy vehicles is solved, and the accuracy of health evaluation and detection efficiency are improved.

CN117634985BActive Publication Date: 2024-11-19CHINA AUTOMOTIVE ENG RES INST +1
View PDF 11 Cites 0 Cited by

Patent Information

Application Number
CN202311842558.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-28
Publication Date
2024-11-19
Estimated Expiration
2043-12-28

Smart Images

  • Figure CN117634985B_ABST
    Figure CN117634985B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of health evaluation of new energy vehicles, and discloses a new energy vehicle health evaluation method and system, comprising obtaining user behavior evaluation index parameters, appearance evaluation index parameters, charging performance evaluation index parameters, vehicle performance evaluation index parameters, battery system electrical performance evaluation index parameters, electric drive appearance performance evaluation index parameters, thermal management performance evaluation index parameters, reliability performance evaluation index parameters and motor performance evaluation index parameters of a target vehicle; according to the evaluation index parameters, calculating the evaluation value of each performance evaluation index parameter, and judging whether the evaluation value is greater than a safety threshold, if so, judging that the evaluation index parameter is abnormal, calling a corresponding correction algorithm for correction, judging whether it can be corrected by the correction algorithm, if not, feeding back the evaluation value and the performance evaluation index parameter to an offline detection platform, and generating a corresponding offline detection report.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of new energy vehicle health evaluation, and in particular to a new energy vehicle health evaluation method and system. Background Art

[0002] With the continuous development of new energy, new energy vehicles have also developed rapidly. In order to be able to monitor and evaluate the performance of new energy vehicles, it is necessary to monitor the performance of new energy vehicles.

[0003] At present, there are generally two methods for new energy vehicles. One is online monitoring, which relies on the accuracy of the corresponding algorithm. The other is offline testing. Although this method can effectively identify faults, the entire testing process is time-consuming and consumes the physical strength of the testers. This single detection method is bound to have loopholes, resulting in inaccurate test results. Summary of the invention

[0004] One of the purposes of the present invention is to provide a method and system for evaluating the health of new energy vehicles, which can solve the problem of low detection accuracy of a single detection method in the prior art, realize the combination of online monitoring and offline detection, and improve the accuracy of health evaluation of new energy vehicles.

[0005] In order to achieve the above object, a new energy vehicle health evaluation method is provided, comprising the following steps:

[0006] S1. Obtaining each user behavior evaluation index parameter corresponding to the target vehicle;

[0007] S2. Obtaining various appearance evaluation index parameters corresponding to the target vehicle;

[0008] S3, obtaining various charging performance evaluation index parameters corresponding to the target vehicle; the charging performance index parameters include the SOH prediction value corresponding to the target vehicle;

[0009] S4, obtaining vehicle performance evaluation index parameters of the target vehicle;

[0010] S5. Obtaining electrical performance evaluation index parameters of the battery system of the target vehicle;

[0011] S6, obtaining electric drive appearance performance evaluation index parameters, thermal management performance evaluation index parameters, reliability performance evaluation index parameters, and motor performance evaluation index parameters of the target vehicle;

[0012] S7. According to the obtained evaluation index parameters, based on the corresponding algorithms and the importance corresponding to the performance evaluation index parameters, the target vehicle is evaluated for health and tested for battery health, the evaluation values ​​corresponding to the performance evaluation index parameters are calculated, and it is determined whether the corresponding evaluation values ​​are greater than the safety threshold. If so, it is determined that the evaluation index parameters are abnormal, and the corresponding correction algorithm is called for correction. It is determined whether the correction algorithm can be used for correction. If so, the correction is automatically performed. Otherwise, the corresponding evaluation values ​​and the corresponding performance evaluation index parameters are fed back to the offline detection platform, and a corresponding offline detection report is generated.

[0013] Technical principle and effect of this scheme: In this scheme, firstly, the various evaluation index parameters of the target vehicle are obtained, including user behavior evaluation index parameters, appearance evaluation index parameters, charging performance evaluation index parameters, vehicle performance evaluation index parameters, electrical performance evaluation index parameters, electric drive appearance performance evaluation index parameters, thermal management performance evaluation index parameters, reliability performance evaluation index parameters and motor performance evaluation index parameters. Then, the evaluation value corresponding to each performance evaluation index parameter is calculated through the algorithm and the importance corresponding to each performance evaluation index parameter. Based on the comparison between the evaluation value and the safety threshold, whether each performance evaluation index parameter is normal is accurately judged. Specifically, if the evaluation value is greater than the safety threshold, it is judged that the evaluation index parameter is abnormal, and the corresponding correction algorithm is called for correction. It is judged whether it can be corrected by the correction algorithm. If so, it is automatically corrected. Otherwise, the corresponding evaluation value and the corresponding performance evaluation index parameter are fed back to the offline detection platform, and the corresponding offline detection report is generated. In this solution, online and offline monitoring can more effectively diagnose vehicle faults. The combination of online and offline detection can reduce the dependence on online algorithms and the time and effort of offline detection, and improve the accuracy of judgment. That is, it solves the problem of low detection accuracy of a single detection method in the existing technology, realizes the combination of online monitoring and offline detection, and improves the accuracy of health evaluation of new energy vehicles.

[0014] Further, the feeding back of the corresponding evaluation value and the corresponding performance evaluation index parameter to the offline testing platform includes the following steps:

[0015] According to the corresponding evaluation value and performance evaluation index parameter, each performance evaluation index parameter is analyzed to determine the priority corresponding to each performance evaluation index parameter, and the highest priority is sent to the offline testing platform as a key offline testing project.

[0016] Beneficial effects: In this scheme, the performance evaluation index parameters are analyzed through the corresponding rating system and performance evaluation index parameters, so as to determine the corresponding priorities and realize the generation of corresponding key offline detection items, which greatly improves the focus of operators corresponding to the offline detection platform during offline detection and the timeliness of fault detection.

[0017] Further, the S3 includes:

[0018] S30, acquiring historical charging data, where the historical charging data includes data of each charging segment;

[0019] S31, based on the data of each charging segment, respectively extract the label capacity and the health feature, and select the health feature with a correlation with the label capacity greater than a preset correlation threshold as input, input it into the pre-stored MLR model for training, and update the hyperparameters in the MLR model based on the SOH corresponding to the label capacity to form a trained MLR model;

[0020] S32, obtaining vehicle data corresponding to the target vehicle to be measured;

[0021] S33, acquiring corresponding health characteristics according to the acquired vehicle data;

[0022] S34. Based on the trained MLR model, the acquired health features are used as input to perform corresponding capacity prediction, form corresponding capacity prediction values, and calculate the SOH prediction value of the target vehicle based on the capacity prediction values.

[0023] Technical principles and effects of this solution: In this solution, the first step is to obtain the historical charging data of new energy vehicles, that is, to know the historical charging data of each new energy vehicle. Based on these charging data, the corresponding label capacity and health characteristics can be extracted and identified, and the pre-stored MLR model can be trained through the label capacity and health characteristics, so that the corresponding model can better meet the requirements of the health evaluation of new energy vehicles. Then the vehicle data of the target vehicle to be measured can be obtained, and the corresponding health characteristics can be obtained through the vehicle data. Based on the health characteristics, it is input into the trained MLR model, and then the capacity of the target vehicle is predicted, and the corresponding capacity prediction value of the target vehicle is output. Based on the capacity prediction value, the health of the target vehicle battery can be reasonably and truly evaluated.

[0024] 1. Through the tag capacity and charging fragment data, the health features with high correlation with the tag capacity can be extracted, providing reliable data support for subsequent model training.

[0025] 2. The MLR model is trained by the health features and the SOH corresponding to the label capacity, so that the trained MLR model can truly and accurately evaluate the battery health status of the vehicle, providing a reliable model for the subsequent evaluation of the battery health status of the target vehicle, and then providing an accurate reference for the evaluation of the health status of the target vehicle. That is, it can achieve an accurate and reliable evaluation of the battery health status of new energy vehicles, thereby achieving the evaluation of the health of new energy vehicles.

[0026] Further, the S31 includes:

[0027] S310, according to each charging segment data, identify the type of each charging segment data, and pre-process according to the corresponding type, and extract the charging segment data corresponding to the pre-processed charging segment data, the difference between the charging end and the charging start SOC of which is greater than a preset threshold;

[0028] S311, calculating the current capacity corresponding to each charging segment data according to the extracted charging segment data, forming a corresponding tag capacity, and calculating the corresponding SOH based on the SOH calculation formula;

[0029] The SOH calculation formula is:

[0030]

[0031] In the formula, Q C is the current capacity of the segment, SOC(t1) is the SOC value at the start of the charging segment, SOC(t2) is the SOC value at the end of the charging segment, and I is the current value of the segment;

[0032] S312, screening and eliminating abnormal data for the tag capacity corresponding to each calculated charging segment data, and saving the processed tag capacity into a variable database;

[0033] S313, according to the charging segment data at this time, intercept the data of the same voltage interval, and based on the correlation analysis, select the feature that is highly correlated with the corresponding tag capacity, and use the feature as the corresponding health feature;

[0034] S314. Input the selected health features into the pre-stored MLR model, use the corresponding SOH as a label, train the MLR model, update the hyperparameters corresponding to the model, and form a new MLR model after training.

[0035] Beneficial effect: In this scheme, firstly, the charging segment data is rationalized by preprocessing the data of each charging segment, and then the current capacity corresponding to each charging segment data is calculated as the corresponding label capacity, and then the corresponding SOH is calculated by the SOH calculation formula, that is, the battery health status corresponding to the charging segment data is calculated, and then the abnormal data of the label capacity is removed by screening and elimination to improve the reliability of the data, and then based on the label capacity and the corresponding correlation analysis, the features that are highly correlated with the label capacity are selected as the corresponding health features, and the preset MLR model is trained through the health features and the corresponding known SOH, so that the recognized MLR model can meet the corresponding battery health evaluation function, greatly improving the accuracy and reliability of battery health assessment.

[0036] Further, the preprocessing according to the corresponding type in S310 includes:

[0037] For the missing data in the charging segment data sampling, the missing data is filled based on the interpolation method;

[0038] Delete data with multiple duplicate records;

[0039] Delete incomplete charging and discharging processes or processes with obvious defects;

[0040] The moving translation method is used to smooth the data of voltage, temperature and current with large fluctuations caused by noise during the sampling process.

[0041] Beneficial effects: In this solution, different types of data are preprocessed differently, so that the processed data can be more complete and authentic, which greatly improves the authenticity and reliability of subsequent data processing and can avoid evaluation errors caused by data errors as much as possible.

[0042] Further, the S312 includes:

[0043] S3120, according to the calculated tag capacities corresponding to the data of each charging segment, based on the first judgment strategy, eliminating obviously abnormal tag capacities;

[0044] The first judgment strategy is: judging the relationship between the corresponding label capacity and the nominal capacity, if 1.2*QRated>=Qlabeld>=0.7*QRated, then the label capacity is a normal capacity value, otherwise it is an obviously abnormal capacity value;

[0045] Where Qlabeld is the calculated label capacity, QRated is the nominal capacity of the battery;

[0046] S3121. According to the LOF algorithm, the tag capacity that has not been eliminated is eliminated and screened for the second time, and the processed tag capacity is saved in the variable database.

[0047] Beneficial effects: In this solution, considering that the sampling interval of real vehicle data is long, the SOC estimation error is large due to the limitation of computing power of the on-board BMS. When the SOC capacity is calculated directly using the charging segment to record the capacity, there are many abnormal values, and the regularity is not obvious. Therefore, in order to ensure that the calculated tag capacity is more credible, two screenings are set up to ensure the efficiency of screening as much as possible while ensuring the accuracy of screening. The first judgment strategy is used to quickly eliminate the obviously abnormal tag capacity, and then the LOF algorithm is used to perform secondary screening and elimination of the tag capacity, which greatly improves the data reliability and validity of the remaining tag capacity.

[0048] Further, the S313 includes:

[0049] S3130, intercepting charging segment data corresponding to the same voltage interval based on the corresponding charging segment data and a preset interception strategy;

[0050] S3131. According to the intercepted charging segment data and the corresponding tag capacity, based on the Pearson correlation coefficient analysis, select a feature whose correlation with the tag capacity is greater than a preset correlation threshold, and use the feature as the corresponding health feature.

[0051] The present invention also provides a new energy vehicle health evaluation system, using the above-mentioned new energy vehicle health evaluation method, including an online monitoring platform and an offline detection platform;

[0052] The online monitoring platform includes:

[0053] A first acquisition module, used to obtain each user behavior evaluation index parameter corresponding to the target vehicle;

[0054] A second acquisition module, used to acquire various appearance evaluation index parameters corresponding to the target vehicle;

[0055] The third acquisition module is used to obtain various charging performance evaluation index parameters corresponding to the target vehicle; the charging performance index parameters include the SOH prediction value corresponding to the target vehicle;

[0056] The fourth acquisition module is used to obtain the vehicle performance evaluation index parameters of the target vehicle;

[0057] A fifth acquisition module, used to obtain electrical performance evaluation index parameters of a battery system of a target vehicle;

[0058] A sixth acquisition module, used to acquire electric drive appearance performance evaluation index parameters, thermal management performance evaluation index parameters, reliability performance evaluation index parameters, and motor performance evaluation index parameters of the target vehicle;

[0059] The processing module is used to perform health assessment and battery health detection on the target vehicle according to the obtained evaluation index parameters, based on the corresponding algorithm and the importance corresponding to each performance evaluation index parameter, calculate the evaluation value corresponding to each performance evaluation index parameter, and determine whether the corresponding evaluation value is greater than the safety threshold. If so, it is determined that the evaluation index parameter is abnormal, and the corresponding correction algorithm is called for correction, and it is determined whether it can be corrected by the correction algorithm. If so, it is automatically corrected. Otherwise, the corresponding evaluation value and the corresponding performance evaluation index parameter are fed back to the offline detection platform, and a corresponding offline detection report is generated.

[0060] Further, the third acquisition module includes:

[0061] A historical data acquisition module, used to acquire historical charging data, wherein the historical charging data includes data of each charging segment;

[0062] A model training module is used to extract label capacity and health features based on the data of each charging segment, and select health features with a correlation with label capacity greater than a preset correlation threshold as input, input them into a pre-stored MLR model for training, and update the hyperparameters in the MLR model based on the SOH corresponding to the label capacity to form a trained MLR model;

[0063] A target data acquisition module, used to acquire vehicle data corresponding to the target vehicle to be measured;

[0064] An identification module, used to obtain corresponding health characteristics based on the acquired vehicle data;

[0065] The health evaluation module is used to perform corresponding capacity prediction based on the trained MLR model, using the acquired health features as input, to form a corresponding capacity prediction value, and to calculate the SOH prediction value of the target vehicle based on the capacity prediction value. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a flow chart of a method for evaluating the health of a new energy vehicle in Embodiment 1 of the present invention;

[0067] Figure 2 This is a logic block diagram of a new energy vehicle health evaluation system in Embodiment 1 of the present invention. DETAILED DESCRIPTION

[0068] The following is further described in detail through specific implementation methods:

[0069] Embodiment 1

[0070] A new energy vehicle health evaluation method is basically as follows Figure 1 As shown, the following steps are included:

[0071] S1. Obtaining various user behavior evaluation index parameters corresponding to the target vehicle; the user behavior evaluation index parameters include charging behavior, driving behavior and risk assessment.

[0072] S2. Acquire various appearance evaluation index parameters corresponding to the target vehicle; the appearance evaluation index parameters include instrument alarm, charging interface, and high and low voltage lines.

[0073] S3, obtaining various charging performance evaluation index parameters corresponding to the target vehicle; the charging performance index parameters include the SOH prediction value corresponding to the target vehicle;

[0074] The S3 includes:

[0075] S30, acquiring historical charging data, where the historical charging data includes data of each charging segment;

[0076] S31, based on the data of each charging segment, respectively extract the label capacity and the health feature, and select the health feature with a correlation with the label capacity greater than a preset correlation threshold as input, input it into the pre-stored MLR model for training, and update the hyperparameters in the MLR model based on the SOH corresponding to the label capacity to form a trained MLR model;

[0077] The S31 includes:

[0078] S310, based on each charging segment data, identify the type of each charging segment data, and pre-process according to the corresponding type, extract the charging segment data corresponding to the pre-processed charging segment data whose charging end and charging start SOC difference is greater than a preset threshold; in this embodiment, the preset threshold is 70%.

[0079] The preprocessing according to the corresponding type in S310 includes:

[0080] For the missing data in the charging segment data sampling, the missing data is filled based on the interpolation method;

[0081] Delete data with multiple duplicate records;

[0082] Delete incomplete charging and discharging processes or processes with obvious defects;

[0083] The moving translation method is used to smooth the data of voltage, temperature and current with large fluctuations caused by noise during the sampling process.

[0084] S311, calculating the current capacity corresponding to each charging segment data according to the extracted charging segment data, forming a corresponding tag capacity, and calculating the corresponding SOH based on the SOH calculation formula;

[0085] The SOH calculation formula is:

[0086]

[0087] In the formula, Q C is the current capacity of the segment, SOC(t1) is the SOC value at the start of the charging segment, SOC(t2) is the SOC value at the end of the charging segment, and I is the current value of the segment;

[0088] S312, screening and eliminating abnormal data for the tag capacity corresponding to each calculated charging segment data, and saving the processed tag capacity into a variable database;

[0089] The S312 includes:

[0090] S3120, according to the calculated tag capacities corresponding to the data of each charging segment, based on the first judgment strategy, eliminating obviously abnormal tag capacities;

[0091] The first judgment strategy is: judging the relationship between the corresponding label capacity and the nominal capacity, if 1.2*QRated>=Qlabeld>=0.7*QRated, then the label capacity is a normal capacity value, otherwise it is an obviously abnormal capacity value;

[0092] Where Qlabeld is the calculated label capacity, QRated is the nominal capacity of the battery;

[0093] S3121. According to the LOF algorithm, the tag capacity that has not been eliminated is eliminated and screened for the second time, and the processed tag capacity is saved in the variable database.

[0094] The S3121 includes:

[0095] S31210, calculating the reachable distance of the neighbor point corresponding to each label capacity based on the reachable distance calculation formula according to the corresponding label capacity;

[0096] The calculation formula of the reachable distance is as follows:

[0097]

[0098] Where d is the distance between two points, k-distance(o) is the kth distance of point o in the domain, and d(p,o) is the distance from point p in the domain o.

[0099] S31211. Calculate the local reachable density corresponding to each corresponding label capacity according to the corresponding reachable distance and based on the local reachable density formula;

[0100] The local reachable density formula is as follows:

[0101]

[0102] Where N(p) is the kth distance neighborhood of p;

[0103] S31212. Calculate the local outlier factor corresponding to each label capacity according to the local reachable density and the local outlier factor calculation formula;

[0104] The local outlier factor calculation formula is as follows:

[0105]

[0106] S31213. According to the local outlier factor corresponding to each label capacity, determine whether the corresponding label capacity is normal. If LOF(p) is greater than or equal to 1, the label capacity is an outlier. If LOF(p) is close to 1, the label capacity is normal data.

[0107] S313. According to the charging segment data at this time, data in the same voltage range is intercepted, and based on correlation analysis, a feature that is highly correlated with the corresponding tag capacity is selected, and the feature is used as the corresponding health feature. In this embodiment, the actual user behavior does not necessarily charge the vehicle according to the charging segment range we selected. Therefore, when extracting the feature, we select the voltage range according to the battery type. The maximum single cell voltage range of the lithium iron phosphate battery selection feature is 3350mV-3500mV, and the maximum single cell voltage range of the ternary battery selection feature is 3800mV-4000mV.

[0108] In this embodiment, it is taken into account that after the corresponding tag capacity is eliminated, the corresponding charging segment data may be interrupted. Therefore, before the specific interception, it is assumed that the corresponding capacity changes in one month are very similar and approximately considered to be unchanged. Based on this assumption, the average capacity of each month is calculated as the true capacity value, and the other capacity values ​​of other corresponding months are filled based on the true capacity value, so as to eliminate the abnormal tag capacity and then fill it.

[0109] The S313 includes:

[0110] S3130, intercepting charging segment data corresponding to the same voltage interval based on the corresponding charging segment data and a preset interception strategy;

[0111] S3131. According to the intercepted charging segment data and the corresponding tag capacity, based on the Pearson correlation coefficient analysis, select a feature whose correlation with the tag capacity is greater than a preset correlation threshold, and use the feature as the corresponding health feature.

[0112] S314. Input the selected health features into the pre-stored MLR model, use the corresponding SOH as a label, train the MLR model, update the hyperparameters corresponding to the model, and form a new MLR model after training.

[0113] S32. Acquire vehicle data corresponding to the target vehicle to be measured. In this embodiment, the corresponding target vehicle and the vehicle corresponding to the historical charging data belong to the same model.

[0114] S33, acquiring corresponding health characteristics according to the acquired vehicle data;

[0115] S34. Based on the trained MLR model, the acquired health features are used as input to perform corresponding capacity prediction, form corresponding capacity prediction values, and calculate the SOH prediction value of the target vehicle based on the capacity prediction values.

[0116] S4. Obtaining vehicle performance evaluation index parameters of the target vehicle; the vehicle performance evaluation index parameters include acceleration abnormality, braking abnormality, coasting safety and power consumption information.

[0117] S5. Obtain electrical performance evaluation index parameters of the battery system of the target vehicle; the electrical performance evaluation index parameters include charging performance, energy efficiency, endurance attenuation and sampling abnormality.

[0118] S6, obtaining electric drive appearance performance evaluation index parameters, thermal management performance evaluation index parameters, reliability performance evaluation index parameters, and motor performance evaluation index parameters of the target vehicle;

[0119] S7. According to the obtained evaluation index parameters, based on the corresponding algorithms and the importance corresponding to the performance evaluation index parameters, the target vehicle is evaluated for health and tested for battery health, the evaluation values ​​corresponding to the performance evaluation index parameters are calculated, and it is determined whether the corresponding evaluation values ​​are greater than the safety threshold. If so, it is determined that the evaluation index parameters are abnormal, and the corresponding correction algorithm is called for correction. It is determined whether the correction algorithm can be used for correction. If so, the correction is automatically performed. Otherwise, the corresponding evaluation values ​​and the corresponding performance evaluation index parameters are fed back to the offline detection platform, and a corresponding offline detection report is generated.

[0120] When correcting the evaluation index parameters, based on the highest threshold and the lowest threshold corresponding to each evaluation index parameter, it is judged whether the corresponding evaluation index parameter is too high or too low. If so, it is judged that the corresponding evaluation index parameter needs to be corrected. At this time, the difference range between the evaluation index parameter and the lowest threshold and the highest threshold is calculated to obtain the parameter range that needs to be corrected for the evaluation index parameter;

[0121] Based on the correlation between the evaluation index parameter and other evaluation index parameters, select other evaluation index parameters with the largest correlation, and based on the correction relationship between the evaluation index parameter and other selected evaluation index parameters, as well as the corresponding parameter range that needs to be corrected, adjust the other selected evaluation index parameters so that the corresponding evaluation index parameter is corrected to within a reasonable threshold.

[0122] In this embodiment, the specific content of the correction algorithm is: if the evaluation index parameter is not within the reasonable threshold, it is considered that its value is abnormal, and then it is corrected through the relevant algorithm. First, it is judged whether each evaluation index parameter is too high or too low. The judgment algorithm is used for judgment. For example, if the evaluation index parameter is less than the minimum threshold, the difference range between the evaluation index parameter value and the minimum and maximum thresholds is calculated to obtain the parameter range that needs to be corrected. Then, by increasing its related evaluation index parameters, for example, if the voltage is too low, the charging voltage is increased to correct its parameters to within the reasonable threshold. Its implementation requires the use of a controller to implant the judgment algorithm into the hardware. According to the calculated value, the controller increases or decreases the value of each parameter to achieve it.

[0123] Feeding back the corresponding evaluation value and the corresponding performance evaluation index parameter to the offline testing platform includes the following steps:

[0124] According to the corresponding evaluation value and performance evaluation index parameter, each performance evaluation index parameter is analyzed, the corresponding priority of each performance evaluation index parameter is determined, and the highest priority is sent to the offline detection platform as the key offline detection project. Compared with the comprehensiveness of the corresponding offline detection report, the setting of the corresponding key offline detection project enables the operator to complete the fault detection as quickly as possible and improve the detection efficiency.

[0125] like Figure 2 As shown, this embodiment also discloses a new energy vehicle health evaluation system, using the above-mentioned new energy vehicle health evaluation method, including an online monitoring platform and an offline detection platform;

[0126] The online monitoring platform includes:

[0127] A first acquisition module, used to obtain each user behavior evaluation index parameter corresponding to the target vehicle;

[0128] A second acquisition module, used to acquire various appearance evaluation index parameters corresponding to the target vehicle;

[0129] The third acquisition module is used to obtain various charging performance evaluation index parameters corresponding to the target vehicle; the charging performance index parameters include the SOH prediction value corresponding to the target vehicle;

[0130] The third acquisition module includes:

[0131] A historical data acquisition module, used to acquire historical charging data, wherein the historical charging data includes data of each charging segment;

[0132] A model training module is used to extract label capacity and health features based on the data of each charging segment, and select health features with a correlation with label capacity greater than a preset correlation threshold as input, input them into a pre-stored MLR model for training, and update the hyperparameters in the MLR model based on the SOH corresponding to the label capacity to form a trained MLR model;

[0133] A target data acquisition module, used to acquire vehicle data corresponding to the target vehicle to be measured;

[0134] An identification module, used to obtain corresponding health characteristics based on the acquired vehicle data;

[0135] The health evaluation module is used to perform corresponding capacity prediction based on the trained MLR model, using the acquired health features as input, to form a corresponding capacity prediction value, and to calculate the SOH prediction value of the target vehicle based on the capacity prediction value.

[0136] The fourth acquisition module is used to obtain the vehicle performance evaluation index parameters of the target vehicle;

[0137] A fifth acquisition module, used to obtain electrical performance evaluation index parameters of a battery system of a target vehicle;

[0138] A sixth acquisition module, used to acquire electric drive appearance performance evaluation index parameters, thermal management performance evaluation index parameters, reliability performance evaluation index parameters, and motor performance evaluation index parameters of the target vehicle;

[0139] The processing module is used to perform health assessment and battery health detection on the target vehicle according to the obtained evaluation index parameters, based on the corresponding algorithm and the importance corresponding to each performance evaluation index parameter, calculate the evaluation value corresponding to each performance evaluation index parameter, and determine whether the corresponding evaluation value is greater than the safety threshold. If so, it is determined that the evaluation index parameter is abnormal, and the corresponding correction algorithm is called for correction, and it is determined whether it can be corrected by the correction algorithm. If so, it is automatically corrected. Otherwise, the corresponding evaluation value and the corresponding performance evaluation index parameter are fed back to the offline detection platform, and a corresponding offline detection report is generated.

[0140] The above is only an embodiment of the present invention. The common sense such as the known specific structure and characteristics in the scheme is described too much here. The ordinary technicians in the relevant field know all the common technical knowledge in the technical field of the invention before the application date or priority date, can know all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent. The protection scope required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A method for evaluating the health of a new energy vehicle, characterized in that: The following steps are involved: S1. Obtaining each user behavior evaluation index parameter corresponding to the target vehicle; S2. Obtaining various appearance evaluation index parameters corresponding to the target vehicle; S3. Obtaining various charging performance evaluation index parameters corresponding to the target vehicle; the charging performance evaluation index parameters include a SOH prediction value corresponding to the target vehicle; S4, obtaining vehicle performance evaluation index parameters of the target vehicle; S5. Obtaining electrical performance evaluation index parameters of the battery system of the target vehicle; S6, obtaining electric drive appearance performance evaluation index parameters, thermal management performance evaluation index parameters, reliability performance evaluation index parameters, and motor performance evaluation index parameters of the target vehicle; S7. According to the obtained evaluation index parameters, based on the corresponding algorithms and the importance corresponding to the performance evaluation index parameters, the target vehicle is evaluated for health and tested for battery health, the evaluation value corresponding to each performance evaluation index parameter is calculated, and it is determined whether the corresponding evaluation value is greater than the safety threshold. If so, it is determined that the evaluation index parameter is abnormal, and the corresponding correction algorithm is called for correction, and it is determined whether it can be corrected by the correction algorithm. If so, it is automatically corrected. Otherwise, the corresponding evaluation value and the corresponding performance evaluation index parameter are fed back to the offline detection platform, and a corresponding offline detection report is generated; When correcting the evaluation index parameters, based on the highest threshold and the lowest threshold corresponding to each evaluation index parameter, it is judged whether the corresponding evaluation index parameter is too high or too low. If so, it is judged that the corresponding evaluation index parameter needs to be corrected. At this time, the difference range between the evaluation index parameter and the lowest threshold and the highest threshold is calculated to obtain the parameter range that needs to be corrected for the evaluation index parameter; Based on the correlation between the evaluation index parameter and other evaluation index parameters, select other evaluation index parameters with the largest correlation, and based on the correction relationship between the evaluation index parameter and other selected evaluation index parameters, as well as the corresponding parameter range that needs to be corrected, adjust the other selected evaluation index parameters so that the corresponding evaluation index parameter is corrected to within a reasonable threshold.

2. A new energy vehicle health evaluation method according to claim 1, characterized in that: Feeding back the corresponding evaluation value and the corresponding performance evaluation index parameter to the offline testing platform includes the following steps: According to the corresponding evaluation value and performance evaluation index parameter, each performance evaluation index parameter is analyzed to determine the priority corresponding to each performance evaluation index parameter, and the highest priority is sent to the offline testing platform as a key offline testing project.

3. A new energy vehicle health evaluation method according to claim 2, characterized in that: The S3 includes: S30, acquiring historical charging data, where the historical charging data includes data of each charging segment; S31, based on the data of each charging segment, respectively extract the label capacity and the health feature, and select the health feature with a correlation with the label capacity greater than a preset correlation threshold as input, input it into the pre-stored MLR model for training, and update the hyperparameters in the MLR model based on the SOH corresponding to the label capacity to form a trained MLR model; S32, obtaining vehicle data corresponding to the target vehicle to be measured; S33, acquiring corresponding health characteristics according to the acquired vehicle data; S34. Based on the trained MLR model, the acquired health features are used as input to perform corresponding capacity prediction, form corresponding capacity prediction values, and calculate the SOH prediction value of the target vehicle based on the capacity prediction values.

4. A new energy vehicle health evaluation method according to claim 3, characterized in that: The S31 includes: S310, according to each charging segment data, identify the type of each charging segment data, and pre-process according to the corresponding type, and extract the charging segment data corresponding to the pre-processed charging segment data, the difference between the charging end and the charging start SOC of which is greater than a preset threshold; S311, calculating the current capacity corresponding to each charging segment data according to the extracted charging segment data, forming a corresponding tag capacity, and calculating the corresponding SOH based on the SOH calculation formula; The SOH calculation formula is: In the formula, is the current capacity of the charging segment, is the SOC value at the start of the charging segment, is the SOC value at the end of the charging segment, and I is the current value of the charging segment; S312, screening and eliminating abnormal data for the tag capacity corresponding to each calculated charging segment data, and saving the processed tag capacity into a variable database; S313, according to the charging segment data at this time, intercept the data of the same voltage interval, and based on the correlation analysis, select the feature that is highly correlated with the corresponding tag capacity, and use the feature as the corresponding health feature; S314. Input the selected health features into the pre-stored MLR model, use the corresponding SOH as a label, train the MLR model, update the hyperparameters corresponding to the model, and form a new MLR model after training.

5. A new energy vehicle health evaluation method according to claim 4, characterized in that: The preprocessing according to the corresponding type in S310 includes: For the missing data in the charging segment data sampling, the missing data is filled based on the interpolation method; Delete data with multiple duplicate records; Delete incomplete charging and discharging processes or processes with obvious defects; The moving translation method is used to smooth the data of voltage, temperature and current with large fluctuations caused by noise during the sampling process.

6. A new energy vehicle health evaluation method according to claim 5, characterized in that: The S312 includes: S3120, according to the calculated tag capacities corresponding to the data of each charging segment, based on the first judgment strategy, eliminating obviously abnormal tag capacities; The first judgment strategy is: judging the relationship between the corresponding tag capacity and the nominal capacity, if , then the tag capacity is a normal capacity value, otherwise it is an obviously abnormal capacity value; In the formula, is the calculated label capacity, is the nominal capacity of the battery; S3121. According to the LOF algorithm, the tag capacity that has not been eliminated is eliminated and screened for the second time, and the processed tag capacity is saved in the variable database.

7. A new energy vehicle health evaluation method according to claim 6, characterized in that: The S313 includes: S3130, intercepting charging segment data corresponding to the same voltage interval based on the corresponding charging segment data and a preset interception strategy; S3131. According to the intercepted charging segment data and the corresponding tag capacity, based on the Pearson correlation coefficient analysis, select a feature whose correlation with the tag capacity is greater than a preset correlation threshold, and use the feature as the corresponding health feature.

8. A new energy vehicle health evaluation system, using a new energy vehicle health evaluation method according to any one of claims 1 to 7, characterized in that: Including online monitoring platform and offline testing platform; The online monitoring platform includes: A first acquisition module, used to obtain each user behavior evaluation index parameter corresponding to the target vehicle; A second acquisition module, used to acquire various appearance evaluation index parameters corresponding to the target vehicle; A third acquisition module is used to acquire various charging performance evaluation index parameters corresponding to the target vehicle; the charging performance evaluation index parameters include the SOH prediction value corresponding to the target vehicle; The fourth acquisition module is used to obtain the vehicle performance evaluation index parameters of the target vehicle; A fifth acquisition module, used to obtain electrical performance evaluation index parameters of a battery system of a target vehicle; A sixth acquisition module, used to acquire electric drive appearance performance evaluation index parameters, thermal management performance evaluation index parameters, reliability performance evaluation index parameters, and motor performance evaluation index parameters of the target vehicle; A processing module, used to perform health assessment and battery health detection on the target vehicle according to the obtained evaluation index parameters, based on the corresponding algorithm and the importance corresponding to each performance evaluation index parameter, calculate the evaluation value corresponding to each performance evaluation index parameter, and determine whether the corresponding evaluation value is greater than the safety threshold. If so, it is determined that the evaluation index parameter is abnormal, and the corresponding correction algorithm is called for correction, and it is determined whether it can be corrected by the correction algorithm. If so, it is automatically corrected. Otherwise, the corresponding evaluation value and the corresponding performance evaluation index parameter are fed back to the offline detection platform, and a corresponding offline detection report is generated; When correcting the evaluation index parameters, based on the highest threshold and the lowest threshold corresponding to each evaluation index parameter, it is judged whether the corresponding evaluation index parameter is too high or too low. If so, it is judged that the corresponding evaluation index parameter needs to be corrected. At this time, the difference range between the evaluation index parameter and the lowest threshold and the highest threshold is calculated to obtain the parameter range that needs to be corrected for the evaluation index parameter; Based on the correlation between the evaluation index parameter and other evaluation index parameters, select other evaluation index parameters with the largest correlation, and based on the correction relationship between the evaluation index parameter and other selected evaluation index parameters, as well as the corresponding parameter range that needs to be corrected, adjust the other selected evaluation index parameters so that the corresponding evaluation index parameter is corrected to within a reasonable threshold.

9. A new energy vehicle health evaluation system according to claim 8, characterized in that: The third acquisition module includes: A historical data acquisition module, used to acquire historical charging data, wherein the historical charging data includes data of each charging segment; A model training module is used to extract label capacity and health features based on the data of each charging segment, and select health features with a correlation with label capacity greater than a preset correlation threshold as input, input them into a pre-stored MLR model for training, and update the hyperparameters in the MLR model based on the SOH corresponding to the label capacity to form a trained MLR model; A target data acquisition module, used to acquire vehicle data corresponding to the target vehicle to be measured; An identification module, used to obtain corresponding health characteristics based on the acquired vehicle data; The health evaluation module is used to perform corresponding capacity prediction based on the trained MLR model, using the acquired health features as input, to form a corresponding capacity prediction value, and to calculate the SOH prediction value of the target vehicle based on the capacity prediction value.

Citation Information

Patent Citations

  • A method and system for health monitor and evaluation of new energy vehicle

    CN109460910A

  • Power battery health degree evaluation method, device and system

    CN110008235A

  • Abnormal data detection method and device, storage medium and electronic equipment

    CN111199018A

  • Vehicle battery health state assessment method and system

    CN113049976A

  • Data screening method and device, electronic equipment and storage medium

    CN114417090A