CN113158947B - Power battery health scoring method, system and storage medium - Google Patents
Power battery health scoring method, system and storage medium Download PDFInfo
- Publication number
- CN113158947B CN113158947B CN202110476738.5A CN202110476738A CN113158947B CN 113158947 B CN113158947 B CN 113158947B CN 202110476738 A CN202110476738 A CN 202110476738A CN 113158947 B CN113158947 B CN 113158947B
- Authority
- CN
- China
- Prior art keywords
- battery
- health
- data
- model
- score
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000036541 health Effects 0.000 title claims abstract description 135
- 238000013077 scoring method Methods 0.000 title claims description 14
- 230000006399 behavior Effects 0.000 claims abstract description 40
- 238000000034 method Methods 0.000 claims abstract description 31
- 230000007613 environmental effect Effects 0.000 claims abstract description 23
- 238000011156 evaluation Methods 0.000 claims abstract description 10
- 238000010801 machine learning Methods 0.000 claims abstract description 6
- 238000012549 training Methods 0.000 claims description 31
- 230000008569 process Effects 0.000 claims description 19
- 238000012360 testing method Methods 0.000 claims description 19
- 230000004927 fusion Effects 0.000 claims description 15
- 230000001419 dependent effect Effects 0.000 claims description 14
- 238000001914 filtration Methods 0.000 claims description 10
- 238000013145 classification model Methods 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000012216 screening Methods 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000007599 discharging Methods 0.000 claims description 4
- 238000000926 separation method Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 238000007689 inspection Methods 0.000 claims description 3
- 238000012417 linear regression Methods 0.000 claims description 3
- 239000013598 vector Substances 0.000 claims description 3
- 238000012423 maintenance Methods 0.000 description 6
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000007477 logistic regression Methods 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 1
- 230000001351 cycling effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/60—Other road transportation technologies with climate change mitigation effect
- Y02T10/70—Energy storage systems for electromobility, e.g. batteries
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Secondary Cells (AREA)
- Tests Of Electric Status Of Batteries (AREA)
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
Abstract
The invention provides a method for scoring the health of an electric vehicle battery, which comprises the following steps: s1, vehicle signal data are collected and preprocessed. And S2, acquiring and analyzing the current battery health state of the vehicle as a basic health score. And S3, analyzing the user behaviors and the environmental working conditions to influence the battery health. And S4, scoring the user behavior of S3 by using a scoring card model. And S5, analyzing the life decay trend of the vehicle battery. S6: and establishing a machine learning model, and predicting the attenuation degree of the vehicle battery along with time. And S7, predicting and scoring the battery attenuation according to the battery health attenuation model obtained in the S6. And S8, setting the weight of each score according to the scoring results of S2, S4 and S7, and carrying out weighted summation to obtain the final battery health score. According to the invention, the SOH of the battery, the driving habits and working conditions of a user, the remaining quality-guaranteed mileage and the future attenuation condition are comprehensively considered when the health score of the battery is evaluated, so that the current health degree of the battery of the vehicle is more objectively reflected, and a more objective basis is provided for the evaluation of the residual value of the battery.
Description
Technical Field
The invention relates to the technical field of power batteries of vehicles, in particular to a power battery health state evaluation technology.
Background
The core power of the electric automobile comes from the battery, and the health degree of the battery directly influences the performance, the endurance and the safety of the whole automobile. The battery has a limited life and will gradually degrade over time and with cycling until it is discarded and unusable. The cost of replacing batteries is enormous, so the health of the batteries becomes the focus of major concern for manufacturers and owners of the vehicles. Meanwhile, it is also an urgent need to be able to effectively and accurately evaluate the health degree of the battery.
The health of a battery is related to a number of factors. On the one hand, the energy comes from the battery, such as capacity, power, internal resistance, charging and discharging depth, recycling times and the like. And on the other hand, the vehicle using habits of a user vehicle owner, such as driving habits, charging habits, ambient temperature, daily maintenance and the like. The correct and standard use habit can keep the health state of the battery and prolong the service life of the battery.
The current method for reflecting the health degree of the battery is to detect and check the SOH value of the battery. The SOH value of a battery is generally calculated from the internal capacity, internal resistance, and the like of the battery. Only can reflect the influence factors of the battery, but not the influence of the vehicle using habit of the vehicle owner on the health of the battery. However, most owner users lack professional knowledge for internal management of the battery, use and maintenance of the battery, and the like. Therefore, the influence of the daily use habit of a user vehicle owner on the health of the battery can be reflected while the health degree of the battery can be effectively and accurately evaluated, and the influence has important significance and value in helping the user improve and improve the use habit, improve the battery maintenance knowledge, prolong the service life of the battery and the like.
Disclosure of Invention
The invention aims to establish a more objective method and a more objective system for grading the battery health of an electric vehicle, which are based on the big data of the Internet of vehicles, comprehensively consider the SOH of a battery, the driving habits and working conditions of a user, the remaining quality protection mileage and the future attenuation condition when evaluating the battery health score, more objectively reflect the current battery health degree of the vehicle, and provide more objective basis for evaluating the residual value of the battery.
The health of a vehicle battery is influenced by the vehicle owner's habits, in addition to the characteristics of the battery itself. Therefore, in the battery health evaluation formula of the electric vehicle, various factors affecting the battery health need to be comprehensively considered, such as: the battery self-factor, the environmental working condition, the user behavior habit of using the vehicle, the daily maintenance and other factors. In the invention, the influence result of the battery self factor is obtained by the SOH value of the battery, and the results of other factors are obtained by the relevant data in the vehicle signal through big data analysis and machine learning modeling. Each factor gets a score and a corresponding weight is assigned and finally weighted and summed to get a final overall score for the battery health.
In order to achieve the above purpose, the present invention proposes the following technical solutions based on the above ideas.
A method for scoring the health of an electric automobile battery is based on Internet of vehicles big data and comprises the following steps:
s1: collecting vehicle signal data and carrying out data preprocessing; the vehicle signal data comprises data of three categories including the inside of a battery, user behaviors and environmental conditions, and is big data obtained through the Internet of vehicles.
S2: the current battery state of health of the vehicle is acquired and analyzed as a base health score.
S3: and analyzing the influence of user behaviors and environmental conditions on the health of the battery.
S4: and scoring the user behavior of S3 by using a scoring card model.
S5: and analyzing the life decay trend of the vehicle battery.
S6: and establishing a machine learning model and predicting the attenuation degree of the vehicle battery along with time.
S7: and predicting and scoring the battery attenuation according to the battery health attenuation model obtained in the step S6.
S8: and setting the weight of each score in the processes of S2, S4 and S7 according to the scoring results of S2, S4 and S7, and weighting and summing to obtain the final battery health score.
Further, the step S1 includes the following steps:
s1-1: the method comprises the steps of collecting signal data such as vehicle types, usage types, battery BMS, vehicle owner behavior data and environmental conditions in a big data mode, wherein the vehicle signal data are big data based on the Internet of vehicles.
S1-2: and classifying the data into three categories, namely the interior of the battery, the user behavior and the environmental working condition according to factors influencing the health of the battery.
S1-3: preprocessing data, and deleting abnormal data such as null values, noise, invalid values and the like.
Further, the step S2 further includes the following steps:
s2-1: and acquiring data of the classified and processed battery internal categories in the S1 process.
S2-2: and checking the SOH value of the current battery health state and analyzing the internal state data of the battery. Including current battery capacity, rated capacity, internal resistance, self-discharge rate, etc.
S2-3: and carrying out basic scoring on the health internal state of the battery according to the current SOH value of the battery and the analysis result of the internal state of the battery.
Further, the step S3 further includes the following steps:
s3-1: and acquiring data of the user behavior category and the environmental working condition category which are classified and processed in the S1 process.
S3-2: and analyzing and counting user behavior data, including accumulated driving mileage, vehicle usage, charging behavior, discharging behavior, DOD depth of discharge, overcharging behavior, usage ratio of fast and slow charging and the like.
S3-3: and analyzing and counting environmental working condition data, including temperature environment, under-voltage and over-voltage conditions, high SOC standing conditions and the like.
Further, the step S4 includes the following steps:
s4-1: constructing a scoring card model sample data set for training and evaluation, wherein the scoring card model sample data set comprises positive and negative sample data;
s4-2: dividing a sample data set into a training set and a test set;
s4-3: processing characteristic engineering, namely performing variable binning and discretization;
s4-4: calculating the WOE (Weight Of evaluation) Value and the IV (Information Value) Value Of the characteristic variable, and performing characteristic screening and binning inspection;
s4-5: training a classification model on a training data set, and verifying the effect of the classification model in a test set;
s4-6: and (4) score conversion is carried out on the scoring card, and scores and model scores of all the variable sub-boxes are output.
Further, the step S5 includes the following steps:
s5-1: acquiring data of three categories which are classified and processed in the S1 process;
s5-2: marking the internal category data of the battery as a dependent variable to form a dependent variable pool;
s5-3: marking the user behavior category and the environmental working condition category data as independent variables to form an independent variable pool;
s5-4: and analyzing the correlation between the variables in the independent variable pool and the dependent variable pool and the influence degree of each variable on each dependent variable.
Further, the step S6 includes the following steps:
s6-1: constructing a sample data set of a battery health decay model for training and evaluation;
s6-2: dividing a sample data set into a training set and a test set;
s6-3: performing feature screening according to the analysis result in the S5 process;
s6-4: performing feature fusion and filtering processing to generate a Health factor HI (Health Index) for representing the Health attenuation of the battery;
s6-5: analyzing the influence and trend of the change of the health factor HI along with time;
s6-6: training a battery health attenuation model on a training data set to generate a battery health attenuation model pool;
s6-7: and verifying the battery health attenuation model on the test data set, and evaluating the model effect.
Further, the step S7 includes the following steps:
s7-1: constructing variables for inputting the battery health decay model from the vehicle data;
s7-2: and inputting the variables into the model, outputting the battery health prediction result by the model, and converting the battery health prediction result into a score.
The invention also provides a battery health scoring system for an electric vehicle, which comprises a memory and a processor, wherein the memory stores instructions for enabling the processor to execute the battery health scoring method for the electric vehicle.
The invention also provides a machine-readable storage medium, wherein the machine-readable storage medium is stored with instructions for enabling a machine to execute the electric vehicle battery health scoring method.
The method utilizes the internal mechanism of the battery and a machine learning model to analyze and mine various factors influencing the health of the battery from multiple angles, and converts the influence degree of the factors into a score form. When the health degree of the battery is accurately evaluated and predicted, the vehicle owner is helped to locate the reasons influencing the reduction of the battery health in a scoring mode, improvement and optimization suggestions of the vehicle owner in the aspects of vehicle using behaviors, environmental working conditions, daily maintenance and the like are given, and finally the vehicle owner is helped to prolong the service life of the vehicle battery.
Drawings
FIG. 1 is a schematic diagram of a scoring method for evaluating the health of a battery of an electric vehicle according to an embodiment of the present invention;
Detailed Description
The invention is further described with reference to the drawings and examples in the following description:
referring to fig. 1, the present embodiment provides a method for scoring the health of a battery of an electric vehicle,
the method comprises the following specific steps:
s1: the method comprises the steps of collecting vehicle signal data and preprocessing the data, wherein the vehicle signal data comprise three categories of battery interior, user behavior and environment working conditions.
S2: obtaining and analyzing the current battery health state of the vehicle as a basic score;
s3: analyzing the influence of user behaviors and environmental conditions on the health of the battery;
s4: scoring the user behavior of S3 using a scoring card model;
s5: analyzing the life decay trend of the vehicle battery;
s6: establishing a machine learning model, and predicting the attenuation degree of the vehicle battery along with time;
s7: predicting and scoring the battery attenuation according to the battery health attenuation model obtained in the S6;
s8: and calculating the total battery health score according to the scoring results of the S2, the S4 and the S7.
In a further embodiment of the present invention, said step S1 further comprises the steps of:
s1-1: collecting signal data such as vehicle type, use type, battery BMS, vehicle use behavior data of a vehicle owner, environmental working conditions and the like;
s1-2: classifying data according to factors influencing the health of the battery, and dividing the data into three categories, namely battery interior, user behavior and environmental working conditions;
s1-3: preprocessing data, and deleting abnormal data such as null values, noise, invalid values and the like.
In a further embodiment of the present invention, said step S2 further comprises the steps of:
s2-1: acquiring data of the internal categories of the battery classified and processed in the S1 process;
s2-2: and checking the SOH value of the current battery state of health and analyzing the internal state data of the battery. Including current capacity, rated capacity, internal resistance, self-discharge rate, etc.;
s2-3: and performing basic scoring on the internal state of health of the battery according to the current SOH value of the battery and the analysis result of the internal state of the battery.
The SOH value in the BMS signal of the vehicle battery indicates the current state of health of the battery, and it can be found by analyzing the relationship between the internal state of the battery and the current SOH value of the battery state of health: the state data of the residual capacity, the internal resistance and the like of the battery are strongly related to the SOH value. For example, the smaller the remaining battery capacity is, the lower the SOH value is, and the larger the internal resistance is, the lower the SOH value is. The SOH value may be used to adequately represent the current internal state of the battery and the battery health. Therefore, in this step, only the SOH value is used to derive the base score, which is expressed as follows:
score base ＝soh*100
wherein, score base Indicates a basal score at 0,100]Within the interval.
In a further embodiment of the present invention, said step S3 further comprises the steps of:
s3-1: acquiring data of the user behavior category and the environmental working condition category which are classified and processed in the S1 process;
s3-2: analyzing and counting user behavior data, including accumulated driving mileage, vehicle usage, charging behavior, discharging behavior, DOD depth of discharge, overcharge behavior, proportion of fast and slow charging, and the like;
s3-3: and analyzing and counting environmental working condition data, including temperature environment, under-voltage and over-voltage conditions, high SOC standing conditions and the like.
Wherein, the environmental condition class data need carry out the discrete of branch case. For example, the temperature data are labeled as discrete bins of [10 ℃,20 ℃ ], [20 ℃,35 ℃ ], [35 ℃,40 ℃ ], and [ others ].
In a further embodiment of the present invention, said step S4 further comprises the steps of:
s4-1: constructing a scoring card model sample data set for training and evaluation, wherein the scoring card model sample data set comprises positive and negative sample data; these sample data are derived from the data relating to user behavior and environmental conditions in step S3.
S4-2: dividing a sample data set into a training set and a test set;
s4-3: processing characteristic engineering, namely performing variable binning and discretization;
and performing box discretization on continuous characteristic variables in all data sets, and processing by using modes such as equal-frequency/equal-distance box separation, chi-square box separation, supervised box separation and the like according to different characteristics.
S4-4: computing WOE (Weight Of event) values and IV (Information Value) values Of characteristic variables, and performing characteristic screening and binning inspection;
WOE (Weight Of Evidence) represents Evidence Weight, and the larger the value is, the larger the difference Of the positive and negative sample proportions in the bin is. The IV (Information Value) is a measure of the Information amount of a variable, and the larger the Value is, the larger the Information amount is. In feature screening, IV values were used for selection. And removing characteristic variables with over-small IV values.
S4-5: training a classification model on a training data set, and verifying the effect of the classification model in a test set.
Further, the specific steps of training and testing in step S4-5 are as follows:
s4-5-1: training a classification model by using a training set and a cross validation mode to find an optimal parameter model;
the classification in this process belongs to a two-classification problem, using a logistic regression classification model:
z＝θ 0 +θ 1 x1 + θ 2 x 2 +…+θ n x n
where θ represents a parameter and x represents a characteristic variable. h is a total of θ Representing Sigmoid function, e, used in logistic regression model -z Is an exponential function with a base number of euler numbers e and an exponent of-z.
S4-5-2: using the test set data, the effect of the model was assessed by observing the AUC values of the confusion matrix, ROC curve.
S4-6: and (4) score conversion is carried out on the scoring card, and scores and model scores of all the variable sub-boxes are output.
The formula of the conversion score of the scoring card is as follows:
A＝P 0 -B*ln(θ 0 )
score＝A+B*ln(odds)
where p represents the negative sample probability and odds represents the occurrence ratio. P is 0 PDO is constant, A and B have no practical meaning, and belong to an intermediate value in calculating the fraction, and can be represented by P 0 And calculating the PDO. score represents the converted score.
In a further embodiment of the present invention, said step S5 further comprises the steps of:
s5-1: acquiring data of three categories classified and processed in the S1 process;
s5-2: marking the internal category data of the battery as a dependent variable to form a dependent variable pool;
s5-3: marking the user behavior type and environment working condition type data as independent variables to form an independent variable pool;
s5-4: and analyzing the correlation between the variables in the independent variable pool and the dependent variable pool and the influence degree of each variable on each dependent variable.
In a further embodiment of the present invention, said step S6 further comprises the steps of:
s6-1: constructing a sample data set of a battery health decay model for training and evaluation;
s6-2: dividing a sample data set into a training set and a test set;
s6-3: performing feature screening according to the analysis result in the S5 process;
s6-4: and performing feature fusion and filtering processing to generate a Health factor HI (Health Index) for representing the Health attenuation of the battery.
Further, the specific steps of feature fusion and filtering processing in step S6-4 are as follows:
s6-4-1: constructing a data set due to feature fusion;
the state of health at the beginning of battery life is defined as 1 and the state of health at the end of battery life is defined as 0. The data of soh >97 in the training set is marked as 1, and the data of soh less than or equal to 80 is marked as 0. The labeled data is then extracted from the training set to form a data set ω for feature fusion:
ω＝{(X，y)}
ω＝{(x i ，0)|soh≤80}∪{(x i ，1)|soh＞97}
s6-4-2: performing feature fusion by using linear regression to generate a health factor HI;
in this process, the health factors were generated using linear regression fusion, the fusion model being as follows:
wherein, y HI Representing the health factor (HI), alpha, beta representing the parameters in the model, x i Representing the feature vector, and epsilon is a noise term to control the overfitting.
S6-4-3: filtering the health factor;
because the health factor (HI) generated after fusion is noisy and has large fluctuation, the Sav _ gol can be further used for filtering processing, and the error of health attenuation is reduced.
S6-5: the influence and trend of the change of the health factor HI along with the time are analyzed.
S6-6: training a battery health attenuation model on a training data set to generate a battery health attenuation model pool;
a plurality of battery attenuation models trained in the process are formed, and a battery health attenuation model pool is formed. The model formula is as follows:
wherein y represents a dependent variable health factor (HI), a, b, c represent parameters in the model, and T represents sdj Representing the time period of the independent variable, epsilon is a noise term to control the overfitting.
S6-7: verifying a battery health attenuation model on a test data set, and evaluating the effect of the model;
the step of validating the model in the test dataset in step S6-7 is as follows:
s6-7-1: filtering soh characteristic variables in the test set;
s6-7-2: performing S6-4 feature fusion and filtering to generate a health factor HI for the test data test ；
S6-7-3: mixing HI test Carrying out similarity comparison with HI of each model in the model pool in the S6-6 step;
similarity comparison uses a way of calculating the euclidean distance of the feature vectors, with closer distances being higher similarities.
S6-7-4: sorting the obtained similarity according to a descending order, and filtering out a model with the similarity smaller than a 3-quantile numerical value;
s6-7-5: respectively predicting by using the rest models, and taking the average value of the predicted values of the models as a final health state value;
s6-7-5: and comparing the obtained predicted value with the soh value which is not filtered in the S6-7-1, wherein the smaller the error is, the better the model effect is.
In a further embodiment of the present invention, said step S7 further comprises the steps of:
s7-1: constructing variables for inputting the battery health decay model according to the vehicle data;
s7-2: inputting the variables into a model, outputting a battery health prediction result by the model, and converting the battery health prediction result into a score;
the conversion score formula is as follows:
score＝y predicted *100
wherein score represents a score, y predioted Representing the model predicted battery state of health value.
In a further embodiment of the present invention, said step S8 further comprises the steps of:
s8-1: obtaining scoring results in the processes of S2, S4 and S7;
s8-2: setting the weight of each score in the S2, S4 and S7 processes, and weighting and summing to obtain a final battery health score;
the final battery health score formula is as follows:
score final ＝w 1 s 1 +w 2 s 2 +w 3 s 3
wherein, score final Score for Battery health, w 1 s 1 Representing base scores and corresponding weights, w 2 s 2 Representing user behavior, environmental condition scores and corresponding weights, w 3 s 3 Representing the battery health decay score and corresponding weight.
According to the detailed content of the embodiment, the method respectively adopts the basic health score, the user vehicle using behavior habit score and the battery attenuation score to obtain the battery health score from different angles, and the comprehensive and comprehensive battery health score is carried out from different angles and different methods. Therefore, the health degree of the battery can be effectively and accurately evaluated, meanwhile, the influence of the daily use habit of a user owner on the health of the battery can be reflected, a basis is provided, and the user is helped to improve the use habit, improve the battery maintenance knowledge, prolong the service life of the battery and the like.
Claims (10)
1. A method for scoring the health of a battery of an electric automobile is characterized by comprising the following steps:
s1: collecting vehicle signal data and carrying out data preprocessing; the vehicle signal data comprises data of three categories including the inside of a battery, user behaviors and environmental working conditions;
s2: obtaining and analyzing the current battery health state of the vehicle as a basic score;
s3: analyzing the influence of user behaviors and environmental conditions on the health of the battery;
s4: scoring the user behavior of S3 using a scoring card model;
s5: analyzing the life decay trend of the vehicle battery;
s6: establishing a machine learning model, and predicting the attenuation degree of the vehicle battery along with time;
s7: predicting and scoring the battery attenuation according to the battery health attenuation model obtained in the S6;
s8: setting the weight of each score in the processes of S2, S4 and S7 according to the score results of S2, S4 and S7, and weighting and summing to obtain the final battery health score;
the step S6 includes:
s6-1: constructing a sample data set of a battery health decay model for training and evaluation;
s6-2: dividing a sample data set into a training set and a test set;
s6-3: performing feature screening according to the analysis result in the S5 process;
s6-4: performing feature fusion and filtering processing to generate a Health factor HI (Health Index) for representing the Health attenuation of the battery;
s6-5: analyzing the influence and trend of the change of the health factor HI along with time;
s6-6: training a battery health attenuation model on a training data set to generate a battery health attenuation model pool;
the model formula is as follows:
wherein y represents a dependent variable health factor (HI), a, b, c represent parameters in the model, and T represents adj Representing the time period of the independent variable, epsilon is a noise term to control the overfitting;
s6-7: verifying a battery health attenuation model on a test data set, and evaluating the effect of the model;
the step S6-4 comprises the following steps:
s6-4-1: constructing a data set for feature fusion;
defining the state of health of the beginning of the battery life as 1, the state of health of the end of the battery life as 0, marking the data of soh >97 in the training set as 1, marking the data of soh less than or equal to 80 as 0, and then extracting the marked data from the training set to form a data set omega for feature fusion:
ω＝{(X,y)}
ω＝{(x i ,0)|soh≤80}∪{(x i ,1)|soh>97}
where X represents a characteristic of the sample in the dataset, y represents a label for the sample (i.e., battery state of health), X i Representing a feature of sample i;
s6-4-2: performing feature fusion by using linear regression to generate a health factor HI;
the fusion model is as follows:
wherein, y HI Indicates health factor (HI), and alpha and beta indicate modulusParameters in the form, x representing the characteristic variables, x i A feature vector representing sample i, ε being a noise term used to control the overfitting;
s6-4-3: and (5) filtering the health factor.
2. The electric vehicle battery health scoring method according to claim 1, wherein the step S1 comprises:
s1-1: collecting vehicle signal data, including vehicle type, use type, battery BMS, vehicle owner use behavior data and environment working condition; the vehicle signal data is big data based on an internet of vehicles;
s1-2: classifying data into three categories, namely battery interior, user behavior and environmental working condition according to factors influencing battery health;
s1-3: and preprocessing data, and deleting null values, noise and invalid values.
3. The electric vehicle battery health scoring method according to claim 1, wherein the step S2 comprises:
s2-1: acquiring data of the internal categories of the battery classified and processed in the S1 process;
s2-2: analyzing internal state data of the battery according to the SOH value of the current battery state of health, wherein the internal state data comprises current battery capacity, rated capacity, internal resistance and self-discharge rate;
s2-3: and performing basic scoring on the internal state of health of the battery according to the current SOH value of the battery and the analysis result of the internal state data of the battery.
4. The electric vehicle battery health scoring method according to claim 3, wherein the base score is derived from an SOH value, and the formula is as follows:
score base ＝soh*100
wherein, score base Indicates a basal score, the basal score being [0,100 ]]Within the interval.
5. The electric vehicle battery health scoring method according to claim 1, wherein the step S3 comprises:
s3-1: acquiring data of the user behavior category and the environmental working condition category which are classified and processed in the S1 process;
s3-2: analyzing and counting user behavior data, including accumulated driving mileage, vehicle usage, charging behavior, discharging behavior, DOD depth of discharge, overcharge behavior, and proportion of charge and discharge;
s3-3: and analyzing and counting environmental working condition data including temperature environment, under-voltage and over-voltage conditions and high SOC standing conditions.
6. The electric vehicle battery health scoring method according to claim 1, wherein the step S4 comprises:
s4-1: constructing a scoring card model sample data set for training and evaluation, wherein the scoring card model sample data set comprises positive and negative sample data;
s4-2: dividing a sample data set into a training set and a test set;
s4-3: performing characteristic engineering treatment, namely performing variable binning and discretization;
s4-4: calculating an evidence weight WOE value and an information quantity IV value of the characteristic variable, and performing characteristic screening and box separation inspection;
s4-5: training a classification model on a training data set, and verifying the effect of the classification model in a test set;
s4-6: and (4) score conversion is carried out on the scoring card, and scores and model scores of all the variable sub-boxes are output.
7. The electric vehicle battery health scoring method according to claim 1, wherein the step S5 comprises:
s5-1: acquiring data of three categories which are classified and processed in the S1 process;
s5-2: marking the internal category data of the battery as a dependent variable to form a dependent variable pool;
s5-3: marking the user behavior type and environment working condition type data as independent variables to form an independent variable pool;
s5-4: and analyzing the correlation between the variables in the independent variable pool and the dependent variable pool and the influence degree of each variable on each dependent variable.
8. The electric vehicle battery health scoring method according to claim 1, wherein the step S7 comprises:
s7-1: constructing variables for inputting the battery health decay model from the vehicle data;
s7-2: inputting the variables into a model, outputting a battery health prediction result by the model, and converting the battery health prediction result into a score;
the conversion score formula is as follows:
score＝y predicted *100
wherein score represents a score, y predicted Representing the model predicted battery state of health value.
9. An electric vehicle battery health scoring system, characterized in that the system comprises a memory and a processor, wherein the memory stores instructions for enabling the processor to execute the electric vehicle battery health scoring method according to any one of claims 1 to 8.
10. A machine-readable storage medium having instructions stored thereon for enabling a machine to perform the electric vehicle battery health scoring method according to any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110476738.5A CN113158947B (en) | 2021-04-29 | 2021-04-29 | Power battery health scoring method, system and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110476738.5A CN113158947B (en) | 2021-04-29 | 2021-04-29 | Power battery health scoring method, system and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113158947A CN113158947A (en) | 2021-07-23 |
| CN113158947B true CN113158947B (en) | 2023-04-07 |
Family
ID=76872810
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110476738.5A Active CN113158947B (en) | 2021-04-29 | 2021-04-29 | Power battery health scoring method, system and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113158947B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112083736B (en) * | 2020-08-11 | 2023-07-25 | 广东电网有限责任公司电力科学研究院 | Unmanned aerial vehicle tracking method |
| CN113884961B (en) * | 2021-09-23 | 2023-09-19 | 中国第一汽车股份有限公司 | SOC calibration method, modeling device, computer equipment and medium |
| CN114415049B (en) * | 2022-01-07 | 2024-03-29 | 中国第一汽车股份有限公司 | Method, device, equipment and storage medium for determining storage battery health degree scoring card |
| CN114895208A (en) * | 2022-04-07 | 2022-08-12 | 合众新能源汽车有限公司 | Estimation method and screening method for SOH (state of health) of automobile power battery |
| CN114910792B (en) * | 2022-04-08 | 2024-10-01 | 中国第一汽车股份有限公司 | Power battery charging depth evaluation device, terminal and storage medium |
| CN114840983A (en) * | 2022-04-18 | 2022-08-02 | 中汽数据(天津)有限公司 | Power storage battery SOH calculation model construction method and calculation method |
| CN114841557A (en) * | 2022-04-29 | 2022-08-02 | 常德市三一机械有限公司 | Engineering equipment health assessment method and system and engineering equipment |
| CN114707908B (en) * | 2022-05-18 | 2022-08-23 | 北京钛镁新技术有限公司 | A kind of power battery rating determination method, device and electronic equipment |
| CN114999134B (en) * | 2022-05-26 | 2024-05-28 | 北京新能源汽车股份有限公司 | Driving behavior early warning method, device and system |
| CN115097345B (en) * | 2022-06-27 | 2024-11-12 | 重庆长安汽车股份有限公司 | Battery health determination method, system, electronic device and readable storage medium |
| CN115166535A (en) * | 2022-08-03 | 2022-10-11 | 江苏华友能源科技有限公司 | SOC correction method based on cloud big data platform |
| CN115564069A (en) * | 2022-09-28 | 2023-01-03 | 北京百度网讯科技有限公司 | Method for determining server maintenance strategy, method for generating model and device thereof |
| CN116265974A (en) * | 2023-03-02 | 2023-06-20 | 尼讯(上海)科技有限公司 | A battery pack capacity detection system |
| CN116430244B (en) * | 2023-06-14 | 2023-08-15 | 聊城大学 | Power battery health state estimation method based on voltage and current characteristics |
| CN117648631B (en) * | 2024-01-29 | 2024-05-28 | 陕西德创数字工业智能科技有限公司 | Power battery health state estimation method for electric automobile group |
| CN118866145B (en) * | 2024-09-25 | 2024-12-17 | 南通国轩新能源科技有限公司 | Method and system for analyzing attenuation of electrochemical performance of energy storage battery based on cloud computing |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108398652A (en) * | 2017-05-26 | 2018-08-14 | 北京航空航天大学 | A kind of lithium battery health state evaluation method merging deep learning based on multilayer feature |
| CN110045288A (en) * | 2019-05-23 | 2019-07-23 | 中山大学 | A kind of capacity of lithium ion battery On-line Estimation method based on support vector regression |
| CN111443294A (en) * | 2020-04-10 | 2020-07-24 | 华东理工大学 | Method and device for indirect prediction of remaining life of lithium ion battery |
| CN111552926A (en) * | 2020-04-28 | 2020-08-18 | 重庆长安新能源汽车科技有限公司 | Driving behavior evaluation method and system based on Internet of vehicles and storage medium |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8785758B2 (en) * | 2010-09-01 | 2014-07-22 | Inmusic Brands, Inc. | Electronic hi-hat cymbal controller |
| CN103954913B (en) * | 2014-05-05 | 2017-06-30 | 哈尔滨工业大学深圳研究生院 | Electric automobile power battery life-span prediction method |
| US20160178706A1 (en) * | 2014-12-17 | 2016-06-23 | National Chung Shan Institute Of Science And Technology | Method and apparatus of detecting states of battery |
| CN109598095B (en) * | 2019-01-07 | 2023-08-08 | 平安科技（深圳）有限公司 | Method and device for establishing scoring card model, computer equipment and storage medium |
| CN110008235A (en) * | 2019-04-15 | 2019-07-12 | 优必爱信息技术（北京）有限公司 | Power battery health degree evaluation method, apparatus and system |
| CN110133508B (en) * | 2019-04-24 | 2022-04-01 | 上海博强微电子有限公司 | Safety early warning method for power battery of electric automobile |
| CN110275119A (en) * | 2019-08-01 | 2019-09-24 | 优必爱信息技术（北京）有限公司 | A kind of cell health state assessment models construction method, appraisal procedure and device |
| CN112083337B (en) * | 2020-10-22 | 2023-06-16 | 重庆大学 | A power battery health prediction method for predictive operation and maintenance |
-
2021
- 2021-04-29 CN CN202110476738.5A patent/CN113158947B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108398652A (en) * | 2017-05-26 | 2018-08-14 | 北京航空航天大学 | A kind of lithium battery health state evaluation method merging deep learning based on multilayer feature |
| CN110045288A (en) * | 2019-05-23 | 2019-07-23 | 中山大学 | A kind of capacity of lithium ion battery On-line Estimation method based on support vector regression |
| CN111443294A (en) * | 2020-04-10 | 2020-07-24 | 华东理工大学 | Method and device for indirect prediction of remaining life of lithium ion battery |
| CN111552926A (en) * | 2020-04-28 | 2020-08-18 | 重庆长安新能源汽车科技有限公司 | Driving behavior evaluation method and system based on Internet of vehicles and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 基于机器学习的新能源汽车电池剩余寿命预测;王正;《机械与电子》;20191224(第12期);全文 * |
| 锂离子电池剩余寿命预测研究;戴海峰等;《电源技术》;20191220(第12期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113158947A (en) | 2021-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113158947B (en) | Power battery health scoring method, system and storage medium | |
| CN112092675B (en) | Battery thermal runaway early warning method, system and server | |
| Song et al. | Intelligent state of health estimation for lithium-ion battery pack based on big data analysis | |
| CN103954913B (en) | Electric automobile power battery life-span prediction method | |
| CN113281671A (en) | Lithium ion battery remaining service life prediction method and system based on IGS-SVM | |
| CN114280479B (en) | Retired battery rapid sorting method based on electrochemical impedance spectroscopy | |
| CN109344517A (en) | A kind of high-voltage isulation method for diagnosing faults of new-energy automobile | |
| CN114460481B (en) | Bi-LSTM and attention mechanism-based energy storage battery thermal runaway early warning method | |
| Yao et al. | Fault identification of lithium-ion battery pack for electric vehicle based on ga optimized ELM neural network | |
| CN117099115A (en) | Carbon emission monitoring method, system, equipment and medium for recycling battery | |
| CN118191628A (en) | Machine learning-based mining lithium ion battery SOC prediction method and system | |
| CN118305097A (en) | Rapid retired power battery sorting method based on interval frequency electrochemical impedance spectrum | |
| CN116881701A (en) | New energy automobile power supply safety detection method | |
| Ezzouhri et al. | A data-driven-based framework for battery remaining useful life prediction | |
| CN118478695B (en) | A safety warning method, device and electronic equipment for power battery | |
| CN119535227A (en) | A lithium battery health state estimation method based on multi-coefficient fusion | |
| CN118425818A (en) | A power battery risk assessment method integrating health decline and consistency | |
| CN118393364A (en) | A power battery system multi-fault diagnosis method, device, medium and product | |
| Wang et al. | A Battery State of Health Estimation Method for Real-World Electric Vehicles Based on Physics-Informed Neural Networks | |
| Zhou et al. | Optimized LSTM based on an improved sparrow search algorithm for power battery fault diagnosis in new energy vehicles | |
| CN119150115B (en) | Classification method and system for re-matching application scenes for retired batteries | |
| CN114676619B (en) | Thermal runaway warning method for energy storage batteries based on autoencoder and ensemble learning | |
| CN113451665B (en) | Vehicle power battery maintenance feature identification method and device | |
| US20240377460A1 (en) | Secondary battery capacity estimation system | |
| Zhang et al. | State of health estimation for lithium batteries using improved weighted grey relational method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 401133 room 208, 2 house, 39 Yonghe Road, Yu Zui Town, Jiangbei District, Chongqing Patentee after: Deep Blue Automotive Technology Co.,Ltd. Address before: 401133 room 208, 2 house, 39 Yonghe Road, Yu Zui Town, Jiangbei District, Chongqing Patentee before: CHONGQING CHANGAN NEW ENERGY AUTOMOBILE TECHNOLOGY Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |