CN118551325A - Battery pack consistency anomaly detection method and system based on big data - Google Patents
Battery pack consistency anomaly detection method and system based on big data Download PDFInfo
- Publication number
- CN118551325A CN118551325A CN202410969533.4A CN202410969533A CN118551325A CN 118551325 A CN118551325 A CN 118551325A CN 202410969533 A CN202410969533 A CN 202410969533A CN 118551325 A CN118551325 A CN 118551325A
- Authority
- CN
- China
- Prior art keywords
- battery pack
- battery
- information
- battery packs
- packs
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/385—Arrangements for measuring battery or accumulator variables
- G01R31/387—Determining ampere-hour charge capacity or SoC
- G01R31/388—Determining ampere-hour charge capacity or SoC involving voltage measurements
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/392—Determining battery ageing or deterioration, e.g. state of health
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/396—Acquisition or processing of data for testing or for monitoring individual cells or groups of cells within a battery
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Probability & Statistics with Applications (AREA)
- Secondary Cells (AREA)
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
Abstract
The invention provides a big data-based battery pack consistency anomaly detection method and a big data-based battery pack consistency anomaly detection system, which relate to the technical field of battery pack anomaly detection, wherein a battery pack set to be compared is determined from multiple dimensions, sample data finally used for training a battery pack anomaly prediction model is determined according to battery usage habit data and driving habit data of the battery pack set to be compared, the inconsistency and fault types of the battery pack to be detected can be predicted based on the battery pack anomaly prediction model, and a first credibility is tested and obtained to represent the credibility of the first inconsistency. According to the invention, the sample data for training the model is determined from the basic information of the battery pack, the automobile brand, the battery using habit, the driving habit and other dimensions, so that the prediction results of the inconsistent condition and the fault type of the battery are more accurate, and the rationality of the maintenance strategy of the battery pack can be further improved through the first credibility.
Description
Technical Field
The invention belongs to the technical field of battery abnormality detection, and particularly relates to a battery pack consistency abnormality detection method and system based on big data.
Background
With the continuous increase of the market preservation amount of new energy automobiles, the performance of the battery pack gradually becomes the development bottleneck of the new energy automobiles, and the existing battery pack consistency detection technology can only be completed through various technical indexes of the battery pack, so that the detection efficiency is low, and the required labor cost is high.
Disclosure of Invention
The invention aims to provide a battery pack consistency anomaly detection method and system based on big data, which are used for solving the technical problems of low battery pack consistency detection efficiency and high labor cost in the prior art.
The battery pack consistency abnormality detection method based on big data comprises the following steps:
s1: determining A first battery packs according to battery pack basic information and automobile brand information of the battery packs to be detected;
s2: for A first battery packs, obtaining A first battery pack usage habit data, and clustering the A first battery pack usage habit data to obtain B first clustering centers;
S3: for A first battery packs, obtaining driving habit data of the A first battery packs, and clustering the driving habit data of the A first battery packs to obtain C second aggregation centers;
s4: obtaining D second battery packs according to the first clustering center and the second clustering center;
s5: taking the battery pack basic information, the first automobile brand information, the first battery pack using habit data and the first battery pack driving habit data of the D second battery packs as inputs, taking inconsistent conditions and fault types of the battery packs as outputs, and training to obtain a battery pack abnormal condition prediction model;
S6: inputting the basic information, the brand information, the using habit data and the driving habit data of the battery pack to be detected into the abnormal condition prediction model of the battery pack to obtain a first inconsistent condition and a first fault type;
s7: performing a first charging test on the battery pack to be detected to obtain a first charging curve, and performing similarity matching on the first charging curve and a standard charging curve to obtain a first similarity score; the standard charging curve is a charging curve obtained after a normal battery pack is subjected to a charging test;
S8: obtaining first credibility of the first inconsistent situation according to the first similarity score, and formulating a battery pack maintenance strategy according to the first credibility, the first inconsistent situation and the first fault type;
Wherein A, B, C, D are natural numbers greater than 1.
Preferably, the step S1 includes the following sub-steps:
S11: generating a first image of the battery pack to be detected according to the basic information of the battery pack;
s12: matching in the basic battery set by using the first portrait to obtain A1 second battery sets;
S13: screening the A1 second battery packs according to the automobile brand information to obtain A first battery packs;
the step S13 specifically comprises the following substeps:
s131: obtaining M first automobile brands to form a first automobile brand set;
S132: for each first automobile brand, acquiring a first market value tag and a first tombstone tag, and forming first automobile brand information according to the first market value tag and the first tombstone tag;
S133: performing similarity matching on the automobile brand information and M pieces of first automobile brand information, and determining the first automobile brand information with the similarity larger than a preset value as second automobile brand information;
s134: determining battery pack information corresponding to each piece of second automobile brand information as a fourth battery pack, and determining intersections of a plurality of fourth battery packs and the A1 second battery packs as A first battery packs;
Both A, A and M are natural numbers greater than 1, and M > A1> A.
Preferably, the step S132 includes the following substeps:
S1321: for each first automobile brand, taking the first automobile brand and battery pack information used by the first automobile brand as keywords, and acquiring a plurality of pieces of first news information;
S1322: inputting each piece of the first news information into a tendency judgment model to obtain a first tendency judgment opinion related to the battery pack of the first automobile brand;
S1323: integrating the plurality of first tendency judgment opinions, extracting a plurality of pieces of first opinion information, determining the first opinion information meeting the preset appearance frequency requirement in the plurality of pieces of first opinion information as second opinion information, and integrating the plurality of pieces of second opinion information to obtain the first public praise label.
Preferably, the step S2 includes the following sub-steps:
S21: acquiring first charging data, first discharging lower limit data and first charging upper limit data of each first battery pack, and forming A first battery pack use habit data;
S22: K-MEANS clustering is carried out on the A first battery pack usage habit data so as to obtain B first clustering centers;
And B is a natural number which is more than or equal to 1 and less than A.
Preferably, the step S3 includes the following sub-steps:
s31: acquiring a first average driving speed, a first starting stage average acceleration and a first braking stage average acceleration of each first battery pack, and forming A first battery pack driving habit data;
S32: K-MEANS clustering is carried out on the driving habit data of the A first battery packs so as to obtain C second aggregation centers;
And C is a natural number which is more than or equal to 1 and less than A.
Preferably, the step S4 includes the following substeps:
S41: respectively acquiring a plurality of third battery packs corresponding to the first clustering center and a plurality of fourth battery packs corresponding to the second clustering center, acquiring a first battery pack set according to the plurality of third battery packs, and acquiring a second battery pack set according to the plurality of fourth battery packs;
S42: determining a plurality of fifth battery packs according to intersections of the first battery pack set and the second battery pack set, and determining battery packs except the plurality of fifth battery packs in the first battery pack set and the second battery pack set as a plurality of sixth battery packs;
S43: and respectively weighting the plurality of fifth battery packs and the plurality of sixth battery packs, and obtaining D second battery packs.
Preferably, the step S5 includes the following substeps:
S51: acquiring inconsistent conditions and fault types of the D second battery packs;
s52: and training a convolutional neural network model by taking the battery pack basic information, the first automobile brand information, the first battery pack using habit data and the first battery pack driving habit data of the D second battery packs as input data and the corresponding inconsistent conditions and fault types as output data so as to obtain the battery pack abnormal condition prediction model.
Preferably, the step S7 includes the following substeps:
s71: discharging the battery pack to be detected to a preset electric quantity range, and then performing a charging test;
s72: recording a first battery power curve, a first current curve and a first voltage curve in the charging test process;
S73: respectively comparing the first battery electric quantity curve, the first current curve and the first voltage curve with the standard battery electric quantity curve, the standard current curve and the standard voltage curve in similarity to obtain a second similarity score, a third similarity score and a fourth similarity score;
s74: and determining an average of the second similarity score, the third similarity score and the fourth similarity score as the first similarity score.
The invention also provides a battery pack consistency abnormality detection system based on big data, which is used for realizing the battery pack consistency abnormality detection method based on the big data.
The beneficial effects are that: according to the battery pack consistency anomaly detection method and system based on big data, a battery pack set to be compared is determined from multiple dimensions according to the basic information of the battery pack to be detected and the brand information of an automobile, the battery pack set to be compared is further screened according to the battery usage habit data and the driving habit data of the battery pack set to be compared, sample data which are finally used for training a battery pack anomaly prediction model are determined, the inconsistency situation and the fault type of the battery pack to be detected can be predicted based on the battery pack anomaly prediction model, the similarity between the battery pack to be detected and a standard charging curve is obtained according to the charging test data of the battery pack to be detected, and the first reliability is obtained to represent the reliability of the first inconsistency situation. According to the invention, the sample data for training the model is determined from the basic information of the battery pack, the automobile brand, the battery using habit, the driving habit and other dimensions, so that the prediction results of the inconsistent condition and the fault type of the battery are more accurate, and the rationality of the maintenance strategy of the battery pack can be further improved through the first credibility.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those of ordinary skill in the art that the drawings in the following description are exemplary only and that other implementations can be obtained from the extensions of the drawings provided without inventive effort.
Fig. 1 is a flowchart illustrating an execution of a battery pack consistency abnormality detection method based on big data in the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The embodiment provides a battery pack consistency anomaly detection method based on big data, and the method flow is shown in fig. 1, and specifically comprises the following steps:
s1: and determining A first battery packs according to the battery pack basic information and the automobile brand information of the battery packs to be detected.
S11: and generating a first image of the battery pack to be detected according to the basic information of the battery pack.
The battery pack basic information includes information for characterizing the battery pack performance, such as battery pack type, rated voltage, maximum capacity, and the like.
S12: and matching in the basic battery set by using the first portrait to obtain A1 second battery packs.
The basic battery set contains a plurality of battery information in use state, and each battery corresponds to an image formed by using the basic information.
That is, in this step, similarity matching may be performed using the first image and the image of each battery pack in the basic battery pack set, so that the battery packs whose similarity satisfies the preset requirement are determined as A1 second battery packs.
S13: and screening the A1 second battery packs according to the automobile brand information to obtain A first battery packs.
The automobile brand information refers to an automobile brand to which the battery pack to be detected belongs.
The step S13 specifically comprises the following substeps:
s131: m first automobile brands are obtained to form a first automobile brand set.
The first automobile brand refers to an automobile brand sold in the market.
S132: and acquiring a first market value tag and a first public praise tag for each first automobile brand, and forming first automobile brand information according to the first market value tag and the first public praise tag.
The first market value tag refers to a location of the first automobile brand in the automobile market, and may be preferably characterized by an average selling price of all on-sale vehicles of the first automobile brand, for example, a low-end vehicle if the average selling price is less than 10 ten thousand yuan, or a high-end vehicle if the average selling price is more than 50 ten thousand yuan.
The first stele tag is obtained by:
S1321: and for each first automobile brand, acquiring a plurality of pieces of first news information by taking the first automobile brand and battery pack information used by the first automobile brand as keywords.
The first news information is news information having as main content the first car brand information and battery pack information used therefor.
S1322: each piece of the first news information is input into a tendency judgment model to obtain a first tendency judgment opinion related to the battery pack of the first automobile brand.
The first tendency judgment opinion includes evaluation information of battery pack information of the first automobile brand, including overall experience, specific defect types, and related information to be further promoted.
The tendency judgment model is obtained through convolutional neural network model training, input data is news information of the brand information of the automobile after the evaluation is completed, and output data is tendency judgment opinion contained in the news information.
S1323: integrating the plurality of first tendency judgment opinions, extracting a plurality of pieces of first opinion information, determining the first opinion information meeting the preset appearance frequency requirement in the plurality of pieces of first opinion information as second opinion information, and integrating the plurality of pieces of second opinion information to obtain the first public praise label.
The integration means to combine and de-duplicate a plurality of the first tendency judgment opinions.
Since each piece of first tendency judgment opinion includes one or more pieces of first opinion information including overall experience, specific defect types, etc., it is necessary to arrange the pieces of first opinion information in descending order according to the occurrence frequency, so as to determine a plurality of pieces of first opinion information, which are ranked first, as second opinion information.
The second opinion information pertains to the first automotive brand comparison commonality rating information. Finally, after merging and de-duplicating the pieces of second opinion information, a first praise label can be obtained, wherein the first praise label contains commonality evaluation opinions extracted from the pieces of second opinion information.
S133: and performing similarity matching on the automobile brand information and the M pieces of first automobile brand information, and determining the first automobile brand information with the similarity larger than a preset value as second automobile brand information.
S134: and determining battery pack information corresponding to each piece of second automobile brand information as a fourth battery pack, and determining intersections of a plurality of pieces of fourth battery pack information and the A1 pieces of second battery packs as A first battery packs.
In S1, both A, A and M are natural numbers greater than 1, and M > A1> a.
S2: and aiming at the A first battery packs, obtaining using habit data of the A first battery packs, and clustering the using habit data of the A first battery packs to obtain B first clustering centers.
The first battery pack usage habit data refers to charge and discharge habit data corresponding to each first battery pack.
The step S2 comprises the following substeps:
s21: and acquiring first charging data, first discharging lower limit data and first charging upper limit data of each first battery pack, and forming A first battery pack use habit data.
The first charge data includes a first charge cycle and a first charge percentage of the first battery pack.
The first charging period is an average charging interval duration of the first battery pack in a preset period. For example, if the first battery pack is charged 10 times in total in one month, the first charging period is 30/10=3 days.
The first charging percentage refers to an average charging percentage of the first battery pack during each charging process in a preset period. For example, if the first battery pack is charged 2 times in one month, each charging percentage is 50% and 60%, respectively, the first charging percentage is (50% +60%)/2=55%.
The first discharging lower limit data and the first charging upper limit data refer to discharging lower limit data of the first battery pack which is average in the using process of the battery and charging upper limit data of the first battery pack which is average in the using process of the battery in a preset time period. For example, if the first battery pack is charged and discharged 2 times in total in one month, each of which has a lower limit of 5% and 9% and each of which has an upper limit of 90% and 100%, the first lower limit of discharge data is (5% +9%)/2=7% and the first upper limit of charge data is (90% +100%)/2=95%, respectively.
S22: and carrying out K-MEANS clustering on the A first battery pack usage habit data to obtain B first clustering centers.
The first battery pack usage habit data is a quaternary vector composed of first charging data, first discharging lower limit data and first charging upper limit data. In this step, K-MEANS clustering is performed on the a quaternary vectors, so that B first clustering centers can be obtained, where a plurality of first battery packs corresponding to each first clustering center have the same or similar battery pack usage habits.
And B is a natural number which is more than or equal to 1 and less than A.
S3: and aiming at the A first battery packs, obtaining driving habit data of the A first battery packs, and clustering the driving habit data of the A first battery packs to obtain C second aggregation centers.
The first battery pack driving habit data refers to driving habit data corresponding to each first battery pack, such as average driving speed, acceleration in a starting stage, acceleration in a braking stage, and the like.
The step S3 comprises the following substeps:
S31: and acquiring the first average driving speed, the first starting stage average acceleration and the first braking stage average acceleration of each first battery pack, and forming A first battery pack driving habit data.
The first average driving speed refers to a ratio of a total driving range of the first battery pack to a total driving time within a preset period of time.
The average acceleration in the first starting stage is an average acceleration value of the first battery pack when the first battery pack is started from a stationary state every time in a preset period.
The average acceleration in the first braking stage is an average acceleration value of the first battery pack in a state of being stationary from a driving state each time in a preset period.
S32: and carrying out K-MEANS clustering on the driving habit data of the A first battery packs to obtain C second aggregation centers.
The first battery pack driving habit data is a ternary vector consisting of a first average driving speed, a first starting stage average acceleration and a first braking stage average acceleration. In this step, K-MEANS clustering is performed on the a ternary vectors, so that C first clustering centers may be obtained, where a plurality of first battery packs corresponding to each second clustering center have the same or similar driving habits.
And C is a natural number which is more than or equal to 1 and less than A.
S4: and D second battery packs are obtained according to the first clustering center and the second clustering center.
The step S4 comprises the following substeps:
S41: and respectively acquiring a plurality of third battery packs corresponding to the first clustering center and a plurality of fourth battery packs corresponding to the second clustering center, acquiring a first battery pack set according to the third battery packs, and acquiring a second battery pack set according to the fourth battery packs.
Wherein each of the first cluster center or the second cluster center corresponds to a plurality of battery packs.
S42: and determining a plurality of fifth battery packs according to the intersection of the first battery pack set and the second battery pack set, and determining the battery packs except the plurality of fifth battery packs in the first battery pack set and the second battery pack set as a plurality of sixth battery packs.
S43: weighting the fifth battery packs and the sixth battery packs respectively, and obtaining D second battery packs
The second battery packs in the first battery pack set correspond to different weights, the weights of the second battery packs belonging to the first battery pack set and the second battery pack set are larger, the weights of the second battery packs belonging to only one of the first battery pack set and the second battery pack set are relatively smaller, and specific assignment conditions can be set according to actual needs.
Wherein D is a natural number greater than zero and less than A.
S5: and taking the battery pack basic information, the first automobile brand information, the first battery pack using habit data and the first battery pack driving habit data of the D second battery packs as inputs, taking the inconsistent condition and the fault type of the battery packs as outputs, and training to obtain a battery pack abnormal condition prediction model.
The battery pack abnormal condition prediction model is obtained through training in the following mode:
s51: and acquiring inconsistent conditions and fault types of the D second battery packs.
The inconsistent condition refers to the inconsistent condition of the battery packs, which occurs in the actual use process of each second battery pack, and the fault type refers to the actual fault type which causes the inconsistent condition of the battery packs.
S52: and training a convolutional neural network model by taking the battery pack basic information, the first automobile brand information, the first battery pack using habit data and the first battery pack driving habit data of the D second battery packs as input data and the corresponding inconsistent conditions and fault types as output data so as to obtain the battery pack abnormal condition prediction model.
Preferably, in the training process, weight values can be set for different sample data, so that differentiated model training is performed according to different weights of different second battery packs.
S6: and inputting the basic information, the brand information, the using habit data and the driving habit data of the battery pack to be detected into the abnormal condition prediction model of the battery pack to obtain a first inconsistent condition and a first fault type.
In this step, the usage habit data of the battery pack and the driving habit data of the battery pack are substantially identical to the acquisition modes of the usage habit data of the first battery pack and the driving habit data of the first battery pack in S2 and S3, and are not described in detail herein.
Preferably, if there is no inconsistency, the first inconsistency is not generated and no problem information is returned.
S7: and carrying out a first charging test on the battery pack to be detected to obtain a first charging curve, and carrying out similarity matching on the first charging curve and a standard charging curve to obtain a first similarity score.
Because the first inconsistent situation and the first fault type in S6 also have certain uncertainty, in this step, according to the actual charging test situation of the battery pack to be detected, a first similarity score is obtained to adjust the first inconsistent situation and the first fault type.
The step S7 comprises the following substeps:
s71: and discharging the battery pack to be detected to a preset electric quantity range, and then performing a charging test.
The preset electric quantity range is preferably between 10% and 20% of the total electric quantity of the battery.
S72: and recording a first battery power curve, a first current curve and a first voltage curve in the charging test process.
S73: and respectively comparing the first battery electric quantity curve, the first current curve and the first voltage curve with the standard battery electric quantity curve, the standard current curve and the standard voltage curve in similarity to obtain a second similarity score, a third similarity score and a fourth similarity score.
The standard battery power curve, the standard current curve and the standard voltage curve refer to the same battery pack with normal functions, and the charging data change curve under the same working condition.
S74: and determining an average of the second similarity score, the third similarity score and the fourth similarity score as the first similarity score.
S8: and obtaining first credibility of the first inconsistent situation according to the first similarity score, and formulating a battery pack maintenance strategy according to the first credibility, the first inconsistent situation and the first fault type.
The first reliability score is inversely proportional to the first similarity score, that is, the higher the similarity of the charging curves of the to-be-detected battery pack and the standard battery pack is, the lower the reliability of the first inconsistent situation is.
The invention also provides a battery pack consistency abnormality detection system based on big data, which is used for executing the battery pack consistency abnormality detection method based on the big data.
According to the battery pack consistency anomaly detection method and system based on big data, a battery pack set to be compared is determined from multiple dimensions according to the basic information of the battery pack to be detected and the brand information of an automobile, the battery pack set to be compared is further screened according to the battery usage habit data and the driving habit data of the battery pack set to be compared, sample data which are finally used for training a battery pack anomaly prediction model are determined, the inconsistency situation and the fault type of the battery pack to be detected can be predicted based on the battery pack anomaly prediction model, the similarity between the battery pack to be detected and a standard charging curve is obtained according to the charging test data of the battery pack to be detected, and the first reliability is obtained to represent the reliability of the first inconsistency situation. According to the invention, the sample data for training the model is determined from the basic information of the battery pack, the automobile brand, the battery using habit, the driving habit and other dimensions, so that the prediction results of the inconsistent condition and the fault type of the battery are more accurate, and the rationality of the maintenance strategy of the battery pack can be further improved through the first credibility.
The foregoing description is only of the preferred embodiments of the invention, and all changes and modifications that come within the meaning and range of equivalency of the structures, features and principles of the invention are therefore intended to be embraced therein.
Claims (9)
1. The battery pack consistency anomaly detection method based on big data is characterized by comprising the following steps:
s1: determining A first battery packs according to battery pack basic information and automobile brand information of the battery packs to be detected;
s2: for A first battery packs, obtaining A first battery pack usage habit data, and clustering the A first battery pack usage habit data to obtain B first clustering centers;
S3: for A first battery packs, obtaining driving habit data of the A first battery packs, and clustering the driving habit data of the A first battery packs to obtain C second aggregation centers;
s4: obtaining D second battery packs according to the first clustering center and the second clustering center;
s5: taking the battery pack basic information, the first automobile brand information, the first battery pack using habit data and the first battery pack driving habit data of the D second battery packs as inputs, taking inconsistent conditions and fault types of the battery packs as outputs, and training to obtain a battery pack abnormal condition prediction model;
S6: inputting the basic information, the brand information, the using habit data and the driving habit data of the battery pack to be detected into the abnormal condition prediction model of the battery pack to obtain a first inconsistent condition and a first fault type;
s7: performing a first charging test on the battery pack to be detected to obtain a first charging curve, and performing similarity matching on the first charging curve and a standard charging curve to obtain a first similarity score; the standard charging curve is a charging curve obtained after a normal battery pack is subjected to a charging test;
S8: obtaining first credibility of the first inconsistent situation according to the first similarity score, and formulating a battery pack maintenance strategy according to the first credibility, the first inconsistent situation and the first fault type;
Wherein A, B, C, D are natural numbers greater than 1.
2. The big data based battery pack consistency anomaly detection method of claim 1, wherein S1 comprises the sub-steps of:
S11: generating a first image of the battery pack to be detected according to the basic information of the battery pack;
s12: matching in the basic battery set by using the first portrait to obtain A1 second battery sets;
S13: screening the A1 second battery packs according to the automobile brand information to obtain A first battery packs;
the step S13 specifically comprises the following substeps:
s131: obtaining M first automobile brands to form a first automobile brand set;
S132: for each first automobile brand, acquiring a first market value tag and a first tombstone tag, and forming first automobile brand information according to the first market value tag and the first tombstone tag;
S133: performing similarity matching on the automobile brand information and M pieces of first automobile brand information, and determining the first automobile brand information with the similarity larger than a preset value as second automobile brand information;
s134: determining battery pack information corresponding to each piece of second automobile brand information as a fourth battery pack, and determining intersections of a plurality of fourth battery packs and the A1 second battery packs as A first battery packs;
Both A, A and M are natural numbers greater than 1, and M > A1> A.
3. The big data based battery pack consistency anomaly detection method of claim 2, wherein S132 comprises the sub-steps of:
S1321: for each first automobile brand, taking the first automobile brand and battery pack information used by the first automobile brand as keywords, and acquiring a plurality of pieces of first news information;
S1322: inputting each piece of the first news information into a tendency judgment model to obtain a first tendency judgment opinion related to the battery pack of the first automobile brand;
S1323: integrating the plurality of first tendency judgment opinions, extracting a plurality of pieces of first opinion information, determining the first opinion information meeting the preset appearance frequency requirement in the plurality of pieces of first opinion information as second opinion information, and integrating the plurality of pieces of second opinion information to obtain the first public praise label.
4. The big data based battery pack consistency anomaly detection method of claim 3, wherein S2 comprises the sub-steps of:
S21: acquiring first charging data, first discharging lower limit data and first charging upper limit data of each first battery pack, and forming A first battery pack use habit data;
S22: K-MEANS clustering is carried out on the A first battery pack usage habit data so as to obtain B first clustering centers;
And B is a natural number which is more than or equal to 1 and less than A.
5. The big data based battery pack consistency anomaly detection method of claim 4, wherein S3 comprises the sub-steps of:
s31: acquiring a first average driving speed, a first starting stage average acceleration and a first braking stage average acceleration of each first battery pack, and forming A first battery pack driving habit data;
S32: K-MEANS clustering is carried out on the driving habit data of the A first battery packs so as to obtain C second aggregation centers;
And C is a natural number which is more than or equal to 1 and less than A.
6. The big data based battery pack consistency anomaly detection method of claim 5, wherein S4 comprises the sub-steps of:
S41: respectively acquiring a plurality of third battery packs corresponding to the first clustering center and a plurality of fourth battery packs corresponding to the second clustering center, acquiring a first battery pack set according to the plurality of third battery packs, and acquiring a second battery pack set according to the plurality of fourth battery packs;
S42: determining a plurality of fifth battery packs according to intersections of the first battery pack set and the second battery pack set, and determining battery packs except the plurality of fifth battery packs in the first battery pack set and the second battery pack set as a plurality of sixth battery packs;
S43: and respectively weighting the plurality of fifth battery packs and the plurality of sixth battery packs, and obtaining D second battery packs.
7. The big data based battery pack consistency anomaly detection method of claim 6, wherein S5 comprises the sub-steps of:
S51: acquiring inconsistent conditions and fault types of the D second battery packs;
s52: and training a convolutional neural network model by taking the battery pack basic information, the first automobile brand information, the first battery pack using habit data and the first battery pack driving habit data of the D second battery packs as input data and the corresponding inconsistent conditions and fault types as output data so as to obtain the battery pack abnormal condition prediction model.
8. The big data based battery pack consistency anomaly detection method of claim 7, wherein S7 comprises the sub-steps of:
s71: discharging the battery pack to be detected to a preset electric quantity range, and then performing a charging test;
s72: recording a first battery power curve, a first current curve and a first voltage curve in the charging test process;
S73: respectively comparing the first battery electric quantity curve, the first current curve and the first voltage curve with the standard battery electric quantity curve, the standard current curve and the standard voltage curve in similarity to obtain a second similarity score, a third similarity score and a fourth similarity score;
s74: and determining an average of the second similarity score, the third similarity score and the fourth similarity score as the first similarity score.
9. A big data based battery pack consistency anomaly detection system for implementing the big data based battery pack consistency anomaly detection method of any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410969533.4A CN118551325B (en) | 2024-07-19 | 2024-07-19 | Battery pack consistency anomaly detection method and system based on big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410969533.4A CN118551325B (en) | 2024-07-19 | 2024-07-19 | Battery pack consistency anomaly detection method and system based on big data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118551325A true CN118551325A (en) | 2024-08-27 |
| CN118551325B CN118551325B (en) | 2024-10-01 |
Family
ID=92453588
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410969533.4A Active CN118551325B (en) | 2024-07-19 | 2024-07-19 | Battery pack consistency anomaly detection method and system based on big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118551325B (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20220065939A1 (en) * | 2020-09-03 | 2022-03-03 | Volkswagen Group Of America, Inc. | Battery state of health estimator |
| US11444338B1 (en) * | 2021-11-30 | 2022-09-13 | Knoetik Solutions, Inc. | Smart battery system |
| US20230145448A1 (en) * | 2021-11-10 | 2023-05-11 | Johnson Controls Tyco IP Holdings LLP | Systems and methods for predicting building faults using machine learning |
| CN116626505A (en) * | 2023-07-21 | 2023-08-22 | 江苏海平面数据科技有限公司 | Battery pack consistency anomaly detection method based on Internet of vehicles big data |
| US20230280411A1 (en) * | 2021-09-13 | 2023-09-07 | Astrolabe Analytics, Inc. | Methods and systems for determining a weak cell of a battery |
| CN116956080A (en) * | 2022-12-15 | 2023-10-27 | 腾讯科技（深圳）有限公司 | Data processing method, device and storage medium |
| CN117554825A (en) * | 2024-01-11 | 2024-02-13 | 深圳市北测检测技术有限公司 | Charging safety performance detection method and system for electric automobile |
| CN117951629A (en) * | 2024-03-26 | 2024-04-30 | 深圳市泰科动力系统有限公司 | A method and system for detecting abnormal charging and discharging temperature behavior of a battery pack |
| CN118209870A (en) * | 2024-05-22 | 2024-06-18 | 深圳市泰科动力系统有限公司 | A lithium battery health level assessment management method and system |
| CN118311465A (en) * | 2024-06-11 | 2024-07-09 | 广东车卫士信息科技有限公司 | Battery health detection method and device, electronic equipment and storage medium |
| CN118343020A (en) * | 2024-04-28 | 2024-07-16 | 国网浙江省电力有限公司 | A method, system, device and storage medium for orderly charging management of electric vehicles |
-
2024
- 2024-07-19 CN CN202410969533.4A patent/CN118551325B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20220065939A1 (en) * | 2020-09-03 | 2022-03-03 | Volkswagen Group Of America, Inc. | Battery state of health estimator |
| US20230280411A1 (en) * | 2021-09-13 | 2023-09-07 | Astrolabe Analytics, Inc. | Methods and systems for determining a weak cell of a battery |
| US20230145448A1 (en) * | 2021-11-10 | 2023-05-11 | Johnson Controls Tyco IP Holdings LLP | Systems and methods for predicting building faults using machine learning |
| US11444338B1 (en) * | 2021-11-30 | 2022-09-13 | Knoetik Solutions, Inc. | Smart battery system |
| CN116956080A (en) * | 2022-12-15 | 2023-10-27 | 腾讯科技（深圳）有限公司 | Data processing method, device and storage medium |
| CN116626505A (en) * | 2023-07-21 | 2023-08-22 | 江苏海平面数据科技有限公司 | Battery pack consistency anomaly detection method based on Internet of vehicles big data |
| CN117554825A (en) * | 2024-01-11 | 2024-02-13 | 深圳市北测检测技术有限公司 | Charging safety performance detection method and system for electric automobile |
| CN117951629A (en) * | 2024-03-26 | 2024-04-30 | 深圳市泰科动力系统有限公司 | A method and system for detecting abnormal charging and discharging temperature behavior of a battery pack |
| CN118343020A (en) * | 2024-04-28 | 2024-07-16 | 国网浙江省电力有限公司 | A method, system, device and storage medium for orderly charging management of electric vehicles |
| CN118209870A (en) * | 2024-05-22 | 2024-06-18 | 深圳市泰科动力系统有限公司 | A lithium battery health level assessment management method and system |
| CN118311465A (en) * | 2024-06-11 | 2024-07-09 | 广东车卫士信息科技有限公司 | Battery health detection method and device, electronic equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 何瑛;钟根鹏;陈翌;: "基于特征组合堆叠融合集成学习的锂离子动力电池SOC估算", 储能科学与技术, no. 05, 5 September 2020 (2020-09-05) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118551325B (en) | 2024-10-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113064939B (en) | New energy vehicle three-electric system safety feature database construction method | |
| CN113696786B (en) | Battery equalization method and system | |
| CN112765772B (en) | A data-driven method for predicting the remaining life of power battery | |
| US20230251324A1 (en) | Systems and methods for determining vehicle battery health | |
| CN112433169B (en) | Cloud power battery health degree evaluation system and method | |
| CN111159533B (en) | An intelligent charging service recommendation method and system based on user portraits | |
| CN109116242B (en) | Data processing method and device for power battery | |
| CN113887601A (en) | Retired power battery recombination method based on cluster sorting | |
| CN113459871A (en) | Battery charging method and device for battery replacement station, storage medium and electronic equipment | |
| CN114833097B (en) | Sorting method and device for gradient utilization of retired power battery | |
| CN111420898B (en) | Retired battery sorting method and system applied by same | |
| CN113866655A (en) | Power battery capacity assessment method fusing vehicle networking operation data and test data | |
| CN119986443A (en) | A Lithium-ion Battery Health State Estimation Method Integrating Physical Information | |
| CN117330964A (en) | A three-interval fusion estimation method of lithium battery state of charge based on fitness value | |
| CN115700717A (en) | A power distribution analysis method based on electric vehicle power demand | |
| CN109946613A (en) | On-line estimation and life detection method of internal resistance of vehicle power battery | |
| CN118551325B (en) | Battery pack consistency anomaly detection method and system based on big data | |
| CN117930027A (en) | Power battery capacity abnormality detection method, device and platform | |
| CN114966412B (en) | Multi-terminal battery balance detection method and device | |
| CN115534756A (en) | A method, system, device and medium for identifying anomalies in electric vehicle cruising range | |
| CN120234616A (en) | A method for rapid detection of health status of retired power lithium batteries | |
| CN109490791A (en) | Determining the Screening Method of Vehicle Retired Batteries by Remote Monitoring Data | |
| CN116794544A (en) | Method for detecting life of power battery, computer-readable storage medium, and vehicle | |
| CN118378888A (en) | Method and device for predicting thermal runaway risk probability of power battery based on working conditions | |
| CN114565283B (en) | Electric vehicle performance demand matching degree evaluation method, system, equipment and medium |
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 |