CN109765490B - Power battery fault detection method and system based on high-dimensional data diagnosis - Google Patents

Power battery fault detection method and system based on high-dimensional data diagnosis Download PDF

Info

Publication number
CN109765490B
CN109765490B CN201811336819.XA CN201811336819A CN109765490B CN 109765490 B CN109765490 B CN 109765490B CN 201811336819 A CN201811336819 A CN 201811336819A CN 109765490 B CN109765490 B CN 109765490B
Authority
CN
China
Prior art keywords
points
matrix
voltage data
threshold
abnormal
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
Application number
CN201811336819.XA
Other languages
Chinese (zh)
Other versions
CN109765490A (en
Inventor
王震坡
刘鹏
王瑾
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Institute Of Technology New Source Information Technology Co ltd
Beijing Institute of Technology BIT
Original Assignee
Beijing Institute Of Technology New Source Information Technology Co ltd
Beijing Institute of Technology BIT
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Institute Of Technology New Source Information Technology Co ltd, Beijing Institute of Technology BIT filed Critical Beijing Institute Of Technology New Source Information Technology Co ltd
Priority to CN201811336819.XA priority Critical patent/CN109765490B/en
Publication of CN109765490A publication Critical patent/CN109765490A/en
Application granted granted Critical
Publication of CN109765490B publication Critical patent/CN109765490B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The invention discloses a power battery fault detection method and system based on high-dimensional data diagnosis, which can be used for diagnosing the faults of a power battery only by adopting an angle-based abnormal value detection method for a primary abnormal point matrix, compared with the detection of all original data sets, the workload of data processing is greatly reduced, the complexity of data processing is reduced, and the detection efficiency is further improved.

Description

Power battery fault detection method and system based on high-dimensional data diagnosis
Technical Field
The invention relates to the field of battery detection, in particular to a power battery fault detection method and system based on high-dimensional data diagnosis.
Background
A power battery system is the most critical component of an electric vehicle, and a plurality of battery cells are generally connected in series to form a battery pack. Although the battery cells are tested and preferably grouped, performance differences still exist, the differences generate new differences in different degrees in the long-term operation process of the battery, the performance of individual battery cells is obviously reduced, the performance of the battery pack is seriously influenced, and even accidents are caused, so the performance reduction and early faults of the battery cells in the battery pack need to be detected so as to be capable of taking measures in time.
The traditional method for detecting the high-dimensional data abnormal detection point based on the angle variance needs to calculate all original data sets, and is high in complexity and long in calculation time.
Disclosure of Invention
The invention aims to provide a power battery fault detection method and system based on high-dimensional data diagnosis, and the effect of improving the power battery fault detection efficiency is achieved.
In order to achieve the purpose, the invention provides the following scheme:
a power battery fault detection method based on high dimensional data diagnostics, the method comprising:
preprocessing voltage data of the battery monomer to obtain a voltage data matrix;
clustering the voltage data matrix by using a clustering algorithm to obtain a clustering result, wherein the clustering result comprises a preliminary abnormal point matrix T formed by boundary points and abnormal points and a core point matrix Q formed by core points;
calculating an abnormal value score by using an abnormal value detection algorithm according to the clustering result;
and comparing the abnormal value fraction with a threshold value, if the abnormal value fraction is smaller than the threshold value, determining that the battery cell corresponding to the abnormal value fraction has a fault, and outputting the number of the battery cell with the fault.
Optionally, the preprocessing is performed on the voltage data matrix of the battery cell to obtain the voltage data matrix, and the method specifically includes:
deleting the data at the moment of repeated recording and the data with errors in the voltage data of the single battery to obtain the voltage data of the single battery after cleaning;
using the formula Δ Ui=Ui+1-UiPerforming first-order difference processing on the voltage data of the cleaned battery monomer to obtain a voltage data matrix;
wherein, UiRepresents the voltage data, Δ U, of the battery cell acquired at the ith timeiAnd i represents the number of times of voltage data acquisition of the battery cell.
Optionally, the clustering the voltage data matrix by using a clustering algorithm to obtain a clustering result specifically includes:
initializing a region threshold and a number threshold;
determining an area in which the number of points in the voltage data matrix is greater than the number threshold within the area threshold as a core area, and forming a core point matrix Q by the points of the core area;
determining a region within the region threshold and outside the core region where the number of points in the voltage data matrix is less than the number threshold as a boundary region whose points are the boundary points;
determining a point of the voltage data matrix outside the region threshold as the outlier; the boundary points and the abnormal points form the preliminary abnormal point matrix T; determining the core point matrix Q and the preliminary abnormal point matrix T as clustering results;
judging whether the clustering result reaches a clustering target or not to obtain a judgment result;
when the judgment result shows yes, outputting the clustering result;
and when the judgment result shows that the voltage data matrix is not the core point matrix, adjusting the area threshold and the number threshold, and returning to the step of determining that the area with the number of points in the voltage data matrix larger than the number threshold in the area threshold is the core area and the points in the core area form the core point matrix Q'.
Optionally, the clustering target is that the number of the preliminary abnormal points accounts for 10% to 20% of the number of the points in the voltage data matrix.
Optionally, the calculating the abnormal value score by using an abnormal value detection algorithm according to the clustering result specifically includes:
mapping the preliminary abnormal point matrix T and the core point matrix Q to a high-dimensional space;
calculating an included angle formed by the point of the preliminary abnormal point matrix T and any two points in the core point matrix Q;
calculating the cosine value of the included angle according to the included angle to obtain a cosine value set of the included angle;
and calculating the variance of the cosine value set, and determining the variance as the abnormal value fraction of the points of the sample data P.
A power battery fault detection system based on high-dimensional data diagnosis comprises
The preprocessing unit is used for preprocessing the voltage data of the battery monomer to obtain a voltage data matrix;
the clustering unit is used for clustering the voltage data matrix by using a clustering algorithm to obtain a clustering result, wherein the clustering result comprises a preliminary abnormal point matrix T formed by boundary points and abnormal points and a core point matrix Q formed by core points;
an abnormal value score calculating unit for calculating an abnormal value score by using an abnormal value detection algorithm according to the clustering result;
and the fault battery determining unit is used for comparing the abnormal value fraction with a threshold value, determining that the battery cell corresponding to the abnormal value fraction has a fault if the abnormal value fraction is smaller than the threshold value, and outputting the number of the battery cell with the fault.
Optionally, the preprocessing unit specifically includes:
the data cleaning subunit is used for deleting the data repeatedly recorded at the moment and the data with errors in the voltage data of the single battery to obtain the voltage data of the single battery after cleaning;
a first order difference subunit for using the formula Δ Ui=Ui+1-UiPerforming first-order difference processing on the voltage data of the cleaned battery monomer to obtain a voltage data matrix;
wherein, UiRepresents the voltage data, Δ U, of the battery cell acquired at the ith timeiAnd i represents the number of times of voltage data acquisition of the battery cell.
Optionally, the clustering unit specifically includes:
a threshold initialization subunit, configured to initialize a region threshold and a number threshold;
a core region determining subunit, configured to determine, as a core region, a region in which the number of points in the voltage data matrix is greater than the number threshold within the region threshold, and the core point matrix Q is formed by the points of the core region;
a boundary region determination subunit, configured to determine, as a boundary region, a region that is within the region threshold and outside the core region, where the number of points in the voltage data matrix is smaller than the number threshold, and the point of the boundary region is the boundary point;
an abnormal point determination subunit, configured to determine a point of the voltage data matrix outside the region threshold as the abnormal point; the boundary points and the abnormal points form the preliminary abnormal point matrix T; determining the core point matrix Q and the preliminary abnormal point matrix T as clustering results;
a clustering result judging subunit, configured to judge whether the clustering result reaches a clustering target, to obtain a judgment result;
a clustering result output subunit, configured to output the clustering result when the determination result indicates yes;
and a threshold adjusting subunit, configured to adjust the area threshold and the number threshold when the determination result indicates no, and return to the step "determine that an area in which the number of points in the voltage data matrix is greater than the number threshold within the area threshold is a core area, and the core point matrix Q is configured by the points of the core area".
Optionally, the clustering target is that the number of the preliminary abnormal points accounts for 10% to 20% of the number of the points in the voltage data matrix.
Optionally, the abnormal value score calculating unit specifically includes:
a mapping subunit, configured to map the preliminary abnormal point matrix T and the core point matrix Q to a high-dimensional space;
an included angle calculating subunit, configured to calculate an included angle formed between the point of the preliminary abnormal point matrix T and any two points in the core point matrix Q;
the included angle cosine value calculating operator unit is used for calculating included angle cosine values according to the included angles to obtain an included angle cosine value set;
an outlier score determining subunit, configured to calculate a variance of the set of cosine values, and determine the variance as an outlier score of a point of the sample data P.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects: according to the method, the voltage data of the single battery is preprocessed to obtain the voltage data matrix, then the voltage data matrix is clustered to obtain the primary abnormal point matrix and the core point matrix, the fault of the power battery can be diagnosed only by adopting an angle-based abnormal value detection method for the primary abnormal data matrix, and compared with the detection of all original data sets, the method greatly reduces the workload of data processing, reduces the complexity of data processing, and further improves the detection efficiency.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a flow chart of a method for detecting a fault of a power battery based on high-dimensional data diagnosis according to an embodiment of the present invention;
FIG. 2 is a block diagram of a power battery fault detection system based on high-dimensional data diagnosis according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a visualized clustering result provided in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a power battery fault detection method and system based on high-dimensional data diagnosis, which can reduce the complexity of data processing and further improve the detection efficiency.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
As shown in fig. 1, the method for detecting a power battery fault based on high-dimensional data diagnosis provided by this embodiment includes:
step 101: and preprocessing the voltage data of the battery monomer to obtain a voltage data matrix.
The step 101 specifically includes:
deleting the data at the moment of repeated recording and the data with errors in the voltage data of the single battery to obtain the voltage data of the single battery after cleaning;
using the formula Δ Ui=Ui+1-UiPerforming first-order difference processing on the voltage data of the cleaned battery monomer to obtain a voltage data matrix;
wherein, UiRepresents the voltage data, Δ U, of the battery cell acquired at the ith timeiAnd i represents the number of times of voltage data acquisition of the battery cell.
The method comprises the steps that voltage data of the single battery is extracted from a platform, data at the moment of repeated recording is deleted, a prison electric vehicle BMS system normally operates, and an error record value is deleted according to a single discharge voltage lower limit threshold and a single charge upper limit threshold, so that the processing amount of the voltage single data can be reduced, and the detection accuracy can be improved; and the data is processed by adopting a first-order difference method, linear trend factors are eliminated, and a stable sequence is obtained.
Step 102: clustering the voltage data matrix by using a clustering algorithm to obtain a clustering result; the clustering result comprises a preliminary abnormal point matrix T formed by boundary points and abnormal points and a core point matrix Q formed by core points.
The step 102 specifically includes:
initializing a region threshold and a number threshold;
determining an area in which the number of points in the voltage data matrix is greater than the number threshold within the area threshold as a core area, and forming a core point matrix Q by the points of the core area;
determining a region within the region threshold and outside the core region where the number of points in the voltage data matrix is less than the number threshold as a boundary region whose points are the boundary points;
determining a point of the voltage data matrix outside the region threshold as the outlier; the boundary points and the abnormal points form the preliminary abnormal point matrix T; determining the core point matrix Q and the preliminary abnormal point matrix T as clustering results;
judging whether the clustering result reaches a clustering target or not to obtain a judgment result; the clustering target is that the number of the initial abnormal points accounts for 10% -20% of the number of the points in the voltage data matrix;
when the judgment result shows yes, outputting the clustering result;
and when the judgment result shows that the voltage data matrix is not the core point matrix, adjusting the area threshold and the number threshold, and returning to the step of determining that the area with the number of points in the voltage data matrix larger than the number threshold in the area threshold is the core area and the points in the core area form the core point matrix Q'.
The number of the initial abnormal data sets obtained after clustering is reduced by 60-70% compared with the total number of data samples of the original data sets, so that the data processing amount is greatly reduced, and the detection efficiency is improved.
Step 102 may be followed by:
performing dimensionality reduction on the clustering result;
and visualizing the clustering result.
The clustering result is subjected to dimension reduction and visualized, whether the clustering result meets the expected requirement or not can be visually seen, and the threshold value is convenient to adjust.
As shown in fig. 3, the large circle points in the figure represent the core point, and the small circle points represent the outlier and the boundary point.
Step 103: and calculating the abnormal value score by using an abnormal value detection algorithm according to the clustering result.
The step 103 specifically includes:
mapping the preliminary abnormal point matrix T and the core point matrix Q to a high-dimensional space;
calculating an included angle formed by the point of the preliminary abnormal point matrix T and any two points in the core point matrix Q;
calculating the cosine value of the included angle according to the included angle to obtain a cosine value set of the included angle;
and calculating the variance of the cosine value set, and determining the variance as the abnormal value fraction of the points of the sample data P.
And calculating an outlier score of the initial outlier relative to the core data point, wherein the outlier score is used for measuring the outlier degree of the data point, and the lower the value is, the greater the outlier degree of the point is, so that the outlier can be initially determined.
Step 104: and comparing the abnormal value fraction with a threshold value, if the abnormal value fraction is smaller than the threshold value, determining that the battery cell corresponding to the abnormal value fraction has a fault, and outputting the number of the battery cell with the fault.
According to the comparison between the abnormal value fraction and the threshold value, the battery cell with the fault can be determined to be the battery cell corresponding to the abnormal value fraction smaller than the threshold value, compared with the comparison between the calculation fractions of all data points and the threshold value, the calculation complexity and the calculation time are obviously reduced, and the fault detection efficiency is improved.
As shown in fig. 2, the power battery fault detection system based on high-dimensional data diagnosis provided by the present embodiment includes: a preprocessing unit 201, a clustering unit 202, an abnormal value score calculating unit 203, and a faulty battery determining unit 204.
The preprocessing unit 201 is configured to preprocess the voltage data of the battery cell to obtain a voltage data matrix.
The preprocessing unit 201 specifically includes:
the data cleaning subunit is used for deleting the data repeatedly recorded at the moment and the data with errors in the voltage data of the single battery to obtain the voltage data of the single battery after cleaning;
a first order difference subunit for using the formula Δ Ui=Ui+1-UiPerforming first-order difference processing on the voltage data of the cleaned battery monomer to obtain a voltage data matrix;
wherein, UiRepresents the voltage data, Δ U, of the battery cell acquired at the ith timeiAnd i represents the number of times of voltage data acquisition of the battery cell.
The method comprises the steps that voltage data of the single battery is extracted from a platform, data at the moment of repeated recording is deleted, a prison electric vehicle BMS system normally operates, and an error record value is deleted according to a single discharge voltage lower limit threshold and a single charge upper limit threshold, so that the processing amount of the voltage single data can be reduced, and the detection accuracy can be improved; and the data is processed by adopting a first-order difference method, linear trend factors are eliminated, and a stable sequence is obtained.
And the clustering unit 202 is configured to perform clustering processing on the voltage data matrix by using a clustering algorithm to obtain a clustering result.
The clustering unit 202 specifically includes:
a threshold initialization subunit, configured to initialize a region threshold and a number threshold;
a core region determining subunit, configured to determine, as a core region, a region in which the number of points in the voltage data matrix is greater than the number threshold within the region threshold, and the core point matrix Q is formed by the points of the core region;
a boundary region determination subunit, configured to determine, as a boundary region, a region that is within the region threshold and outside the core region, where the number of points in the voltage data matrix is smaller than the number threshold, and the point of the boundary region is the boundary point;
an abnormal point determination subunit, configured to determine a point of the voltage data matrix outside the region threshold as the abnormal point; the boundary points and the abnormal points form the preliminary abnormal point matrix T; determining the core point matrix Q and the preliminary abnormal point matrix T as clustering results;
a clustering result judging subunit, configured to judge whether the clustering result reaches a clustering target, to obtain a judgment result; the clustering target is that the number of the initial abnormal points accounts for 10% -20% of the number of the points in the voltage data matrix;
a clustering result output subunit, configured to output the clustering result when the determination result indicates yes;
and a threshold adjusting subunit, configured to adjust the area threshold and the number threshold when the determination result indicates no, and return to the step "determine that an area in which the number of points in the voltage data matrix is greater than the number threshold within the area threshold is a core area, and the core point matrix Q is configured by the points of the core area".
The number of the initial abnormal data sets obtained after clustering is reduced by 60-70% compared with the total number of data samples of the original data sets, so that the data processing amount is greatly reduced, and the detection efficiency is improved.
The clustering processing is carried out on the voltage data matrix by utilizing a clustering algorithm, and after a clustering result is obtained, the method further comprises the following steps:
the dimensionality reduction processing unit is used for carrying out dimensionality reduction processing on the clustering result;
and the visualization unit is used for visualizing the clustering result.
The clustering result is subjected to dimension reduction and visualized, whether the clustering result meets the expected requirement or not can be visually seen, and the threshold value is convenient to adjust.
As shown in fig. 3, the large circle points in the figure represent the core point, and the small circle points represent the outlier and the boundary point.
An abnormal value score calculating unit 203 for calculating an abnormal value score using an abnormal value detection algorithm according to the clustering result.
The abnormal value score calculation unit 203 specifically includes:
a mapping subunit, configured to map the preliminary abnormal point matrix T and the core point matrix Q to a high-dimensional space;
an included angle calculating subunit, configured to calculate an included angle formed between the point of the preliminary abnormal point matrix T and any two points in the core point matrix Q;
the included angle cosine value calculating operator unit is used for calculating included angle cosine values according to the included angles to obtain an included angle cosine value set;
an outlier score determining subunit, configured to calculate a variance of the set of cosine values, and determine the variance as an outlier score of a point of the sample data P.
And calculating an outlier score of the initial outlier relative to the core data point, wherein the outlier score is used for measuring the outlier degree of the data point, and the lower the value is, the greater the outlier degree of the point is, so that the outlier can be initially determined.
And the faulty battery determining unit 204 is used for comparing the abnormal value fraction with a threshold value, determining that the battery cell corresponding to the abnormal value fraction is faulty if the abnormal value fraction is smaller than the threshold value, and outputting the number of the faulty battery cell.
According to the comparison between the abnormal value fraction and the threshold value, the battery cell with the fault can be determined to be the battery cell corresponding to the abnormal value fraction smaller than the threshold value, compared with the comparison between the calculation fractions of all data points and the threshold value, the calculation complexity and the calculation time are obviously reduced, and the fault detection efficiency is improved.
For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.

Claims (8)

1. A power battery fault detection method based on high-dimensional data diagnosis is characterized by comprising the following steps:
preprocessing voltage data of the battery monomer to obtain a voltage data matrix;
clustering the voltage data matrix by using a clustering algorithm to obtain a clustering result, wherein the clustering result comprises a preliminary abnormal point matrix T formed by boundary points and abnormal points and a core point matrix Q formed by core points;
calculating an abnormal value score by using an abnormal value detection algorithm according to the clustering result;
calculating the abnormal value score by using an abnormal value detection algorithm according to the clustering result, which specifically comprises the following steps:
mapping the preliminary abnormal point matrix T and the core point matrix Q to a high-dimensional space;
calculating included angles formed by any two points in the core point matrix Q and the points of the preliminary abnormal point matrix T;
calculating the cosine value of the included angle according to the included angle to obtain a cosine value set of the included angle;
calculating the variance of the cosine value set, and determining the variance as the abnormal value fraction of the points of the preliminary abnormal point matrix T;
and comparing the abnormal value fraction with a threshold value, if the abnormal value fraction is smaller than the threshold value, determining that the battery cell corresponding to the abnormal value fraction has a fault, and outputting the number of the battery cell with the fault.
2. The method for detecting faults of power batteries based on high-dimensional data diagnosis according to claim 1, wherein the step of preprocessing the voltage data matrix of the battery cells to obtain the voltage data matrix specifically comprises:
deleting the data at the moment of repeated recording and the data with errors in the voltage data of the single battery to obtain the voltage data of the single battery after cleaning;
using the formula Δ Ui=Ui+1-UiPerforming first-order difference processing on the voltage data of the cleaned battery monomer to obtain a voltage data matrix;
wherein, UiRepresents the voltage data, Δ U, of the battery cell acquired at the ith timeiAnd i represents the number of times of voltage data acquisition of the battery cell.
3. The method for detecting the fault of the power battery based on the high-dimensional data diagnosis, according to claim 1, wherein the clustering processing is performed on the voltage data matrix by using a clustering algorithm to obtain a clustering result, and specifically comprises:
initializing a region threshold and a number threshold;
determining an area in which the number of points in the voltage data matrix is greater than the number threshold within the area threshold as a core area, and forming a core point matrix Q by the points of the core area;
determining a region within the region threshold and outside the core region where the number of points in the voltage data matrix is less than the number threshold as a boundary region whose points are the boundary points;
determining a point of the voltage data matrix outside the region threshold as the outlier; the boundary points and the abnormal points form the preliminary abnormal point matrix T; determining the core point matrix Q and the preliminary abnormal point matrix T as clustering results;
judging whether the clustering result reaches a clustering target or not to obtain a judgment result;
when the judgment result shows yes, outputting the clustering result;
and when the judgment result shows that the voltage data matrix is not the core point matrix, adjusting the area threshold and the number threshold, and returning to the step of determining that the area with the number of points in the voltage data matrix larger than the number threshold in the area threshold is the core area and the points in the core area form the core point matrix Q'.
4. The power battery fault detection method based on high-dimensional data diagnosis according to claim 3, characterized in that the clustering target is that the number of the preliminary abnormal points accounts for 10% -20% of the number of the points in the voltage data matrix.
5. A power battery fault detection system based on high-dimensional data diagnosis is characterized by comprising a preprocessing unit, a voltage data matrix and a data processing unit, wherein the preprocessing unit is used for preprocessing voltage data of battery monomers to obtain the voltage data matrix;
the clustering unit is used for clustering the voltage data matrix by using a clustering algorithm to obtain a clustering result, wherein the clustering result comprises a preliminary abnormal point matrix T formed by boundary points and abnormal points and a core point matrix Q formed by core points;
an abnormal value score calculating unit for calculating an abnormal value score by using an abnormal value detection algorithm according to the clustering result;
the abnormal value score calculating unit specifically includes:
a mapping subunit, configured to map the preliminary abnormal point matrix T and the core point matrix Q to a high-dimensional space;
an included angle calculating subunit, configured to calculate an included angle formed between any two points in the core point matrix Q and a point of the preliminary abnormal point matrix T;
the included angle cosine value calculating operator unit is used for calculating included angle cosine values according to the included angles to obtain an included angle cosine value set;
an abnormal value score determining subunit, configured to calculate a variance of the set of cosine values, and determine the variance as an abnormal value score of a point of the preliminary abnormal point matrix T;
and the fault battery determining unit is used for comparing the abnormal value fraction with a threshold value, determining that the battery cell corresponding to the abnormal value fraction has a fault if the abnormal value fraction is smaller than the threshold value, and outputting the number of the battery cell with the fault.
6. The power battery fault detection system based on high-dimensional data diagnosis according to claim 5, wherein the preprocessing unit specifically comprises:
the data cleaning subunit is used for deleting the data repeatedly recorded at the moment and the data with errors in the voltage data of the single battery to obtain the voltage data of the single battery after cleaning;
a first order difference subunit for using the formula Δ Ui=Ui+1-UiPerforming first-order difference processing on the voltage data of the cleaned battery monomer to obtain a voltage data matrix;
wherein, UiRepresents the voltage data, Δ U, of the battery cell acquired at the ith timeiAnd i represents the number of times of voltage data acquisition of the battery cell.
7. The power battery fault detection system based on high-dimensional data diagnosis according to claim 5, wherein the clustering processing unit specifically comprises:
a threshold initialization subunit, configured to initialize a region threshold and a number threshold;
a core region determining subunit, configured to determine, as a core region, a region in which the number of points in the voltage data matrix is greater than the number threshold within the region threshold, and the core point matrix Q is formed by the points of the core region;
a boundary region determining subunit, configured to determine, as a boundary region, a region within the region threshold and outside the core region where the number of points in the voltage data matrix is smaller than the number threshold, where the point of the boundary region is the boundary point;
an abnormal point determination subunit, configured to determine a point of the voltage data matrix outside the region threshold as the abnormal point; the boundary points and the abnormal points form the preliminary abnormal point matrix T; determining the core point matrix Q and the preliminary abnormal point matrix T as clustering results;
a clustering result judging subunit, configured to judge whether the clustering result reaches a clustering target, to obtain a judgment result;
a clustering result output subunit, configured to output the clustering result when the determination result indicates yes;
and a threshold adjusting subunit, configured to adjust the area threshold and the number threshold when the determination result indicates no, and return to the step "determine that an area in which the number of points in the voltage data matrix is greater than the number threshold within the area threshold is a core area, and the core point matrix Q is configured by the points of the core area".
8. The high-dimensional data diagnosis based power battery fault detection system according to claim 7, wherein the clustering target is that the number of the preliminary abnormal points accounts for 10% to 20% of the number of the points in the voltage data matrix.
CN201811336819.XA 2018-11-12 2018-11-12 Power battery fault detection method and system based on high-dimensional data diagnosis Active CN109765490B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811336819.XA CN109765490B (en) 2018-11-12 2018-11-12 Power battery fault detection method and system based on high-dimensional data diagnosis

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811336819.XA CN109765490B (en) 2018-11-12 2018-11-12 Power battery fault detection method and system based on high-dimensional data diagnosis

Publications (2)

Publication Number Publication Date
CN109765490A CN109765490A (en) 2019-05-17
CN109765490B true CN109765490B (en) 2020-08-04

Family

ID=66449573

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811336819.XA Active CN109765490B (en) 2018-11-12 2018-11-12 Power battery fault detection method and system based on high-dimensional data diagnosis

Country Status (1)

Country Link
CN (1) CN109765490B (en)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110781569B (en) * 2019-11-08 2023-12-19 桂林电子科技大学 Abnormality detection method and system based on multi-resolution grid division
CN111157850B (en) * 2020-01-15 2022-06-21 上海电力大学 Mean value clustering-based power grid line fault identification method
EP4166963A4 (en) * 2020-08-03 2023-11-29 LG Energy Solution, Ltd. Battery diagnosis device, battery pack, battery system, and battery diagnosis method
CN111929591A (en) * 2020-08-21 2020-11-13 彩虹无线(北京)新技术有限公司 Fault battery detection method, device, equipment and computer storage medium
CN112213657A (en) * 2020-09-29 2021-01-12 江西江铃集团新能源汽车有限公司 Battery early warning method and system
CN112319309B (en) * 2020-11-04 2021-12-31 北京理工大学 Fault diagnosis method and system for power battery
CN113376534B (en) * 2021-05-12 2022-04-05 山东大学 Phase plane diagnosis method for early fault of power battery and advanced early warning system
CN114264957A (en) * 2021-12-02 2022-04-01 东软集团股份有限公司 Abnormal monomer detection method and related equipment thereof

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106021789A (en) * 2016-06-01 2016-10-12 北京交通大学 Fuzzy-intelligence-based rail car suspension system fault classification method and system
US20160349330A1 (en) * 2015-06-01 2016-12-01 Verizon Patent And Licensing Inc. Systems and methods for determining vehicle battery health
JP2017092028A (en) * 2015-11-02 2017-05-25 三星電子株式会社Samsung Electronics Co.,Ltd. Battery management method and battery management apparatus
CN107025384A (en) * 2015-10-15 2017-08-08 赵乐平 A kind of construction method of complex data forecast model
CN107340475A (en) * 2016-04-29 2017-11-10 株式会社日立制作所 Battery fault detection method and battery fault detection device
CN107755295A (en) * 2017-10-19 2018-03-06 杭州电子科技大学 A kind of lead-acid accumulator sorting technique based on charging and discharging curve

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160349330A1 (en) * 2015-06-01 2016-12-01 Verizon Patent And Licensing Inc. Systems and methods for determining vehicle battery health
CN107025384A (en) * 2015-10-15 2017-08-08 赵乐平 A kind of construction method of complex data forecast model
JP2017092028A (en) * 2015-11-02 2017-05-25 三星電子株式会社Samsung Electronics Co.,Ltd. Battery management method and battery management apparatus
CN107340475A (en) * 2016-04-29 2017-11-10 株式会社日立制作所 Battery fault detection method and battery fault detection device
CN106021789A (en) * 2016-06-01 2016-10-12 北京交通大学 Fuzzy-intelligence-based rail car suspension system fault classification method and system
CN107755295A (en) * 2017-10-19 2018-03-06 杭州电子科技大学 A kind of lead-acid accumulator sorting technique based on charging and discharging curve

Also Published As

Publication number Publication date
CN109765490A (en) 2019-05-17

Similar Documents

Publication Publication Date Title
CN109765490B (en) Power battery fault detection method and system based on high-dimensional data diagnosis
CN114430080B (en) Power battery cell abnormal self-discharge identification method based on operation data
CN113219361B (en) Abnormal self-discharge diagnosis method and system for lithium ion battery pack
EP4231027A1 (en) Lithium ion battery life span prediction method, discharge capacity retention rate prediction method, life span prediction program, discharge capacity retention rate prediction program, and information processing device
CN113406524B (en) Inconsistent fault diagnosis method and system for power battery system
CN112630660B (en) Battery fault identification method based on support vector machine
CN112345955B (en) Multi-fault online diagnosis method and system for power battery
CN114035086B (en) Multi-fault diagnosis method for battery pack based on signal processing
CN114559819A (en) Electric vehicle battery safety early warning method based on signal processing
CN114942387A (en) Real data-based power battery fault online detection method and system
CN112485695A (en) Detection method and device for power battery
CN114994543A (en) Energy storage power station battery fault diagnosis method and device and storage medium
CN115097319A (en) Power battery pack fault online diagnosis method and system
CN113791351B (en) Lithium battery life prediction method based on transfer learning and difference probability distribution
CN115327417A (en) Early warning method and system for abnormity of power battery monomer and electronic equipment
CN112009252B (en) Fault diagnosis and fault-tolerant control method for power battery system
CN112578296B (en) Battery capacity estimation method and apparatus, and computer storage medium
CN116400249A (en) Detection method and device for energy storage battery
CN107561452A (en) A kind of appraisal procedure of batteries health degree
CN116203450A (en) Method and device for detecting battery short-circuit fault, electronic equipment and storage medium
CN116008847A (en) Online lithium ion battery lithium separation detection method based on vehicle cloud cooperation
CN116400231B (en) Battery multi-fault detection method and device of energy storage system and electronic equipment
CN115436827A (en) Insulation resistance detection method, device and equipment
JP7435071B2 (en) Diagnostic equipment, diagnostic methods and programs
CN114879054A (en) Battery safety monitoring method and device, electronic equipment and storage 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