CN111025153A - Electric vehicle battery fault diagnosis method and device - Google Patents
Electric vehicle battery fault diagnosis method and device Download PDFInfo
- Publication number
- CN111025153A CN111025153A CN201811173222.8A CN201811173222A CN111025153A CN 111025153 A CN111025153 A CN 111025153A CN 201811173222 A CN201811173222 A CN 201811173222A CN 111025153 A CN111025153 A CN 111025153A
- Authority
- CN
- China
- Prior art keywords
- battery
- fault
- classifier
- diagnosed
- fault type
- 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.)
- Pending
Links
- 238000003745 diagnosis Methods 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000010801 machine learning Methods 0.000 claims abstract description 10
- 238000010276 construction Methods 0.000 claims description 4
- 208000032953 Device battery issue Diseases 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Landscapes
- Secondary Cells (AREA)
- Electric Propulsion And Braking For Vehicles (AREA)
Abstract
The embodiment of the application discloses a battery fault diagnosis method for an electric vehicle, which comprises the steps of acquiring historical battery characteristic data in advance, selecting a positive sample and a negative sample from the acquired historical battery characteristic data for one fault type, training a weak classifier corresponding to the fault type based on the positive sample and the negative sample of the fault type according to a machine learning algorithm, and constructing a battery fault classifier based on the weak classifier; when the battery fault classifier is used for diagnosing the battery fault, the battery characteristic data of the battery to be diagnosed is obtained, the battery characteristic data is input into the battery fault classifier, the battery fault classifier at least comprises one classifier, one classifier corresponds to one fault type, the battery fault classifier can automatically determine the fault type of the battery to be diagnosed according to the battery characteristic data, an engineer does not need to manually determine the fault type of the battery to be diagnosed, and the workload of the engineer in diagnosing the battery fault of the electric automobile is greatly reduced.
Description
Technical Field
The application relates to the field of battery fault diagnosis, in particular to a method and a device for diagnosing battery faults of an electric vehicle.
Background
The battery of the electric automobile is a power source on the electric automobile, the starting and running of the electric automobile need to depend on the electric energy provided by the battery, and once the battery of the electric automobile breaks down, the battery of the electric automobile has great influence on the work of the electric automobile. Therefore, it is very important to diagnose the fault of the battery of the electric vehicle accurately in time.
In the prior art, when a Battery of an electric vehicle fails, an engineer is usually required to read a fault code sent by a Battery Monitoring Unit (BMU), and analyze a current fault condition of the Battery by combining data such as Battery voltage and temperature stored in a Non-volatile memory (NVM), so as to obtain a Battery fault diagnosis result.
When the prior art is adopted to detect the faults of the electric automobile battery, engineers are often required to have abundant diagnosis experience, and have strong data analysis capability and fault judgment capability, so that accurate fault diagnosis results can be obtained.
Disclosure of Invention
In order to solve the technical problem, the application provides a battery fault diagnosis method for an electric vehicle, which can automatically perform battery fault diagnosis on a battery to be diagnosed, provide certain reference information for fault analysis and processing of an engineer, and reduce the workload required by the engineer to perform battery fault diagnosis.
The embodiment of the application discloses the following technical scheme:
in a first aspect, an embodiment of the present application provides a method for diagnosing a battery fault of an electric vehicle, where the method includes:
acquiring historical battery characteristic data;
for one fault type, selecting battery characteristic data which belong to the fault type from the historical battery characteristic data as a positive sample, and selecting battery characteristic data which do not belong to the fault type from the historical battery characteristic data as a negative sample;
training a positive sample and a negative sample of the fault type according to a machine learning algorithm to obtain a weak classifier corresponding to the fault type;
constructing a battery fault classifier according to the weak classifier;
acquiring battery characteristic data of a battery to be diagnosed;
inputting battery characteristic data of the battery to be diagnosed into the battery fault classifier, wherein the battery fault classifier at least comprises one classifier, and one classifier corresponds to one fault type;
and determining the fault type of the battery to be diagnosed according to the diagnosis data output by the battery fault classifier.
Optionally, the constructing a battery fault classifier according to the weak classifier includes:
training one weak classifier corresponding to one fault type, wherein the weak classifier is the battery fault classifier.
Optionally, the determining, according to the diagnostic data output by the battery fault classifier, the fault type to which the battery to be diagnosed belongs includes:
if the diagnosis data output by the battery is larger than a first preset threshold value, the battery to be diagnosed belongs to the fault type corresponding to the battery fault classifier;
and if the diagnosis data output by the battery fault classifier is smaller than the first preset threshold, the battery to be diagnosed does not belong to the fault type corresponding to the battery fault classifier.
Optionally, the constructing a battery fault classifier according to the weak classifier includes:
training at least two weak classifiers corresponding to a fault type, and constructing a strong classifier corresponding to the fault type according to the at least two weak classifiers, wherein the strong classifier is the battery fault classifier.
Optionally, the determining, according to the diagnostic data output by the battery fault classifier, the fault type to which the battery to be diagnosed belongs includes:
if the diagnosis data output by the battery fault classifier is larger than a second preset threshold value, the battery to be diagnosed belongs to the fault type corresponding to the battery fault classifier;
and if the diagnosis data output by the battery fault classifier is smaller than the second preset threshold, the battery to be diagnosed does not belong to the fault type corresponding to the battery fault classifier.
Optionally, the constructing a battery fault classifier according to the weak classifier includes:
training at least two weak classifiers corresponding to a fault type, and constructing a strong classifier corresponding to the fault type according to the at least two weak classifiers;
and operating the strong classifiers corresponding to the fault types in parallel, wherein the strong classifiers corresponding to the fault types in parallel form the battery fault classifier.
Optionally, the determining, according to the diagnostic data output by the battery fault classifier, the fault type to which the battery to be diagnosed belongs includes:
if the diagnosis data output by the strong classifier in the battery fault classifier is larger than a third preset threshold, the battery to be diagnosed belongs to the fault type corresponding to the strong classifier;
and if the diagnosis data output by the strong classifier in the battery fault classifier is smaller than the third preset threshold, the battery to be diagnosed does not belong to the fault type corresponding to the strong classifier.
In a second aspect, an embodiment of the present application provides an electric vehicle battery fault diagnosis device, which includes:
the historical data acquisition module is used for acquiring historical battery characteristic data;
the sample selection module is used for selecting battery characteristic data which belong to a fault type from the historical battery characteristic data as a positive sample and selecting battery characteristic data which do not belong to the fault type from the historical battery characteristic data as a negative sample for the fault type;
the training module is used for training the positive sample and the negative sample of the fault type according to a machine learning algorithm to obtain a weak classifier corresponding to the fault type;
the construction module is used for constructing a battery fault classifier according to the weak classifier;
the acquisition module is used for acquiring battery characteristic data of the battery to be diagnosed;
the input module is used for inputting the battery characteristic data of the battery to be diagnosed into a pre-trained battery fault classifier, wherein the battery fault classifier at least comprises one classifier, and one classifier corresponds to one fault type;
and the determining module is used for determining the fault type of the power-off battery to be diagnosed according to the diagnosis data output by the battery fault classifier.
According to the technical scheme, the battery fault diagnosis method for the electric vehicle, provided by the application, comprises the steps of acquiring historical battery characteristic data in advance, selecting a positive sample and a negative sample from the acquired historical battery characteristic data for one fault type, training a weak classifier corresponding to the fault type based on the positive sample and the negative sample of the fault type according to a machine learning algorithm, and constructing a battery fault classifier based on the weak classifier; when the battery fault classifier is used for diagnosing the battery fault, the battery characteristic data of the battery to be diagnosed is obtained, the battery characteristic data is input into the battery fault classifier, the battery fault classifier at least comprises one classifier, one classifier corresponds to one fault type, the battery fault classifier can automatically determine the fault type of the battery to be diagnosed according to the battery characteristic data, and an engineer does not need to determine the fault type of the battery to be diagnosed by manually analyzing the fault code sent by the BMU and the battery data such as the battery temperature, the battery voltage and the like stored in the NVM, so that the workload of the engineer in diagnosing the battery fault of the electric automobile is greatly reduced.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
Fig. 1 is a schematic flowchart of a method for diagnosing a battery fault of an electric vehicle according to an embodiment of the present application;
fig. 2 is a schematic flowchart of a training method of a battery fault classifier according to an embodiment of the present disclosure;
FIG. 3 is a diagram illustrating a strong classifier constructed by weak classifiers according to an embodiment of the present application;
FIG. 4 is a schematic diagram of determining a fault type of a battery to be diagnosed by using a strong classifier according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of an electric vehicle battery fault diagnosis device provided in an embodiment of the present application.
Detailed Description
The embodiments of the present application will be described below with reference to the accompanying drawings.
The method for diagnosing the battery fault of the electric automobile can automatically determine the fault type of the battery to be diagnosed according to the battery characteristic data of the battery to be diagnosed, and does not need to determine the fault diagnosis result of the battery to be diagnosed by manually analyzing fault codes sent by a BMU of the battery to be diagnosed and battery data such as battery voltage, temperature and the like stored in an NVM through an engineer, so that the workload of the engineer is greatly reduced.
The core technical idea of the electric vehicle battery fault diagnosis method provided by the application is introduced as follows:
the method for diagnosing the battery fault of the electric vehicle comprises the steps of obtaining historical battery characteristic data in advance, selecting a positive sample and a negative sample from the obtained historical battery characteristic data according to a fault type, training a weak classifier corresponding to the fault type according to a machine learning algorithm and the positive sample and the negative sample of the fault type, and constructing a battery fault classifier based on the weak classifier; when the battery fault classifier is used for diagnosing the battery fault, the battery characteristic data of the battery to be diagnosed is obtained, the battery characteristic data are input into the battery fault classifier, the battery fault classifier at least comprises one classifier, one classifier corresponds to one fault type, the battery fault classifier correspondingly outputs diagnosis data according to the input battery characteristic data, and the fault type of the battery to be diagnosed can be determined according to the diagnosis data.
According to the method for diagnosing the battery fault of the electric automobile, the battery characteristic data of the battery to be diagnosed is obtained, and the battery characteristic data is input into the pre-trained battery fault classifier, so that the battery fault classifier can automatically determine the fault type of the battery to be diagnosed according to the battery characteristic data, an engineer does not need to determine the fault type of the battery to be diagnosed by manually analyzing the fault code sent by the BMU and the battery data such as the temperature and the voltage of the battery stored in the NVM, and therefore the workload of the engineer in diagnosing the battery fault of the electric automobile is greatly reduced.
The method for diagnosing the battery fault of the electric vehicle provided by the application is introduced by the following embodiments:
referring to fig. 1, fig. 1 is a schematic flowchart of a method for diagnosing a battery fault of an electric vehicle according to an embodiment of the present application, and as shown in fig. 1, the method for diagnosing a battery fault of an electric vehicle includes:
step 101: and acquiring battery characteristic data of the battery to be diagnosed.
In some cases, a user finds that the electric vehicle is working abnormally, and thinks that the reason causing the electric vehicle to work abnormally is that the electric vehicle battery is out of order, and accordingly, the battery characteristic data of the battery to be diagnosed can be acquired by taking the electric vehicle battery as the battery to be diagnosed.
In addition, in some cases, an engineer needs to perform factory detection on a factory-shipped electric vehicle battery, detect whether the factory-shipped electric vehicle battery has a fault, and accordingly, may also use the electric vehicle battery that needs to be detected as a battery to be diagnosed, and obtain battery characteristic data of the battery to be diagnosed.
It should be noted that the battery characteristic data of the battery to be diagnosed may specifically include: the BMU state data of the battery to be diagnosed, and the voltage data, the current data, the temperature data, and other data stored in the battery NVM to be diagnosed, however, the battery characteristic data of the battery to be diagnosed may also include other data, and no limitation is made on the type of data specifically included in the battery characteristic data.
Step 102: and inputting the battery characteristic data of the battery to be diagnosed into a pre-trained battery fault classifier, wherein the battery fault classifier at least comprises one classifier, and one classifier corresponds to one fault type.
After battery characteristic data of a battery to be diagnosed are obtained, the battery characteristic data are input into a pre-trained battery fault classifier, and the battery fault classifier determines a fault type of the battery to be diagnosed according to the input battery characteristic data.
It should be noted that the battery fault classifier includes at least one classifier, and one classifier corresponds to one fault type.
It can be understood that, if only one classifier is included in the battery fault classifier, according to the battery characteristic data of the battery to be diagnosed, the battery fault classifier can determine whether the battery to be diagnosed belongs to the fault type corresponding to the classifier included in the battery fault classifier; if the battery fault classifier comprises a plurality of classifiers and the plurality of classifiers correspond to different fault types respectively, that is, the battery fault classifier can diagnose the fault types corresponding to the plurality of classifiers included in the battery fault classifier, then, according to the battery characteristic data of the battery to be diagnosed, the battery fault classifier can determine whether the battery to be diagnosed belongs to the fault types which can be diagnosed by the battery fault classifier, and if the battery to be diagnosed belongs to the fault types which can be diagnosed by the battery fault classifier, the battery fault classifier can further determine which fault type the battery to be diagnosed belongs to specifically according to the battery characteristic data of the battery to be diagnosed.
Step 103: and determining the fault type of the battery to be diagnosed according to the diagnosis data output by the battery fault classifier.
After the battery characteristic data of the battery to be diagnosed is input to the battery fault classifier, the battery fault classifier correspondingly outputs corresponding diagnosis data aiming at the battery characteristic data of the battery to be diagnosed according to the battery characteristic data of the battery to be diagnosed, and the fault type of the battery to be diagnosed can be determined according to the diagnosis data.
It can be understood that, if only one classifier is included in the battery fault classifier, whether the battery to be diagnosed belongs to the fault type corresponding to the classifier included in the battery fault classifier can be judged according to the diagnosis data output by the battery fault classifier; if the battery fault classifier comprises a plurality of classifiers, whether the power-off battery to be diagnosed belongs to the fault type which can be detected by the battery fault classifier or not can be judged according to the diagnosis data output by the battery fault classifier, and if the power-off battery to be diagnosed belongs to the fault type which can be detected by the battery fault classifier, which fault type the power-off battery to be diagnosed belongs to can be further judged.
According to the battery fault diagnosis method for the electric automobile, the battery characteristic data of the battery to be diagnosed are obtained, and the battery characteristic data are input into the battery fault classifier which is trained in advance, the battery fault classifier can automatically determine the fault type of the battery to be diagnosed according to the battery characteristic data, an engineer does not need to determine the fault type of the battery to be diagnosed through manually analyzing the fault code sent by the BMU and the battery data such as the battery temperature and the battery voltage stored in the NVM, and therefore the workload of the engineer in battery fault diagnosis of the electric automobile is greatly reduced.
In the embodiment corresponding to fig. 1, the battery fault classifier for determining the fault type of the battery to be diagnosed is pre-trained, and the battery fault classifier is constructed based on the weak classifier, and the training method of the weak classifier is described below with reference to fig. 2:
referring to fig. 2, fig. 2 is a schematic flowchart of a training method of a battery fault classifier according to an embodiment of the present disclosure. As shown in fig. 2, the training method of the battery fault classifier includes the following steps:
step 201: historical battery characteristic data is obtained.
In this embodiment, the historical battery characteristic data is battery characteristic data used for constructing the training sample, and the historical battery characteristic data may specifically include BMU state data of the battery, and data such as voltage data, current data, and temperature data stored in the battery NVM. The historical battery characteristic data may also include non-fault battery characteristic data, and a weak classifier corresponding to the non-fault battery characteristic data may be trained according to the non-fault battery characteristic data.
It is understood that the data type included in the historical battery characteristic data may determine the data type included in the battery characteristic data input to the battery fault classifier, specifically, if the historical battery characteristic data used to construct the training sample includes: correspondingly, when the BMU state data of the battery, the voltage data, the current data and the temperature data stored in the NVM of the battery are utilized to carry out fault diagnosis on the battery to be diagnosed by utilizing the battery fault classifier, the data type contained in the battery characteristic data input into the battery fault classifier needs to be consistent with the data type in the historical battery characteristic data, namely the BMU state data of the battery to be diagnosed, the voltage data, the current data and the temperature data stored in the NVM of the battery to be diagnosed need to be input into the battery fault classifier.
It can be understood that the more the historical battery feature data are acquired, the more accurate the classification result of the battery fault classifier trained according to the historical battery feature data is, but the number of the acquired historical battery feature data also affects the efficiency of training the battery fault classifier, so the acquired historical battery feature data are not limited in this embodiment, and can be determined according to actual requirements during specific implementation.
Step 202: and for one fault type, selecting battery characteristic data belonging to the fault type from the historical battery characteristic data as a positive sample, and selecting battery characteristic data not belonging to the fault type from the historical battery characteristic data as a negative sample.
For a fault type, the battery feature data belonging to the fault type is selected from the historical battery feature data obtained in step 201 as a positive sample, the battery feature data not belonging to the fault type is selected from the historical battery feature data as a negative sample, of course, the remaining battery feature data not belonging to the fault type in the historical battery feature data may also be all used as negative samples, and the positive sample and the negative sample for the fault type are combined to form a training sample set for training the classifier corresponding to the fault type.
For ease of understanding, the specific forms of the above-described positive and negative examples are illustrated below:
as shown in table 1 below, for the fault type that the battery voltage of the electric vehicle is high, a training sample set for the fault type is constructed. Selecting battery characteristic data belonging to a battery voltage high fault type from the obtained historical battery characteristic data as a positive sample, wherein as shown in table 1, a sample 2 and a sample 3 all belong to the battery characteristic data of the battery voltage high fault type, so that a classification label '1' is added to the sample 1, the sample 2 and the sample 3, namely, the sample 1, the sample 2 and the sample 3 are determined to be positive samples of the battery voltage high fault type; accordingly, the remaining historical battery characteristic data which is not of the battery voltage high fault type is taken as a negative sample, and as shown in table 1, a classification label "-1" is added to the sample 4, the sample 5 and the sample 6, namely, the sample 4, the sample 5 and the sample 6 are determined to be negative samples of the battery voltage high fault type.
TABLE 1
Step 203: and training the positive sample and the negative sample of the fault type according to a machine learning algorithm to obtain a weak classifier corresponding to the fault type.
And training positive samples and negative samples corresponding to the fault types by adopting a machine learning algorithm to obtain weak classifiers corresponding to the fault types. In specific implementation, a BP neural network algorithm can be adopted to train a weak classifier corresponding to a certain fault type. The weak classifier can determine whether the battery to be diagnosed belongs to the fault type corresponding to the weak classifier or not according to the input battery characteristic data of the battery to be diagnosed.
On the basis of constructing the weak classifier, the battery fault classifier can be further determined according to the constructed weak classifier, and the following methods for constructing the battery fault classifier are introduced on the basis of the weak classifier:
in a possible implementation manner, the method for constructing the weak classifier may be adopted, only one weak classifier is constructed for one fault type, the weak classifier is used as a battery fault classifier, and the battery fault classifier may judge whether the battery to be diagnosed belongs to the fault type corresponding to the battery fault classifier according to the input battery characteristic data of the battery to be diagnosed. When the battery fault classifier is implemented specifically, after battery characteristic data of a battery to be diagnosed is input into the battery fault classifier, the battery fault classifier outputs corresponding diagnosis data according to the battery characteristic data of the battery to be diagnosed, and when the diagnosis data output by the battery fault classifier is larger than a first preset threshold value, the battery to be diagnosed belongs to a fault type corresponding to the battery fault classifier, namely the battery to be diagnosed belongs to a fault type corresponding to the weak classifier; otherwise, when the diagnosis data output by the battery fault classifier is smaller than the first preset threshold, it is indicated that the battery to be diagnosed does not belong to the fault type corresponding to the battery fault classifier, that is, the battery to be diagnosed does not belong to the fault type corresponding to the weak classifier.
It should be noted that the first preset threshold may be set according to actual requirements, and the first preset threshold is not specifically limited herein.
For ease of understanding, the above method using the weak classifier as the battery fault classifier is exemplified below:
if a weak classifier corresponding to a high battery voltage fault is trained by using the acquired historical battery data, the weak classifier is used as a battery fault classifier, and accordingly, the fault type corresponding to the battery fault classifier is the high battery voltage. The battery fault classifier outputs corresponding diagnosis data according to the input battery characteristic data, when the diagnosis data output by the battery fault classifier is larger than a first preset threshold value, the fault type of the battery to be diagnosed belongs to the fault type with high battery voltage, and when the diagnosis data output by the battery fault classifier is smaller than the first preset threshold value, the fault type of the battery to be diagnosed does not belong to the fault type with high battery voltage.
In a possible implementation manner, corresponding weak classifiers can be respectively constructed for multiple fault types, different weak classifiers can identify different fault types, the weak classifiers corresponding to different fault types are operated in parallel, a plurality of the weak classifiers which are operated in parallel are used as battery fault classifiers, the battery fault classifier can determine whether a battery to be diagnosed belongs to a fault type which can be detected per se according to input battery characteristic data of the battery to be diagnosed, and if the battery to be diagnosed belongs to the fault type which can be detected per se, the battery fault classifier can also determine which fault type the battery to be diagnosed belongs to. When the battery fault classifier is implemented specifically, battery characteristic data of a battery to be diagnosed is input into the battery fault classifier, the battery fault classifier can judge whether the battery to be diagnosed belongs to a fault type which can be detected by the battery fault classifier according to the input battery characteristic data, and if the diagnosis data output by each weak classifier in the battery fault classifier is smaller than a first preset threshold value, the battery to be diagnosed does not belong to the fault type which can be detected by the battery fault classifier, or the battery to be diagnosed does not have faults; if the diagnosis data output by any weak classifier in the battery fault classifier is greater than the first preset threshold, it is indicated that the power-off battery to be diagnosed belongs to the fault type which can be detected by the battery fault classifier and belongs to the fault type corresponding to the weak classifier, it is indicated that the power-off battery to be diagnosed has multiple faults if the diagnosis data output by a plurality of weak classifiers in the battery fault classifier are greater than the first preset threshold, and the fault type to which the power-off battery to be diagnosed belongs is the fault type corresponding to the weak classifier whose output diagnosis data are greater than the first preset threshold.
It should be noted that the weak classifier has a weak classification capability and a low classification accuracy, and therefore, the weak classifier is used as a battery fault classifier, and a fault result obtained by diagnosing with the battery fault classifier has a poor accuracy.
In a possible implementation manner, in order to improve the accuracy of the fault diagnosis of the battery fault classifier, at least two weak classifiers are generally trained for one fault type, specifically, a training sample set composed of the historical battery data is trained for one fault type by adopting different training methods such as a gradient descent method, a conjugate method and the like, so as to obtain at least two weak classifiers for the fault type; or, at least two weak classifiers for the fault type are obtained by changing training parameters of a neural network for training the weak classifiers. And constructing a strong classifier corresponding to the fault type according to at least two weak classifiers, and using the strong classifier as a battery fault classifier corresponding to the fault type.
The following describes a specific implementation method for constructing a strong classifier according to at least two weak classifiers with reference to fig. 3:
as shown in fig. 3, at least two weak classifiers, such as weak classifier 1, weak classifier 2, weak classifier 3, … …, and weak classifier n, are constructed for a certain fault type by using historical battery feature data, where the number of trained weak classifiers can be determined according to actual requirements, and the number of trained weak classifiers is not specifically limited. When the method is concretely implemented, calibration data can be used for training a weight value corresponding to each weak classifier, the weight value corresponding to each weak classifier is used for carrying out weighting processing on data output by each weak classifier, as shown in fig. 3, the weight value corresponding to the weak classifier 1 is w1, the weight value corresponding to the weak classifier 2 is w2, the weight value corresponding to the weak classifier 3 is w3, … …, the weight value corresponding to the weak classifier n is wn, and data obtained after weighting processing on the data output by each weak classifier is data output by the strong classifier, namely diagnostic data output by the battery fault classifier.
Correspondingly, when the fault type of the battery to be diagnosed is determined according to the diagnosis data output by the battery fault classifier formed by strong classification, whether the diagnosis data output by the battery fault classifier is larger than a second preset threshold value needs to be judged, if the diagnosis data output by the battery fault classifier is larger than the second preset threshold value, the battery to be diagnosed belongs to the fault type corresponding to the battery fault classifier, namely the battery to be diagnosed belongs to the fault type corresponding to the strong classifier; on the contrary, if the diagnosis data output by the battery fault classifier is smaller than the second preset threshold, it is indicated that the battery to be diagnosed does not belong to the fault type corresponding to the battery fault classifier, that is, the battery to be diagnosed does not belong to the fault type corresponding to the strong classifier.
It should be noted that the second preset threshold may be set according to actual requirements, and the second preset threshold is not specifically limited herein.
Because the performance of the strong classifier is better than that of the weak classifier, the strong classifier can be used as a battery fault classifier to more accurately identify the fault type of the battery to be diagnosed, and the obtained diagnosis result is more accurate.
In a possible implementation manner, corresponding strong classifiers can be respectively constructed for multiple fault types, that is, at least two weak classifiers corresponding to a certain fault type are constructed for the certain fault type, a strong classifier corresponding to the fault type is constructed according to the at least two weak classifiers corresponding to the fault type, a corresponding strong classifier is respectively constructed for each fault type according to the above method, the strong classifiers corresponding to the fault types are operated in parallel, the strong classifiers corresponding to the fault types and operated in parallel are used as battery fault classifiers, and the battery fault classifier can determine which fault type the battery to be diagnosed specifically belongs to according to the input battery characteristic data of the battery to be diagnosed.
The method for determining the fault type of the battery to be diagnosed is specifically described below with reference to fig. 4:
as shown in fig. 4, the battery feature data of the battery to be diagnosed is input into the battery fault classifier, a strong classifier corresponding to each fault type, such as the strong classifier 1, the strong classifier 2, the strong classifier 3, … …, and the strong classifier n, is operated in parallel in the battery fault classifier, each strong classifier in the battery fault classifier correspondingly outputs the diagnostic data according to the input battery feature data, such as the strong classifier 1 outputs the diagnostic data 1, the strong classifier 2 outputs the diagnostic data 2, the strong classifier 3 outputs the diagnostic data 3, … …, the strong classifier 4 outputs the diagnostic data 4, determines whether each diagnostic data is greater than a third preset threshold, if the diagnostic data output by any strong classifier in the battery fault classifier is greater than the third preset threshold, the battery to be diagnosed belongs to the fault type corresponding to the strong classifier, if the diagnostic data output by a plurality of strong classifiers in the battery fault classifier is greater than the third preset threshold, the power failure type of the power failure pool to be diagnosed is the corresponding fault type of the strong classifiers; if the diagnostic data output by each strong classifier in the battery fault classifier is smaller than the third preset threshold, it is indicated that the power-off battery to be diagnosed does not belong to the fault type which can be diagnosed by the battery fault classifier, and the power-off battery to be diagnosed may not have any fault.
It should be noted that the third preset threshold may be set according to actual requirements, and the third preset threshold is not specifically limited herein.
The battery fault classifier is constructed by adopting the construction method of the battery fault classifier, and the battery fault classifier can determine whether the battery to be diagnosed belongs to a certain fault type or not according to the input battery characteristic data of the battery to be diagnosed, or determine the fault type to which the battery to be diagnosed specifically belongs. Therefore, the fault type of the battery to be diagnosed can be determined by using the battery fault classifier, and an engineer does not need to determine the fault type of the battery to be diagnosed by manually analyzing the fault code sent by the BMU and the battery data such as the battery temperature, the battery voltage and the like stored in the NVM, so that the workload of the engineer in the battery fault diagnosis of the electric automobile is greatly reduced.
In addition, the present application further provides an electric vehicle battery fault diagnosis device, referring to fig. 5, fig. 5 is a schematic structural diagram of an electric vehicle battery fault diagnosis device 500, and the device includes:
a historical data obtaining module 501, configured to obtain historical battery characteristic data;
a sample selecting module 502, configured to, for a fault type, select battery characteristic data belonging to the fault type from the historical battery characteristic data as a positive sample, and select battery characteristic data not belonging to the fault type from the historical battery characteristic data as a negative sample;
a training module 503, configured to train the positive sample and the negative sample of the fault type according to a machine learning algorithm to obtain a weak classifier corresponding to the fault type;
a construction module 504, configured to construct a battery fault classifier according to the weak classifier;
an obtaining module 505, configured to obtain battery characteristic data of a battery to be diagnosed;
an input module 506, configured to input battery characteristic data of the battery to be diagnosed into a pre-trained battery fault classifier, where the battery fault classifier includes at least one classifier, and one classifier corresponds to one fault type;
and the determining module 507 is configured to determine a fault type to which the battery to be diagnosed belongs according to the diagnostic data output by the battery fault classifier.
The battery fault diagnosis device for the electric vehicle, provided by the embodiment of the application, acquires the battery characteristic data of the battery to be diagnosed, and inputs the battery characteristic data into the pre-trained battery fault classifier, the battery fault classifier can automatically determine the fault type of the battery to be diagnosed according to the battery characteristic data, an engineer does not need to determine the fault type of the battery to be diagnosed by manually analyzing the fault code sent by the BMU and the battery data such as the battery temperature and the battery voltage stored in the NVM, and therefore the workload of the engineer in battery fault diagnosis of the electric vehicle is greatly reduced.
It should be noted that, in the present specification, all the embodiments are described in a progressive manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the apparatus and system embodiments, since they are substantially similar to the method embodiments, they are described in a relatively simple manner, and reference may be made to some of the descriptions of the method embodiments for related points. The above-described embodiments of the apparatus and system are merely illustrative, and the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
The above description is only one specific embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (8)
1. An electric vehicle battery fault diagnosis method, characterized in that the method comprises:
acquiring historical battery characteristic data;
for one fault type, selecting battery characteristic data which belong to the fault type from the historical battery characteristic data as a positive sample, and selecting battery characteristic data which do not belong to the fault type from the historical battery characteristic data as a negative sample;
training a positive sample and a negative sample of the fault type according to a machine learning algorithm to obtain a weak classifier corresponding to the fault type;
constructing a battery fault classifier according to the weak classifier;
acquiring battery characteristic data of a battery to be diagnosed;
inputting battery characteristic data of the battery to be diagnosed into the battery fault classifier, wherein the battery fault classifier at least comprises one classifier, and one classifier corresponds to one fault type;
and determining the fault type of the battery to be diagnosed according to the diagnosis data output by the battery fault classifier.
2. The method of claim 1, wherein said constructing a battery fault classifier from the weak classifiers comprises:
training one weak classifier corresponding to one fault type, wherein the weak classifier is the battery fault classifier.
3. The method according to claim 2, wherein the determining the fault type of the battery to be diagnosed belongs to according to the diagnosis data output by the battery fault classifier comprises:
if the diagnosis data output by the battery fault classifier is larger than a first preset threshold value, the battery to be diagnosed belongs to the fault type corresponding to the battery fault classifier;
and if the diagnosis data output by the battery fault classifier is smaller than the first preset threshold, the battery to be diagnosed does not belong to the fault type corresponding to the battery fault classifier.
4. The method of claim 1, wherein said constructing a battery fault classifier from the weak classifiers comprises:
training at least two weak classifiers corresponding to a fault type, and constructing a strong classifier corresponding to the fault type according to the at least two weak classifiers, wherein the strong classifier is the battery fault classifier.
5. The method according to claim 4, wherein the determining the fault type of the battery to be diagnosed belongs to according to the diagnosis data output by the battery fault classifier comprises:
if the diagnosis data output by the battery fault classifier is larger than a second preset threshold value, the battery to be diagnosed belongs to the fault type corresponding to the battery fault classifier;
and if the diagnosis data output by the battery fault classifier is smaller than the second preset threshold, the battery to be diagnosed does not belong to the fault type corresponding to the battery fault classifier.
6. The method of claim 1, wherein said constructing a battery fault classifier from the weak classifiers comprises:
training at least two weak classifiers corresponding to a fault type, and constructing a strong classifier corresponding to the fault type according to the at least two weak classifiers;
and operating the strong classifiers corresponding to the fault types in parallel, wherein the strong classifiers corresponding to the fault types in parallel form the battery fault classifier.
7. The method according to claim 6, wherein the determining the fault type of the battery to be diagnosed belongs to according to the diagnosis data output by the battery fault classifier comprises:
if the diagnosis data output by the strong classifier in the battery fault classifier is larger than a third preset threshold, the battery to be diagnosed belongs to the fault type corresponding to the strong classifier;
and if the diagnosis data output by the strong classifier in the battery fault classifier is smaller than the third preset threshold, the battery to be diagnosed does not belong to the fault type corresponding to the strong classifier.
8. An electric vehicle battery failure diagnosis apparatus, characterized by comprising:
the historical data acquisition module is used for acquiring historical battery characteristic data;
the sample selection module is used for selecting battery characteristic data which belong to a fault type from the historical battery characteristic data as a positive sample and selecting battery characteristic data which do not belong to the fault type from the historical battery characteristic data as a negative sample for the fault type;
the training module is used for training the positive sample and the negative sample of the fault type according to a machine learning algorithm to obtain a weak classifier corresponding to the fault type;
the construction module is used for constructing a battery fault classifier according to the weak classifier;
the acquisition module is used for acquiring battery characteristic data of the battery to be diagnosed;
the input module is used for inputting the battery characteristic data of the battery to be diagnosed into the battery fault classifier, the battery fault classifier at least comprises one classifier, and one classifier corresponds to one fault type;
and the determining module is used for determining the fault type of the power-off battery to be diagnosed according to the diagnosis data output by the battery fault classifier.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811173222.8A CN111025153A (en) | 2018-10-09 | 2018-10-09 | Electric vehicle battery fault diagnosis method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811173222.8A CN111025153A (en) | 2018-10-09 | 2018-10-09 | Electric vehicle battery fault diagnosis method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111025153A true CN111025153A (en) | 2020-04-17 |
Family
ID=70190950
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811173222.8A Pending CN111025153A (en) | 2018-10-09 | 2018-10-09 | Electric vehicle battery fault diagnosis method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111025153A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111157898A (en) * | 2020-01-07 | 2020-05-15 | 清华大学深圳国际研究生院 | New energy vehicle online battery fault detection and analysis method and device |
| CN112067997A (en) * | 2020-08-14 | 2020-12-11 | 杭州小电科技股份有限公司 | Diagnosis method and system for mobile power supply leasing equipment, electronic equipment and storage medium |
| CN112710956A (en) * | 2020-12-17 | 2021-04-27 | 四川虹微技术有限公司 | Battery management system fault detection system and method based on expert system |
| CN115230475A (en) * | 2022-07-28 | 2022-10-25 | 中国汽车工程研究院股份有限公司 | Fault diagnosis method for abnormal electrical connection of battery system based on fault data characteristics |
| CN116429339A (en) * | 2023-06-07 | 2023-07-14 | 海斯坦普汽车组件(北京)有限公司 | Leakage detection method and system for new energy battery box |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2482474A2 (en) * | 2011-01-31 | 2012-08-01 | MediaTek Inc. | Apparatus for communicating another device |
| CN103633938A (en) * | 2013-12-10 | 2014-03-12 | 中国矿业大学 | Photovoltaic array fault positioning method |
| CN103995237A (en) * | 2014-05-09 | 2014-08-20 | 南京航空航天大学 | Satellite power supply system online fault diagnosis method |
| CN104021238A (en) * | 2014-03-25 | 2014-09-03 | 重庆邮电大学 | Lead-acid power battery system fault diagnosis method |
| CN106168799A (en) * | 2016-06-30 | 2016-11-30 | 常伟 | A kind of method carrying out batteries of electric automobile predictive maintenance based on big data machine learning |
| CN107633266A (en) * | 2017-09-07 | 2018-01-26 | 西安交通大学 | A kind of electric locomotive OCS and pantograph arc method for measuring |
| CN108344947A (en) * | 2018-01-11 | 2018-07-31 | 西南交通大学 | A kind of fuel cell diagnostic method of non-intrusion type |
-
2018
- 2018-10-09 CN CN201811173222.8A patent/CN111025153A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2482474A2 (en) * | 2011-01-31 | 2012-08-01 | MediaTek Inc. | Apparatus for communicating another device |
| CN103633938A (en) * | 2013-12-10 | 2014-03-12 | 中国矿业大学 | Photovoltaic array fault positioning method |
| CN104021238A (en) * | 2014-03-25 | 2014-09-03 | 重庆邮电大学 | Lead-acid power battery system fault diagnosis method |
| CN103995237A (en) * | 2014-05-09 | 2014-08-20 | 南京航空航天大学 | Satellite power supply system online fault diagnosis method |
| CN106168799A (en) * | 2016-06-30 | 2016-11-30 | 常伟 | A kind of method carrying out batteries of electric automobile predictive maintenance based on big data machine learning |
| CN107633266A (en) * | 2017-09-07 | 2018-01-26 | 西安交通大学 | A kind of electric locomotive OCS and pantograph arc method for measuring |
| CN108344947A (en) * | 2018-01-11 | 2018-07-31 | 西南交通大学 | A kind of fuel cell diagnostic method of non-intrusion type |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111157898A (en) * | 2020-01-07 | 2020-05-15 | 清华大学深圳国际研究生院 | New energy vehicle online battery fault detection and analysis method and device |
| CN112067997A (en) * | 2020-08-14 | 2020-12-11 | 杭州小电科技股份有限公司 | Diagnosis method and system for mobile power supply leasing equipment, electronic equipment and storage medium |
| CN112067997B (en) * | 2020-08-14 | 2023-11-14 | 杭州小电科技股份有限公司 | Diagnosis method and system for portable power source leasing equipment, electronic equipment and storage medium |
| CN112710956A (en) * | 2020-12-17 | 2021-04-27 | 四川虹微技术有限公司 | Battery management system fault detection system and method based on expert system |
| CN112710956B (en) * | 2020-12-17 | 2023-08-04 | 四川虹微技术有限公司 | Expert system-based battery management system fault detection system and method |
| CN115230475A (en) * | 2022-07-28 | 2022-10-25 | 中国汽车工程研究院股份有限公司 | Fault diagnosis method for abnormal electrical connection of battery system based on fault data characteristics |
| CN116429339A (en) * | 2023-06-07 | 2023-07-14 | 海斯坦普汽车组件(北京)有限公司 | Leakage detection method and system for new energy battery box |
| CN116429339B (en) * | 2023-06-07 | 2023-09-01 | 海斯坦普汽车组件(北京)有限公司 | Leakage detection method and system for new energy battery box |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111025153A (en) | Electric vehicle battery fault diagnosis method and device | |
| WO2023024851A1 (en) | Battery equalization method and system | |
| CN110161425B (en) | A Prediction Method of Remaining Service Life Based on Degradation Stages of Lithium Batteries | |
| CN113870260B (en) | Welding defect real-time detection method and system based on high-frequency time sequence data | |
| CN108931725B (en) | A method and device for battery fault detection | |
| CN106772082B (en) | Voltage acquisition line disconnection detection method for single battery | |
| US20190115778A1 (en) | Autonomous screening and optimization of battery formation and cycling procedures | |
| CN115267557B (en) | Lithium battery electrolyte leakage fault diagnosis method and device and electronic equipment | |
| CN105453141B (en) | Apparatus and method for detecting the failure in electronic system | |
| CN105116870B (en) | Fault analysis method, device and system of air conditioning unit | |
| US11733309B2 (en) | Method and system for self-discharge prognostics for vehicle battery cells with an internal short circuit | |
| CN111191400B (en) | Vehicle part life prediction method and system based on user fault reporting data | |
| CN112183780B (en) | Fault repair guidance method, device, system, and storage medium | |
| KR20200024637A (en) | Method, apparatus and computer program for diagnosing an internal short of a batterty pack | |
| CN117150276B (en) | Machine learning model construction method, vehicle driving risk prediction method and device | |
| CN115291113A (en) | Unified joint estimation method and system for SOC, SOH and RUL of lithium ion battery | |
| CN118483593A (en) | Battery fault diagnosis method, model training method, system and program product | |
| CN108226790A (en) | Equalization circuit state detection device and method, battery pack and battery management system | |
| CN115480944A (en) | Black screen fault analysis method and device of vehicle-mounted entertainment terminal, vehicle and medium | |
| CN117648612B (en) | Parallel battery pack fault detection method, device, electronic equipment and storage medium | |
| CN119377790A (en) | A lithium-ion battery pack multi-fault diagnosis method, device, terminal equipment and computer-readable storage medium | |
| CN112327804A (en) | Fault diagnosis method, device and system of traction control unit and train | |
| CN114964467B (en) | Equipment detection method and device, sound collection equipment and storage medium | |
| CN117059844A (en) | Fuel cell fault diagnosis method, device, equipment and storage medium | |
| KR102656339B1 (en) | Device for measuring service life of reusable battery using artificial intelligence |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200417 |