CN114994543A - Energy storage power station battery fault diagnosis method and device and storage medium - Google Patents

Energy storage power station battery fault diagnosis method and device and storage medium Download PDF

Info

Publication number
CN114994543A
CN114994543A CN202210917698.8A CN202210917698A CN114994543A CN 114994543 A CN114994543 A CN 114994543A CN 202210917698 A CN202210917698 A CN 202210917698A CN 114994543 A CN114994543 A CN 114994543A
Authority
CN
China
Prior art keywords
voltage
battery
predicted
prediction
energy storage
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
Application number
CN202210917698.8A
Other languages
Chinese (zh)
Inventor
欧阳志国
夏向阳
马芳
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.)
Hunan Huada Electrician Hi Tech Co ltd
Original Assignee
Hunan Huada Electrician Hi Tech Co ltd
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 Hunan Huada Electrician Hi Tech Co ltd filed Critical Hunan Huada Electrician Hi Tech Co ltd
Priority to CN202210917698.8A priority Critical patent/CN114994543A/en
Publication of CN114994543A publication Critical patent/CN114994543A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y02E60/10—Energy storage using batteries

Abstract

The invention discloses a method and a device for diagnosing battery faults of an energy storage power station and a storage medium, wherein the diagnosis method comprises the following steps: acquiring battery historical data of an energy storage power station; respectively inputting battery historical data into a battery voltage prediction model and a battery ECM model based on LSTM to respectively obtain a first prediction voltage and a second prediction voltage; calculating a predicted normal voltage by using an improved self-adaptive boosting method according to the first predicted voltage and the second predicted voltage; and comparing the predicted normal voltage with the actual voltage of the battery to realize battery fault diagnosis. The method can realize potential fault risk assessment and diagnose the battery fault in time, and has good robustness and high prediction precision.

Description

Energy storage power station battery fault diagnosis method and device and storage medium
Technical Field
The invention relates to the technical field of batteries, in particular to a method and a device for diagnosing battery faults of an energy storage power station and a storage medium.
Background
To meet the challenges of fossil oil consumption and environmental pollution, energy storage power stations are being actively developed and widely adopted on a global scale. The battery system is an indispensable component of the energy storage power station, and determines the working efficiency and the cost benefit of the energy storage power station to a great extent.
Multiple battery cells are typically connected in series and/or parallel configurations to meet voltage and capacity requirements. These constituent batteries may malfunction due to battery degradation, electrical failure, or misuse. If these faults are left untreated, thermal runaway may be triggered.
Therefore, accurate and timely battery fault diagnosis and safety warning issuance are critical to preventing thermal runaway from occurring and ensuring safe operation of energy storage power stations. However, the existing method for diagnosing the battery fault of the energy storage power station has low accuracy and needs to be further improved.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides a method, a device and a storage medium for diagnosing the battery fault of an energy storage power station, wherein a coupling module based on improved self-adaptive boosting is adopted by combining a long-time memory recurrent neural network (LSTM) and an Equivalent Circuit Model (ECM). A real-world fault diagnosis strategy is used to implement online fault diagnosis.
In a first aspect, a method for diagnosing battery faults of an energy storage power station is provided, which includes:
acquiring battery historical data of an energy storage power station;
respectively inputting battery historical data into a battery voltage prediction model and a battery ECM model based on LSTM to respectively obtain a first prediction voltage and a second prediction voltage;
calculating a predicted normal voltage by using an improved self-adaptive boosting method according to the first predicted voltage and the second predicted voltage;
and comparing the predicted normal voltage with the actual voltage of the battery to realize battery fault diagnosis.
Further, an input matrix of the LSTM-based battery voltage prediction model
Figure 557217DEST_PATH_IMAGE001
Is represented as follows:
Figure 130149DEST_PATH_IMAGE002
wherein n is the number of input features;
Figure 378728DEST_PATH_IMAGE003
is the time step;
Figure 977200DEST_PATH_IMAGE004
is a step of time
Figure 506751DEST_PATH_IMAGE005
A battery unit
Figure 63635DEST_PATH_IMAGE006
The voltage of (a) is set to be,
Figure 533930DEST_PATH_IMAGE007
；
Figure 998410DEST_PATH_IMAGE008
is a step of timetTo be atjAn input feature;
the output matrix B of the LSTM-based battery voltage prediction model is represented as follows:
Figure 136130DEST_PATH_IMAGE009
wherein the content of the first and second substances,
Figure 582024DEST_PATH_IMAGE010
is a step of time
Figure 805195DEST_PATH_IMAGE011
A battery unit
Figure 11048DEST_PATH_IMAGE006
The predicted voltage of (2).
Further, in the battery ECM model, the battery voltage is predicted by assuming that the battery SOC, temperature, and aging level remain constant for three consecutive time steps by the following equation:
Figure 800013DEST_PATH_IMAGE012
in the formula (I), the compound is shown in the specification,
Figure 151228DEST_PATH_IMAGE013
is a step of time
Figure 127275DEST_PATH_IMAGE014
A battery unit
Figure 871240DEST_PATH_IMAGE006
The predicted voltage of (a);
Figure 983552DEST_PATH_IMAGE015
is that
Figure 36828DEST_PATH_IMAGE016
The data matrix of (2), wherein,
Figure 234591DEST_PATH_IMAGE017
is the voltage of the battery pack at a time step t,
Figure 516668DEST_PATH_IMAGE018
is the current of the battery at time step t;
Figure 280224DEST_PATH_IMAGE019
is a parameter matrix in which
Figure 255134DEST_PATH_IMAGE020
Wherein
Figure 926811DEST_PATH_IMAGE021
Is a battery cell
Figure 74895DEST_PATH_IMAGE006
OCV at time step t;
Figure 896221DEST_PATH_IMAGE022
、
Figure 510873DEST_PATH_IMAGE023
are ECM model parameters.
Further, calculating the predicted normal voltage by using an improved adaptive boosting method according to the first predicted voltage and the second predicted voltage, wherein the method comprises the following steps:
predicting normal voltage
Figure 932496DEST_PATH_IMAGE024
Calculated by the following formula:
Figure 25217DEST_PATH_IMAGE025
in the formula (I), the compound is shown in the specification,
Figure 966628DEST_PATH_IMAGE026
is the first predicted voltage to be applied to the first transistor,
Figure 798187DEST_PATH_IMAGE027
is the second predicted voltage to be the second predicted voltage,
Figure 192259DEST_PATH_IMAGE028
and
Figure 885408DEST_PATH_IMAGE029
respectively, the weights of the first predicted voltage and the second predicted voltage.
Further, the weights of the first predicted voltage and the second predicted voltage
Figure 212485DEST_PATH_IMAGE028
And
Figure 965677DEST_PATH_IMAGE029
obtained by the following method:
using the LSTM-based battery voltage prediction model and the battery ECM model to predict the voltage, the following voltage prediction matrix is obtained
Figure 96313DEST_PATH_IMAGE030
：
Figure 593153DEST_PATH_IMAGE031
In the formula (I), the compound is shown in the specification,
Figure 243577DEST_PATH_IMAGE032
the class of the model is represented by,
Figure 636513DEST_PATH_IMAGE032
for LSTM, corresponding to the LSTM-based battery voltage prediction model,
Figure 251515DEST_PATH_IMAGE032
when the battery is the ECM, the battery corresponds to a battery ECM model;
Figure 286467DEST_PATH_IMAGE033
is a battery cell
Figure 853715DEST_PATH_IMAGE006
Step of time
Figure 683130DEST_PATH_IMAGE005
The predicted voltage of (a);
Figure 273512DEST_PATH_IMAGE034
the total step length is the corresponding prediction step length in the voltage prediction matrix, and b is the total number of the battery units;
the rate of regression of the voltage prediction matrices for the LSTM-based battery voltage prediction model and the battery ECM model predicted voltages was calculated by the following formula:
Figure 361422DEST_PATH_IMAGE035
in the formula (I), the compound is shown in the specification,
Figure 517597DEST_PATH_IMAGE036
is a battery cell
Figure 721177DEST_PATH_IMAGE006
Step of time of
Figure 782542DEST_PATH_IMAGE005
The actual voltage at (c); e is the maximum error of the error signal,
Figure 424876DEST_PATH_IMAGE037
(ii) a m is the number of samples of m = a × b;
calculating initial weights of predicted voltages of the LSTM-based battery voltage prediction model and the battery ECM model by the following formulas
Figure 701137DEST_PATH_IMAGE038
And
Figure 872355DEST_PATH_IMAGE039
：
Figure 437329DEST_PATH_IMAGE040
obtaining the weight of the first prediction voltage and the second prediction voltage by the following formula
Figure 414512DEST_PATH_IMAGE028
And
Figure 732230DEST_PATH_IMAGE029
：
Figure 339929DEST_PATH_IMAGE041
Figure 188936DEST_PATH_IMAGE042
。
further, weights of the first predicted voltage and the second predicted voltage are calculated
Figure 641914DEST_PATH_IMAGE028
And
Figure 96029DEST_PATH_IMAGE029
according to the online update dataTo last
Figure 126826DEST_PATH_IMAGE034
And (4) calculating each time step.
Further, the comparing the predicted normal voltage with the actual voltage of the battery to realize the battery fault diagnosis includes:
and calculating the difference value between the actual voltage and the predicted normal voltage of the battery, judging whether the voltage of the battery is in a normal, overvoltage or undervoltage state according to the comparison result of the difference value and a preset fault level voltage threshold value, and if the voltage of the battery is in the overvoltage or undervoltage state, sending out a warning of a corresponding level according to the comparison result.
In a second aspect, an apparatus for diagnosing battery faults of an energy storage power station is provided, which includes:
the data acquisition module is used for acquiring battery historical data of the energy storage power station;
the voltage prediction module is used for respectively inputting battery historical data into a battery voltage prediction model and a battery ECM model based on LSTM to respectively obtain a first predicted voltage and a second predicted voltage;
the coupling module is used for calculating the predicted normal voltage by using an improved self-adaptive boosting method according to the first predicted voltage and the second predicted voltage;
and the fault diagnosis module is used for comparing the predicted normal voltage with the actual voltage of the battery to realize battery fault diagnosis.
In a third aspect, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the energy storage power station battery fault diagnosis method as described above.
The invention provides a method and a device for diagnosing battery faults of an energy storage power station and a storage medium, and a battery voltage prediction model and a battery ECM model based on LSTM are good in battery voltage prediction. The prediction accuracy of the battery ECM model is better than that of the LSTM-based battery voltage prediction model, but the battery ECM model cannot be directly used to distinguish between a faulty battery and a normal battery. In contrast, LSTM-based battery voltage prediction models are better able to capture voltage variation trends and are suitable for fault diagnosis based on data collected from energy storage power stations. Therefore, the battery voltage predicted by the LSTM-based battery voltage prediction model and the battery ECM model is combined by using the self-adaptive boosting method, so that the prediction error is reduced, the accuracy of fault diagnosis is improved, the potential fault risk assessment can be realized, and an early thermal runaway warning can be sent out. The scheme of the invention has good robustness and high prediction precision.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic diagram of a fault diagnosis scheme for a battery of an energy storage power station according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the examples given herein without any inventive step, are within the scope of the present invention.
Example 1
As shown in fig. 1, the present embodiment provides a method for diagnosing a battery fault of an energy storage power station, including:
s1: and acquiring historical battery data of the energy storage power station.
Preprocessing historical data of the battery by using a preprocessing module, wherein the preprocessing process comprises the step of standardizing data of different units so as to facilitate subsequent dimensionless calculation; filling the data vacancy by utilizing methods such as interpolation, fitting, simulation and the like; clustering the data and removing abnormal values; and the principal component analysis method is utilized to reduce the dimension of data, so that the calculation speed is conveniently accelerated.
S2: and respectively inputting the processed battery historical data into a battery voltage prediction model and a battery ECM model based on the LSTM to respectively obtain a first predicted voltage and a second predicted voltage.
In particular, the LSTM model performs well as a recurrent neural network in avoiding gradient extinction and explosion. During the actual operation of the energy storage power station, the voltage is a time-dependent parameter, and an LSTM-based battery voltage prediction model can be established to predict the change of the voltage along with the time. The historical data plays an important role in predicting the battery state, so that a 'many-to-one' structure of the LSTM is selected, and an input matrix of a battery voltage prediction model based on the LSTM is selected
Figure 463130DEST_PATH_IMAGE001
Is represented as follows:
Figure 985378DEST_PATH_IMAGE002
wherein n is the number of input features;
Figure 28420DEST_PATH_IMAGE003
is the time step;
Figure 774660DEST_PATH_IMAGE004
is a step of time
Figure 535942DEST_PATH_IMAGE005
A battery unit
Figure 845570DEST_PATH_IMAGE006
The voltage of (a) is set to be,
Figure 805435DEST_PATH_IMAGE007
；
Figure 660259DEST_PATH_IMAGE008
is a step of time
Figure 174417DEST_PATH_IMAGE005
To be at
Figure 835205DEST_PATH_IMAGE043
An input feature;
the output matrix B of the LSTM-based battery voltage prediction model is represented as follows:
Figure 852840DEST_PATH_IMAGE009
wherein the content of the first and second substances,
Figure 393412DEST_PATH_IMAGE010
is a step of time
Figure 926024DEST_PATH_IMAGE011
A battery unit
Figure 328187DEST_PATH_IMAGE006
The predicted voltage of (2).
The LSTM-based battery voltage prediction model may be implemented in an off-line training mode and an on-line training mode, during the training, a large amount of data collected from a general energy storage power station may be used to perform the LSTM-based battery voltage prediction model training, and AOM (approximate optimization method) and Pearson Correlation Coefficient (PCC) are respectively used to optimize the hyperparameter and select the input features, in this embodiment, the selected input features are preferably selected from the voltage of each battery cell, the total voltage of the battery pack, and the SOC of the battery pack.
The WEIGAN model (ECM model) stands out for its high modeling accuracy and low computational complexity and is used to describe battery dynamics. It is reasonable to assume that in the battery ECM model, the battery SOC, temperature and aging level remain constant for three consecutive time steps and the battery voltage is predicted by the following equation:
Figure 200328DEST_PATH_IMAGE012
in the formula，
Figure 724850DEST_PATH_IMAGE013
Is a step of time
Figure 682442DEST_PATH_IMAGE014
A battery unit
Figure 137563DEST_PATH_IMAGE006
The predicted voltage of (a);
Figure 926527DEST_PATH_IMAGE015
is that
Figure 559634DEST_PATH_IMAGE016
The data matrix of (2), wherein,
Figure 4521DEST_PATH_IMAGE017
is the voltage of the battery at time step t,
Figure 810803DEST_PATH_IMAGE018
is the current of the battery at time step t;
Figure 657537DEST_PATH_IMAGE019
is a parameter matrix in which
Figure 543820DEST_PATH_IMAGE020
Wherein
Figure 741583DEST_PATH_IMAGE021
Is a battery cell
Figure 289239DEST_PATH_IMAGE006
OCV at time step t;
Figure 52796DEST_PATH_IMAGE022
、
Figure 762126DEST_PATH_IMAGE023
are ECM model parameters.
S3: and calculating the predicted normal voltage by using an improved self-adaptive boosting method according to the first predicted voltage and the second predicted voltage.
Adaptive boosting (adaboost) is an algorithm that combines weak classifiers into a strong classifier. Improved AdaBoost may be used to solve the regression problem. The voltages predicted by the two models are combined using a weighted sum method.
Firstly, weight calculation is carried out, and the voltage is predicted by using a battery voltage prediction model based on LSTM and a battery ECM model to obtain the following voltage prediction matrix
Figure 243923DEST_PATH_IMAGE030
：
Figure 844537DEST_PATH_IMAGE031
In the formula (I), the compound is shown in the specification,
Figure 400284DEST_PATH_IMAGE032
the class of the model is represented by,
Figure 811673DEST_PATH_IMAGE032
for LSTM, corresponding to the LSTM-based battery voltage prediction model,
Figure 515187DEST_PATH_IMAGE032
when the battery is the ECM, the battery corresponds to a battery ECM model;
Figure 139066DEST_PATH_IMAGE044
（t=1,2,…,
Figure 329745DEST_PATH_IMAGE034
(ii) a c =1,2, …, b) is a battery cell
Figure 443195DEST_PATH_IMAGE006
Step of time
Figure 571688DEST_PATH_IMAGE005
The predicted voltage of (a);
Figure 61575DEST_PATH_IMAGE034
the total step length is the corresponding prediction step length in the voltage prediction matrix, and b is the total number of the battery units;
calculating a regression error rate of a voltage prediction matrix of the LSTM-based battery voltage prediction model and the battery ECM model predicted voltages by the following formula
Figure 857493DEST_PATH_IMAGE045
：
Figure 345106DEST_PATH_IMAGE035
In the formula (I), the compound is shown in the specification,
Figure 475742DEST_PATH_IMAGE036
is a battery cell
Figure 503741DEST_PATH_IMAGE006
Step of time of
Figure 888586DEST_PATH_IMAGE005
The actual voltage at (c); e is the maximum error of the error signal,
Figure 609417DEST_PATH_IMAGE037
(ii) a m is the number of samples of m = a × b;
calculating initial weights of predicted voltages of the LSTM-based battery voltage prediction model and the battery ECM model by the following formulas
Figure 978081DEST_PATH_IMAGE038
And
Figure 747454DEST_PATH_IMAGE039
：
Figure 49123DEST_PATH_IMAGE040
obtaining the weight of the first prediction voltage and the second prediction voltage by the following formula
Figure 396315DEST_PATH_IMAGE028
And
Figure 986696DEST_PATH_IMAGE029
：
Figure 559760DEST_PATH_IMAGE041
Figure 981514DEST_PATH_IMAGE042
。
in addition, the weights of the first predicted voltage and the second predicted voltage are calculated
Figure 434361DEST_PATH_IMAGE028
And
Figure 246459DEST_PATH_IMAGE029
according to the end of online update data
Figure 685531DEST_PATH_IMAGE034
The calculation of each time step is carried out,
Figure 165054DEST_PATH_IMAGE034
the equivalent value of 10, 15 and 20 can be taken according to actual needs.
To obtain
Figure 585539DEST_PATH_IMAGE028
And
Figure 619355DEST_PATH_IMAGE029
thereafter, the normal voltage is predicted
Figure 534221DEST_PATH_IMAGE024
Calculated by the following formula:
Figure 930567DEST_PATH_IMAGE025
in the formula (I), the compound is shown in the specification,
Figure 521954DEST_PATH_IMAGE026
is the first predicted voltage to be applied to the first transistor,
Figure 574224DEST_PATH_IMAGE027
is the second predicted voltage.
S4: and comparing the predicted normal voltage with the actual voltage of the battery to realize the battery fault diagnosis.
Specifically, in order to implement fault diagnosis in actual energy storage power station operation, an LSTM-based battery voltage prediction model is trained using test battery pack data, and different fault levels and their respective voltage thresholds are set to perform field fault diagnosis. And calculating the difference value between the actual voltage and the predicted normal voltage of the battery, judging whether the voltage of the battery belongs to a normal state, an overvoltage state or an undervoltage state according to the comparison result of the difference value and a preset fault level voltage threshold value, and if the voltage of the battery belongs to the overvoltage state or the undervoltage state, sending a warning of a corresponding level according to the comparison result. And positioning the potential fault and thermal runaway battery units according to the alarm level, repeating the steps on new data during the operation of the energy storage power station, and supplementing the prediction model in real time according to input data of the operation of the energy storage power station.
Since sensor and prediction errors may affect the setting of the threshold, statistical methods calculate the maximum prediction error of the test battery, which is used as the threshold for the primary warning. It is worth mentioning that the proposed fault diagnosis method is also applicable to other battery systems based on appropriate threshold settings. The voltage abnormality is divided into an overvoltage and an undervoltage, and in this embodiment, each case is divided into three levels. A primary warning is a relatively safe state, but requires timely intervention to avoid potential failure. A second level of warning means that the associated battery is in a dangerous state and needs to be carefully checked. When triggering a third level of warning, the relevant technician is required to be particularly addressed and contacted. First and second level warnings are used to assess the risk of a potential failure, while a third level warning is used for early warning of a thermal runaway event. In fig. 1, the failure prediction strategy corresponds to the determination of the failure type, and the risk assessment strategy corresponds to the determinationThe alarm level is cut off, and the battery failure frequency refers to the frequency when the battery fails, so that reference is provided for subsequent improvement; and positioning the fault battery unit according to the fault diagnosis result. As shown in table 1, the alarm levels and threshold conditions for fault diagnosis under two application conditions are provided for this embodiment. In the table, the number of the first and second,
Figure 89519DEST_PATH_IMAGE017
which represents the actual voltage of a certain battery cell,
Figure 278055DEST_PATH_IMAGE046
indicating the predicted normal voltage of a certain cell, and the "/" indicates "or", and the threshold values before and after the "or" indicates "or", correspond to two application conditions.
Figure 56655DEST_PATH_IMAGE048
Example 2
The embodiment provides an energy storage power station battery fault diagnosis device, includes:
the data acquisition module is used for acquiring battery historical data of the energy storage power station;
the voltage prediction module is used for respectively inputting battery historical data into a battery voltage prediction model and a battery ECM model based on LSTM to respectively obtain a first predicted voltage and a second predicted voltage;
the coupling module is used for calculating the predicted normal voltage by using a self-adaptive boosting method according to the first predicted voltage and the second predicted voltage;
and the fault diagnosis module is used for comparing the predicted normal voltage with the actual voltage of the battery to realize battery fault diagnosis.
It should be understood that the functional unit modules in the embodiments may be integrated into one processing unit, or each unit module may exist alone physically, or two or more unit modules are integrated into one unit module, and may be implemented in the form of hardware or software.
Example 3
The present embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the energy storage power station battery fault diagnosis method according to embodiment 1.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It is understood that the same or similar parts in the above embodiments may be mutually referred to, and the same or similar parts in other embodiments may be referred to for the content which is not described in detail in some embodiments.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps in the process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
Although embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are exemplary and not to be construed as limiting the present invention, and that changes, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.

Claims (8)

1. A method for diagnosing battery faults of an energy storage power station is characterized by comprising the following steps:
acquiring battery historical data of an energy storage power station;
respectively inputting battery historical data into a battery voltage prediction model and a battery ECM model based on LSTM to respectively obtain a first prediction voltage and a second prediction voltage;
calculating a predicted normal voltage by using an improved self-adaptive boosting method according to the first predicted voltage and the second predicted voltage;
comparing the predicted normal voltage with the actual voltage of the battery to realize battery fault diagnosis;
battery voltage prediction model output based on LSTMInto a matrix
Figure 395934DEST_PATH_IMAGE001
Is represented as follows:
Figure 42816DEST_PATH_IMAGE002
wherein the content of the first and second substances,
Figure 351438DEST_PATH_IMAGE003
is the number of input features;
Figure 446081DEST_PATH_IMAGE004
is the time step;
Figure 207363DEST_PATH_IMAGE005
is a step of time
Figure 657936DEST_PATH_IMAGE006
A battery unit
Figure 617802DEST_PATH_IMAGE007
The voltage of (a) is set to be,
Figure 738205DEST_PATH_IMAGE008
；
Figure 376996DEST_PATH_IMAGE009
is a step of time
Figure 241047DEST_PATH_IMAGE006
To the first
Figure 383316DEST_PATH_IMAGE010
An input feature;
the output matrix B of the LSTM-based battery voltage prediction model is represented as follows:
Figure 940199DEST_PATH_IMAGE011
wherein, the first and the second end of the pipe are connected with each other,
Figure 535128DEST_PATH_IMAGE012
is a step of time
Figure 937291DEST_PATH_IMAGE013
A battery unit
Figure 199645DEST_PATH_IMAGE007
The predicted voltage of (2).
2. The energy storage power station battery fault diagnosis method of claim 1, characterized in that battery voltage is predicted in a battery ECM model by the following formula:
Figure 927429DEST_PATH_IMAGE014
in the formula (I), the compound is shown in the specification,
Figure 9655DEST_PATH_IMAGE015
is a time step
Figure 277825DEST_PATH_IMAGE016
A battery unit
Figure 270052DEST_PATH_IMAGE007
The predicted voltage of (a);
Figure 27792DEST_PATH_IMAGE017
is that
Figure 738259DEST_PATH_IMAGE018
The data matrix of (2), wherein,
Figure 875367DEST_PATH_IMAGE019
is the voltage of the battery pack at a time step t,
Figure 722100DEST_PATH_IMAGE020
is the current of the battery at time step t;
Figure 650742DEST_PATH_IMAGE021
is a parameter matrix in which
Figure 582926DEST_PATH_IMAGE022
In which
Figure 520795DEST_PATH_IMAGE023
Is a battery cell
Figure 487614DEST_PATH_IMAGE007
OCV at time step t;
Figure 587157DEST_PATH_IMAGE024
、
Figure 272216DEST_PATH_IMAGE025
are ECM model parameters.
3. The energy storage power station battery fault diagnosis method according to any one of claims 1 to 2, characterized in that the calculation of the predicted normal voltage by using the improved adaptive boosting method according to the first predicted voltage and the second predicted voltage comprises:
predicting normal voltage
Figure 748197DEST_PATH_IMAGE026
Calculated by the following formula:
Figure 569523DEST_PATH_IMAGE027
in the formula (I), the compound is shown in the specification,
Figure 839967DEST_PATH_IMAGE028
is the first one of the predicted voltages to be,
Figure 746743DEST_PATH_IMAGE029
is the second predicted voltage to be the second predicted voltage,
Figure 26415DEST_PATH_IMAGE030
and
Figure 498984DEST_PATH_IMAGE031
respectively, the weights of the first predicted voltage and the second predicted voltage.
4. The energy storage power station battery fault diagnostic method of claim 3, wherein the weights of the first predicted voltage and the second predicted voltage
Figure 81275DEST_PATH_IMAGE030
And
Figure 599981DEST_PATH_IMAGE031
obtained by the following method:
using the LSTM-based battery voltage prediction model and the battery ECM model to predict the voltage, the following voltage prediction matrix is obtained
Figure 293131DEST_PATH_IMAGE032
：
Figure 210753DEST_PATH_IMAGE033
In the formula (I), the compound is shown in the specification,
Figure 963945DEST_PATH_IMAGE034
the class of the model is represented by,
Figure 438789DEST_PATH_IMAGE034
is LSTM, corresponding to the LSTM-based cell voltage prediction model,
Figure 935629DEST_PATH_IMAGE034
when the battery is the ECM, the battery corresponds to a battery ECM model;
Figure 710687DEST_PATH_IMAGE035
is a battery cell
Figure 634781DEST_PATH_IMAGE007
Step of time
Figure 128079DEST_PATH_IMAGE006
The predicted voltage of (2);
Figure 163031DEST_PATH_IMAGE036
the total step length is the corresponding prediction step length in the voltage prediction matrix, and b is the total number of the battery units;
calculating a regression error rate of a voltage prediction matrix of the LSTM-based battery voltage prediction model and the battery ECM model predicted voltages by the following formula
Figure 792596DEST_PATH_IMAGE037
：
Figure 684328DEST_PATH_IMAGE038
In the formula (I), the compound is shown in the specification,
Figure 274710DEST_PATH_IMAGE039
is a battery cell
Figure 237986DEST_PATH_IMAGE007
Step of time of
Figure 863003DEST_PATH_IMAGE006
The actual voltage at (c); e is the maximum error of the optical signals,
Figure 253533DEST_PATH_IMAGE040
(ii) a m is the number of samples of m = a × b;
calculating initial weights of predicted voltages of the LSTM-based battery voltage prediction model and the battery ECM model by the following formulas
Figure 331210DEST_PATH_IMAGE041
And
Figure 832599DEST_PATH_IMAGE042
：
obtaining the weight of the first prediction voltage and the second prediction voltage by the following formula
Figure 312122DEST_PATH_IMAGE030
And
Figure 876483DEST_PATH_IMAGE031
：
Figure 441456DEST_PATH_IMAGE043
Figure 480956DEST_PATH_IMAGE044
。
5. the energy storage power station battery fault diagnosis method of claim 4, characterized in that weights for the first predicted voltage and the second predicted voltage are calculated
Figure 80565DEST_PATH_IMAGE030
And
Figure 812898DEST_PATH_IMAGE031
according to the end of online update data
Figure 599588DEST_PATH_IMAGE036
And (4) calculating each time step.
6. The energy storage power station battery fault diagnosis method of claim 1, wherein comparing the predicted normal voltage with the actual battery voltage to perform battery fault diagnosis comprises:
and calculating the difference value between the actual voltage and the predicted normal voltage of the battery, judging whether the voltage of the battery is in a normal, overvoltage or undervoltage state according to the comparison result of the difference value and a preset fault level voltage threshold value, and if the voltage of the battery is in the overvoltage or undervoltage state, sending out a warning of a corresponding level according to the comparison result.
7. An energy storage power station battery fault diagnosis device, characterized by comprising:
the data acquisition module is used for acquiring battery historical data of the energy storage power station;
the voltage prediction module is used for respectively inputting battery historical data into a battery voltage prediction model and a battery ECM model based on LSTM to respectively obtain a first prediction voltage and a second prediction voltage;
the coupling module is used for calculating the predicted normal voltage by using an improved self-adaptive boosting method according to the first predicted voltage and the second predicted voltage;
the fault diagnosis module is used for comparing the predicted normal voltage with the actual voltage of the battery to realize battery fault diagnosis;
input matrix of LSTM-based battery voltage prediction model
Figure 708358DEST_PATH_IMAGE001
Is represented as follows:
Figure 959211DEST_PATH_IMAGE002
wherein the content of the first and second substances,
Figure 737811DEST_PATH_IMAGE003
is the number of input features;
Figure 136432DEST_PATH_IMAGE004
is the time step;
Figure 658680DEST_PATH_IMAGE005
is a step of time
Figure 91935DEST_PATH_IMAGE006
A battery unit
Figure 775857DEST_PATH_IMAGE007
The voltage of (a) is set to be,
Figure 927353DEST_PATH_IMAGE008
；
Figure 987713DEST_PATH_IMAGE009
is a step of time
Figure 541054DEST_PATH_IMAGE006
To be at
Figure 661457DEST_PATH_IMAGE010
An input feature;
the output matrix B of the LSTM-based battery voltage prediction model is represented as follows:
Figure 706773DEST_PATH_IMAGE011
wherein the content of the first and second substances,
Figure 895790DEST_PATH_IMAGE012
is a step of time
Figure 975742DEST_PATH_IMAGE013
A battery unit
Figure 391680DEST_PATH_IMAGE007
The predicted voltage of (2).
8. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method of energy storage power station battery fault diagnosis according to any one of claims 1 to 6.
CN202210917698.8A 2022-08-01 2022-08-01 Energy storage power station battery fault diagnosis method and device and storage medium Pending CN114994543A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210917698.8A CN114994543A (en) 2022-08-01 2022-08-01 Energy storage power station battery fault diagnosis method and device and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210917698.8A CN114994543A (en) 2022-08-01 2022-08-01 Energy storage power station battery fault diagnosis method and device and storage medium

Publications (1)

Publication Number Publication Date
CN114994543A true CN114994543A (en) 2022-09-02

Family

ID=83021627

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210917698.8A Pending CN114994543A (en) 2022-08-01 2022-08-01 Energy storage power station battery fault diagnosis method and device and storage medium

Country Status (1)

Country Link
CN (1) CN114994543A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115825756A (en) * 2023-02-16 2023-03-21 中国华能集团清洁能源技术研究院有限公司 Distributed energy storage power station fault multi-stage diagnosis method and system
CN115877222A (en) * 2023-02-14 2023-03-31 国网浙江省电力有限公司宁波供电公司 Energy storage power station fault detection method and device, medium and energy storage power station
CN115963408A (en) * 2022-12-19 2023-04-14 北京双登慧峰聚能科技有限公司 Fault early warning system and method for single battery of energy storage power station

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170242077A1 (en) * 2016-02-19 2017-08-24 Johnson Controls Technology Company Systems and methods for real-time estimation of capacity of a rechargeable battery
CN108279383A (en) * 2017-11-30 2018-07-13 深圳市科列技术股份有限公司 battery life predicting method, battery data server and battery data processing system
CN110824364A (en) * 2019-10-24 2020-02-21 重庆邮电大学 Lithium battery SOH estimation and RUL prediction method based on AST-LSTM neural network
WO2020155233A1 (en) * 2019-01-28 2020-08-06 东北大学 Method for diagnosing external short circuit fault of lithium-ion battery pack on basis of two-stage model prediction
CN111624494A (en) * 2020-04-20 2020-09-04 北京航空航天大学 Battery analysis method and system based on electrochemical parameters
WO2020191800A1 (en) * 2019-03-27 2020-10-01 东北大学 Method for predicting remaining service life of lithium-ion battery employing wde-optimized lstm network
US20210088591A1 (en) * 2019-09-19 2021-03-25 Samsung Electronics Co., Ltd. Method and system for battery-management in devices
CN113687251A (en) * 2021-08-23 2021-11-23 重庆邮电大学 Dual-model-based lithium ion battery pack voltage abnormity fault diagnosis method
CN114036647A (en) * 2021-11-30 2022-02-11 武汉理工大学 Power battery safety risk assessment method based on real vehicle data
US20220179009A1 (en) * 2020-12-03 2022-06-09 Robert Bosch Gmbh Method and Apparatus for Operating a System for Providing Predicted States of Health of Electrical Energy Stores for a Device Using Machine Learning Methods

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170242077A1 (en) * 2016-02-19 2017-08-24 Johnson Controls Technology Company Systems and methods for real-time estimation of capacity of a rechargeable battery
CN108279383A (en) * 2017-11-30 2018-07-13 深圳市科列技术股份有限公司 battery life predicting method, battery data server and battery data processing system
WO2020155233A1 (en) * 2019-01-28 2020-08-06 东北大学 Method for diagnosing external short circuit fault of lithium-ion battery pack on basis of two-stage model prediction
WO2020191800A1 (en) * 2019-03-27 2020-10-01 东北大学 Method for predicting remaining service life of lithium-ion battery employing wde-optimized lstm network
US20210088591A1 (en) * 2019-09-19 2021-03-25 Samsung Electronics Co., Ltd. Method and system for battery-management in devices
CN110824364A (en) * 2019-10-24 2020-02-21 重庆邮电大学 Lithium battery SOH estimation and RUL prediction method based on AST-LSTM neural network
CN111624494A (en) * 2020-04-20 2020-09-04 北京航空航天大学 Battery analysis method and system based on electrochemical parameters
US20220179009A1 (en) * 2020-12-03 2022-06-09 Robert Bosch Gmbh Method and Apparatus for Operating a System for Providing Predicted States of Health of Electrical Energy Stores for a Device Using Machine Learning Methods
CN113687251A (en) * 2021-08-23 2021-11-23 重庆邮电大学 Dual-model-based lithium ion battery pack voltage abnormity fault diagnosis method
CN114036647A (en) * 2021-11-30 2022-02-11 武汉理工大学 Power battery safety risk assessment method based on real vehicle data

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
DA LI等: "Battery Fault Diagnosis for Electric Vehicles Based on Voltage Abnormality by Combining the Long Short-Term Memory Neural Network and the Equivalent Circuit Model", 《IEEE TRANSACTIONS ON POWER ELECTRONICS》 *
周仁等: "基于实车数据和BP-AdaBoost算法的电动汽车动力电池健康状态估计", 《科学技术与工程》 *
费陈等: "基于XGBoost算法的锂离子电池健康状态估算", 《浙江电力》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115963408A (en) * 2022-12-19 2023-04-14 北京双登慧峰聚能科技有限公司 Fault early warning system and method for single battery of energy storage power station
CN115963408B (en) * 2022-12-19 2024-04-16 北京双登慧峰聚能科技有限公司 Energy storage power station single battery fault early warning system and method
CN115877222A (en) * 2023-02-14 2023-03-31 国网浙江省电力有限公司宁波供电公司 Energy storage power station fault detection method and device, medium and energy storage power station
CN115825756A (en) * 2023-02-16 2023-03-21 中国华能集团清洁能源技术研究院有限公司 Distributed energy storage power station fault multi-stage diagnosis method and system

Similar Documents

Publication Publication Date Title
CN114994543A (en) Energy storage power station battery fault diagnosis method and device and storage medium
CN107329088B (en) Battery state of health diagnostic device and method
CN108205114B (en) Method and system for predicting service life of battery
US10613153B2 (en) Method and apparatus for defect pre-warning of power device
JP4922597B2 (en) Diagnostic method and diagnostic apparatus for fuel cell system
CN109752646B (en) Vehicle-mounted direct current contactor evaluation and early warning method, device and equipment
CN112098915B (en) Method for evaluating secondary errors of multiple voltage transformers under double-bus segmented wiring
CN102880170A (en) System failure early warning method based on baseline model and Bayesian factor
CN114464906A (en) Power battery early warning method and device
CN111796233A (en) Method for evaluating secondary errors of multiple voltage transformers in double-bus connection mode
DE102021006561A1 (en) Big data for error detection in battery systems
CN112529464A (en) Health degree evaluation method and device of battery pack and related product
CN112668415A (en) Aircraft engine fault prediction method
KR102110319B1 (en) System for generating learning data
CN103235206A (en) Transformer fault diagnosis method
CN116893366A (en) Method and device for operating a system to detect anomalies in an electrical energy store of a device
CN115563868A (en) Fault diagnosis method and device for oil circuit system of aviation alternating-current generator
CN114879054A (en) Battery safety monitoring method and device, electronic equipment and storage medium
CN115792634A (en) Battery monomer voltage sampling fault identification method based on cloud online data
CN105138755A (en) Design method of fault response generator for testbed of liquid propellant rocket engine
WO2022162060A1 (en) Big data for fault identification in battery systems
Wu et al. Research on short-circuit fault-diagnosis strategy of lithium-ion battery in an energy-storage system based on voltage cosine similarity
CN113312758B (en) Health state evaluation method and device for wind generating set
CN109739210A (en) The appraisal procedure and device of part of appliance health status
CN109492913A (en) A kind of modularization Risk Forecast Method of distribution terminal, device and can 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
CB02 Change of applicant information
CB02 Change of applicant information

Address after: Room 401A19, Entrepreneurship Building, Hunan University Science Park, No. 186 Guyuan Road, Changsha Hi tech Development Zone, Changsha, Hunan 410005

Applicant after: HUNAN HUADA ELECTRICIAN HI-TECH Co.,Ltd.

Address before: Room 908, Building 8, Sunshine West Group, California, No. 38, Guyuan Road, Changsha High-tech Development Zone, Changsha, Hunan Province, 410005

Applicant before: HUNAN HUADA ELECTRICIAN HI-TECH Co.,Ltd.

RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20220902