CN115932592A - Electric vehicle battery fault diagnosis method based on digital twin model - Google Patents
Electric vehicle battery fault diagnosis method based on digital twin model Download PDFInfo
- Publication number
- CN115932592A CN115932592A CN202211153359.3A CN202211153359A CN115932592A CN 115932592 A CN115932592 A CN 115932592A CN 202211153359 A CN202211153359 A CN 202211153359A CN 115932592 A CN115932592 A CN 115932592A
- Authority
- CN
- China
- Prior art keywords
- data
- model
- twin
- electric vehicle
- vehicle battery
- 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
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/60—Other road transportation technologies with climate change mitigation effect
- Y02T10/70—Energy storage systems for electromobility, e.g. batteries
Landscapes
- Secondary Cells (AREA)
Abstract
A battery fault diagnosis method of an electric automobile based on a digital twin model belongs to the technical field of battery management. The method comprises the following steps: firstly, determining electric vehicle battery parameters, and then obtaining electric vehicle battery fault data and corresponding fault categories through an electric vehicle battery simulation model; secondly, constructing a data twin mapping model which consists of a physical entity, a virtual entity, twin data and information, and forming a set of closed-loop mapping mechanism with the physical entity, wherein the key lies in the construction of the electric automobile battery pack twin model; then fusing the data with the twin model; for the electric vehicle battery fault diagnosis, a data-driven fault diagnosis method is provided by combining the time sequence characteristics of data and the advantages of deep learning. The method is mainly based on a bidirectional long-short-time memory network, and a deep bidirectional long-short-time memory network is constructed to realize fault diagnosis and separation of the electric vehicle battery.
Description
Technical Field
The invention relates to the technical field of battery management, in particular to a battery fault diagnosis method for an electric automobile based on a digital twin model.
Background
With the increasing demand of the operation and maintenance of the equipment on digitization and intellectualization, the digital twin technology and the machine learning algorithm in the cyber-physical system become a hotspot of academic and industrial research.
At present, fault diagnosis methods mainly include three types, namely a fault diagnosis method based on an analysis model, a fault diagnosis method based on empirical knowledge, and a fault diagnosis method based on data driving. With the rapid development of sensing technology and computer technology, fault diagnosis methods based on data driving are becoming the focus of research of experts and scholars. The data-driven fault diagnosis method mainly comprises a signal processing-based method and an artificial intelligence-based method.
The digital twin is rapidly developed as a key technology of information physical fusion, is applied to the Apollo project in the American space engineering at the earliest and lays a solid foundation for later development. With the development of new generation network information technology, the concept of digital twin technology is continuously expanded, from the earliest health maintenance and guarantee to the whole process of design, manufacture and operation and maintenance.
Although a certain research result is achieved by the intelligent fault diagnosis method based on digital twins and machine learning, some defects still exist: most data-driven fault diagnosis methods are based on off-line data development research, and lack certain real-time performance, cooperativity and interactivity. The information is delayed due to poor real-time performance, so that the information lacks timeliness; poor cooperativity and interactivity results in failure lacking real-time visualization. Therefore, real-time mapping of the physical space and the information space in the cyber-physical system, failure prediction, and failure information feedback cannot be achieved.
Disclosure of Invention
In view of the above situation, in order to solve the problems in the above technologies, the present invention provides a method for diagnosing a battery fault of an electric vehicle based on a digital twin model, which is applied to intelligent fault diagnosis of a battery pack of an electric vehicle, so that a physical entity in a physical space and the digital twin model in an information space are synchronized, and real-time mapping, fault prediction and fault information feedback of the physical space and the information space are realized.
A battery fault diagnosis method of an electric vehicle based on a digital twin model integrates a deep learning algorithm and a model simulation technology. Firstly, constructing a digital twin model integrating model drive and data drive, connecting data transmitted from a physical space to a digital space based on an initial model, and constructing an operation and maintenance digital twin model mapped in real time; fusing the initial model and the multi-data acquired in the physical space to enable the model to have behavior characteristics and form an operation and maintenance digital twin model of the equipment; real-time data actually acquired by equipment is combined with a performance model to form a self-adaptive model which changes along with the operating environment and the equipment performance, so that the aims of monitoring the local state and the overall performance of the equipment are fulfilled; introducing a fault mode containing historical maintenance data into a physical model and a performance model, and constructing a fault model for fault diagnosis and maintenance; combining historical data with a performance model, establishing a performance prediction model under the drive of the data, and evaluating the performance and the service life prediction of equipment; introducing a local linear model into the operating environment to form a control optimization model and provide a strategy for equipment optimization; the models jointly describe a digital twin model containing multiple behavior attributes, and monitoring diagnosis and performance optimization of the equipment are achieved.
Specifically, the method comprises the following steps:
s1, collecting data: the method comprises the steps of collecting overcharge fault data of an electric automobile battery, collecting overdischarge fault data of the electric automobile battery and collecting aging fault data of the electric automobile battery;
in order to ensure that the twin model can be subjected to real-time iterative optimization, the standards of data communication and conversion need to be set to collect multi-source heterogeneous data, and unified conversion and encapsulation of data among different communication interfaces are realized, so that unified standard processing can be performed on the data, and integration and fusion of the multi-source heterogeneous data are realized.
S2, establishing a model and fusing the model: constructing a data twin mapping model according to a physical entity, wherein the data twin mapping model consists of the physical entity, a virtual entity, twin data and information, and a closed-loop mapping mechanism is formed between the data twin mapping model and the physical entity; the models are fused structurally and functionally by establishing the relation of the models of each layer;
and establishing a real mapping from the electric automobile battery to the twin, and analyzing the correlation and mapping between models in multiple dimensions. The omnibearing modeling of the electric automobile battery follows multiple dimensions such as geometry, physics, behaviors, rules and the like, the models are fused structurally and functionally by establishing the relation of the models of all layers, and the generated models and virtual simulation are visually displayed in a three-dimensional form.
S3, data fusion: the method comprises the steps of denoising and modeling real-time data of an electric vehicle battery entity, classifying and analyzing results, and finally iterating, evolving and fusing equipment real-time data and model data to realize data fusion of a physical model and a virtual model, so that the virtual entity can truly reflect the running state of all elements of the physical entity in the whole working process;
s4, failure prediction: and after the twin model is built, the physical entity and the virtual entity are synchronously mapped, and are interacted in real time under the interaction of twin data, so that a basis is provided for the fault prediction of the electric vehicle battery.
According to the electric vehicle battery fault diagnosis method based on the digital twin model, when data are collected in the step S1, the total voltage, the total current, the charge state, the highest temperature of the battery pack, the lowest temperature of the battery pack, the sampling time, the vehicle speed parameter and the control instruction of the electric vehicle battery are obtained in real time from the data collection and monitoring system and the state monitoring system of the electric vehicle battery pack through a TCP/IP (transmission control protocol)/UDP (user datagram protocol) and/or UDP (user datagram protocol) communication mode, and the real-time perception of the digital twin system on the running state of the electric vehicle battery pack is achieved.
And in the step S1, key operation parameters are obtained when the battery pack of the electric automobile operates. On the premise of ensuring communication safety, the digital twin system can obtain related parameters and control instructions such as total voltage, total current, state of charge, maximum battery temperature, minimum battery temperature, sampling time, vehicle speed and the like of the electric vehicle battery pack in real time from a supervisory control and data acquisition (SCADA) system and a state monitoring system (CMS) of the electric vehicle battery pack by utilizing communication modes such as TCP/IP (transmission control protocol/internet protocol), UDP (user datagram protocol) and the like, so that the digital twin system can realize real-time perception of the running state of the electric vehicle battery pack.
According to the method for diagnosing the battery fault of the electric automobile based on the digital twin model, the digital twin mapping model is constructed according to the following formula:
Sap DT ＝(E p ，N DT ，D DT ，I))
among them, sap DT Represents a numerical twin map, E p Representing the battery entity of an electric vehicle, N DT Representing twin models of batteries of electric vehicles, D DT Represents twin data of an electric vehicle battery pack, I represents E p 、N DT 、 D DT To each other. Electric vehicle battery (E) p ) An entity is a precondition for implementing a twin mapping model, thereby creating a real to imaginary mapping mechanism.
According to the electric vehicle battery fault diagnosis method based on the digital twin model, the electric vehicle battery pack twin model is constructed according to the following formula:
N DT ＝(N G ，N A ，N E )
wherein N is DT Representing electricityTwin model of a battery pack for a motor vehicle, N G Representing a geometric model, N A Representative analytical model, N E Is an evolution model.
According to the electric vehicle battery fault diagnosis method based on the digital twin model, the established digital twin mapping models are completely combined and constructed, and perfect fusion of physical space, information space and twin data is achieved.
According to the electric vehicle battery fault diagnosis method based on the digital twin model, the fault prediction in the step S4 is that a deep bidirectional long-term and short-term memory network is constructed on the basis of the bidirectional long-term and short-term memory network, so that the fault diagnosis and the separation of the electric vehicle battery are realized.
After the technology provided by the invention is adopted, the fault diagnosis method has the advantages that firstly, a fault diagnosis model based on deep learning (a deep bidirectional long-term and short-term memory network) is constructed for prediction, and a fault result verification method based on model simulation is used for verifying a prediction result. Secondly, analyzing the consistency of the simulation data and the actual data, taking the consistency as a judgment condition for judging whether the model is self-adjusted, and constructing a feedback mechanism to realize the self-adjustment of the model. And then, determining whether the digital twin model is corrected or not by taking the consistency of the simulation data and the actual data as a judgment condition, and feeding back the evolution process of the battery pack of the electric automobile through continuous correction of the twin model, so that the physical entity in the physical space and the digital twin model in the information space are synchronous. And finally, verifying the digital twin-driven fault diagnosis system through a fault diagnosis example of the electric automobile battery pack. The real-time mapping, the fault prediction and the fault information feedback of the physical space and the information space are realized.
Drawings
FIG. 1 is a diagram illustrating a mapping mechanism of digital twinning in an embodiment of the present invention;
FIG. 2 is a block diagram of a battery fault diagnosis method for an electric vehicle based on a DB-LSTM-RNN neural network in an embodiment of the invention;
fig. 3 is a flowchart of intelligent fault diagnosis according to an embodiment of the present invention.
Detailed Description
Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings. The following description with reference to the accompanying drawings is provided to assist in understanding the exemplary embodiments of the invention as defined by the claims and their equivalents. It includes various specific details to assist understanding, but they are to be construed as merely illustrative. Accordingly, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present invention. Also, in order to make the description clearer and simpler, a detailed description of functions and configurations well known in the art will be omitted.
The invention provides a digital twinning model-based electric vehicle battery fault diagnosis method, which is realized based on a digital twinning technology and specifically comprises the following steps of:
s1: and (6) collecting data. In order to ensure that the twin model can be subjected to real-time iterative optimization, the data communication and conversion standard is set to collect multi-source heterogeneous data, and uniform conversion and encapsulation of data among different communication interfaces are realized, so that uniform standard processing can be performed on the data, and integration and fusion of the multi-source heterogeneous data are realized.
S2: and establishing and fusing the model. And establishing real mapping from the electric automobile battery to the twin, and analyzing the correlation and mapping between models in multiple dimensions. The omnibearing modeling of the electric automobile battery follows multiple dimensions such as geometry, physics, behaviors, rules and the like, the models are fused structurally and functionally by establishing the relation of the models of all layers, and the generated models and virtual simulation are visually displayed in a three-dimensional form.
S3: and (4) fusing data. The method comprises the steps of denoising and modeling real-time data of an electric vehicle battery entity, classifying and analyzing results, and finally iterating, evolving and fusing the real-time data of equipment and model data to realize data fusion of a physical model and a virtual model, so that the virtual entity can truly reflect the running state of all elements of the physical entity in the whole working process.
S4: and (4) predicting the fault. And after the twin model is built, the physical entity and the virtual entity are synchronously mapped, and under the interaction of twin data, the real-time interaction between the physical entity and the virtual entity provides a basis for the fault prediction of the electric vehicle battery.
The step S1 mainly includes three types of fault data acquisition, which are respectively: the method comprises the steps of acquiring overcharge fault data of the electric automobile battery, acquiring overdischarge fault data of the electric automobile battery and acquiring aging fault data of the electric automobile battery.
In the embodiment, in order to obtain the overcharge fault data of the battery of the electric automobile, the overcharge test steps are as follows:
(1) The battery was subjected to normal charge-discharge cycles 5 times, and the average of 5 discharge capacities was used as an initial capacity.
(2) Normal charge and discharge: and (3) constant current-constant voltage charging is adopted, the rated current is used for constant current charging until the voltage reaches the rated voltage, constant voltage charging is continued until the current is reduced to the set current value, the constant current charging is carried out for 2 hours, the constant current discharging is carried out until the set voltage value is reached, and the voltage, the current, the temperature and the capacity data in the charging and discharging process are recorded.
(3) Charging the battery according to different degrees of overcharge voltage, wherein the charging adopts a constant current-constant voltage mode, constant current charging is carried out by using rated current until the voltage reaches different overcharge voltages, constant voltage charging is continued until the current is reduced to a set current value, standing for 2 hours is carried out, so that constant current discharging is carried out to the set voltage value, and voltage, current, temperature and capacity data in the charging and discharging process are recorded.
(4) Repeating the step 3 for 2 times, and then performing the step 2 for 1 time to obtain discharge capacity calibration.
(5) Repeating the step 4 until the battery capacity is reduced to be below 80% of the average capacity measured in the step 1 or the battery is damaged and cannot be used.
Wherein, the voltage and the current are data collected once in 2 seconds, data collected once in 1 second at the temperature, and the charge capacity and the discharge capacity are collected once in one charge-discharge cycle. Meanwhile, in the same environment, the normal charge and discharge experiment is carried out on the other group of lithium batteries according to the step 2, and the experiment is used as a comparison experiment. The data of the battery during overcharge and the data of the battery during normal charge and discharge were collected according to the above experimental procedure.
In order to obtain the overdischarge fault data of the battery of the electric automobile, the overdischarge experimental steps are as follows:
(1) The battery was normally charged and discharged 5 times, and the average of 5 discharge capacities was used as an initial capacity.
(2) Normal charge and discharge: and (3) constant current-constant voltage charging is adopted, rated current constant current charging is adopted until the voltage reaches rated voltage, constant voltage charging is continued until the current is reduced to a set current value, the constant current charging is carried out for 2 hours, constant current discharging is carried out by using a certain current until the set voltage value is reached, and voltage, current, temperature and capacity data in the charging and discharging period are recorded.
(3) Discharging the battery according to overdischarge operations of different degrees to reach a set overdischarge value: (1) discharging the battery with a constant current of a certain current until the voltage value is lower than the voltage value set in the step (2); (2) the battery is discharged for 3 hours by using a constant current with a certain current. And (3) constant current-constant voltage charging is adopted, the rated current is used for constant current charging until the voltage reaches the rated voltage, constant voltage charging is continued until the current is reduced to the set current value, the constant voltage charging is shelved for 2 hours, discharging operation is carried out, and voltage, current, temperature and capacity data in the charging and discharging process are recorded.
(4) Repeating the step 3 for 20 times, and then performing the step 2 for 1 time to obtain the discharge capacity calibration.
(5) Repeating the step 4 until the battery capacity is reduced to be below 80% of the average capacity measured in the step 1 or the battery is damaged and cannot be used.
The voltage and the current are data acquired once in 2 seconds, data acquired once at the temperature of 1 second, and the charge capacity and the discharge capacity are acquired once in one charge-discharge cycle, wherein the capacity mentioned in the text is the discharge capacity. Meanwhile, in the same environment, the normal charge and discharge experiment according to the step 2 is carried out on the other group of lithium batteries as a comparison experiment. And (3) collecting the data of the battery during the overdischarge of the battery according to the overdischarge experimental steps, wherein the data are called that the overdischarge is slight, and the data are called that the data are serious when the voltage is lower than the voltage set in the step (2), and the data are called that the data are serious after 3 hours.
In order to obtain the aging fault data of the battery of the electric automobile, the aging experiment of the battery of the electric automobile comprises the following steps:
(1) Under a certain environment, four lithium batteries (No. 5, no. 6, no. 7 and No. 8) are firstly charged at a constant current of a rated current until the voltage reaches the rated voltage, and then are charged at a constant voltage until the current is reduced to 20mA.
(2) The four batteries are discharged at a constant current until the voltages of the batteries 5, 6, 7 and 8 are respectively reduced to different voltage values.
(3) And (4) repeating the steps 1 and 2 until the end-of-life standard of the battery is reached, namely the discharge capacity of the battery is reduced to 70 percent of the initial capacity.
Wherein, the voltage and the current are data collected once in 2 seconds, data collected once in 1 second at the temperature, and the charge capacity and the discharge capacity are collected once in one charge-discharge cycle.
Referring to fig. 1, the step S2 primarily works to construct a data twin mapping model according to physical entities. The data twin mapping model mainly comprises a physical entity, a virtual entity, twin data and information, and a set of closed-loop mapping mechanism is formed between the data twin mapping model and the physical entity. The method mainly comprises the steps of constructing a data twin mapping model and constructing an electric automobile battery pack twin model.
Step S21: the construction formula of the digital twin mapping model is as follows:
Sap DT ＝(E p ，N DT ，D DT ，I)
among them, sap DT Representing a digital twin map, E p Representing the battery entity of an electric vehicle, N DT Representing twin models of batteries of electric vehicles, D DT Represents twin data of an electric vehicle battery pack, I represents E p 、N DT 、 D DT Are connected with each other. Electric vehicle battery (E) p ) An entity is a precondition for implementing a twin mapping model, thereby creating a real to imaginary mapping mechanism.
Step S22: twin model of electric automobile battery (N) DT ) The construction formula is as follows: n is a radical of DT ＝(N G ，N A ，N E )
Wherein N is DT Representing twin models of batteries of electric vehicles, N G Representing a geometric model, N A Representative analytical model, N E Is an evolution model.
The geometric model is constructed according to the following formula:
N G ＝(p g ，r g )
wherein p is g Representing the geometric parameter of the battery of the electric vehicle, r g Representing the geometric relationship of the battery pack of the electric automobile.
N A ＝(n d ，n n )
Wherein n is d Representing a data-driven fault analysis model, n n Representing a model-driven fault analysis model.
The evolution model is constructed according to the following formula:
N E ＝(t f ，l f ，n f )
wherein, t f Indicates the type of failure,/ f Indicates the location of the fault, n f A fault model is represented.
Twin data (D) DT ) The method is the root of health management driven by data twin, and the twin data fuses multi-angle and multi-dimensional data, so that the health state of a physical entity is described more comprehensively. The twin data of the electric automobile battery pack mainly comprise related data, acquisition device parameters, external environment data, knowledge data and the like of the electric automobile battery pack.
The related data of the electric vehicle battery pack mainly comprises technical parameters, material attributes, operation data, fault data and the like, such as related parameters of total voltage, total current, state of charge, highest temperature of the battery pack, lowest temperature of the battery pack, sampling time, vehicle speed and the like. The sensor is the core equipment for data acquisition and is also the key for realizing data driving. Therefore, the parameters of the data acquisition device mainly comprise sensor parameter information, data acquisition frequency, layout information of measuring points and the like. The external environmental data mainly includes ambient temperature, humidity, pressure, and the like. The knowledge data mainly comprises a fault diagnosis model base, an expert knowledge rule base and the like.
And the data fusion in the step S3 is the connection between the physical entity, the twin data and the virtual model. The connection (I) mainly comprises the connection between an electric automobile battery pack entity and an electric automobile battery pack twin model, the connection between the electric automobile battery pack entity and twin data and the connection between the electric automobile battery pack twin model and twin data. And after calling the fault diagnosis algorithm in the twin data to predict, the electric automobile battery pack twin model transmits the prediction result to the twin data. And reading the fault information from the twin data by the battery pack entity of the electric automobile, performing edge calculation on the fault information, and then driving the physical entity to make corresponding adjustment by utilizing edge control. In the process, the fault is adjusted and processed according to the actual condition of the fault.
Referring to fig. 2, in the step S4, according to the development trend of the diagnosis process intelligence, the fault diagnosis method based on artificial intelligence in the data drive is a great trend, and therefore, the fault diagnosis method based on the deep bidirectional long-term and short-term memory neural network is mainly used in the present patent.
The bidirectional long-short time memory network (BLSTM) is a comprehensive product of the long-short time memory network (LSTM) and the Bidirectional Recurrent Neural Network (BRNN). The traditional recurrent neural network has the problems of gradient disappearance and gradient explosion in the training process, and the defects in the training process are overcome by using the network. The long and short time memory network is provided with a memory unit, and the memory unit consists of an input gate, a forgetting gate and an output gate. Such networks are therefore suitable for processing data having a time series character. The structure of the bidirectional long-time and short-time memory network model comprises a forward propagation mode and a backward propagation mode, so that forward and backward events can be correlated through forward and backward reasoning.
The input gate obtains new input from the outside and processes new data; a forgetting gate determines when a particular output result is selectively forgotten, thereby selecting an optimal time lag for the input sequence; the output gate computes all results and generates an output for the LSTM unit. The calculation formulas of the input gate, the forgetting gate and the output gate are as follows:
an input gate:
i t ＝σ(W i [h t-1 ，x t ]+b i )
forget the door:
f t ＝σ(W f [h t-1 ，x t ]+b f )
an output gate:
o t ＝σ(W o [h t-1 ，x t ]+b o )
h t ＝o t ×tanh(C t )
the traditional intelligent fault diagnosis method mainly utilizes a signal processing technology and a pattern recognition technology, firstly utilizes the signal processing technology to extract fault characteristics, and then uses the fault characteristics as the input of a shallow machine learning model to finally realize fault pattern recognition. When a large amount of high-dimensional data is faced, deep features of the signal cannot be excavated, and the fault diagnosis capability is reduced.
At present, a data-driven fault diagnosis method is provided by combining the time-sequence characteristics of data and the advantages of deep learning. A novel electric automobile battery fault diagnosis method based on a digital twin model is generated by referring to an intelligent fault diagnosis flow chart in FIG. 3. The method is mainly based on a bidirectional long-short-time memory network, a deep bidirectional long-short-time memory network is constructed, the defects of the traditional method are overcome, and the fault type of the equipment is identified in a self-adaptive mode. The input of the deep neural network is mainly a time domain signal, the hidden layer comprises three bidirectional long-time memory networks, a full-connection layer and a classifier, and the output is a predicted fault type.
Aiming at the problem of reliability of a prediction result of a data-driven deep learning model and the problem of real-time cooperativity between a physical entity and a virtual model, a fault result verification method based on model simulation is provided for verifying a data-driven fault diagnosis result, on one hand, the consistency of the results of the two is used as a condition for regulating the fault prediction model, a feedback mechanism is constructed, and the functions of model automatic regulation under the condition that the fault is known and model automatic learning under the condition that the fault is unknown are realized; on the other hand, the digital twin model is driven to update by the modified digital twin model. Through the mode, the intelligent diagnosis of the electric vehicle battery pack is realized.
According to the electric vehicle battery fault diagnosis method based on the digital twin model, the physical entity transmits twin data such as state data to the virtual system in real time, the running state of virtual entity equipment is synchronous with the physical entity, and new data such as fault prediction data, maintenance decision data and the like are continuously generated in the process. The newly generated real-time data of the physical entity equipment and the virtual entity is further fused with the existing twin data, and the service system evaluates the running state of the equipment according to the obtained fused data so as to quickly sense a fault event, accurately position the fault reason and provide a reasonable maintenance strategy.
Wherein, the three types of fault data acquisition that mainly include do respectively: the method comprises the steps of acquiring overcharge fault data of the electric automobile battery, acquiring overdischarge fault data of the electric automobile battery and acquiring aging fault data of the electric automobile battery.
According to the electric vehicle battery fault diagnosis method based on the digital twin model, firstly, a fault diagnosis model based on deep learning (a deep bidirectional long-and-short-term memory network) is constructed for prediction, and a fault result verification method based on model simulation is used for verifying a prediction result. Secondly, analyzing the consistency of the simulation data and the actual data, taking the consistency as a judgment condition for judging whether the model is self-adjusted, and constructing a feedback mechanism to realize the self-adjustment of the model. And then, determining whether the digital twin model is corrected or not by taking the consistency of the simulation data and the actual data as a judgment condition, and feeding back the evolution process of the battery pack of the electric automobile through continuous correction of the twin model, so that the physical entity in the physical space and the digital twin model in the information space are synchronized. And finally, verifying the digital twin-driven fault diagnosis system through the electric automobile battery pack fault diagnosis example. The real-time mapping, the fault prediction and the fault information feedback of the physical space and the information space are realized.
The present invention has been described in detail, and the principle and embodiments of the present invention are explained herein by using specific examples, which are only used to help understand the method and the core idea of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
From the above description of the embodiments, it will be apparent to those skilled in the art that the present invention may be practiced. Of course, the above listed cases are only examples, and the present invention is not limited thereto. It should be understood by those skilled in the art that other modifications or simplifications according to the technical solution of the present invention may be appropriately applied to the present invention and should be included in the scope of the present invention.
Claims (6)
1. A battery fault diagnosis method of an electric automobile based on a digital twin model is characterized by comprising the following steps:
s1, collecting data: the method comprises the steps of collecting overcharge fault data of an electric automobile battery, collecting overdischarge fault data of the electric automobile battery and collecting aging fault data of the electric automobile battery;
s2, establishing a model and fusing the model: constructing a data twin mapping model according to a physical entity, wherein the data twin mapping model consists of the physical entity, a virtual entity, twin data and information, and a closed-loop mapping mechanism is formed between the data twin mapping model and the physical entity; the models are fused structurally and functionally by establishing the relation of the models of each layer;
s3, data fusion: the method comprises the steps of denoising and modeling real-time data of an electric vehicle battery entity, classifying and analyzing results, and finally iterating, evolving and fusing equipment real-time data and model data to realize data fusion of a physical model and a virtual model, so that the virtual entity can truly reflect the running state of all elements of the physical entity in the whole working process;
s4, failure prediction: and after the twin model is built, the physical entity and the virtual entity are synchronously mapped, and are interacted in real time under the interaction of twin data, so that a basis is provided for the fault prediction of the electric vehicle battery.
2. The electric vehicle battery fault diagnosis method based on the digital twin model according to claim 1, wherein the total voltage, the total current, the state of charge, the highest temperature of the battery pack, the lowest temperature of the battery pack, the sampling time, the vehicle speed parameters and the control instructions of the electric vehicle battery pack are obtained in real time from a data acquisition and monitoring system and a state monitoring system of the electric vehicle battery pack through a TCP/IP and/or UDP communication mode when data are acquired in the step S1, so that the real-time perception of the digital twin system on the running state of the electric vehicle battery pack is realized.
3. The method for diagnosing the battery fault of the electric automobile based on the digital twin model according to claim 1,
the construction formula of the digital twin mapping model is as follows:
Sap DT ＝(E p ，N DT ，D DT ，I)
among them, sap DT Representing a digital twin map, E p Representing the battery entity of an electric vehicle, N DT Representing twin models of batteries of electric vehicles, D DT Represents twin data of an electric vehicle battery pack, I represents E p 、N DT 、D DT Are connected with each other; electric vehicle battery (E) p ) An entity is a precondition for implementing a twin mapping model, thereby creating a real to imaginary mapping mechanism.
4. The electric vehicle battery fault diagnosis method based on the digital twin model according to claim 3,
the electric automobile battery pack twin model is constructed according to the following formula:
N DT ＝(N G ，N A ，N E )
wherein N is DT Representing twin models of batteries of electric vehicles, N G Representing a geometric model, N A Representative analytical model, N E Is an evolution model.
5. The electric vehicle battery fault diagnosis method based on the digital twin model as claimed in claim 1, wherein the established digital twin mapping models are completely combined and constructed to realize perfect fusion of physical space, information space and twin data.
6. The electric vehicle battery fault diagnosis method based on the digital twin model as claimed in claim 1, wherein the fault prediction in step S4 is further characterized in that a deep bidirectional long-term and short-term memory network is constructed based on the bidirectional long-term and short-term memory network, so as to implement fault diagnosis and separation of the electric vehicle battery.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211153359.3A CN115932592A (en) | 2022-09-19 | 2022-09-19 | Electric vehicle battery fault diagnosis method based on digital twin model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211153359.3A CN115932592A (en) | 2022-09-19 | 2022-09-19 | Electric vehicle battery fault diagnosis method based on digital twin model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115932592A true CN115932592A (en) | 2023-04-07 |
Family
ID=86649556
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211153359.3A Pending CN115932592A (en) | 2022-09-19 | 2022-09-19 | Electric vehicle battery fault diagnosis method based on digital twin model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115932592A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116740293A (en) * | 2023-06-13 | 2023-09-12 | 西安速度时空大数据科技有限公司 | Digital twinning-based three-dimensional terrain model acquisition method, device and storage medium |
| CN117113453A (en) * | 2023-08-28 | 2023-11-24 | 上海智租物联科技有限公司 | Battery problem diagnosis method and storage medium based on 3D and big data technology |
| CN117388708A (en) * | 2023-10-30 | 2024-01-12 | 暨南大学 | Power battery system and thermal runaway monitoring method thereof |
-
2022
- 2022-09-19 CN CN202211153359.3A patent/CN115932592A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116740293A (en) * | 2023-06-13 | 2023-09-12 | 西安速度时空大数据科技有限公司 | Digital twinning-based three-dimensional terrain model acquisition method, device and storage medium |
| CN117113453A (en) * | 2023-08-28 | 2023-11-24 | 上海智租物联科技有限公司 | Battery problem diagnosis method and storage medium based on 3D and big data technology |
| CN117113453B (en) * | 2023-08-28 | 2024-04-26 | 上海智租物联科技有限公司 | Battery problem diagnosis method and storage medium based on 3D and big data technology |
| CN117388708A (en) * | 2023-10-30 | 2024-01-12 | 暨南大学 | Power battery system and thermal runaway monitoring method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115932592A (en) | Electric vehicle battery fault diagnosis method based on digital twin model | |
| CN114154429B (en) | Digital twin body construction method and device of energy storage system and storage medium | |
| CN110658462B (en) | Lithium battery online service life prediction method based on data fusion and ARIMA model | |
| CN114372417A (en) | Electric vehicle battery health state and remaining life evaluation method based on charging network | |
| CN105302848A (en) | Evaluation value calibration method of equipment intelligent early warning system | |
| CN113656938A (en) | Monitoring method, control device and storage medium for distribution network transformer | |
| CN110941929A (en) | Battery health state assessment method based on ARMA and Elman neural network combined modeling | |
| CN109375116B (en) | Battery system abnormal battery identification method based on self-encoder | |
| CN107436409B (en) | Intelligent SOC prediction device for power battery of electric automobile | |
| KR102297343B1 (en) | Battery Output Voltage Response and State-of-Charge Forecasting Method using Hybrid VARMA and LSTM | |
| CN112838946B (en) | Method for constructing intelligent sensing and early warning model based on communication network faults | |
| AU2021101964A4 (en) | Artificial intelligence based smart electric vehicle battery management system | |
| CN116093459B (en) | Intelligent monitoring method and system for lithium battery energy storage box | |
| CN116683588B (en) | Lithium ion battery charge and discharge control method and system | |
| CN114035098A (en) | Lithium battery health state prediction method integrating future working condition information and historical state information | |
| CN110516813A (en) | A method of batteries of electric automobile RDR prediction is carried out based on big data machine learning | |
| CN114919559A (en) | System and method for predicting residual service life of braking system based on digital twinning | |
| CN117154263A (en) | Lithium battery cascade utilization charging and discharging system and control method | |
| CN117554846B (en) | Lithium battery life prediction method and system considering constraint conditions | |
| CN102095953B (en) | A kind of performance of accumulator charger online test method | |
| US20230305073A1 (en) | Method and apparatus for providing a predicted aging state of a device battery based on a predicted usage pattern | |
| CN116794545A (en) | Yun Bian cooperation-based battery health prediction method and system | |
| CN116901707A (en) | Power battery pack fault early warning method, system and vehicle | |
| CN116699435A (en) | Method and apparatus for identifying critical anomalies in a battery pack based on a machine learning method | |
| CN114492578A (en) | Electrical equipment operation data classification method based on BP neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |