CN117261599A - Fault detection method and device of electric automobile, electronic equipment and electric automobile - Google Patents
Fault detection method and device of electric automobile, electronic equipment and electric automobile Download PDFInfo
- Publication number
- CN117261599A CN117261599A CN202311345272.0A CN202311345272A CN117261599A CN 117261599 A CN117261599 A CN 117261599A CN 202311345272 A CN202311345272 A CN 202311345272A CN 117261599 A CN117261599 A CN 117261599A
- Authority
- CN
- China
- Prior art keywords
- electric automobile
- data
- power battery
- fault detection
- battery data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 79
- 238000000034 method Methods 0.000 claims abstract description 59
- 238000005070 sampling Methods 0.000 claims abstract description 31
- 239000013598 vector Substances 0.000 claims description 55
- 238000009826 distribution Methods 0.000 claims description 50
- 238000013528 artificial neural network Methods 0.000 claims description 25
- 230000008569 process Effects 0.000 claims description 24
- 230000006870 function Effects 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 10
- 230000002159 abnormal effect Effects 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 4
- 238000007476 Maximum Likelihood Methods 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000002360 preparation method Methods 0.000 claims description 3
- 238000004422 calculation algorithm Methods 0.000 description 53
- 210000004027 cell Anatomy 0.000 description 14
- 238000003860 storage Methods 0.000 description 10
- 239000000178 monomer Substances 0.000 description 9
- 238000012216 screening Methods 0.000 description 8
- 238000000354 decomposition reaction Methods 0.000 description 6
- 238000000605 extraction Methods 0.000 description 5
- 230000002829 reductive effect Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 230000001960 triggered effect Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000007599 discharging Methods 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 239000002131 composite material Substances 0.000 description 2
- 238000009413 insulation Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 208000032953 Device battery issue Diseases 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000002547 anomalous effect Effects 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 230000002269 spontaneous effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
- B60L3/00—Electric devices on electrically-propelled vehicles for safety purposes; Monitoring operating variables, e.g. speed, deceleration or energy consumption
- B60L3/0023—Detecting, eliminating, remedying or compensating for drive train abnormalities, e.g. failures within the drive train
- B60L3/0046—Detecting, eliminating, remedying or compensating for drive train abnormalities, e.g. failures within the drive train relating to electric energy storage systems, e.g. batteries or capacitors
-
- 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
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- Sustainable Development (AREA)
- Sustainable Energy (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Secondary Cells (AREA)
Abstract
The invention relates to the technical field of electric automobiles, in particular to a fault detection method and device of an electric automobile, electronic equipment and the electric automobile. Wherein the method comprises the following steps: the method comprises the steps of removing sampling faults from original power battery data of each electric automobile collected by each vehicle-mounted BMS to obtain target power battery data of each electric automobile; wherein the power battery data includes voltage data, current data, temperature data, and time data; and inputting the target power battery data of each electric automobile into a pre-trained first fault detection model to determine the electric automobile with faults. The scheme of the invention can effectively detect faults of the electric automobile.
Description
Technical Field
The invention relates to the technical field of electric automobiles, in particular to a fault detection method and device of an electric automobile, electronic equipment and the electric automobile.
Background
Power cells are one of the key parts of electric vehicles, and the risk of failure directly affects the safety and reliability of the entire vehicle. With the use of an electric automobile, the performance nonlinearity of the power battery is reduced due to the severe road conditions, the environmental temperature and the dynamic change of the load, and further the problems of liquid leakage, insulation damage, partial short circuit and the like are caused.
If the internal reaction mechanism of the power battery cannot be clarified, the fault characteristics are monitored and the health state is evaluated in time, the power battery is accelerated to age, spontaneous combustion, explosion and other serious safety accidents are caused, and serious threat is caused to life and property safety of people, so that the fault detection of the electric automobile becomes more important.
Therefore, a method and device for detecting faults of an electric vehicle, an electronic device and an electric vehicle are needed to solve the above technical problems.
Disclosure of Invention
The embodiment of the invention provides a fault detection method and device for an electric automobile, electronic equipment and the electric automobile, and can effectively detect faults of the electric automobile.
In a first aspect, an embodiment of the present invention provides a fault detection method for an electric automobile, including:
the method comprises the steps of removing sampling faults from original power battery data of each electric automobile collected by each vehicle-mounted BMS to obtain target power battery data of each electric automobile; wherein the power battery data includes voltage data, current data, temperature data, and time data;
and inputting the target power battery data of each electric automobile into a pre-trained first fault detection model to determine the electric automobile with faults.
In a second aspect, an embodiment of the present invention provides a fault detection device for an electric vehicle, including:
the fault eliminating module is used for eliminating sampling faults of the original power battery data of each electric automobile collected by each vehicle-mounted BMS to obtain target power battery data of each electric automobile; wherein the power battery data includes voltage data, current data, temperature data, and time data;
the first detection module is used for inputting the target power battery data of each electric automobile into a pre-trained first fault detection model so as to determine the electric automobile with the fault.
In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the processor implements the method according to any embodiment of the present invention.
In a fourth aspect, an embodiment of the present invention further provides an electric automobile, including the electronic device according to the foregoing embodiment of the present invention.
The embodiment of the invention provides a fault detection method and device for electric vehicles, electronic equipment and the electric vehicles, wherein the fault detection method and device for the electric vehicles is characterized in that firstly, sampling faults are removed from original power battery data of each electric vehicle collected by each vehicle-mounted BMS, so that false alarms caused by the sampling faults can be avoided; and then inputting the target power battery data of each electric automobile into a pre-trained first fault detection model, so that the electric automobiles can be quickly and effectively subjected to fault detection.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a fault detection method of an electric automobile according to an embodiment of the present invention;
fig. 2 is a hardware architecture diagram of an electronic device according to an embodiment of the present invention;
fig. 3 is a structural diagram of a fault detection device of an electric automobile according to an embodiment of the present invention;
fig. 4 is a graph of basic reliability distribution of a power cell in one of its characteristic dimensions according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments, and all other embodiments obtained by those skilled in the art without making any inventive effort based on the embodiments of the present invention are within the scope of protection of the present invention.
Referring to fig. 1, an embodiment of the present invention provides a fault detection method for an electric vehicle, including:
step 100: the method comprises the steps of removing sampling faults from original power battery data of each electric automobile collected by each vehicle-mounted BMS to obtain target power battery data of each electric automobile; the power battery data comprises voltage data, current data, temperature data and time data;
step 102: and inputting the target power battery data of each electric automobile into a pre-trained first fault detection model to determine the electric automobile with faults.
In the embodiment of the invention, firstly, the original power battery data of each electric automobile collected by each vehicle-mounted BMS is subjected to sampling fault elimination, so that false alarm caused by sampling faults can be avoided; and then inputting the target power battery data of each electric automobile into a pre-trained first fault detection model, so that the electric automobiles can be quickly and effectively subjected to fault detection.
In some embodiments, the original power battery data of each electric vehicle collected by each vehicle-mounted BMS (Battery Management System ) is uploaded to a cloud platform (i.e. a remote monitoring platform) through a T-BOX (remote communication terminal), the cloud platform can read, clean and store the original power battery data, and the read uploaded battery data structure can be sorted and stored according to the national standard 32960 format.
In some embodiments, the rejection of sampling faults may be achieved by: carrying out differential processing on voltage data of all power batteries in a certain time period collected by the cloud platform; calculating the Z fraction of the voltage differential data at each sampling moment, and identifying the voltage data of which the Z fraction of the battery cell is larger than a preset threshold value; for the battery cell voltage data with the fraction Z of the battery cell being larger than a preset threshold value, calculating the difference value between the cell voltage data and the average voltage at each moment, and judging whether the difference value is larger than the preset threshold value; if the difference value is larger than the threshold value, judging that the monomer data are outlier monomer data; judging whether two battery monomer numbers with highest outlier degree are adjacent to the data judged as outlier monomers; for the data of judging that the adjacent monomer is outlier, judging whether the voltage differential data of the adjacent monomer is symmetrical or approximately symmetrical; if the sampling is symmetrical or approximately symmetrical, the sampling abnormality of the adjacent battery cells is judged.
The process of fault detection of the target power cell data of all electric vehicles using the pre-trained first fault detection model (which may be referred to herein as coarse screening) is described below.
In one embodiment of the invention, the first fault detection model comprises an encoder, a sampling layer and a decoder connected in sequence, the encoder and the decoder each comprising an LSTM or a GRU, the encoder being configured to convert input sequence data into a first potential vector, the sampling layer being configured to map the first potential vector into a gaussian distribution and to sample a second potential vector from the gaussian distribution, the decoding layer being configured to convert the second potential vector into output sequence data, the input sequence data and the output sequence data being consecutive vectors of the same type.
In this embodiment, the first failure detection model generates new data (i.e., output sequence data) by learning the gaussian distribution of the input sequence data, so that the output sequence data can be made more continuous and diversified; also, both the encoder and decoder include LSTM or GRU, which better captures the time-dependent nature of the target power cell.
In addition, when the encoder and the decoder both comprise the LSTM, the LSTM has the structure of the memory unit, so that the first fault detection model can decompose the first potential vector into state vectors of various time steps, each dimension of the first potential vector has a more definite meaning, and the interpretability of the model is enhanced.
In one embodiment of the invention, the first fault detection model is trained by:
inputting sample power battery data serving as a positive sample into a target neural network to be trained; the target neural network comprises an encoder, a sampling layer and a decoder which are connected in sequence;
And when the training times reach the preset times or the preset loss function is smaller than a preset value, obtaining a first fault detection model after training.
Generally, the first fault detection model is trained using normal data (i.e., power battery data when the power battery is in a normal state) to detect new data using the trained first fault detection model. Specifically, new data is input into the first fault detection model, resulting in a second potential vector for the new data. If the difference between the new data and the second potential vector of normal data is small, then the new data is considered normal; otherwise, the new data is considered anomalous (i.e., there is a fault).
However, the inventors found during the development process that: if only the power battery data when the power battery is in a normal state is used as a positive sample for model training, the final detection accuracy is not high. For this reason, the inventor creatively mixes and uses the power battery data when a small part (for example, less than 10%) of the power batteries are in abnormal states with the other normal data to jointly serve as a positive sample of model training, so that the final experimental result shows that the model detection precision of the latter is higher than that of the former.
In one embodiment of the invention, the loss function uses the following formula:
in the method, in the process of the invention,Las a function of the loss,sampling the mean value of the Gaussian distribution obtained by layer mapping for each training process, < >>Sampling the variance of the layer mapping obtained gaussian distribution for each training process +.>To output the first in the sequence dataiThe number of vectors is the number of vectors,x i to input the first in the sequence dataiThe number of vectors is the number of vectors,nthe number of vectors in the input sequence data and the output sequence data.
In this embodiment, the loss function includes two terms, the first term being used to make the first potential vector generated by the encoder as conform to the gaussian distribution as possible, and the second term being used to make the input sequence data and the output sequence data of the first failure detection model as similar as possible.
For the established target neural network, the calculation precision and efficiency are better than those of a machine learning algorithm, but the internal super parameters are numerous, and long period and cost are required for parameter adjustment by means of expert experience, so that the established neural network is optimized by combining an intelligent optimization algorithm.
In one embodiment of the invention, the superparameter of the target neural network is determined by:
coding the initial super parameters of the target neural network to obtain an initial population; the initial super-parameters are individuals of an initial population, and the initial super-parameters comprise the number of layers of the neural network, the number of neurons and the weight;
Setting a population scale, a maximum evolution frequency, a crossover probability, a variation probability, a learning rate, a speed range and a position range of an individual;
initializing the speed and the position of each individual in the initial population, and calculating the fitness value of each individual in the initial population;
updating the speed and the position of each individual in the initial population, and calculating the fitness value of each updated individual;
selecting, crossing and mutating the individuals of the initial population based on the updated fitness value of each individual to obtain a next generation population, circularly executing the computation, selection, crossing and mutating of the fitness value of the individuals in each generation population until the iteration of the maximum evolution times is completed, and outputting the globally optimal individual with the highest fitness value;
and decoding the globally optimal individual to obtain the optimal super-parameters.
In this embodiment, by converting a problem to be optimized into a search problem in a multidimensional space, the solution of the problem is regarded as a certain position in the multidimensional space, and then by continuously moving the position of an individual, the optimal solution is gradually found. Each individual has its own position and velocity and can be adjusted by its own experience and group experience to achieve global optimization.
The following is an explanation of some of the operations described above:
mutation operation: for each individual, k individuals different from themselves are randomly selected, and the differences between them are calculated, resulting in a new vector.
Crossover operation: the new vector is intersected with the original vector to generate a offspring vector.
Selection operation: and comparing the fitness value of the offspring vector with the fitness value of the parent vector, and selecting the vector with higher fitness as the population of the next generation.
Note that, the fitness value may be calculated using a preset fitness function, and the fitness function specifically selected is not limited herein.
In one embodiment of the invention, the learning rate is adjusted by the following formula:
in the method, in the process of the invention,for learning rate->Is a constant within the range of (0, 1, ">Is a constant in the range of (1, 2),d j is the firstjThe expected network output values of the individual samples,a j is the firstjThe actual network output values of the individual samples,mse(k)representing target neural network operationkThe mean square error between the next actual network output value and the desired network output value,kfor the current number of runs of the target neural network,kis a positive integer which is used for the preparation of the high-voltage power supply,Nis the number of samples.
In this embodiment, the adaptive learning rate mode is adopted in the target neural network, so that efficient learning and rapid convergence of the network can be further promoted, and the defects of low convergence speed, easy trapping of local minima and the like of the common neural network are effectively overcome.
In one embodiment of the present invention, the step of inputting the target power battery data of each electric vehicle into the first failure detection model trained in advance to determine the failed electric vehicle may specifically include:
inputting target power battery data of each electric automobile into a pre-trained first fault detection model to obtain a second potential vector of each electric automobile;
determining an anomaly threshold value for each electric vehicle based on the following formula:
in the method, in the process of the invention,zas a result of the abnormality threshold value,t98% fraction of the second potential vector for the current electric car,and->To follow the maximum likelihood estimates of the pareto distribution,nas the number of the second potential vectors of the current electric automobile,N t for peak value greater thantVector number of (2);
and when the target power battery data exceeds an abnormal threshold value, determining that the current electric automobile fails.
In this embodiment, the real world data is difficult to generalize with a known distribution, for example, for certain extreme events (anomalies), a probability model (e.g., a gaussian distribution) would give its probability of 0. The above formula can infer the distribution of extreme events without any distribution assumption based on the raw data. Thus, the above formula does not require manual setting of thresholds and assuming a distribution of data compliance, so that in cases where the original data distribution is very complex, it is still possible to estimate extreme events.
The technical scheme is that the first failure detection model trained in advance is utilized to carry out the process of coarse screening on all electric automobiles. In order to further effectively reduce the false alarm rate of the model on the premise of ensuring the recall ratio, the inventor creatively considers performing fine screening on the basis of coarse screening. The process of performing fine screening on all suspected faulty electric vehicles (i.e., the first fault detection model can determine which electric vehicles are suspected faulty electric vehicles) by using the pre-trained second fault detection model is described below.
In one embodiment of the present invention, after step 102, the method may specifically further include:
step 104: extracting the characteristics of the target power battery data of each suspected fault electric vehicle to obtain target characteristic values of each power battery in different characteristic dimensions of each suspected fault electric vehicle;
step 106: inputting target feature values of each power battery in each suspected fault electric vehicle in different feature dimensions into a pre-trained second fault detection model to determine the number of the power battery with fault of each suspected fault electric vehicle; the second fault detection model comprises an isolated forest algorithm, a first unsupervised learning algorithm based on distance and a second unsupervised learning algorithm based on density;
Step 108: and aiming at each suspected fault electric vehicle with the power battery fault, determining the predicted fault level of the current electric vehicle based on the power battery number of the fault electric vehicle and the target characteristic values of each power battery in different characteristic dimensions of the current electric vehicle.
In the embodiment, the characteristic extraction is carried out on the target power battery data of each electric automobile, and the target characteristic values of each power battery in each electric automobile in different characteristic dimensions are input into a pre-trained second fault detection model, so that the false alarm rate of the model can be effectively reduced on the premise of ensuring the recall rate; and finally, determining the predicted fault level of the current electric automobile based on the number of the power battery with the fault of the current electric automobile and the target characteristic values of each power battery in different characteristic dimensions of the current electric automobile.
In one embodiment of the invention, the feature dimension includes at least one of: internal resistance, open circuit voltage, capacity increment curve, differential capacity curve, charge current difference, charge voltage singular value decomposition, charge cutoff voltage, discharge current difference, discharge voltage difference and discharge voltage singular value decomposition.
It should be noted that, the target feature values of some feature dimensions may be one or more, for example, the target feature values of the internal resistance, the open circuit voltage, the charging current difference, the charging voltage difference, the charging cut-off voltage, the discharging current difference, and the discharging voltage difference may be one, and the target feature values of the charging voltage singular value decomposition and the discharging voltage singular value decomposition may be more than one.
In addition, the specific manner of feature extraction of the target power battery data of each electric vehicle is well known to those skilled in the art, and the specific manner of feature extraction is not described herein.
Because the second fault detection model comprises an isolated forest algorithm, a first unsupervised learning algorithm based on distance and a second unsupervised learning algorithm based on density, the abnormal detection principle of the second fault detection model can be guaranteed to relate to three aspects (namely tree-based, distance-based and density-based), and therefore the false alarm rate of the model can be effectively reduced on the premise of guaranteeing the recall rate.
In one embodiment of the invention, the first unsupervised learning algorithm is a K-means algorithm or a Canopy algorithm;
the second unsupervised learning algorithm is a DBSCAN algorithm, an OPTICS algorithm, or a DENCLUE algorithm.
It is appreciated that the K-means algorithm, the Canopy algorithm, the DBSCAN algorithm, the OPTICS algorithm, and the DENCLUE algorithm are well known to those skilled in the art and will not be described in detail herein.
In one embodiment of the present invention, step 106 may specifically include:
inputting target feature values of each power battery in different feature dimensions of each suspected fault electric vehicle into a pre-trained isolated forest algorithm, a first unsupervised learning algorithm based on distance and a second unsupervised learning algorithm based on density, and determining the number of feature dimensions of outliers of each power battery in each algorithm;
for each algorithm, if the number of the characteristic dimensions of the outliers in the current algorithm exceeds the preset proportion of the total number of the characteristic dimensions, determining the current power battery as a suspected fault power battery;
and performing intersection calculation on all suspected fault power batteries determined in the three algorithms to determine the number of the power battery with fault of each electric automobile.
In this embodiment, through setting the second fault detection model including three algorithms, suspected fault power batteries of electric vehicles output by the three algorithms can be respectively screened out, and then intersection operation is performed on all suspected fault power batteries determined in the three algorithms, so that the number of the power battery actually having a fault in each suspected fault electric vehicle can be accurately and effectively screened out, that is, the false alarm rate of the model can be effectively reduced on the premise of ensuring the recall rate.
In some embodiments, the preset ratio may be adaptively adjusted according to specific situations, which is not limited herein.
As shown in fig. 4, in one embodiment of the present invention, step 108 may specifically include:
for each electric vehicle suspected of failure of the power battery, the following operations are executed:
calculating the centroid of the target characteristic values of all power batteries in each characteristic dimension in the current electric automobile so as to calculate the Euclidean distance from each target characteristic value to the centroid;
based on the calculated euclidean distance of all the target feature values to the centroid, a mean value of all the target feature values for each feature dimension is determined (i.e. in fig. 4μ) And standard deviation (i.e. in FIG. 4σ);
Fitting to obtain a first confidence distribution curve (i.e. in fig. 4) supporting normal of each feature dimension based on a preset lower confidence distribution limit (e.g. 0.1 in fig. 4) and an upper confidence distribution limit (e.g. 0.8 in fig. 4) and the mean and standard deviation of all target feature valuesm(N)) A second confidence score for supporting anomalies (i.e., in FIG. 4m(A)) And support for an unknown third confidence score (i.e., in FIG. 4m({N,A}));
And determining the predicted fault level of the current electric automobile based on the number of the power battery with the fault of the current electric automobile and the first reliability distribution curve, the second reliability distribution curve and the third reliability distribution curve in different characteristic dimensions.
In this embodiment, by calculating the centroid of the target feature values of all the power batteries in each feature dimension in the current electric vehicle and calculating the euclidean distance from each target feature value to the centroid, statistics deviating from mathematical expectations can be performed on the target feature values of all the power batteries in each feature dimension, that is, the mean value and standard deviation of all the target feature values of each feature dimension are determined, so that a gaussian distribution curve (i.e., a third confidence distribution curve) of all the target feature values of all the power batteries in each feature dimension in the current electric vehicle can be obtained; then fitting the data to obtain a first credibility distribution curve and a second credibility distribution curve; and finally, according to the number of the power battery with the fault of the current electric automobile, obtaining a reliability value corresponding to the number of the power battery, and further determining the predicted fault level of the current electric automobile.
With continued reference to fig. 4, in one embodiment of the present invention, the step of determining the predicted failure level of the current electric vehicle based on the number of the failed power battery of the current electric vehicle and the first, second and third confidence distribution curves in different feature dimensions may specifically include:
Determining a first reliability value, a second reliability value and a third reliability value of the current electric vehicle in different characteristic dimensions (namely the first reliability value, the second reliability value and the third reliability value of the current electric vehicle in different characteristic dimensions corresponding to the failed power battery number) based on the failed power battery number of the current electric vehicle and the first reliability distribution curve, the second reliability distribution curve and the third reliability distribution curve in different characteristic dimensions;
determining the upper limit of the fault probability of the current electric automobile based on the first reliability value, the second reliability value and the third reliability value of the current electric automobile in different characteristic dimensions;
and determining the predicted fault level of the current electric automobile based on the upper limit of the fault probability of the current electric automobile.
It should be noted that, the upper limit of the fault probability of the current electric automobile may be obtained by setting a custom formula, and inputting the first confidence value, the second confidence value and the third confidence value of the current electric automobile in different feature dimensions into the custom formula, where the custom formula is not limited.
It will be appreciated that the above process of determining the predicted fault level of the current electric vehicle is a process of predicting a future situation based on current data, and in practical application, the fault level of the electric vehicle in the future half an hour to one hour may be predicted. However, this method needs to use a cloud platform, that is, the calculation force of the vehicle end may not be enough, and of course, the vehicle end may perform the method, which is not specifically limited herein.
In one embodiment of the present invention, after obtaining the target power battery data of each electric vehicle, the method further includes:
for each electric car, the following operations are performed:
determining all alarm types triggered by the current electric automobile based on target power battery data of the current electric automobile and preset alarm rules; the alarm rule comprises a plurality of alarm types and threshold judgment rules, at least one alarm type comprises at least two threshold judgment rules aiming at different battery data types, the battery data types comprise voltage, current and temperature, and each alarm type is given a preset weight;
and determining the real-time fault level of the current electric automobile based on the sum of the weights of all alarm types triggered by the current electric automobile.
In this embodiment, by setting the judging mode of the composite threshold, the fault level of the current electric automobile can be detected at the vehicle end in real time.
In some embodiments, alarm types include, but are not limited to, a SOC false high alarm, a SOC too low alarm, an adjacent cell acquisition anomaly, a temperature difference too large alarm, an insulation resistance alarm, a cell consistency alarm, a static pressure difference alarm, a cell overpressure alarm, a cell underpressure alarm, a cell overtemperature alarm, a voltage anomaly alarm.
In some embodiments, for example, the threshold discrimination rule for neighboring monomer acquisition anomalies is: delta V is more than 0.3V, the pressure difference of adjacent monomers is more than or equal to 95 percent delta V, and the average value is less than or equal to the average value of other monomers plus or minus 0.05V.
In summary, three barriers of a coarse screening algorithm (namely a first fault detection model), a fine screening algorithm (namely a second fault detection model) and a composite threshold algorithm (namely a detection algorithm of real-time fault level) based on data reconstruction comparison are established in the technical scheme provided by the embodiment of the invention, all monitoring and key screening of all online electric vehicles by a cloud platform are realized, the application of a battery early warning algorithm in the fields of vehicles, energy storage and the like can be realized only by simple training and limited parameter adjustment, the recall ratio of the algorithm is effectively improved, false alarm is reduced, and early warning of thermal runaway at least half an hour in advance can be realized.
As shown in fig. 2 and 3, the embodiment of the invention provides a fault detection device for an electric automobile. The apparatus embodiments may be implemented by software, or may be implemented by hardware or a combination of hardware and software. In terms of hardware, as shown in fig. 2, a hardware architecture diagram of an electronic device where a fault detection device for an electric automobile is located according to an embodiment of the present invention is shown, where in addition to a processor, a memory, a network interface, and a nonvolatile memory shown in fig. 2, the electronic device where the device is located in an embodiment may generally include other hardware, such as a forwarding chip responsible for processing a message, and so on. Taking a software implementation as an example, as shown in fig. 3, the device in a logic sense is formed by reading a corresponding computer program in a nonvolatile memory into a memory by a CPU of an electronic device where the device is located and running the computer program.
As shown in fig. 3, the fault detection device for an electric vehicle provided in this embodiment includes:
the fault rejection module 300 is configured to reject sampling faults of the original power battery data of each electric vehicle collected by each vehicle-mounted BMS, so as to obtain target power battery data of each electric vehicle; wherein the power battery data includes voltage data, current data, temperature data, and time data;
the first detection module 302 is configured to input the target power battery data of each electric vehicle into a first failure detection model trained in advance, so as to determine the electric vehicle that has failed.
In an embodiment of the present invention, the fault rejection module 300 may be configured to perform step 100 in the above-described method embodiment, and the first detection module 302 may be configured to perform step 102 in the above-described method embodiment.
In one embodiment of the invention, the first fault detection model comprises an encoder, a sampling layer and a decoder connected in sequence, the encoder and the decoder each comprising an LSTM or a GRU, the encoder being configured to convert input sequence data into a first potential vector, the sampling layer being configured to map the first potential vector into a gaussian distribution and to sample a second potential vector from the gaussian distribution, the decoding layer being configured to convert the second potential vector into output sequence data, the input sequence data and the output sequence data being consecutive vectors of the same type.
In one embodiment of the invention, the first fault detection model is trained by:
inputting sample power battery data serving as a positive sample into a target neural network to be trained; the target neural network comprises an encoder, a sampling layer and a decoder which are sequentially connected, wherein the sample power battery data comprise first sample power battery data when more than a preset percentage of power batteries in the electric automobile are in a normal state and second sample power battery data when the rest power batteries in the electric automobile are in an abnormal state;
and when the training times reach the preset times or the preset loss function is smaller than a preset value, obtaining a first fault detection model after training.
In one embodiment of the invention, the loss function uses the following formula:
in the method, in the process of the invention,Las a function of the loss in question,for each training process the mean value of the gaussian distribution obtained by mapping the sampling layer is +.>Variance of gaussian distribution mapped for the sampling layer for each training process +.>Is the first in the output sequence dataiThe number of vectors is the number of vectors,x i is the first in the input sequence dataiThe number of vectors is the number of vectors, nThe number of vectors in the input sequence data and the output sequence data.
In one embodiment of the invention, the superparameter of the target neural network is determined by:
coding the initial super parameters of the target neural network to obtain an initial population; the initial super-parameters are individuals of the initial population, and the initial super-parameters comprise the number of layers of the neural network, the number of neurons and the weight;
setting a population scale, a maximum evolution frequency, a crossover probability, a variation probability, a learning rate, a speed range and a position range of an individual;
initializing the speed and the position of each individual in the initial population, and calculating the fitness value of each individual in the initial population;
updating the speed and the position of each individual in the initial population, and calculating the fitness value of each updated individual;
selecting, crossing and mutating the individuals of the initial population based on the updated fitness value of each individual to obtain a next generation population, circularly executing the computation, selection, crossing and mutating of the fitness value of the individuals in each generation population until the iteration of the maximum evolution times is completed, and outputting the globally optimal individual with the highest fitness value;
And decoding the globally optimal individual to obtain the optimal super-parameters.
In one embodiment of the invention, the learning rate is adjusted by the following formula:
in the method, in the process of the invention,for the learning rate, ++>Is a constant within the range of (0, 1, ">Is a constant in the range of (1, 2),d j is the firstjThe expected network output values of the individual samples,a j is the firstjThe actual network output values of the individual samples,mse(k)representing target neural network operationkThe mean square error between the next actual network output value and the desired network output value,kfor the current number of runs of the target neural network,kis a positive integer which is used for the preparation of the high-voltage power supply,Nis the number of samples.
In one embodiment of the present invention, the inputting the target power battery data of each electric vehicle into the first failure detection model trained in advance to determine the failed electric vehicle includes:
inputting target power battery data of each electric automobile into a pre-trained first fault detection model to obtain a second potential vector of each electric automobile;
determining an anomaly threshold value for each electric vehicle based on the following formula:
in the method, in the process of the invention,zas a result of the said anomaly threshold value,t98% fraction of the second potential vector for the current electric car, And->To follow the maximum likelihood estimates of the pareto distribution,nas the number of the second potential vectors of the current electric automobile,N t for peak value greater thantVector number of (2);
and when the target power battery data exceeds the abnormal threshold value, determining that the current electric automobile fails.
In one embodiment of the present invention, further comprising:
the feature extraction module is used for carrying out feature extraction on the target power battery data of each suspected fault electric automobile to obtain target feature values of each power battery in different feature dimensions in each suspected fault electric automobile;
the second detection module is used for inputting target characteristic values of each power battery in each suspected fault electric vehicle in different characteristic dimensions into a pre-trained second fault detection model so as to determine the number of the power battery with fault of each suspected fault electric vehicle; the second fault detection model comprises an isolated forest algorithm, a first unsupervised learning algorithm based on distance and a second unsupervised learning algorithm based on density;
the fault prediction module is used for determining a predicted fault level of the current electric automobile based on the number of the power battery of the current electric automobile with faults and target characteristic values of each power battery in different characteristic dimensions of the current electric automobile aiming at each suspected fault electric automobile with faults of the power battery.
In one embodiment of the invention, the feature dimension includes at least one of: internal resistance, open circuit voltage, capacity increment curve, differential capacity curve, charge current difference, charge voltage singular value decomposition, charge cutoff voltage, discharge current difference, discharge voltage difference and discharge voltage singular value decomposition.
In one embodiment of the present invention, the first unsupervised learning algorithm is a K-means algorithm or a Canopy algorithm:
the second unsupervised learning algorithm is a DBSCAN algorithm, an OPTICS algorithm or a DENCLUE algorithm.
In one embodiment of the present invention, the second detection module is configured to perform the following operations:
inputting target characteristic values of each power battery in different characteristic dimensions of each electric automobile into a pre-trained isolated forest algorithm, a first unsupervised learning algorithm based on distance and a second unsupervised learning algorithm based on density, and determining the number of characteristic dimensions of outliers of each power battery in each algorithm;
for each algorithm, if the number of the characteristic dimensions of the outliers in the current algorithm exceeds the preset proportion of the total number of the characteristic dimensions, determining the current power battery as a suspected fault power battery;
And performing intersection calculation on all suspected fault power batteries determined in the three algorithms to determine the number of the power battery with fault of each electric automobile.
In one embodiment of the present invention, the fault prediction module is configured to perform the following operations:
for each electric automobile with power battery failure, the following operations are executed:
calculating the centroid of the target characteristic values of all power batteries in each characteristic dimension in the current electric automobile so as to calculate the Euclidean distance from each target characteristic value to the centroid;
determining the mean value and standard deviation of all the target feature values of each feature dimension based on the Euclidean distance from all the target feature values to the centroid;
fitting to obtain a first normal-supporting reliability distribution curve, a second abnormal-supporting reliability distribution curve and an unknown third reliability distribution curve of each characteristic dimension based on a preset lower reliability distribution limit and an preset upper reliability distribution limit and the average value and standard deviation of all target characteristic values;
and determining the predicted fault level of the current electric automobile based on the number of the power battery with the fault of the current electric automobile and the first reliability distribution curve, the second reliability distribution curve and the third reliability distribution curve in different characteristic dimensions.
In one embodiment of the present invention, the fault prediction module is configured to, when executing the power battery number based on the fault of the current electric vehicle and the first, second and third reliability distribution curves in different feature dimensions, determine a predicted fault level of the current electric vehicle, execute the following operations:
determining a first reliability value, a second reliability value and a third reliability value of the current electric automobile in different feature dimensions based on the number of the power battery with the fault of the current electric automobile and the first reliability distribution curve, the second reliability distribution curve and the third reliability distribution curve in different feature dimensions;
determining the upper limit of the fault probability of the current electric automobile based on the first reliability value, the second reliability value and the third reliability value of the current electric automobile in different characteristic dimensions;
and determining the predicted fault level of the current electric automobile based on the upper limit of the fault probability of the current electric automobile.
In one embodiment of the present invention, further comprising:
a fault diagnosis module configured to:
for each electric car, it is used to perform the following operations:
determining all alarm types triggered by the current electric automobile based on target power battery data of the current electric automobile and preset alarm rules; the alarm rule comprises a plurality of alarm types and a threshold judgment rule, at least one alarm type comprises at least two threshold judgment rules aiming at different battery data types, the battery data types comprise voltage, current and temperature, and each alarm type is given a preset weight;
And determining the real-time fault level of the current electric automobile based on the sum of the weights of all alarm types triggered by the current electric automobile.
It will be appreciated that the structure illustrated in the embodiments of the present invention does not constitute a specific limitation on a fault detection device for an electric vehicle. In other embodiments of the invention, a fault detection device for an electric vehicle may include more or fewer components than shown, or may combine certain components, or may split certain components, or may have a different arrangement of components. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.
The content of information interaction and execution process between the modules in the device is based on the same conception as the embodiment of the method of the present invention, and specific content can be referred to the description in the embodiment of the method of the present invention, which is not repeated here.
The embodiment of the invention also provides electronic equipment, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the fault detection method of the electric automobile in any embodiment of the invention when executing the computer program.
The embodiment of the invention also provides a computer readable storage medium, and the computer readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor is caused to execute the fault detection method of the electric automobile in any embodiment of the invention.
Specifically, a system or apparatus provided with a storage medium on which a software program code realizing the functions of any of the above embodiments is stored, and a computer (or CPU or MPU) of the system or apparatus may be caused to read out and execute the program code stored in the storage medium.
In this case, the program code itself read from the storage medium may realize the functions of any of the above-described embodiments, and thus the program code and the storage medium storing the program code form part of the present invention.
Examples of the storage medium for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (e.g., CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a nonvolatile memory card, and a ROM. Alternatively, the program code may be downloaded from a server computer by a communication network.
Further, it should be apparent that the functions of any of the above-described embodiments may be implemented not only by executing the program code read out by the computer, but also by causing an operating system or the like operating on the computer to perform part or all of the actual operations based on the instructions of the program code.
Further, it is understood that the program code read out by the storage medium is written into a memory provided in an expansion board inserted into a computer or into a memory provided in an expansion module connected to the computer, and then a CPU or the like mounted on the expansion board or the expansion module is caused to perform part and all of actual operations based on instructions of the program code, thereby realizing the functions of any of the above embodiments.
In addition, the embodiment of the invention also discloses an electric automobile, which comprises: such as the electronic device disclosed above. In some embodiments, the electric vehicle may be a pure electric vehicle or a hybrid electric vehicle, without limitation.
It is noted that relational terms such as first and second, and the like, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, where the program, when executed, performs steps including the above method embodiments; and the aforementioned storage medium includes: various media in which program code may be stored, such as ROM, RAM, magnetic or optical disks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. The fault detection method for the electric automobile is characterized by comprising the following steps of:
the method comprises the steps of removing sampling faults from original power battery data of each electric automobile collected by each vehicle-mounted BMS to obtain target power battery data of each electric automobile; wherein the power battery data includes voltage data, current data, temperature data, and time data;
And inputting the target power battery data of each electric automobile into a pre-trained first fault detection model to determine the electric automobile with faults.
2. The method of claim 1, wherein the first fault detection model comprises an encoder, a sampling layer and a decoder connected in sequence, each of the encoder and the decoder comprising an LSTM or a GRU, the encoder for converting input sequence data into a first potential vector, the sampling layer for mapping the first potential vector into a gaussian distribution and sampling a second potential vector from the gaussian distribution, the decoding layer for converting the second potential vector into output sequence data, the input sequence data and the output sequence data being consecutive vectors of the same type.
3. The method of claim 2, wherein the first fault detection model is trained by:
inputting sample power battery data serving as a positive sample into a target neural network to be trained; the target neural network comprises an encoder, a sampling layer and a decoder which are sequentially connected, wherein the sample power battery data comprise first sample power battery data when more than a preset percentage of power batteries in the electric automobile are in a normal state and second sample power battery data when the rest power batteries in the electric automobile are in an abnormal state;
And when the training times reach the preset times or the preset loss function is smaller than a preset value, obtaining a first fault detection model after training.
4. A method according to claim 3, wherein the loss function uses the formula:
in the method, in the process of the invention,Las a function of the loss in question,for each training process the sampling layer maps the resulting mean value of the gaussian distribution,variance of gaussian distribution mapped for the sampling layer for each training process +.>Is the first in the output sequence dataiThe number of vectors is the number of vectors,x i is the first in the input sequence dataiThe number of vectors is the number of vectors,nthe number of vectors in the input sequence data and the output sequence data.
5. A method according to claim 3, wherein the hyper-parameters of the target neural network are determined by:
coding the initial super parameters of the target neural network to obtain an initial population; the initial super-parameters are individuals of the initial population, and the initial super-parameters comprise the number of layers of the neural network, the number of neurons and the weight;
setting a population scale, a maximum evolution frequency, a crossover probability, a variation probability, a learning rate, a speed range and a position range of an individual;
Initializing the speed and the position of each individual in the initial population, and calculating the fitness value of each individual in the initial population;
updating the speed and the position of each individual in the initial population, and calculating the fitness value of each updated individual;
selecting, crossing and mutating the individuals of the initial population based on the updated fitness value of each individual to obtain a next generation population, circularly executing the computation, selection, crossing and mutating of the fitness value of the individuals in each generation population until the iteration of the maximum evolution times is completed, and outputting the globally optimal individual with the highest fitness value;
and decoding the globally optimal individual to obtain the optimal super-parameters.
6. The method of claim 5, wherein the learning rate is adjusted by the formula:
in the method, in the process of the invention,for the learning rate, ++>Is a constant within the range of (0, 1, ">Is a constant in the range of (1, 2),d j is the firstjThe expected network output values of the individual samples,a j is the firstjThe actual network output values of the individual samples,mse(k)representing target neural network operationkThe mean square error between the next actual network output value and the desired network output value, kFor the current number of runs of the target neural network,kis a positive integer which is used for the preparation of the high-voltage power supply,Nis the number of samples.
7. The method of any one of claims 2-6, wherein the inputting the target power battery data for each electric vehicle into the pre-trained first fault detection model to determine a faulty electric vehicle comprises:
inputting target power battery data of each electric automobile into a pre-trained first fault detection model to obtain a second potential vector of each electric automobile;
determining an anomaly threshold value for each electric vehicle based on the following formula:
in the method, in the process of the invention,zas a result of the said anomaly threshold value,t98% fraction of the second potential vector for the current electric car,and->To follow the maximum likelihood estimates of the pareto distribution,nas the number of the second potential vectors of the current electric automobile,N t for peak value greater thantVector number of (2);
and when the target power battery data exceeds the abnormal threshold value, determining that the current electric automobile fails.
8. A fault detection device for an electric vehicle, comprising:
the fault eliminating module is used for eliminating sampling faults of the original power battery data of each electric automobile collected by each vehicle-mounted BMS to obtain target power battery data of each electric automobile; wherein the power battery data includes voltage data, current data, temperature data, and time data;
The first detection module is used for inputting the target power battery data of each electric automobile into a pre-trained first fault detection model so as to determine the electric automobile with the fault.
9. An electronic device comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the method of any of claims 1-7 when the computer program is executed.
10. An electric vehicle comprising the electronic device according to claim 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311345272.0A CN117261599B (en) | 2023-10-18 | 2023-10-18 | Fault detection method and device of electric automobile, electronic equipment and electric automobile |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311345272.0A CN117261599B (en) | 2023-10-18 | 2023-10-18 | Fault detection method and device of electric automobile, electronic equipment and electric automobile |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117261599A true CN117261599A (en) | 2023-12-22 |
CN117261599B CN117261599B (en) | 2024-05-03 |
Family
ID=89212285
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311345272.0A Active CN117261599B (en) | 2023-10-18 | 2023-10-18 | Fault detection method and device of electric automobile, electronic equipment and electric automobile |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117261599B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117907845A (en) * | 2024-03-20 | 2024-04-19 | 山东泰开电力电子有限公司 | Electrochemical energy storage system insulation detection method based on electrical parameter analysis |
Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108416103A (en) * | 2018-02-05 | 2018-08-17 | 武汉大学 | A kind of method for diagnosing faults of electric automobile of series hybrid powder AC/DC convertor |
CN109978229A (en) * | 2019-02-12 | 2019-07-05 | 常伟 | The method that the full battery core multi-point temperature of a kind of pair of power battery pack and tie point temperature carry out thermal runaway prediction |
CN110929798A (en) * | 2019-11-29 | 2020-03-27 | 重庆邮电大学 | Image classification method and medium based on structure optimization sparse convolution neural network |
CN111007401A (en) * | 2019-12-16 | 2020-04-14 | 国网江苏省电力有限公司电力科学研究院 | Electric vehicle battery fault diagnosis method and device based on artificial intelligence |
CN111861013A (en) * | 2020-07-23 | 2020-10-30 | 长沙理工大学 | Power load prediction method and device |
CN113987129A (en) * | 2021-11-08 | 2022-01-28 | 重庆邮电大学 | Digital media protection text steganography method based on variational automatic encoder |
CN114065613A (en) * | 2021-10-27 | 2022-02-18 | 中国华能集团清洁能源技术研究院有限公司 | Multi-working-condition process industrial fault detection and diagnosis method based on deep migration learning |
CN114155270A (en) * | 2021-11-10 | 2022-03-08 | 南方科技大学 | Pedestrian trajectory prediction method, device, equipment and storage medium |
CN115356636A (en) * | 2022-07-31 | 2022-11-18 | 常伟 | Data-driven new energy automobile battery fault alarm and fault early warning model |
CN116434741A (en) * | 2023-03-09 | 2023-07-14 | 平安科技(深圳)有限公司 | Speech recognition model training method, device, computer equipment and storage medium |
CN116626505A (en) * | 2023-07-21 | 2023-08-22 | 江苏海平面数据科技有限公司 | Battery pack consistency anomaly detection method based on Internet of vehicles big data |
CN116842337A (en) * | 2023-06-13 | 2023-10-03 | 国网甘肃省电力公司 | Transformer fault diagnosis method based on LightGBM (gallium nitride based) optimal characteristics and COA-CNN (chip on board) model |
CN117067920A (en) * | 2023-10-18 | 2023-11-17 | 北京航空航天大学 | Fault detection method and device of power battery, electronic equipment and electric automobile |
CN117067921A (en) * | 2023-10-18 | 2023-11-17 | 北京航空航天大学 | Fault detection method of electric automobile and electric automobile |
-
2023
- 2023-10-18 CN CN202311345272.0A patent/CN117261599B/en active Active
Patent Citations (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190242936A1 (en) * | 2018-02-05 | 2019-08-08 | Wuhan University | Fault diagnosis method for series hybrid electric vehicle ac/dc converter |
CN108416103A (en) * | 2018-02-05 | 2018-08-17 | 武汉大学 | A kind of method for diagnosing faults of electric automobile of series hybrid powder AC/DC convertor |
CN109978229A (en) * | 2019-02-12 | 2019-07-05 | 常伟 | The method that the full battery core multi-point temperature of a kind of pair of power battery pack and tie point temperature carry out thermal runaway prediction |
CN110929798A (en) * | 2019-11-29 | 2020-03-27 | 重庆邮电大学 | Image classification method and medium based on structure optimization sparse convolution neural network |
CN111007401A (en) * | 2019-12-16 | 2020-04-14 | 国网江苏省电力有限公司电力科学研究院 | Electric vehicle battery fault diagnosis method and device based on artificial intelligence |
CN111861013A (en) * | 2020-07-23 | 2020-10-30 | 长沙理工大学 | Power load prediction method and device |
CN114065613A (en) * | 2021-10-27 | 2022-02-18 | 中国华能集团清洁能源技术研究院有限公司 | Multi-working-condition process industrial fault detection and diagnosis method based on deep migration learning |
CN113987129A (en) * | 2021-11-08 | 2022-01-28 | 重庆邮电大学 | Digital media protection text steganography method based on variational automatic encoder |
CN114155270A (en) * | 2021-11-10 | 2022-03-08 | 南方科技大学 | Pedestrian trajectory prediction method, device, equipment and storage medium |
CN115356636A (en) * | 2022-07-31 | 2022-11-18 | 常伟 | Data-driven new energy automobile battery fault alarm and fault early warning model |
CN116434741A (en) * | 2023-03-09 | 2023-07-14 | 平安科技(深圳)有限公司 | Speech recognition model training method, device, computer equipment and storage medium |
CN116842337A (en) * | 2023-06-13 | 2023-10-03 | 国网甘肃省电力公司 | Transformer fault diagnosis method based on LightGBM (gallium nitride based) optimal characteristics and COA-CNN (chip on board) model |
CN116626505A (en) * | 2023-07-21 | 2023-08-22 | 江苏海平面数据科技有限公司 | Battery pack consistency anomaly detection method based on Internet of vehicles big data |
CN117067920A (en) * | 2023-10-18 | 2023-11-17 | 北京航空航天大学 | Fault detection method and device of power battery, electronic equipment and electric automobile |
CN117067921A (en) * | 2023-10-18 | 2023-11-17 | 北京航空航天大学 | Fault detection method of electric automobile and electric automobile |
Non-Patent Citations (1)
Title |
---|
刘宏阳;杨林;李济霖;: "基于长短期记忆网络的电动汽车电池故障诊断", 机电一体化, no. 1, 15 April 2020 (2020-04-15), pages 18 - 24 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117907845A (en) * | 2024-03-20 | 2024-04-19 | 山东泰开电力电子有限公司 | Electrochemical energy storage system insulation detection method based on electrical parameter analysis |
CN117907845B (en) * | 2024-03-20 | 2024-05-17 | 山东泰开电力电子有限公司 | Electrochemical energy storage system insulation detection method based on electrical parameter analysis |
Also Published As
Publication number | Publication date |
---|---|
CN117261599B (en) | 2024-05-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Li et al. | Battery fault diagnosis for electric vehicles based on voltage abnormality by combining the long short-term memory neural network and the equivalent circuit model | |
Gan et al. | Data-driven fault diagnosis of lithium-ion battery overdischarge in electric vehicles | |
CN112858919B (en) | Battery system online fault diagnosis method and system based on cluster analysis | |
Hannan et al. | Neural network approach for estimating state of charge of lithium-ion battery using backtracking search algorithm | |
CN117067920B (en) | Fault detection method and device of power battery, electronic equipment and electric automobile | |
CN117067921B (en) | Fault detection method of electric automobile and electric automobile | |
CN117261599B (en) | Fault detection method and device of electric automobile, electronic equipment and electric automobile | |
CN111257753B (en) | Battery system fault diagnosis method | |
CN115859058B (en) | UPS fault prediction method and system based on width learning network | |
Zhang et al. | State‐of‐charge estimation of lithium‐ion battery pack based on improved RBF neural networks | |
CN112329350A (en) | Airplane lead-acid storage battery abnormity detection semi-supervision method based on isolation | |
Li et al. | Lithium-ion battery remaining useful life prognostics using data-driven deep learning algorithm | |
CN113255764A (en) | Method, system and device for detecting electrochemical energy storage system fault by using machine learning | |
Harippriya et al. | Battery management system to estimate battery aging using deep learning and machine learning algorithms | |
CN116298936A (en) | Intelligent lithium ion battery health state prediction method in incomplete voltage range | |
CN114881410A (en) | Model-data hybrid driven power system transient stability online evaluation method | |
CN116643178A (en) | SOC estimation method and related device of battery management system | |
Li et al. | Multi-dimension statistical analysis and selection of safety-representing features for battery pack in real-world electric vehicles | |
Machlev | EV battery fault diagnostics and prognostics using deep learning: Review, challenges & opportunities | |
CN116859259A (en) | Battery health state estimation method based on integrated learning and real vehicle big data | |
CN116679213A (en) | SOH estimation method for electric vehicle power battery based on integrated deep learning | |
CN115865421A (en) | Intrusion detection method and system for power distribution network information system | |
Li et al. | An ensemble hybrid model with outlier detection for prediction of lithium-ion battery remaining useful life | |
CN115718260A (en) | Battery fault prediction method and system based on combination of CNN-LSTM and correlation coefficient | |
CN114881162A (en) | Method, apparatus, device and medium for predicting failure of metering automation master station |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |