An early warning method and device for a battery

Abstract

An embodiment of the present invention discloses a warning method and device for a battery. The method includes: obtaining target battery data corresponding to a battery to be detected; for each battery to be detected, using the target battery data corresponding to the battery to be detected and a preset battery fault detection algorithm to determine first fault detection information corresponding to the battery to be detected; using the target battery data corresponding to all batteries to be detected and a target battery data detection model to determine second fault detection information corresponding to each battery to be detected; based on the first fault detection information and the second fault detection information, determining fault verification redundancy information corresponding to each battery to be detected; and outputting the first fault detection information, the second fault detection information, and the fault verification redundancy information to achieve accurate and safe warning of a power battery and improve the safety of battery use.

Description

Technical Field

[0001] The present invention relates to the technical field of battery safety detection, and in particular, to a warning method and device for a battery. Background Art

[0002] Under the dual pressures of energy shortage and environmental pollution problems, the application of new energy has become an irreversible trend in the development of science and technology. The electrification of the automotive power system has gradually become the main trend in the future development of automotive technology. One of the main characteristics of the electrification of the automotive power system is the use of electrical energy instead of chemical energy as the main driving energy source for vehicles. Correspondingly, secondary batteries such as lithium-ion power batteries have become the main source of power for electric vehicles.

[0003] However, in recent years, with the gradual demonstration and application of electric vehicles, safety accidents of power batteries characterized by thermal runaway have occurred from time to time. Accidents of power batteries such as lithium-ion power batteries usually manifest as phenomena such as a sudden increase in temperature, smoking, fire, and even explosion centered around thermal runaway. Thermal runaway accidents will hit the public's confidence in accepting electric vehicles and hinder the popularization of electric vehicles.

[0004] The occurrence of thermal runaway accidents in power batteries may be triggered by various inducements, such as mechanical abuse, thermal abuse, and electrical abuse, which cause abnormal electrochemical potentials inside the power battery during the charge and discharge process, and then induce abnormal growth of metal dendrites. The growth of dendrites will eventually pierce the separator, resulting in an internal short circuit in the power battery, and then a thermal runaway accident occurs.

[0005] For self-initiated internal short circuits caused by dendrite growth, manufacturing defects, or metal impurities, etc., there is a long development and evolution process before they trigger thermal runaway, which provides the possibility for detecting the internal short circuit of power batteries, and thus provides the possibility for the safety warning of power batteries. Summary of the Invention

[0006] The present invention provides a warning method and device for a battery to achieve accurate safety warning of power batteries and improve the safety of battery use. The specific technical solutions are as follows:

[0007] In a first aspect, an embodiment of the present invention provides a warning method for a battery, the method including:

[0008] Obtaining target battery data corresponding to a battery to be detected;

[0009] For each battery to be detected, using the target battery data corresponding to the battery to be detected and a preset battery fault detection algorithm to determine first fault detection information corresponding to the battery to be detected;

[0010] Using the target battery data corresponding to all the batteries to be detected and the target battery data detection model, determine the second fault detection information corresponding to each battery to be detected, where the target battery data detection model is: a target battery data detection model obtained by training the training data obtained after performing a preset data cleaning operation on each sample battery data and its corresponding calibration information, the sample dimension and the feature dimension of each training data are the same, and the preset cleaning operation at least includes: a deletion and supplementation operation on outliers and / or missing values in the sample battery data;

[0011] Based on the first fault detection information and the second fault detection information, determine the fault verification redundancy information corresponding to each battery to be detected;

[0012] Output the first fault detection information, the second fault detection information, and the fault verification redundancy information.

[0013] Optionally, the step of, for each battery to be detected, using the target battery data corresponding to the battery to be detected and a preset battery fault detection algorithm to determine the first fault detection information corresponding to the battery to be detected includes:

[0014] For each battery to be detected, use the target battery data corresponding to the battery to be detected and a preset three-level early warning algorithm to determine the first fault information corresponding to the battery to be detected;

[0015] For each battery to be detected, use the first specified feature data in the target battery data corresponding to the battery to be detected and a preset performance evaluation algorithm to determine the performance score information corresponding to the battery to be detected;

[0016] For each battery to be detected, use the second specified feature data in the target battery data corresponding to the battery to be detected and a preset short-circuit detection algorithm to determine the short-circuit detection information corresponding to the battery to be detected;

[0017] For each battery to be detected, based on the performance score information and the short-circuit detection information corresponding to the battery to be detected, determine the second fault information corresponding to the battery to be detected to obtain the first fault detection information.

[0018] Optionally, the target battery data corresponding to the battery to be detected includes: the original battery data groups corresponding to each charging process or each discharging process of the battery to be detected;

[0019] The step of using the target battery data corresponding to all the batteries to be detected and the target battery data detection model to determine the second fault detection information corresponding to each battery to be detected includes:

[0020] For each original battery data group corresponding to each battery to be detected, perform the preset data cleaning operation on the original battery data group to obtain the data to be utilized corresponding to the original battery data group;

[0021] Input the data to be utilized corresponding to each battery to be detected into the target battery data detection model to determine the third fault information corresponding to each data to be utilized, so as to determine the second fault detection information corresponding to the battery to be detected.

[0022] Optionally, the original battery data group includes: battery data corresponding to the basic features generated at each moment corresponding to each timestamp during the charging or discharging process of the battery to be detected; the preset data cleaning operation further includes: a feature construction operation;

[0023] The step of performing the preset data cleaning operation on the original battery data group to obtain the data to be utilized corresponding to the original battery data group includes:

[0024] Traverse the battery data corresponding to each timestamp in the original battery data group, fill or delete the battery data corresponding to the timestamp with data missing and / or data anomaly situations, to obtain the intermediate battery data corresponding to the original battery data group, wherein, if the number of missing and / or abnormal data of the battery data corresponding to the timestamp with data missing and / or data anomaly situations is not higher than the first preset value, fill the battery data corresponding to the timestamp with data missing and / or data anomaly situations; if the number of missing and / or abnormal data of the battery data corresponding to the timestamp with data missing and / or data anomaly situations is higher than the first preset value, delete the battery data corresponding to the timestamp with data missing and / or data anomaly situations;

[0025] Based on the data division intervals corresponding to the specified basic features and the battery data corresponding to the specified basic features in the intermediate battery data, divide the intermediate battery data to obtain the battery data segments corresponding to each data division interval;

[0026] For each battery data segment, based on the battery data corresponding to other basic features in the battery data segment, determine the data corresponding to each preset extended feature as the feature data corresponding to the battery data segment, wherein the other basic features are the features other than the specified basic feature among the basic features;

[0027] Based on the feature data corresponding to the battery data segment, determine the data to be utilized corresponding to the original battery data group.

[0028] Optionally, for each battery data segment, the step of determining the battery data corresponding to each preset extended feature based on the battery data corresponding to other basic features in the battery data segment as the feature data corresponding to the battery data segment includes:

[0029] For each battery data segment, based on whether there is battery data in the battery data segment, determine the feature data corresponding to the indicative feature corresponding to the battery data segment;

[0030] For each other basic feature in each battery data segment, based on the battery data corresponding to the other basic feature in the battery data segment, determine the feature data corresponding to the statistical feature corresponding to the other basic feature, where the statistical feature includes: a feature indicating calculating the mean value and / or variance of the battery data corresponding to each other basic feature based on the battery data corresponding to each other basic feature;

[0031] For each battery data segment, based on the battery data corresponding to the current feature and the battery data corresponding to the voltage feature in the battery data segment, determine the feature data corresponding to the resistance feature corresponding to the battery data segment to obtain the feature data corresponding to the battery data segment, where the current feature and the voltage feature belong to other basic features.

[0032] Optionally, before the step of inputting each piece of data to be utilized corresponding to each battery to be detected into the target battery data detection model to determine the third fault information corresponding to each piece of data to be utilized to determine the second fault detection information corresponding to the battery to be detected, the method further includes:

[0033] The process of training the target battery data detection model, where the process includes:

[0034] Obtain various sample battery data and their corresponding calibration information, where the sample battery data is: data generated by each sample battery during the charging process or the discharging process; when the sample battery data is data generated during the charging process, the original battery data group is data generated during the charging process; when the sample battery data is data generated during the discharging process, the original battery data group is data generated during the discharging process, and the calibration information is used to calibrate whether the corresponding sample battery data represents whether the corresponding sample battery has a fault and the type of fault that occurs;

[0035] For each sample battery data, perform the preset data cleaning operation on the sample battery data to obtain the sample battery data after the preset data cleaning operation as training data;

[0036] Obtain multiple initial battery data detection models;

[0037] For each initial battery data detection model, use the first training data in the training data to train the initial battery data detection model to obtain a trained battery data detection model, where the first training data is part of the training data;

[0038] Use the second training data in the training data to determine the battery data detection model with the best detection result from multiple trained battery data detection models as the target battery data detection model, where the second training data is part of the training data.

[0039] Optionally, the first fault detection information includes: determining the first fault information corresponding to each battery to be detected based on the target battery data corresponding to each battery to be detected and a three - level early warning algorithm, and determining the second fault information corresponding to the battery to be detected based on the target battery data corresponding to each battery to be detected, a preset performance evaluation algorithm, and a preset short - circuit detection algorithm;

[0040] The step of determining the fault verification redundancy information corresponding to each battery to be detected based on the first fault detection information and the second fault detection information includes:

[0041] For each battery to be detected, determine the verification redundancy information between the first fault information and the second fault information based on the first fault information and the second fault information corresponding to the battery to be detected;

[0042] For each battery to be detected, determine the verification redundancy information between the second fault information and the second fault detection information based on the second fault information and the second fault detection information corresponding to the battery to be detected to determine the fault verification redundancy information corresponding to each battery to be detected.

[0043] In a second aspect, an embodiment of the present invention provides a warning device for a battery, and the device includes;

[0044] A first acquisition module configured to acquire target battery data corresponding to a battery to be detected;

[0045] A first determination module configured to, for each battery to be detected, determine the first fault detection information corresponding to the battery to be detected by using the target battery data corresponding to the battery to be detected and a preset battery fault detection algorithm;

[0046] A second determination module, configured to use the target battery data corresponding to all the batteries to be detected and a target battery data detection model to determine the second fault detection information corresponding to each battery to be detected, where the target battery data detection model is: a target battery data detection model obtained by training the training data obtained after performing a preset data cleaning operation on each sample battery data and its corresponding calibration information, and the sample dimensions and feature dimensions of each training data are the same, and the preset cleaning operation at least includes: a deletion and supplementation operation for outliers and / or missing values in the sample battery data;

[0047] A third determination module, configured to determine the fault verification redundancy information corresponding to each battery to be detected based on the first fault detection information and the second fault detection information;

[0048] An output module, configured to output the first fault detection information, the second fault detection information, and the fault verification redundancy information.

[0049] Optionally, the first determination module is specifically configured to, for each battery to be detected, use the target battery data corresponding to the battery to be detected and a preset three-level early warning algorithm to determine the first fault information corresponding to the battery to be detected;

[0050] For each battery to be detected, use the first specified feature data in the target battery data corresponding to the battery to be detected and a preset performance evaluation algorithm to determine the performance score information corresponding to the battery to be detected;

[0051] For each battery to be detected, use the second specified feature data in the target battery data corresponding to the battery to be detected and a preset short-circuit detection algorithm to determine the short-circuit detection information corresponding to the battery to be detected;

[0052] For each battery to be detected, based on the performance score information and the short-circuit detection information corresponding to the battery to be detected, determine the second fault information corresponding to the battery to be detected, so as to obtain the first fault detection information.

[0053] Optionally, the target battery data corresponding to the battery to be detected includes: an original battery data set corresponding to each charging process or each discharging process of the battery to be detected;

[0054] The second determination module includes:

[0055] A cleaning unit, configured to, for each original battery data set corresponding to each battery to be detected, perform the preset data cleaning operation on the original battery data set to obtain the data to be used corresponding to the original battery data set;

[0056] An input determination unit, configured to input each piece of data to be utilized corresponding to each battery to be detected into a target battery data detection model, determine third fault information corresponding to each piece of data to be utilized, so as to determine second fault detection information corresponding to the battery to be detected.

[0057] Optionally, the original battery data group includes: battery data corresponding to basic features generated at each moment corresponding to each timestamp during the charging process or discharging process of the battery to be detected; the preset data cleaning operation further includes: a feature construction operation.

[0058] The cleaning unit is specifically configured to traverse the battery data corresponding to each timestamp in the original battery data group, fill or delete the battery data corresponding to the timestamp with data missing and / or data anomaly, to obtain intermediate battery data corresponding to the original battery data group, wherein, if the number of missing and / or abnormal battery data corresponding to the timestamp with data missing and / or data anomaly is not higher than a first preset value, then fill the battery data corresponding to the timestamp with data missing and / or data anomaly; if the number of missing and / or abnormal battery data corresponding to the timestamp with data missing and / or data anomaly is higher than the first preset value, then delete the battery data corresponding to the timestamp with data missing and / or data anomaly.

[0059] Based on the data division intervals corresponding to the specified basic features and the battery data corresponding to the specified basic features in the intermediate battery data, divide the intermediate battery data to obtain battery data segments corresponding to each data division interval.

[0060] For each battery data segment, based on the battery data corresponding to other basic features in the battery data segment, determine the data corresponding to each preset extended feature as the feature data corresponding to the battery data segment, wherein the other basic features are the features in the basic features except the specified basic feature.

[0061] Based on the feature data corresponding to the battery data segment, determine the data to be utilized corresponding to the original battery data group.

[0062] Optionally, the cleaning unit is specifically configured to, for each battery data segment, based on whether there is battery data in the battery data segment, determine the feature data corresponding to the indicative feature corresponding to the battery data segment.

[0063] For each other basic feature in each battery data segment, based on the battery data corresponding to the other basic feature in the battery data segment, determine the feature data corresponding to the statistical feature corresponding to the other basic feature, where the statistical feature includes: a feature indicating to calculate the mean and / or variance of the battery data corresponding to each other basic feature based on the battery data corresponding to each other basic feature;

[0064] For each battery data segment, based on the battery data corresponding to the current feature and the battery data corresponding to the voltage feature in the battery data segment, determine the feature data corresponding to the resistance feature corresponding to the battery data segment, so as to obtain the feature data corresponding to the battery data segment, where the current feature and the voltage feature belong to other basic features.

[0065] Optionally, the device further includes:

[0066] A model training module, configured to train the target battery data detection model before inputting each piece of data to be utilized corresponding to each battery to be detected into the target battery data detection model to determine the third fault information corresponding to each piece of data to be utilized, so as to determine the second fault detection information corresponding to the battery to be detected, where the model training module is specifically configured to obtain each sample battery data and its corresponding calibration information, where the sample battery data is: the data generated by each sample battery during the charging process or the discharging process; when the sample battery data is the data generated during the charging process, the original battery data group is the data generated during the charging process; when the sample battery data is the data generated during the discharging process, the original battery data group is the data generated during the discharging process, and the calibration information is used to calibrate whether the corresponding sample battery data represents whether the corresponding sample battery has a fault and the type of the fault that occurs;

[0067] For each sample battery data, perform the preset data cleaning operation on the sample battery data to obtain the sample battery data after the preset data cleaning operation as training data;

[0068] Obtain multiple initial battery data detection models;

[0069] For each initial battery data detection model, use the first training data in the training data to train the initial battery data detection model to obtain the trained battery data detection model, where the first training data is a part of the training data;

[0070] Use the second training data in the training data to determine the battery data detection model with the best detection result from multiple trained battery data detection models as the target battery data detection model, where the second training data is a part of the training data.

[0071] Optionally, the first fault detection information includes: determining first fault information corresponding to each battery to be detected based on the target battery data corresponding to each battery to be detected and a three-level early warning algorithm, and determining second fault information corresponding to the battery to be detected based on the target battery data corresponding to each battery to be detected, a preset performance evaluation algorithm, and a preset short-circuit detection algorithm;

[0072] The third determination module is specifically configured to, for each battery to be detected, determine check redundancy information between the first fault information and the second fault information based on the first fault information and the second fault information corresponding to the battery to be detected;

[0073] For each battery to be detected, determine check redundancy information between the second fault information and the second fault detection information based on the second fault information and the second fault detection information corresponding to the battery to be detected, so as to determine check redundancy information of faults corresponding to each battery to be detected.

[0074] As can be seen from the above, an early warning method and device for a battery provided by an embodiment of the present invention obtain target battery data corresponding to a battery to be detected; for each battery to be detected, determine first fault detection information corresponding to the battery to be detected by using the target battery data corresponding to the battery to be detected and a preset battery fault detection algorithm; determine second fault detection information corresponding to each battery to be detected by using the target battery data corresponding to all batteries to be detected and a target battery data detection model, where the target battery data detection model is: a target battery data detection model obtained by training training data obtained after preset data cleaning operations on each sample battery data and its corresponding calibration information, the sample dimensions and feature dimensions of each training data are the same, and the preset cleaning operation at least includes: deleting and supplementing outliers and / or missing values in the sample battery data; determining check redundancy information of faults corresponding to each battery to be detected based on the first fault detection information and the second fault detection information; outputting the first fault detection information, the second fault detection information, and the check redundancy information of faults.

[0075] Applying the embodiments of the present invention, the training of the target battery data detection model is realized through training data with consistent sample dimensions and feature dimensions. Then, through different fault detection algorithms including the target battery data detection model and the battery data corresponding to the battery to be detected collected, the fault detection of the battery to be detected is realized, and multiple fault detection results are obtained. The multiple fault detection results are mutually verified to obtain and output fault verification redundancy information, so as to better reflect the fault detection results of the battery to be detected, and then realize the safety warning of the battery, and improve the accuracy of the safety warning to a certain extent, and improve the safety of battery use. Of course, any product or method implementing the present invention does not necessarily need to achieve all the above-mentioned advantages at the same time.

[0076] The innovation points of the embodiments of the present invention include:

[0077] 1. Through different fault detection algorithms including the target battery data detection model trained by training data with consistent sample dimensions and feature dimensions, and the battery data corresponding to the battery to be detected collected, the fault detection of the battery to be detected is realized, and multiple fault detection results are obtained. The multiple fault detection results are mutually verified to obtain and output fault verification redundancy information, so as to better reflect the fault detection results of the battery to be detected, and then realize the safety warning of the battery, and improve the accuracy of the safety warning to a certain extent, and improve the safety of battery use.

[0078] 2. Fill or delete the battery data in the original battery data group, and divide the interval segments. Feature expansion is performed on the battery data segments in different interval segments to obtain the data to be utilized with sample dimensions and feature dimensions respectively consistent with those of the training data, providing a basis for the detection using the target battery data detection model, and increasing the features of the battery data through feature expansion, and improving the accuracy of the detection results of the target battery data detection model to a certain extent.

[0079] 3. Provide specific feature types for feature expansion to realize the feature expansion of the battery data from different angles, so as to comprehensively represent the battery to be detected through the expanded feature data.

[0080] 4. Train different battery data detection models through training data with consistent sample dimensions and feature dimensions obtained after a preset data cleaning operation, and obtain the battery data detection model with the best detection result as the target battery data detection model, providing a basis for the accuracy of the fault detection of the battery.

[0081] 5. Cross-check among multiple different fault detection messages to obtain fault check redundancy information, so as to accurately obtain the true fault information of the battery to be detected. Furthermore, combined with the true fault information of the battery, the accuracy of safety warning can be improved to a certain extent, and the safety of battery use can be enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0083] Figure 1 It is a schematic flowchart of a method for warning of a battery provided by an embodiment of the present invention;

[0084] Figure 2 It is a schematic diagram of a training process of a target battery data detection model provided by an embodiment of the present invention;

[0085] Figure 3 It is a schematic structural diagram of a warning device for a battery provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0086] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0087] It should be noted that the terms "include" and "have" and any variations thereof in the embodiments of the present invention and the accompanying drawings are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices.

[0088] The present invention provides a method and device for warning of a battery to achieve accurate safety warning of a power battery and improve the safety of battery use. The following will elaborate on the embodiments of the present invention.

[0089] Figure 1 It is a schematic flowchart of a method for warning of a battery provided by an embodiment of the present invention. The method may include the following steps:

[0090] S101: Obtain the target battery data corresponding to the battery to be detected.

[0091] The battery warning method provided by the embodiments of the present invention is applied to a cloud platform. In one implementation, the functional software for implementing the battery warning method may exist in the form of a separate client software or in the form of a plugin of the current relevant client software, which is acceptable.

[0092] The cloud platform can communicate with multiple batteries to be detected, that is, obtain the data generated by the multiple batteries to be detected during the charging and discharging processes. In one implementation manner, the battery to be detected may be a battery set for objects such as electric vehicles, electric bicycles, and charging piles that require electric energy to provide power. Correspondingly, the cloud platform can communicate with the terminal platform set for the object that requires electric energy to provide power. That is, the terminal platform can monitor the charging and discharging processes of the battery to be detected, obtain the data generated during the charging and discharging processes and the data calculated based on the data generated during its charging and discharging processes as the initial battery data, and send it to the cloud platform; the cloud platform can obtain the initial battery data, perform pre-cleaning and store it. Among them, the pre-cleaning may be to classify and store the obtained battery data.

[0093] The initial battery data corresponding to the battery to be detected may include, but is not limited to: the charging ID number corresponding to the battery to be detected, the timestamp, and the total voltage of the battery pack, the total current of the battery pack, the cumulative mileage (odometer / GPS information), the SOC (State of Charge) of the battery pack, the charging and discharging state, the lowest temperature corresponding to the battery cells in the battery pack, the highest temperature corresponding to the battery cells in the battery pack, the lowest value of the battery cell voltage, the highest value of the battery cell voltage, the code of the battery cell with the lowest voltage, the code of the battery cell with the highest voltage, the code of the temperature probe with the lowest temperature, the code of the temperature probe with the highest temperature, the voltage values of all battery cells, and the temperature values of all temperature probes. If the battery to be detected is set in a vehicle, the battery data may further include: the vehicle Vin code of the vehicle where the battery pack is located. The timestamp may refer to the acquisition moment of the battery data.

[0094] In one case, the initial battery data corresponding to the battery to be detected may further include the initial fault information detected by the corresponding terminal platform. The fault types of the initial fault information may include, but are not limited to: overcharge fault, over-discharge fault, over-temperature fault, and insulation fault, etc. Among them, the overcharge fault may refer to that the charging current or voltage is instantaneously too large, the over-discharge fault may refer to that the discharge current or voltage is instantaneously too large, and the over-temperature fault may refer to that the battery temperature is too high during the charging and discharging processes.

[0095] The battery to be detected may be a rechargeable battery such as a lithium battery.

[0096] The cloud platform can obtain the target battery data corresponding to the battery to be detected in real time or periodically. Among them, the target battery data can be part or all of the above initial battery data. In one implementation, the target battery data corresponding to the battery to be detected may include: the battery data of the specified feature obtained within the current time and a preset duration forward. The specified feature may include, but is not limited to, timestamp, the total voltage of the battery to be detected (i.e., the battery pack), the total current of the battery pack, the SOC (State of Charge) of the battery pack, the charge and discharge state, the lowest temperature corresponding to the battery cells in the battery pack, the highest temperature corresponding to the battery cells in the battery pack, the lowest value of the battery cell voltage, the highest value of the battery cell voltage, the code of the battery cell with the lowest voltage, the code of the battery cell with the highest voltage, the code of the lowest temperature probe, the code of the highest temperature probe, the voltage values of all battery cells, and the temperature values of all temperature probes. The data in the target battery data corresponding to the battery to be detected are sorted in the order of the timestamps corresponding to them.

[0097] S102: For each battery to be detected, use the target battery data corresponding to the battery to be detected and a preset battery fault detection algorithm to determine the first fault detection information corresponding to the battery to be detected.

[0098] In one case, the preset battery fault detection algorithm may include, but is not limited to, a direct fault detection algorithm and a fault detection algorithm based on battery mechanism and related battery models.

[0099] Among them, the battery mechanism may include, but is not limited to, other battery mechanisms such as the double-tank mechanism, and the related battery models may include, but is not limited to, other battery models such as the equivalent circuit model, the electrochemical model, and the empirical model. Through the above battery mechanism and related battery models, the corresponding operating process of the corresponding battery can be simulated. Furthermore, the corresponding operating process of the battery can be monitored to determine the fault result of the battery through the corresponding fault detection algorithm.

[0100] In one case, the fault detection algorithm based on battery mechanism and related battery models may include, but is not limited to, other algorithms such as the internal short circuit diagnosis algorithm, the lithium plating detection algorithm, the consistency scoring algorithm, and the thermal runaway warning algorithm in the related technologies.

[0101] In one implementation, the preset battery fault detection algorithm includes a direct fault detection algorithm. The target battery data corresponding to each battery to be detected may include: the initial fault information reported by the terminal platform corresponding to the battery to be detected. The cloud platform can directly screen out the initial fault information from the target battery data corresponding to each battery to be detected for each battery to be detected; and use the preset three-level early warning algorithm and the target battery data corresponding to the battery to be detected to determine other fault information, so as to determine the direct fault detection information corresponding to the battery to be detected, that is, the first fault detection information. Among them, the preset three-level early warning algorithm can be any type of three-level early warning algorithm in the related art that can implement multi-layer detection of the safety of the battery to be detected, and this embodiment does not make any limitations.

[0102] In one embodiment, the preset battery fault detection algorithm includes a fault detection algorithm based on battery mechanism and related battery models. Correspondingly, for each battery to be detected, the cloud platform uses the first specified feature data in the target battery data corresponding to the battery to be detected and the preset performance evaluation algorithm to determine the performance score information corresponding to the battery to be detected. For each battery to be detected, the second specified feature data in the target battery data corresponding to the battery to be detected and the preset short-circuit detection algorithm are used to determine the short-circuit detection information corresponding to the battery to be detected. For each battery to be detected, based on the performance score information and short-circuit detection information corresponding to the battery to be detected, the second fault information corresponding to the battery to be detected is determined. That is, the first fault detection information including the first fault information and the second fault information is obtained.

[0103] Among them, the first specified feature data includes but is not limited to: time stamps and the voltages, currents, temperatures, internal resistances, capacities, and powers of each battery cell in the battery to be detected corresponding to the time stamps, etc. The second specified feature data includes but is not limited to: the capacity, voltage, state of charge of the battery cell, and the corresponding time stamps, etc. The battery capacity, power, and internal resistance of the battery cell can be determined by using the corresponding current and / or voltage of the battery cell and its corresponding time stamp.

[0104] Through the preset performance evaluation algorithm and the first specified feature data corresponding to the battery to be detected, the consistency scores corresponding to the first specified feature data corresponding to the battery to be detected can be determined. Furthermore, based on the consistency thresholds corresponding to the first specified feature data and the consistency scores corresponding to the first specified feature data corresponding to the battery to be detected, information indicating whether a failure occurs in the first specified feature data corresponding to the battery to be detected is determined, that is, performance scoring information. For example, the consistency scores corresponding to the first specified features corresponding to the first specified feature data may include: voltage consistency scores, temperature consistency scores, internal resistance consistency scores, capacity consistency scores, and power consistency scores of each single battery in the battery to be detected. Among them, the greater the difference between the first specified feature data corresponding to each battery cell in the battery to be detected, correspondingly, the lower the consistency score corresponding to the first specified feature; conversely, the smaller the difference between the first specified feature data corresponding to each battery cell in the battery to be detected, correspondingly, the higher the consistency score corresponding to the first specified feature. Among them, the preset performance evaluation algorithm may be: an evaluation algorithm for the consistency scores of the first specified features of each battery cell in the battery to be detected in the related art, and the embodiments of the present invention do not make any limitations.

[0105] Based on the preset short-circuit detection algorithm and the second specified feature data corresponding to the battery to be detected, it can be determined whether the battery to be detected has a short circuit. Among them, the preset short-circuit detection algorithm may include, but is not limited to, algorithms for detecting micro short circuits of the battery and algorithms for detecting internal short circuits of the battery, etc., which can implement short-circuit detection of the battery, and the embodiments of the present invention do not make any limitations.

[0106] The performance scoring information and short-circuit detection information corresponding to the battery to be detected are determined as the second fault information corresponding to the battery to be detected, and the first fault information and the second fault information are determined as the first fault detection information.

[0107] S103: Using the target battery data corresponding to all batteries to be detected and the target battery data detection model, determine the second fault detection information corresponding to each battery to be detected.

[0108] Among them, the target battery data detection model is: a target battery data detection model obtained by training the training data obtained after the preset data cleaning operation on each sample battery data and its corresponding calibration information, the sample dimensions and feature dimensions of each training data are the same, and the preset cleaning operation at least includes: deletion and supplementation operations for outliers and / or missing values in the sample battery data.

[0109] The target battery data corresponding to each battery to be detected may include the data corresponding to each charging process or each discharging process of the battery to be detected. By using the data corresponding to each charging or discharging process in the target battery data corresponding to all batteries to be detected and the target battery data detection model, it can be determined whether a fault occurs in each charging or discharging process of each battery to be detected, and the type of the fault that occurs, so as to obtain the second fault detection information corresponding to each battery to be detected.

[0110] The fault detection types corresponding to the second fault detection information may include, but are not limited to: battery temperature anomaly detection, battery terminal high temperature detection, battery rapid temperature rise detection, battery pack overvoltage detection, battery pack voltage increase detection, battery pack undervoltage detection, battery cell overvoltage and undervoltage detection, and SOC low and too low and extremely low detection.

[0111] In another embodiment of the present invention, the target battery data corresponding to the battery to be detected includes: the original battery data set corresponding to each charging process or each discharging process of the battery to be detected;

[0112] The step S103 may include the following steps 021-022:

[0113] 021: For each original battery data set corresponding to each battery to be detected, perform a preset data cleaning operation on the original battery data set to obtain the data to be utilized corresponding to the original battery data set.

[0114] 022: Input the data to be utilized corresponding to each battery to be detected into the target battery data detection model, determine the third fault information corresponding to each data to be utilized, so as to determine the second fault detection information corresponding to the battery to be detected.

[0115] In this implementation manner, the preset data cleaning operation may include a feature deletion and supplementation operation, or a feature deletion and supplementation operation and a feature construction operation. When the preset data cleaning operation includes a feature deletion and supplementation operation, the cloud platform may perform a filling or deletion operation on each original battery data set corresponding to each battery to be detected, so as to obtain the data to be utilized corresponding to the original battery data set whose feature dimension and sample dimension are respectively consistent with the feature dimension and sample dimension of the training data for training the target battery data detection model. When the preset data cleaning operation includes a feature deletion and supplementation operation and a feature construction operation, the cloud platform may perform a filling or deletion operation on each original battery data set corresponding to each battery to be detected, and then perform a feature construction operation, so as to obtain the data to be utilized corresponding to the original battery data set whose feature dimension and sample dimension are respectively consistent with the feature dimension and sample dimension of the training data for training the target battery data detection model.

[0116] Furthermore, input each piece of data to be utilized corresponding to each battery to be detected into the target battery data detection model to obtain the abnormal probability corresponding to each piece of data to be utilized. Use the abnormal probability corresponding to each piece of data to be utilized and the corresponding abnormal probability threshold to determine the fault result corresponding to each piece of data to be utilized. This fault result can characterize whether a fault occurs and the type of fault during a certain charging process or a certain discharging process of the battery to be detected, so as to determine the third fault information corresponding to each piece of data to be utilized, that is, determine the second fault detection information corresponding to each battery to be detected.

[0117] S104: Based on the first fault detection information and the second fault detection information, determine the fault verification redundancy information corresponding to each battery to be detected.

[0118] In this step, the first fault detection information and the second fault detection information can be compared to determine the fault results corresponding to various types of fault detections, so as to obtain the fault verification redundancy information corresponding to each battery to be detected.

[0119] The fault detection of the battery in the embodiment of the present invention includes but is not limited to: battery temperature abnormality detection, battery terminal post high temperature detection, battery rapid temperature rise detection, battery pack overvoltage detection, battery pack voltage increase detection, battery pack undervoltage detection, battery single cell overvoltage and undervoltage detection, and SOC low, too low, and extremely low detection.

[0120] S105: Output the first fault detection information, the second fault detection information, and the fault verification redundancy information.

[0121] The cloud platform outputs the first fault detection information, the second fault detection information, and the fault verification redundancy information. Specifically, it can be: display the first fault detection information, the second fault detection information, and the fault verification redundancy information through the connected display device. Among them, the first fault detection information, the second fault detection information, and the fault verification redundancy information can be output in the form of a table to facilitate subsequent staff to view the fault problems of each battery to be detected. Or, it can be: the cloud platform sends the first fault detection information, the second fault detection information, and the fault verification redundancy information corresponding to each battery to be detected to the terminal platform corresponding to each battery to be detected for display through the display device of the terminal platform, so as to facilitate the users of each battery to be detected to view the fault problems of their batteries to be detected. Realize phased, hierarchical, and multi-signal early warning for the users of the batteries to be detected.

[0122] In one case, the second fault detection information determined based on the target battery data detection model can be used as the first-level fault warning. The output of the first-level fault warning requires a high detection rate, allows a high false alarm rate, a low fault level, and early warning. The second fault information is used as the second-level fault warning. The output of the second-level warning requires a high detection rate, a low false alarm rate, a higher fault level, and early warning, which plays a guiding role in the fault diagnosis result based on the big data algorithm and reduces the fault false alarm rate. The first fault information is used as the third-level fault warning. The output of the third-level fault warning requires a high detection rate, allows a high false alarm rate, and a lower fault level, and can achieve real-time alarm.

[0123] In one implementation, among the above first fault detection information and second fault detection information, for the fault detection results of the same fault type, if two of the fault information indicate that the battery to be detected has a fault of this fault type, while the other fault information indicates that the battery to be detected does not have a fault of this fault type, it can be considered that the parameters used in the fault detection process corresponding to this fault information are inaccurate, and the corresponding parameters can be appropriately adjusted and updated.

[0124] Applying the embodiments of the present invention, the training of the target battery data detection model is realized through training data with consistent sample dimensions and feature dimensions. Then, through different fault detection algorithms including the target battery data detection model and the battery data corresponding to the battery to be detected collected, the fault detection of the battery to be detected is realized, and multiple fault detection results are obtained. The multiple fault detection results are mutually verified to obtain and output fault verification redundancy information to better reflect the fault detection results of the battery to be detected, thereby realizing the safety warning of the battery and improving the accuracy of the safety warning to a certain extent, and improving the safety of battery use.

[0125] In another embodiment of the present invention, the original battery data group includes: battery data corresponding to the basic features generated at each time stamp during the charging process or discharging process of the battery to be detected; the preset data cleaning operation further includes: feature construction operation;

[0126] The 021 may include the following steps 0211-0214:

[0127] 0211: Traverse the battery data corresponding to each time stamp in the original battery data group, fill or delete the battery data corresponding to the time stamps with data missing and / or data abnormal conditions to obtain the intermediate battery data corresponding to the original battery data group.

[0128] Among them, if the number of missing and / or abnormal battery data corresponding to the timestamps with data missing and / or abnormal conditions does not exceed a first preset value, the battery data corresponding to the timestamps with data missing and / or abnormal conditions is filled; if the number of missing and / or abnormal battery data corresponding to the timestamps with data missing and / or abnormal conditions exceeds the first preset value, the battery data corresponding to the timestamps with data missing and / or abnormal conditions is deleted.

[0129] 0212: Based on the data division intervals corresponding to the specified basic features and the battery data corresponding to the specified basic features in the intermediate battery data, the intermediate battery data is divided to obtain battery data segments corresponding to each data division interval.

[0130] 0213: For each battery data segment, based on the battery data corresponding to other basic features in the battery data segment, the data corresponding to each preset extended feature is determined as the feature data corresponding to the battery data segment.

[0131] Among them, the other basic features are the features in the basic features except the specified basic feature.

[0132] 0214: Based on the feature data corresponding to the battery data segment, the data to be utilized corresponding to the original battery data group is determined.

[0133] Among them, the basic features include but are not limited to: battery pack voltage feature, battery pack current feature, battery pack SOC feature, highest temperature feature corresponding to a battery cell in the battery pack, lowest temperature feature corresponding to a battery cell in the battery pack, maximum voltage feature corresponding to a battery cell in the battery pack, and minimum voltage feature corresponding to a battery cell in the battery pack.

[0134] In this implementation manner, the cloud platform can traverse the battery data corresponding to each timestamp in sequence according to the chronological order of the timestamps in the original battery data group, and determine whether there is data missing and / or data anomaly in the battery data corresponding to each traversed timestamp. For example: The battery data corresponding to timestamp M includes the voltage value generated at timestamp M, the current value generated at timestamp M, and the SOC value generated at timestamp M. If the voltage value generated at timestamp M, the current value generated at timestamp M, and / or the SOC value generated at timestamp M is missing, it can be considered that there is data missing in the battery data corresponding to timestamp M. If the voltage value generated at timestamp M, the current value generated at timestamp M, and the SOC value generated at timestamp M all exist, it is considered that there is no data missing in the battery data corresponding to timestamp M. Another example: Taking the data corresponding to the voltage feature as an example for illustration, the battery data corresponding to timestamp N includes the voltage value generated at timestamp N. If the value of the voltage value generated at timestamp N is greater than the first voltage threshold or less than the second voltage threshold, it can be determined that there is a data anomaly in the voltage value generated at timestamp N. Among them, the first voltage threshold is greater than the second voltage threshold.

[0135] Fill or delete the battery data corresponding to the timestamp with data missing and / or data anomaly. Among them, if the number of missing and / or abnormal data in the battery data corresponding to the timestamp with data missing and / or data anomaly is lower than the first preset value, fill the battery data corresponding to the timestamp with data missing and / or data anomaly; if the number of missing and / or abnormal data in the battery data corresponding to the timestamp with data missing and / or data anomaly is not lower than the first preset value, delete the battery data corresponding to the timestamp with data missing and / or data anomaly. Among them, the operation of filling the battery data corresponding to the timestamp with data anomaly is to modify the abnormal value with data anomaly in the battery data corresponding to the timestamp with data anomaly to the specified value corresponding to its corresponding feature.

[0136] For example, the battery data corresponding to the timestamp M includes the voltage value generated at the timestamp M, the current value generated at the timestamp M, and the SOC value generated at the timestamp M; the first preset value is set to 1. If any one of the battery data corresponding to the voltage value generated at the timestamp M, the current value generated at the timestamp M, or the SOC value generated at the timestamp M is missing and / or abnormal, that is, the number of missing and / or abnormal battery data corresponding to the timestamp with data missing and / or data abnormal conditions does not exceed the first preset value of 1, the missing and / or abnormal battery data corresponding to this feature with missing or abnormal timestamp M can be filled. In the case of data missing, it can be based on the specific value of the battery data with this data missing corresponding to the previous timestamp of the timestamp M and the specific value of the battery data with this data missing corresponding to the next timestamp of the timestamp M to calculate the specific value of the battery data with this data missing corresponding to the timestamp M. In the case of data abnormal, the value of the battery data with data abnormal corresponding to the timestamp M is modified to the specified value corresponding to this abnormal feature.

[0137] If at least two of the battery data corresponding to the voltage value generated at the timestamp M, the current value generated at the timestamp M, and the SOC value generated at the timestamp M are missing and / or abnormal, that is, the number of missing battery data corresponding to the timestamp with data missing and / or data abnormal conditions is higher than the first preset value of 1, the cloud platform can delete all the battery data corresponding to the timestamp M from the original battery data.

[0138] Among them, the specific value of the first preset value can be set based on the actual situation.

[0139] Subsequently, based on the data division intervals corresponding to the specified basic features and the battery data corresponding to the specified basic features in the intermediate battery data, the intermediate battery data is divided to obtain the battery data segments corresponding to each data division interval. For example, the specified basic feature is the SOC feature, and the data division intervals include five, which can be 0 - 20%, 20% - 40%, 40% - 60%, 60% - 80%, and 80% - 100% respectively. Based on the specific values of the battery data corresponding to the SOC feature in the intermediate battery data, the intermediate battery data is divided to obtain the battery data segments corresponding to the five data division intervals respectively, that is, five battery data segments.

[0140] For each battery data segment, the cloud platform determines the data corresponding to each preset expansion feature based on the battery data corresponding to other basic features in this battery data segment as the feature data corresponding to this battery data segment, where other basic features are the dimensions except the specified basic feature in the basic features.

[0141] Due to factors such as different charging durations, variable charged amounts, and variable remaining battery levels at the start of charging during different charging processes, there may be a situation where there is no battery data in a certain interval or certain intervals during the process of dividing the intermediate battery data based on data intervals. That is, in the five battery data segments, there may be a situation where one or some of the battery data segments have no battery data, while one or some of the battery data segments have battery data.

[0142] The preset expansion features may include, but are not limited to: an indicative feature for identifying whether there is battery data in the battery data segment, statistical features corresponding to other basic features, and quadratic non-linear features.

[0143] In another embodiment of the present invention, the 0213 may include the following steps:

[0144] For each battery data segment, based on whether there is battery data in the battery data segment, determine the feature data corresponding to the indicative feature corresponding to the battery data segment.

[0145] For each other basic feature in each battery data segment, based on the battery data corresponding to the other basic feature in the battery data segment, determine the feature data corresponding to the statistical feature corresponding to the other basic feature. The statistical features include: features indicating calculating the mean value and / or variance of the battery data corresponding to each other basic feature based on the battery data corresponding to each other basic feature.

[0146] For each battery data segment, based on the battery data corresponding to the current feature and the battery data corresponding to the voltage feature in the battery data segment, determine the feature data corresponding to the resistance feature corresponding to the battery data segment to obtain the feature data corresponding to the battery data segment, where the current feature and the voltage feature belong to other basic features.

[0147] To ensure the accuracy of the detection results of battery data, for each battery data segment, based on whether there is battery data in the battery data segment, determine the feature data corresponding to the indicative feature corresponding to the battery data segment. Among them, when there is battery data in the battery data segment, the feature data corresponding to the indicative feature corresponding to the battery data segment may be the first indicative feature data; when there is no battery data in the battery data segment, the feature data corresponding to the indicative feature corresponding to the battery data segment may be the second indicative feature data.

[0148] For each other basic feature in each battery data segment, based on the battery data corresponding to the other basic feature, determine the feature data corresponding to the statistical feature corresponding to the other basic feature, that is, calculate the mean and / or variance of the battery data corresponding to the other basic feature. For example: For the voltage values corresponding to the voltage feature in the battery data segment, calculate its mean and / or variance; for the current values corresponding to the current feature in the battery data segment, calculate its mean and / or variance.

[0149] And for each battery data segment, based on the battery data corresponding to the other basic features in the battery data segment, determine the feature data corresponding to the corresponding quadratic non-linear features. For example: Based on the battery data corresponding to the current feature in the battery data segment and the battery data corresponding to the voltage feature, determine the feature data corresponding to the resistance feature corresponding to the battery data segment. For example: Using the voltage values and current values corresponding to the same timestamp in the battery data segment, calculate the feature data corresponding to the resistance feature corresponding to the battery data segment, that is, the resistance value.

[0150] In another implementation, the quadratic non-linear feature can also include other types of features, which can be set according to requirements. For example: Power values, etc.

[0151] Correspondingly, the feature data corresponding to the indicative feature corresponding to the above battery data segment, the feature data corresponding to the statistical feature corresponding to each other basic feature of the battery data segment, and the feature data corresponding to the resistance feature corresponding to the battery data segment, that is, the resistance value, serve as the feature data corresponding to the battery data segment.

[0152] Among them, if there is no battery data in the battery data segment, the feature data corresponding to the statistical feature corresponding to each other basic feature corresponding to it and the feature data corresponding to the resistance feature corresponding to the battery data segment, that is, the resistance value, can be represented by a preset value.

[0153] Furthermore, arrange the feature data corresponding to each battery data segment in the original battery data in a one-dimensional sequence in a preset order to obtain the data to be utilized corresponding to the original battery data group.

[0154] In another embodiment of the present invention, before the step of 022, the method may further include:

[0155] The process of training the target battery data detection model, where, as Figure 2 shown, the process may include:

[0156] S201: Obtain each sample battery data and its corresponding calibration information.

[0157] Among them, the sample battery data is the data generated by each sample battery during the charging process or the discharging process; when the sample battery data is the data generated during the charging process, the original battery data set is the data generated during the charging process; when the sample battery data is the data generated during the discharging process, the original battery data set is the data generated during the discharging process, and the calibration information is used to calibrate whether the corresponding sample battery data represents whether the corresponding sample battery has a fault and the type of fault that occurs.

[0158] S202: For each sample battery data, perform the preset data cleaning operation on the sample battery data to obtain the sample battery data after the preset data cleaning operation as training data.

[0159] S203: Obtain multiple initial battery data detection models.

[0160] S204: For each initial battery data detection model, use the first training data in the training data to train the initial battery data detection model to obtain the trained battery data detection model.

[0161] Among them, the first training data is part of the training data.

[0162] S205: Use the second training data in the training data to determine the battery data detection model with the best detection result from multiple trained battery data detection models as the target battery data detection model.

[0163] Among them, the second training data is part of the training data.

[0164] Among them, the second training data is part of the training data, which can include the same training data as the first training data or can include training data different from the first training data.

[0165] In this implementation manner, in order to ensure the accuracy of the detection result of the battery data, a battery data detection model with a sufficiently accurate detection result needs to be obtained. Correspondingly, the cloud platform can first obtain the battery data generated by each sample battery during the charging process or the discharging process as the sample battery data, and obtain the corresponding calibration information for calibrating whether the sample battery data has a fault and the type of fault that occurs for each sample battery data. The sample battery can be set on an electric bicycle, an electric motor vehicle, a charging pile, and other devices powered by electric energy.

[0166] Among them, the calibration information can be calibrated by staff manually for each sample battery data or can be calibrated for each sample battery data using a specific application program.

[0167] In one implementation, each sample battery data is battery data generated by the sample battery during a charging process. The battery data detection model trained using such sample battery data can be used to determine whether the battery fails during this charging process using the battery data corresponding to each battery charging process. Each sample battery data is battery data generated by the sample battery during a discharging process. The battery data detection model trained using such sample battery data can be used to determine whether the battery fails during this discharging process using the battery data corresponding to each battery discharging process.

[0168] Due to the different charging or discharging processes, there are factors such as different charging or discharging time, unfixed charging or discharging power, and unfixed remaining power at the beginning of charging or discharging, so the amount of information of the obtained sample battery data is different, that is, the sample dimensions of different sample battery data are different. For example, taking the charging process as an example: the battery of target vehicle A has a long charging time and is fully charged, and the SOC value is charged from 20% to 100%, which contains 300 records of battery data corresponding to basic characteristics such as current value and voltage value, but the battery of target vehicle B has a short charging time and the SOC value is charged from 50% to 60%, which only contains 50 records of battery data corresponding to basic characteristics such as current value and voltage value. It can be seen that the sample battery data generated by the current charging of the battery of target vehicle A contains much more information than the sample battery data generated by the current charging of the battery of target vehicle B. However, data with different sample dimensions, that is, data with unfixed sample dimensions, generally cannot be trained to obtain a detection model using traditional methods.

[0169] In view of this, the cloud platform obtains each sample battery data, performs a preset data cleaning operation on each sample battery data, and obtains the sample battery data after the preset data cleaning operation as training data.

[0170] In one implementation, when the preset data cleaning operation includes a deletion and a patching operation and a feature construction operation, the cloud platform traverses each sample battery data, fills or deletes the battery data corresponding to the timestamp where data is missing and / or where data is missing, and obtains the intermediate sample battery data, wherein if the number of missing and / or abnormal battery data corresponding to the timestamp where data is missing and / or where data is missing is lower than a first preset value, the battery data corresponding to the timestamp where data is missing and / or where data is missing is filled; if the number of missing and / or abnormal battery data corresponding to the timestamp where data is missing and / or where data is missing is not lower than a first preset value, the battery data corresponding to the timestamp where data is missing and / or where data is missing is deleted. For details, refer to the deletion and patching operation process of the original battery data group.

[0171] Furthermore, for each intermediate sample battery data, based on the data division intervals corresponding to the specified basic features and the battery data corresponding to the specified basic features in the intermediate sample battery data, the intermediate sample battery data is divided to obtain the sample battery data segments corresponding to each data division interval; for each sample battery data segment of each intermediate sample battery data, based on the battery data corresponding to other basic features in the sample battery data segment, the data corresponding to each preset extended feature is determined as the feature data corresponding to the battery data segment. That is, for whether there is sample battery data in the sample battery data segment, the feature data corresponding to the indicative feature corresponding to the sample battery data segment is determined; for each other basic feature, based on the sample battery data corresponding to the other basic feature in the sample battery data segment, the feature data corresponding to the statistical feature corresponding to the other basic feature is determined, that is, the mean and / or variance, etc. of the sample battery data corresponding to the other basic feature. Based on the sample battery data corresponding to the current feature and the sample battery data corresponding to the voltage feature in the sample battery data segment, the feature data corresponding to the resistance feature corresponding to the sample battery data segment is determined, that is, the resistance value, so as to obtain the feature data corresponding to each sample battery data segment corresponding to each sample battery data. Specifically, refer to the feature construction operation process of the original battery data group.

[0172] It can be understood that the sample battery data after each deletion and supplementation operation is segmented according to the data division intervals corresponding to the specified basic features. Furthermore, constructing new features for the battery data in the sample battery data segments corresponding to each data division interval can not only expand the features of the battery data, but also the feature data corresponding to the expanded features contains the features of the feature data corresponding to the basic features, and can control the data corresponding to each data division interval corresponding to each sample battery data to the same sample dimension, that is, the same data volume, and the features are the same.

[0173] For each sample battery data, based on the feature data corresponding to all sample battery data segments corresponding to the sample battery data, arranged in a one-dimensional sequence in a preset order, that is, a row vector is obtained, so as to obtain the training data corresponding to the sample battery data. This can greatly increase the number of features of the training data. For example, after the deletion and supplementation operation and feature construction operation on the sample battery data, data with a feature of x and a sample dimension of y is obtained, that is, a y * x matrix; arranging the y * x matrix into a one-dimensional sequence, that is, a 1 * yx row vector is obtained. At this time, the data is converted from data with a feature of x to data with a feature of yx, increasing the number of features of the training data.

[0174] In another implementation, the preset data cleaning operation includes deletion and supplementation operations. Correspondingly, the cloud platform determines the sample dimension of each sample battery data for each sample battery data, and determines the target sample dimension based on the sample dimension of the sample battery. Among them, the sample dimension with the largest quantity can be used as the target sample dimension, or the sample dimension with the intermediate size can be used as the target sample dimension. For each sample battery data, the cloud platform traverses the sample battery data. When the sample dimension corresponding to the sample battery data exceeds the target sample dimension, the sample battery data corresponding to each timestamp in the sample battery data is deleted to obtain the training data corresponding to the sample battery data with the sample dimension being the target sample dimension. Among them, the sample battery data corresponding to the timestamp with data missing and / or data anomaly in the sample battery data is preferentially deleted. When the sample dimension corresponding to the sample battery data is lower than the target sample dimension, the sample battery data corresponding to each timestamp in the sample battery data is filled. Among them, the sample battery data corresponding to the timestamp with data missing and / or data anomaly in the sample battery data is preferentially filled. Furthermore, in a difference manner, the intermediate data corresponding to the sample battery data with the sample dimension being the target sample dimension is filled.

[0175] For example, the battery data includes the battery data corresponding to m basic features, that is, the number of basic features in the battery data is m. For example, taking the charging process as an example, during a charging process of sample vehicle A, sample battery data A is obtained. Sample battery data A has 100 records. Correspondingly, sample battery data A can be represented as a 100*m matrix; during a charging process of sample vehicle B, sample battery data B is obtained. Sample battery data B has 80 records. Correspondingly, sample battery data B can be represented as an 80*m matrix; during a charging process of sample vehicle C, sample battery data C is obtained. Sample battery data C has 60 records. Correspondingly, sample battery data C can be represented as a 60*m matrix. It can be seen from this that the sample dimensions of different sample battery data are inconsistent. By selecting a target sample dimension, the sample battery data with a long time, that is, a large sample dimension, can be deleted, and the sample battery data with a short time, that is, a small sample dimension, can be filled. For example, Arima (Autoregressive Integrated Moving Average model) is used to fill the sample battery data with a small sample dimension. The sample dimension of the training data corresponding to all sample battery data is made to be n. For example, when n = 80, for sample battery data A, appropriate deletion is performed and 80 records are retained. For sample battery data B, no operation is performed. For sample battery data C, by using the time series method, it is reasonably filled so that the number of records reaches 80. In this way, the sample dimensions corresponding to all intermediate data are all 80*m.

[0176] Convert each 80*m data matrix into a 1*80m row vector to describe the battery data of a charging process, and obtain the training data corresponding to each sample battery data. This can greatly enhance the features, that is, change the sample dimension from m to 80m.

[0177] Furthermore, obtain multiple initial battery data detection models. Among them, each initial battery data detection model is different. The initial battery data detection model can include, but is not limited to, XGBoost model, LightGBM model, etc.

[0178] For each initial battery data detection model, the cloud platform uses the first training data in the training data to train the initial battery data detection model to obtain the trained battery data detection model; and uses the second training data in the training data to determine the battery data detection model with the best detection result from multiple trained battery data detection models as the target battery data detection model. To determine a target battery data detection model with relatively accurate detection results for subsequent battery data detection processes.

[0179] Among them, for the training process of the initial battery data detection model, reference can be made to the training process of the model in related technologies, which will not be elaborated here.

[0180] Subsequently, the cloud platform can use the trained target battery data detection model and each original battery data group in the target battery data corresponding to each battery to be detected to detect each charging process or discharging process of each battery to be detected, and obtain the probabilities of various fault types occurring in each charging process or discharging process of each battery to be detected. Furthermore, based on the obtained probabilities of various fault types and the corresponding preset fault probability thresholds, determine whether each charging process or discharging process of each battery to be detected has a fault, that is, the type of fault that occurs, that is, whether a fault occurs, and the type of fault that occurs.

[0181] In another embodiment of the present invention, the first fault detection information includes: determining the first fault information corresponding to each battery to be detected based on the target battery data corresponding to each battery to be detected and a preset three-level early warning algorithm, and determining the second fault information corresponding to the battery to be detected based on the target battery data corresponding to each battery to be detected, a preset performance evaluation algorithm, and a preset short-circuit detection algorithm;

[0182] The S104 may include the following steps 031-032:

[0183] 031: For each battery to be detected, determine the verification redundancy information between the first fault information and the second fault information based on the first fault information and the second fault information corresponding to the battery to be detected.

[0184] 032: For each battery to be detected, based on the second fault information and the second fault detection information corresponding to the battery to be detected, determine the check redundancy information between the second fault information and the second fault detection information, so as to determine the fault check redundancy information corresponding to each battery to be detected.

[0185] In this implementation manner, the cloud platform compares the first fault information corresponding to each battery to be detected with the second fault information to determine the fault detection results for the same fault type in the first fault information and the second fault information. If the fault detection results for the same fault type are the same, it can be determined that the detection results of the fault type with the same fault detection results are relatively accurate, and the probability that the fault of this fault type actually exists is high. If the fault detection results for the same fault type are different, the fault detection result indicating a fault for this unified fault type can be used as a fault supplement for the fault detection result indicating no fault, indicating that the probability that the fault of this fault type actually exists is relatively low. And the fault detection results for other different fault types are used as corresponding fault supplements to obtain the check redundancy information between the first fault information and the second fault information.

[0186] The cloud platform compares the second fault information and the second fault detection information corresponding to each battery to be detected to determine the fault detection results for the same fault type in the second fault information and the second fault detection information. If the fault detection results for the same fault type are the same, it can be determined that the detection results of the fault type with the same fault detection results are relatively accurate, and the probability that the fault of this fault type actually exists is high. If the fault detection results for the same fault type are different, the fault detection result indicating a fault for this unified fault type can be used as a fault supplement for the fault detection result indicating no fault, indicating that the probability that the fault of this fault type actually exists is relatively low. And the fault detection results for other different fault types are used as corresponding fault supplements to obtain the check redundancy information between the second fault information and the second fault detection information, so as to determine the fault check redundancy information corresponding to each battery to be detected.

[0187] For the fault information corresponding to the fault type with a high probability of battery occurrence in the fault check redundancy information, key warnings can be given, so that subsequent staff can pay more attention to this fault information. Among them, the key warning can be highlighting, displaying this type of fault information in a special color, or warning in combination with a sound prompt method, etc.

[0188] Corresponding to the above method embodiments, an embodiment of the present invention provides a warning device for a battery, as Figure 3 shown, the device includes:

[0189] The first acquisition module 310 is configured to acquire target battery data corresponding to the battery to be detected;

[0190] The first determination module 320 is configured to, for each battery to be detected, use the target battery data corresponding to the battery to be detected and a preset battery fault detection algorithm to determine first fault detection information corresponding to the battery to be detected;

[0191] The second determination module 330 is configured to use the target battery data corresponding to all the batteries to be detected and a target battery data detection model to determine second fault detection information corresponding to each battery to be detected, where the target battery data detection model is: a target battery data detection model obtained by training using training data obtained after performing a preset data cleaning operation on each sample battery data and its corresponding calibration information, the sample dimensions and feature dimensions of each training data are the same, and the preset cleaning operation at least includes: a deletion and supplementation operation for outliers and / or missing values in the sample battery data;

[0192] The third determination module 340 is configured to determine fault verification redundancy information corresponding to each battery to be detected based on the first fault detection information and the second fault detection information;

[0193] The output module 350 is configured to output the first fault detection information, the second fault detection information, and the fault verification redundancy information.

[0194] Applying the embodiments of the present invention, the training of the target battery data detection model is realized through training data with the same sample dimensions and features, and then through different fault detection algorithms including the target battery data detection model and the battery data corresponding to the battery to be detected collected, the fault detection of the battery to be detected is realized, multiple fault detection results are obtained, and the multiple fault detection results are mutually verified to obtain and output fault verification redundancy information, so as to better reflect the fault detection results of the battery to be detected, and then realize the safety warning of the battery, and to a certain extent improve the accuracy of the safety warning and improve the safety of battery use.

[0195] In another embodiment of the present invention, the first determination module 320 is specifically configured to, for each battery to be detected, use the target battery data corresponding to the battery to be detected and a preset three-level warning algorithm to determine first fault information corresponding to the battery to be detected;

[0196] For each battery to be detected, use first specified feature data in the target battery data corresponding to the battery to be detected and a preset performance evaluation algorithm to determine performance scoring information corresponding to the battery to be detected;

[0197] For each battery to be detected, the short - circuit detection information corresponding to the battery to be detected is determined by using the second specified feature data in the target battery data corresponding to the battery to be detected and a preset short - circuit detection algorithm;

[0198] For each battery to be detected, based on the performance scoring information and short - circuit detection information corresponding to the battery to be detected, the second fault information corresponding to the battery to be detected is determined to obtain the first fault detection information.

[0199] In another embodiment of the present invention, the target battery data corresponding to the battery to be detected includes: the original battery data group corresponding to each charging process or each discharging process of the battery to be detected;

[0200] The second determination module 330 includes:

[0201] A cleaning unit (not shown in the figure), configured to perform the preset data cleaning operation on each original battery data group corresponding to each battery to be detected to obtain the data to be utilized corresponding to the original battery data group;

[0202] An input determination unit (not shown in the figure), configured to input each data to be utilized corresponding to each battery to be detected into the target battery data detection model, determine the third fault information corresponding to each data to be utilized, and determine the second fault detection information corresponding to the battery to be detected.

[0203] In another embodiment of the present invention, the original battery data group includes: the battery data corresponding to the basic features generated at each moment corresponding to each timestamp during the charging process or discharging process of the battery to be detected; the preset data cleaning operation further includes: a feature construction operation;

[0204] The cleaning unit is specifically configured to traverse the battery data corresponding to each timestamp in the original battery data group, fill or delete the battery data corresponding to the timestamp with data missing and / or data anomaly, and obtain the intermediate battery data corresponding to the original battery data group. Among them, if the number of missing and / or abnormal data of the battery data corresponding to the timestamp with data missing and / or data anomaly is not higher than the first preset value, the battery data corresponding to the timestamp with data missing and / or data anomaly is filled; if the number of missing and / or abnormal data of the battery data corresponding to the timestamp with data missing and / or data anomaly is higher than the first preset value, the battery data corresponding to the timestamp with data missing and / or data anomaly is deleted;

[0205] Divide the intermediate battery data based on the data division intervals corresponding to the specified basic features and the battery data corresponding to the specified basic features in the intermediate battery data, to obtain battery data segments corresponding to each data division interval;

[0206] For each battery data segment, determine the data corresponding to each preset extended feature based on the battery data corresponding to other basic features in the battery data segment, as the feature data corresponding to the battery data segment, where the other basic features are the features in the basic features except the specified basic feature;

[0207] Determine the data to be utilized corresponding to the original battery data group based on the feature data corresponding to the battery data segment.

[0208] In another embodiment of the present invention, the cleaning unit is specifically configured to, for each battery data segment, determine the feature data corresponding to the indicative feature corresponding to the battery data segment based on whether there is battery data in the battery data segment;

[0209] For each other basic feature in each battery data segment, determine the feature data corresponding to the statistical feature corresponding to the other basic feature based on the battery data corresponding to the other basic feature in the battery data segment, where the statistical feature includes: a feature indicating calculating the mean and / or variance of the battery data corresponding to each other basic feature based on the battery data corresponding to each other basic feature;

[0210] For each battery data segment, determine the feature data corresponding to the resistance feature corresponding to the battery data segment based on the battery data corresponding to the current feature and the battery data corresponding to the voltage feature in the battery data segment, to obtain the feature data corresponding to the battery data segment, where the current feature and the voltage feature belong to other basic features.

[0211] In another embodiment of the present invention, the device further includes:

[0212] A model training module (not shown in the figure) is configured to train the target battery data detection model before inputting each piece of data to be utilized corresponding to each battery to be detected into the target battery data detection model to determine the third fault information corresponding to each piece of data to be utilized, so as to determine the second fault detection information corresponding to the battery to be detected. The model training module is specifically configured to obtain various sample battery data and their corresponding calibration information. The sample battery data is data generated by each sample battery during the charging process or the discharging process. When the sample battery data is data generated during the charging process, the original battery data group is data generated during the charging process. When the sample battery data is data generated during the discharging process, the original battery data group is data generated during the discharging process. The calibration information is used to calibrate whether the corresponding sample battery data represents whether the corresponding sample battery has a fault and the type of the fault that occurs.

[0213] For each sample battery data, perform the preset data cleaning operation on the sample battery data to obtain the sample battery data after the preset data cleaning operation as training data.

[0214] Obtain multiple initial battery data detection models.

[0215] For each initial battery data detection model, use the first training data in the training data to train the initial battery data detection model to obtain a trained battery data detection model. The first training data is part of the training data.

[0216] Use the second training data in the training data to determine the battery data detection model with the best detection result from multiple trained battery data detection models as the target battery data detection model. The second training data is part of the training data.

[0217] In another embodiment of the present invention, the first fault detection information includes: determining the first fault information corresponding to each battery to be detected based on the target battery data corresponding to each battery to be detected and a three-level early warning algorithm, and determining the second fault information corresponding to the battery to be detected based on the target battery data corresponding to each battery to be detected, a preset performance evaluation algorithm, and a preset short-circuit detection algorithm.

[0218] The third determination module 340 is specifically configured to, for each battery to be detected, determine the verification redundancy information between the first fault information and the second fault information based on the first fault information and the second fault information corresponding to the battery to be detected.

[0219] For each battery to be detected, based on the second fault information and the second fault detection information corresponding to the battery to be detected, determine the check redundancy information between the second fault information and the second fault detection information, so as to determine the fault check redundancy information corresponding to each battery to be detected.

[0220] The above system and device embodiments correspond to the system embodiment and have the same technical effects as the method embodiment. For specific descriptions, refer to the method embodiment. The device embodiment is obtained based on the method embodiment. For specific descriptions, refer to the method embodiment section and will not be elaborated here. Those of ordinary skill in the art can understand that the drawings are only schematic diagrams of one embodiment, and the modules or processes in the drawings are not necessarily essential for implementing the present invention.

[0221] Those of ordinary skill in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description in the embodiment, or can be correspondingly changed and located in one or more devices different from this embodiment. The modules in the above embodiments can be combined into one module, or further split into multiple sub-modules.

[0222] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A warning method for a battery, characterized in that, the method includes: obtaining target battery data corresponding to the battery to be detected; for each battery to be detected, using the target battery data corresponding to the battery to be detected and a preset battery fault detection algorithm to determine first fault detection information corresponding to the battery to be detected, including: for each battery to be detected, using the target battery data corresponding to the battery to be detected and a preset three-level warning algorithm to determine first fault information corresponding to the battery to be detected; for each battery to be detected, using first specified feature data in the target battery data corresponding to the battery to be detected and a preset performance evaluation algorithm to determine performance scoring information corresponding to the battery to be detected; for each battery to be detected, using second specified feature data in the target battery data corresponding to the battery to be detected and a preset short-circuit detection algorithm to determine short-circuit detection information corresponding to the battery to be detected; for each battery to be detected, based on the performance scoring information and short-circuit detection information corresponding to the battery to be detected, determining second fault information corresponding to the battery to be detected, so as to obtain first fault detection information including the first fault information and the second fault information; using the target battery data corresponding to all the batteries to be detected and a target battery data detection model to determine second fault detection information corresponding to each battery to be detected, including: when the target battery data corresponding to the battery to be detected includes an original battery data group corresponding to each charging process or each discharging process of the battery to be detected, for each original battery data group corresponding to each battery to be detected, performing the preset data cleaning operation on the original battery data group to obtain data to be utilized corresponding to the original battery data group, inputting the data to be utilized corresponding to each battery to be detected into the target battery data detection model to obtain an abnormal probability corresponding to each data to be utilized, using the abnormal probability corresponding to each data to be utilized and a corresponding abnormal probability threshold to determine a fault result corresponding to each data to be utilized, where the fault result can characterize whether a fault occurs and the type of the fault during a certain charging process or a certain discharging process of the battery to be detected, so as to determine third fault information corresponding to each data to be utilized, and obtaining second fault detection information including the third fault information, where the target battery data detection model is: a target battery data detection model trained by using training data obtained after performing the preset data cleaning operation on each sample battery data and its corresponding calibration information, the sample dimension and the feature dimension of each training data are the same, and the preset cleaning operation at least includes: a deletion and supplementation operation on outliers and / or missing values in the sample battery data; based on the first fault detection information and the second fault detection information, determining fault verification redundancy information corresponding to each battery to be detected; outputting the first fault detection information, the second fault detection information, and the fault verification redundancy information.

2. The method according to claim 1, characterized in that, The original battery data set includes: battery data corresponding to the basic features generated at the moments corresponding to each time stamp during the charging or discharging process of the battery to be detected; the preset data cleaning operation further includes: a feature construction operation; The step of performing the preset data cleaning operation on the original battery data set to obtain the data to be utilized corresponding to the original battery data set includes: Traverse the battery data corresponding to each time stamp in the original battery data set, fill or delete the battery data corresponding to the time stamps with data missing and / or data anomaly situations, to obtain the intermediate battery data corresponding to the original battery data set, wherein, if the number of missing and / or abnormal data of the battery data corresponding to the time stamps with data missing and / or data anomaly situations is not higher than a first preset value, fill the battery data corresponding to the time stamps with data missing and / or data anomaly situations; if the number of missing and / or abnormal data of the battery data corresponding to the time stamps with data missing and / or data anomaly situations is higher than the first preset value, delete the battery data corresponding to the time stamps with data missing and / or data anomaly situations; Based on the data division intervals corresponding to the specified basic features and the battery data corresponding to the specified basic features in the intermediate battery data, divide the intermediate battery data to obtain the battery data segments corresponding to each data division interval; For each battery data segment, based on the battery data corresponding to other basic features in the battery data segment, determine the data corresponding to each preset extended feature as the feature data corresponding to the battery data segment, wherein the other basic features are the features other than the specified basic feature among the basic features; Based on the feature data corresponding to the battery data segment, determine the data to be utilized corresponding to the original battery data set.

3. The method according to claim 2, characterized in that, The step of, for each battery data segment, based on the battery data corresponding to other basic features in the battery data segment, determining the battery data corresponding to each preset extended feature as the feature data corresponding to the battery data segment includes: For each battery data segment, based on whether there is battery data in the battery data segment, determine the feature data corresponding to the indicative feature corresponding to the battery data segment; For each other basic feature in each battery data segment, based on the battery data corresponding to the other basic feature in the battery data segment, determine the feature data corresponding to the statistical feature corresponding to the other basic feature, the statistical feature including: a feature indicating calculating the mean value and / or variance of the battery data corresponding to each other basic feature based on the battery data corresponding to each other basic feature; For each battery data segment, based on the battery data corresponding to the current feature and the battery data corresponding to the voltage feature in the battery data segment, determine the feature data corresponding to the resistance feature corresponding to the battery data segment, so as to obtain the feature data corresponding to the battery data segment, wherein the current feature and the voltage feature belong to other basic features.

4. The method according to claim 2, characterized in that, Before the step of inputting each piece of data to be utilized corresponding to each battery to be detected into the target battery data detection model to determine the third fault information corresponding to each piece of data to be utilized, so as to determine the second fault detection information corresponding to the battery to be detected, the method further includes: The process of training the target battery data detection model, where the process includes: Obtaining various sample battery data and their corresponding calibration information, where the sample battery data is: data generated by each sample battery during the charging process or the discharging process; when the sample battery data is data generated during the charging process, the original battery data group is data generated during the charging process; when the sample battery data is data generated during the discharging process, the original battery data group is data generated during the discharging process, and the calibration information is used to calibrate whether the corresponding sample battery data represents whether the corresponding sample battery has a fault and the type of fault that occurs; For each sample battery data, performing the preset data cleaning operation on the sample battery data to obtain the sample battery data after the preset data cleaning operation as training data; Obtaining a plurality of initial battery data detection models; For each initial battery data detection model, using the first training data in the training data to train the initial battery data detection model to obtain a trained battery data detection model, where the first training data is part of the training data; Using the second training data in the training data to determine the battery data detection model with the best detection result from among the plurality of trained battery data detection models as the target battery data detection model, where the second training data is part of the training data.

5. The method according to any one of claims 1-4, characterized in that, The step of determining the fault verification redundancy information corresponding to each battery to be detected based on the first fault detection information and the second fault detection information includes: For each battery to be detected, determining the verification redundancy information between the first fault information and the second fault information based on the first fault information and the second fault information corresponding to the battery to be detected; For each battery to be detected, determining the verification redundancy information between the second fault information and the second fault detection information based on the second fault information and the second fault detection information corresponding to the battery to be detected, so as to determine the fault verification redundancy information corresponding to each battery to be detected.

6. An early warning device for a battery, characterized in that, The device includes: A first obtaining module configured to obtain target battery data corresponding to a battery to be detected; A first determining module configured to, for each battery to be detected, use the target battery data corresponding to the battery to be detected and a preset battery fault detection algorithm to determine the first fault detection information corresponding to the battery to be detected; A second determination module, configured to use the target battery data corresponding to all the batteries to be detected and a target battery data detection model to determine second fault detection information corresponding to each battery to be detected, where the target battery data detection model is: a target battery data detection model obtained by training the training data obtained after performing a preset data cleaning operation on each sample battery data and its corresponding calibration information, the sample dimensions and feature dimensions of each training data are the same, and the preset cleaning operation at least includes: a deletion and supplementation operation on outliers and / or missing values in the sample battery data; A third determination module, configured to determine fault verification redundancy information corresponding to each battery to be detected based on the first fault detection information and the second fault detection information; An output module, configured to output the first fault detection information, the second fault detection information, and the fault verification redundancy information; The first determination module is specifically configured to, for each battery to be detected, use the target battery data corresponding to the battery to be detected and a preset three-level early warning algorithm to determine first fault information corresponding to the battery to be detected; For each battery to be detected, use first specified feature data in the target battery data corresponding to the battery to be detected and a preset performance evaluation algorithm to determine performance score information corresponding to the battery to be detected; For each battery to be detected, use second specified feature data in the target battery data corresponding to the battery to be detected and a preset short-circuit detection algorithm to determine short-circuit detection information corresponding to the battery to be detected; For each battery to be detected, based on the performance score information and the short-circuit detection information corresponding to the battery to be detected, determine second fault information corresponding to the battery to be detected, so as to obtain first fault detection information including the first fault information and the second fault information; The target battery data corresponding to the battery to be detected includes: an original battery data group corresponding to the battery to be detected in each charging process or each discharging process; The second determination module includes: A cleaning unit, configured to perform the preset data cleaning operation on each original battery data group corresponding to each battery to be detected to obtain data to be utilized corresponding to the original battery data group; An input determination unit, configured to input the data to be utilized corresponding to each battery to be detected into the target battery data detection model to determine third fault information corresponding to each data to be utilized, so as to determine second fault detection information corresponding to the battery to be detected, including: inputting the data to be utilized corresponding to each battery to be detected into the target battery data detection model to obtain an abnormal probability corresponding to each data to be utilized, and using the abnormal probability corresponding to each data to be utilized and a corresponding abnormal probability threshold to determine a fault result corresponding to each data to be utilized, where the fault result can represent whether a fault occurs and the type of the fault in a certain charging process or a certain discharging process of the battery to be detected, so as to determine third fault information corresponding to each data to be utilized and obtain second fault detection information including the third fault information.

Images