Battery data detection method and device, equipment, storage medium and program product

Abstract

The invention discloses a battery data detection method and device, equipment, a storage medium and a program product, and the method comprises the steps: determining initial fragmentation data with missing information, the initial fragmentation data comprising battery information corresponding to a plurality of battery parameter fields; at least one iteration detection model in a plurality of iteration detection models matched with the plurality of battery parameter fields in a one-to-one manner is utilized to at least carry out supplementary prediction processing on battery information corresponding to the plurality of battery parameter fields in the initial fragmentation data, and target fragmentation data corresponding to the initial fragmentation data is obtained; and performing accuracy detection on the supplementary information and the prediction information in the target fragmented data by using at least one detection model in a plurality of detection models matched with the plurality of battery parameter fields one by one to obtain a detection result of the target fragmented data.

Description

Technical Field

[0001] The present application relates to the technical field of welding equipment detection, and in particular to a battery data detection method, apparatus, equipment, storage medium, and program product. Background Art

[0002] New energy batteries are being used more and more widely in life and industry, for example, in the fields of energy storage and new energy vehicles.

[0003] Currently, battery anomaly detection often requires collecting relevant battery data and then testing the battery based on the collected data. However, the quality of the collected data is often low, and using low-quality data for battery anomaly detection will not produce accurate detection results. To improve data quality, related technologies determine whether data is missing based on the date, thereby supplementing the missing data and improving data quality. However, the problem with this solution is that the data integrity detection is inefficient and cannot verify the accuracy of the data. Summary of the Invention

[0004] In view of this, embodiments of the present application provide at least one method, apparatus, device, storage medium, and program product for detecting battery data.

[0005] The technical solution of the embodiment of the present application is implemented as follows:

[0006] In a first aspect, an embodiment of the present application provides a method for detecting battery data, the method comprising:

[0007] Determine initial fragmented data with missing information, where the initial fragmented data includes battery information corresponding to a plurality of battery parameter fields;

[0008] performing at least supplementary prediction processing on battery information corresponding to each of the plurality of battery parameter fields in the initial sharded data using at least one iterative detection model from a plurality of iterative detection models that are matched one-to-one with the plurality of battery parameter fields, to obtain target sharded data corresponding to the initial sharded data; the target sharded data including supplementary information for the missing information and predicted information for the non-missing information;

[0009] For the target slice data, at least one detection model from a plurality of detection models that match the plurality of battery parameter fields one by one is used to perform accuracy detection on the supplementary information and the prediction information in the target slice data to obtain a detection result of the target slice data.

[0010] In an embodiment of the present application, at least one of the multiple iterative detection models that match multiple battery parameter fields one by one can be used to perform supplementary prediction processing on the initial sliced ​​data with missing information, thereby achieving the supplementation of missing information and the prediction of non-missing information; by using at least one of the multiple detection models that match multiple battery parameter fields one by one, the accuracy of the labeled sliced ​​data output by the iterative detection model is detected, thereby achieving the accuracy detection of the supplementary information of the missing information and the prediction information of the non-missing information. In this way, the integrity detection of the data and the supplementation of the missing information are completed by the iterative detection model corresponding to the battery parameter field, so that the interfering data in the initial sliced ​​data can be eliminated. And the accuracy detection of the data is completed by the detection model corresponding to the battery parameter field, thereby achieving the verification of the output data of the iterative detection model. Because the above two models both correspond to the battery parameter field, the accuracy of the integrity detection and the accuracy detection can also be improved.

[0011] In some embodiments, the method further includes: sharding the battery data set of the battery of the electrical device to obtain a plurality of sharded data belonging to different time periods; each of the sharded data includes battery information corresponding to a plurality of battery parameter fields; in the case where there is missing information in the sharded data, determining the proportion information of the missing information in the sharded data according to the battery parameter field; determining the initial sharded data with missing information includes: determining the sharded data whose proportion information of the missing information is less than or equal to a preset first proportion threshold as the initial sharded data.

[0012] In the embodiments of the present application, by sharding the battery data set of the battery of the electric device, the amount of data input to the model can be reduced, thereby improving the computational efficiency of the model. Furthermore, by determining the proportion of missing information in each battery parameter field for the sharded data with missing information, the proportion of missing information in the input of the iterative detection model (i.e., the initial sharded data) can be made less than or equal to a preset first proportion threshold. This can improve the accuracy of the supplementary prediction of the iterative detection model.

[0013] In some embodiments, the method further includes: determining the shard data whose proportion of missing information is greater than a preset first proportion threshold as the first shard data; based on the target battery parameter field corresponding to the missing information of the first shard data and the identifier of the first shard data, obtaining the second shard data that is associated in time series and does not have missing information in the target battery parameter field from the multiple shard data; sorting the first shard data and the second shard data in time series to form a first shard data set; and sharding the first shard data set so that the missing information in the first shard data is distributed in different shard data, and the proportion of missing information in the different shard data is less than or equal to the first proportion threshold.

[0014] In an embodiment of the present application, for the first shard data whose proportion of missing information is greater than a preset first proportion threshold, the second shard data that is associated with the first shard data in time sequence and does not have missing information on the battery parameter field of the missing information of the first shard data can be obtained, and then the first shard data and the second shard data (i.e., the first shard data set) that are sorted in time sequence are re-sharded so that the proportion of missing information on the multiple shard data after sharding is less than or equal to the first proportion threshold. In this way, by obtaining the second shard data, the proportion of missing information in the combined data of the first shard data and the second shard data can be effectively reduced, and the battery information in the first shard data set is continuous in time sequence, and the battery information in the shard data after re-sharding is also continuous in time, so that the iterative detection model can perform supplementary prediction processing on the continuous battery information, thereby improving the accuracy of the iterative detection model.

[0015] In some embodiments, the method also includes: determining the deviation ratio between the target shard data and the initial shard data; using at least one detection model from a plurality of detection models that match the plurality of battery parameter fields one-to-one to perform accuracy detection on the supplementary information in the target shard data and the prediction information to obtain the detection result of the target shard data, including: for the first target shard data in which the deviation ratio in the target shard data is less than or equal to a preset threshold, using at least one detection model from a plurality of detection models that match the plurality of battery parameter fields one-to-one to perform accuracy detection on the supplementary information in the first target shard data and the prediction information to obtain the detection result of the first target shard data.

[0016] In an embodiment of the present application, it is possible to first determine whether the deviation ratio of the target slice data is less than or equal to a preset threshold value. If it is less than, the first target slice data whose deviation ratio is less than or equal to the preset threshold value can be input into the detection model, so that the detection model performs accuracy detection on the supplementary information and the prediction information in the first target slice data. In this way, by calculating the deviation ratio of the target slice data, the accuracy of the target slice data can be pre-detected. When the pre-detection passes (that is, the deviation ratio of the target slice data is less than or equal to the preset threshold value), the first target slice data is input into the detection model, and the accuracy detection is performed on the supplementary information and the prediction information in the first target slice data. In this way, the amount of calculation of the detection model can be reduced, and the efficiency of the detection model in performing accuracy detection on the target slice data can be improved.

[0017] In some embodiments, the method further includes: for the second target shard data in the target shard data whose deviation ratio is greater than a preset threshold, based on the identifier of the second target shard data, obtaining the fourth shard data associated in time series from the initial shard data; sorting the initial shard data and the fourth shard data corresponding to the second target shard data in time series to form a third shard data set; performing supplementary prediction processing on the third shard data set based on the iterative detection model until the deviation ratio between the processed shard data and the third shard data set is less than or equal to the preset threshold, and using the processed shard data as the target shard data of the third shard data set.

[0018] In an embodiment of the present application, for second target slice data whose deviation ratio is greater than a preset threshold, the fourth slice data associated in time sequence can be obtained through the identifier of the second target slice data, thereby expanding the input data of the iterative detection model, thereby making the deviation ratio of the target slice data output by the iterative detection model less than or equal to the preset threshold. In this way, by adjusting the size of the input data of the iterative detection model, the deviation ratio between the output data of the iterative detection model and the input data can be reduced, thereby reducing the number of manual interventions and improving the efficiency of battery information quality testing.

[0019] In some embodiments, the method further includes: receiving a data processing request sent by a back-end detection module of the battery; the data processing request is used to request detection of battery information of at least one battery parameter field to be processed; based on the data processing request, determining a battery data set corresponding to the at least one battery parameter field to be processed; based on the at least one battery parameter field to be processed, determining the at least one iterative detection model and the at least one detection model from a plurality of iterative detection models that match the multiple battery parameter fields one-to-one and a plurality of detection models that match the multiple battery parameter fields one-to-one; if the detection result of the target shard data meets a preset condition, sending the target shard data to the back-end detection module of the battery; the battery detection module is used to perform abnormality detection on the battery of the electric device based on the target shard data to determine whether the battery of the electric device has an abnormality.

[0020] In an embodiment of the present application, based on a data processing request sent by a back-end detection module for requesting detection of battery information of at least one battery parameter field to be processed, a battery data set corresponding to the at least one battery parameter field to be processed can be determined, and the at least one iterative detection model and the at least one detection model can be determined from a plurality of iterative detection models that match the multiple battery parameter fields one-to-one and a plurality of detection models that match the multiple battery parameter fields one-to-one. In this way, the corresponding battery data set, iterative detection model, and detection model can be determined according to different types of data processing requests. And the target slice data whose detection results meet the preset conditions can be sent to the back-end detection module, so that the battery detection module performs abnormality detection on the battery of the electrical device based on the target slice data. In this way, the battery detection module can obtain more accurate battery detection results through battery information with higher data quality.

[0021] In some embodiments, the method further includes: when it is detected through binary logs that a battery source data set is stored in the database of the data warehouse, extracting the battery source data set into the operational data storage layer of the data warehouse; the battery source data set is a data set obtained by performing attribute detection on the battery of the electrical equipment; in the operational data storage layer, logically processing the battery source data set to obtain the battery data set.

[0022] In the embodiment of the present application, the binary log can be used to promptly detect whether a battery source data set exists in the database. This allows for timely logical processing of the battery source data set to obtain a battery data set, thereby performing data quality testing on the battery information in the battery data set.

[0023] In some embodiments, the method further includes: when the detection result does not meet the preset conditions, determining the abnormal data in the target shard data and storing the abnormal data in the database; generating an alarm label based on the abnormal data, and generating a data abnormality alarm table based on the abnormal data and the alarm label; the data abnormality alarm table is used to adjust the logical processing.

[0024] In this embodiment of the present application, if the test results do not meet the preset conditions, abnormal data in the target slice data is identified and stored in the database. Storing only abnormal data in the database reduces storage space overhead. Furthermore, by generating a data anomaly alarm table, an effective basis for adjusting logical processing is provided, thereby improving the accuracy of the generated battery data set.

[0025] In a second aspect, an embodiment of the present application provides a battery data detection device, the battery data detection device comprising:

[0026] A sharding unit, configured to shard a battery data set of a battery of an electric device to obtain a plurality of shard data belonging to different time periods; each of the shard data includes battery information corresponding to a plurality of battery parameter fields;

[0027] a first detection unit, configured to, for initial slice data having missing information among the plurality of slice data, perform at least supplementary prediction processing on battery information corresponding to each of the plurality of battery parameter fields in the initial slice data using at least one iterative detection model from among a plurality of iterative detection models that are matched one-to-one with the plurality of battery parameter fields, to obtain target slice data corresponding to the initial slice data; the target slice data including supplementary information for the missing information and predicted information for the non-missing information;

[0028] The second detection unit is used to perform accuracy detection on the target slice data by using at least one detection model from multiple detection models that match the multiple battery parameter fields one by one, on the supplementary information in the target slice data and the predicted information, to obtain a detection result of the target slice data.

[0029] In a third aspect, an embodiment of the present application provides a battery data detection device, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the program, some or all of the steps in the above method are implemented.

[0030] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which implements some or all of the steps in the above method when executed by a processor.

[0031] In a fifth aspect, an embodiment of the present application provides a computer program product, comprising a computer program or instructions, which implement some or all of the steps in the above method when executed by a processor.

[0032] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and do not limit the technical solutions of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The drawings herein are incorporated into and constitute a part of the specification. These drawings illustrate embodiments consistent with the present application and, together with the specification, are used to illustrate the technical solutions of the present application.

[0034] Figure 1 A schematic diagram of the implementation process of a battery data detection method provided in an embodiment of the present application Figure 1 ;

[0035] Figure 2 A schematic diagram of the implementation process of a battery data detection method provided in an embodiment of the present application Figure 2 ;

[0036] Figure 3 A schematic diagram of the implementation process of a battery data detection method provided in an embodiment of the present application Figure 3 ;

[0037] Figure 4 A schematic diagram of the implementation process of a battery data detection method provided in an embodiment of the present application Figure 4 ;

[0038] Figure 5 A schematic diagram of the implementation process of a battery data detection method provided in an embodiment of the present application Figure 5 ;

[0039] Figure 6 A schematic diagram of the implementation process of a battery data detection method provided in an embodiment of the present application Figure 6 ;

[0040] Figure 7 A schematic diagram of the implementation process of a battery data detection method provided in an embodiment of the present application Figure 7 ;

[0041] Figure 8 A schematic diagram of the implementation process of a battery data detection method provided in an embodiment of the present application Figure 8 ;

[0042] Figure 9 A schematic diagram of the implementation process of a battery data detection method provided in an embodiment of the present application Figure 9 ;

[0043] Figure 10 A schematic diagram of the structure of a battery data detection device provided in an embodiment of the present application;

[0044] Figure 11 This is a hardware entity diagram of a battery data detection device in an embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions of this application are further elaborated in detail below with reference to the accompanying drawings and embodiments. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0046] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0047] The terms "first / second / third" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It is understandable that "first / second / third" can be interchanged with a specific order or sequence where permitted so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing this application only and are not intended to limit this application.

[0049] Currently, new energy batteries are increasingly being used in everyday life and industry. They are not only used in energy storage systems such as hydropower, thermal power, wind power, and solar power plants, but are also widely used in electric vehicles like electric bicycles, electric motorcycles, and electric cars, as well as in a variety of fields such as aerospace. As the application of power batteries continues to expand, market demand is also growing.

[0050] In the embodiments of the present disclosure, the battery may be a battery cell. A battery cell refers to a basic unit that can realize the mutual conversion of chemical energy and electrical energy, and can be used to make a battery module or battery pack, thereby being used to supply power to an electrical device. The battery cell may be a secondary battery, which refers to a battery cell that can be recharged to activate the active material after the battery cell is discharged and continue to be used. The battery cell may be a lithium-ion battery, a sodium-ion battery, a sodium-lithium-ion battery, a lithium metal battery, a sodium metal battery, a lithium-sulfur battery, a magnesium-ion battery, a nickel-metal hydride battery, a nickel-cadmium battery, a lead-acid battery, etc., and the embodiments of the present disclosure are not limited to this.

[0051] In the embodiments of the present disclosure, the battery may also be a single physical module including one or more battery cells to provide higher voltage and capacity. When there are multiple battery cells, the multiple battery cells are connected in series, in parallel or in hybrid via a busbar.

[0052] With the increase in the number of battery application scenarios, various products have also accumulated a large amount of historical data. However, the amount of data is too large (for example, if a battery generates one frame of data per second, there will be 86,400 pieces of data per day, which will accumulate a huge amount of data over a few years), and there are problems such as poor data quality (for example, when the device is not working or is overloaded due to special reasons, communication transmission failure uploads invalid values, transmission interruption causes loss, etc.), which makes it impossible to use it directly; in actual tests, it is also impossible to fully reproduce the actual usage cycle and working conditions. Therefore, how to process historical data to improve data quality is an urgent problem to be solved. In related technologies, whether the data is missing is determined based on the date of the data, so that the missing data can be supplemented, thereby improving the quality of the data. However, the problem with this solution is that the efficiency of data integrity detection is low, and the accuracy of the data cannot be detected.

[0053] In order to solve the technical problems in related technologies of low efficiency of data integrity detection and inability to detect data accuracy, an embodiment of the present application provides a battery data detection method, which can be applied to electronic devices, including but not limited to fixed devices and / or mobile devices. For example, fixed devices include but are not limited to: personal computers (PCs), or servers, etc. The server can be a cloud server or an ordinary server. Mobile devices include but are not limited to: one or more of mobile phones, tablet computers or wearable devices. Figure 1 As shown, the method includes steps S101 to S103, wherein:

[0054] Step S101: determining initial fragmented data with missing information, wherein the initial fragmented data includes battery information corresponding to a plurality of battery parameter fields.

[0055] Here, the battery parameter field may include at least one of the following: battery voltage, battery current, battery resistance, battery temperature, battery charge state, battery operating state, battery flag, and vehicle type on which the battery is installed.

[0056] In some embodiments, the data type of the battery information may include at least one of the following: character type, numeric type, and Boolean type.

[0057] In the embodiment of the present application, a battery data set may be first obtained, and then the battery data set may be segmented according to a preset time period to obtain multiple segmented data belonging to different time periods. The segmented data with missing information among the multiple segmented data may be used as the initial segmented data.

[0058] Step S102, using at least one iterative detection model from a plurality of iterative detection models that match the plurality of battery parameter fields one-to-one, at least supplementary prediction processing is performed on the battery information corresponding to the plurality of battery parameter fields in the initial sharded data to obtain target sharded data corresponding to the initial sharded data; the target sharded data includes supplementary information for the missing information and prediction information for the non-missing information.

[0059] Here, the initial sharded data may be sharded data in which there is missing information in a plurality of sharded data and the proportion of the missing information is less than or equal to a preset first proportion threshold. Exemplarily, the first proportion threshold may be 10%. One-to-one matching of multiple iterative detection models with multiple battery parameter fields means that the battery parameter fields of the battery information processed by each iterative detection model are different. Exemplarily, when the plurality of battery parameter fields include battery voltage, battery current, battery resistance, and battery temperature, the plurality of iterative detection models include an iterative detection model for processing battery information of battery voltage, an iterative detection model for processing battery information of battery current, an iterative detection model for processing battery information of battery resistance, and an iterative detection model for processing battery information of battery temperature.

[0060] Here, at least one iterative detection model refers to an iterative detection model corresponding to at least one battery parameter field to be processed. Because the number of battery parameter fields in the initial fragmented data is large, in some application scenarios, it is not necessary to detect the battery information of all battery parameter fields. Therefore, it is necessary to determine at least one battery parameter field from multiple iterative detection models that match multiple battery parameter fields one by one. For example, if only the battery information of the battery voltage and battery current needs to be detected, then the iterative detection model corresponding to the battery voltage and the iterative detection model corresponding to the battery information can be determined from the multiple iterative detection models.

[0061] It should be noted that for the same initial sliced ​​data, when testing different battery parameter fields, the input to different iterative detection models is the same initial sliced ​​data. For example, if battery information such as battery voltage and battery current needs to be tested, the initial sliced ​​data is input into the iterative detection model corresponding to battery voltage and the iterative detection model corresponding to battery information, respectively, to obtain the corresponding target sliced ​​data.

[0062] In an embodiment of the present application, the training process for the missing detection processing of the iterative detection model for each battery parameter field may be as follows: constructing a first sample set, wherein the first sample set includes the battery information corresponding to the battery parameter field, and each battery information includes corresponding label information. For example, when the battery information is missing, the corresponding label information is the first label information indicating that the battery information is not missing; when the battery information is not missing, the corresponding label information is the second label information indicating that the battery information is missing. The iterative detection model is then trained using this first sample set, so that the iterative detection model can perform missing detection on the initial segmented data. In some embodiments, the battery information in the first sample set used to train the iterative detection models for different battery parameter fields can be the same, except that the battery parameter fields corresponding to the label information in different first sample sets are different. For example, the first sample set for battery voltage can include battery information for all battery parameter fields, but only the battery information for battery voltage includes label information. In other embodiments, the battery information in the first sample set used to train the iterative detection models for different battery parameter fields is different. Illustratively, the first sample set of the battery voltage only includes the battery information of the battery voltage; it may also include the battery information of the battery voltage and battery information related to the battery voltage.

[0063] In some embodiments, the iterative detection model for each battery parameter field can also be trained for predictive processing. The predictive processing training process includes: establishing a second sample set, which includes both non-missing battery information and missing battery information, and the missing battery information contains a label of the correct battery information. Similar to the first sample set, the battery information in the second sample set can be the same or different. Training the iterative detection model with the second sample set enables the iterative detection model to supplement the missing battery information and predict the non-missing battery information.

[0064] In some embodiments, the iterative detection model may be a model constructed using an eXtreme Gradient Boosting (xgboost) algorithm.

[0065] In an embodiment of the present application, initial shard data can be first determined from a plurality of shard data, and then at least one iterative detection model can be used to supplement the missing information in the battery information corresponding to the plurality of battery parameter fields in the initial shard data, and to perform prediction processing on the non-missing information, thereby obtaining target shard data containing supplementary information for the missing information and predicted information for the non-missing information.

[0066] In some embodiments, if each iterative detection model includes a missing complement model and a prediction model, step S102 may include: using the missing complement model to complement the missing information in the initial fragmented data to obtain supplementary information for the missing information; and using the prediction model to predict the non-missing information and the supplementary information in the initial fragmented data to obtain target fragmented data corresponding to the initial fragmented data. The training process for the missing complement model can refer to the training process for missing detection processing of the iterative detection model; the training process for the prediction model can refer to the training process for prediction processing of the iterative detection model.

[0067] In some embodiments, if each iterative detection model only includes one model, the above-mentioned step S102 may include: using the iterative detection model to supplement the missing information in the initial sharded data to obtain supplementary information of the missing information; using the iterative detection model to predict the non-missing information and supplementary information in the initial sharded data to obtain the target sharded data corresponding to the initial sharded data.

[0068] Step S103: For the target slice data, use at least one detection model from a plurality of detection models that match the plurality of battery parameter fields one by one to perform accuracy detection on the supplementary information and the prediction information in the target slice data to obtain a detection result of the target slice data.

[0069] Here, the one-to-one matching of multiple detection models with multiple battery parameter fields means that the battery parameter fields of the battery information processed by each detection model are different. For example, when the multiple battery parameter fields include battery voltage, battery current, battery resistance and battery temperature, then the multiple detection models include a detection model for processing battery voltage, a detection model for processing battery current, a detection model for processing battery resistance and a detection model for processing battery temperature. At least one detection model refers to a detection model corresponding to at least one battery parameter field to be processed. The detection result is used to characterize the proportion of abnormal data in the target shard data. In some embodiments, the detection result may also include label information of the abnormal data. In other words, the abnormal data in the target shard data can be determined through the detection result.

[0070] In an embodiment of the present application, after the iterative detection model outputs the target slice data, the detection model corresponding to the battery parameter field of the iterative detection model can be used to perform accuracy detection on the supplementary information in the target slice data and the predicted information. For example, when at least one battery parameter field to be processed is battery current and battery voltage, the iterative detection model for battery current can output target slice data corresponding to the battery current, and the iterative detection model for battery voltage can output target slice data corresponding to the battery voltage. The target slice data corresponding to the battery current is then input into the detection model for battery current to obtain a detection result for the target slice data for the battery current. The target slice data corresponding to the battery voltage is then input into the detection model for battery voltage to obtain a detection result for the target slice data for the battery voltage.

[0071] In some embodiments, the detection model may be a model constructed by a bootstrap aggregating (bagging) algorithm.

[0072] In some embodiments, the detection model of each battery parameter field can also be trained for accuracy processing, and the accuracy processing training includes: obtaining a battery data set for the battery parameter field; randomly obtaining multiple sample data in the battery data set for the battery parameter field, repeating the step of obtaining sample data multiple times to obtain multiple sample training sets for the battery parameter field; based on the multiple sample training sets of the battery parameter field, respectively obtaining multiple learners, and using the multiple learners as detection verification models for the battery parameter field.

[0073] In an embodiment of the present application, at least one of the multiple iterative detection models that match multiple battery parameter fields one by one can be used to perform supplementary prediction processing on the initial sliced ​​data with missing information, thereby achieving the supplementation of missing information and the prediction of non-missing information; by using at least one of the multiple detection models that match multiple battery parameter fields one by one, the accuracy of the labeled sliced ​​data output by the iterative detection model is detected, thereby achieving the accuracy detection of the supplementary information of the missing information and the prediction information of the non-missing information. In this way, the integrity detection of the data and the supplementation of the missing information are completed by the iterative detection model corresponding to the battery parameter field, so that the interfering data in the initial sliced ​​data can be eliminated. And the accuracy detection of the data is completed by the detection model corresponding to the battery parameter field, thereby achieving the verification of the output data of the iterative detection model. Because the above two models both correspond to the battery parameter field, the accuracy of the integrity detection and the accuracy detection can also be improved.

[0074] In some embodiments, as Figure 2As shown, the above-mentioned battery data detection method further includes step S201 and step S202; the above-mentioned step S101 can be implemented through step S203:

[0075] Step S201 : Slice a battery data set of a battery of an electric device to obtain a plurality of sliced ​​data belonging to different time periods; each of the sliced ​​data includes battery information corresponding to a plurality of battery parameter fields.

[0076] Here, the power-consuming device can be a device that converts electrical energy into other forms of energy, for example, an electric vehicle, aircraft, ship, electric bicycle, or other electronic device with a power battery. The present embodiment of the application does not limit the type of power-consuming device. The battery data set can be a data table, illustratively, the data table can be a hive table. In the embodiment of the present application, the rows of the data table represent the collection time of the battery information, and the columns represent the battery parameter fields of the battery information.

[0077] In some embodiments, the data type of the battery information may include at least one of the following: character type, numeric type, and Boolean type.

[0078] In an embodiment of the present application, the battery data set can be segmented according to a preset time period to obtain multiple segmented data belonging to different time periods. For example, if the battery data set includes battery information for a battery within a month, the preset time period can be one week, that is, the one-month battery data set is divided into four segmented data, where the time range of the battery information in each segmented data is one week. In some embodiments, the time range of the battery information in each segmented data can also be different, but the time corresponding to the multiple segmented data needs to be continuous.

[0079] In some embodiments, a preset time window can be used to slide through the battery data set to obtain multiple slices of data belonging to different time periods. The size of the time window can be greater than or equal to the number of sliding steps of the time window.

[0080] Step S202: When there is missing information in the slice data, determine the proportion of the missing information in the slice data according to the battery parameter field.

[0081] In an embodiment of the present application, each shard data can be traversed to determine the shard data with missing information, and then the proportion of missing information under each battery parameter field can be determined. Here, the proportion of missing information refers to the proportion of missing information to the battery information under the battery parameter field. For example, shard data 1 includes field 1 and field 2, and field 1 and field 2 both include 10 battery information. Among them, there are 2 missing information in field 1 and 4 missing information in field 2. Then the proportion of missing information to the battery information under field 1 is 20%, and the proportion of missing information to the battery information under field 2 is 40%.

[0082] Step S203 : determining the fragmented data whose missing information ratio is less than or equal to a preset first ratio threshold as the initial fragmented data.

[0083] It is understandable that the more missing information there is in the input of the iterative detection model, the lower the accuracy of the iterative detection model's supplementary prediction will be. Therefore, in order to improve the accuracy of the iterative detection model's supplementary prediction, it is necessary to control the proportion of missing information in the initial sharded data.

[0084] In an embodiment of the present application, for each battery parameter field, it can be determined whether there is any fragmented data in the multiple fragmented data whose proportion of missing information under the battery parameter field is less than or equal to a preset first proportion threshold. If so, the fragmented data can be used as the input of the iterative detection model of the battery parameter field, that is, the initial fragmented data of the battery parameter field. For example, if the proportion of missing information of fragmented data 1 under field 1 is 10%, the proportion of missing information of fragmented data 1 under field 2 is 5%, and the first proportion threshold is 10%, then fragmented data 1 can be used as the initial fragmented data of the iterative detection model of field 1, and fragmented data 2 can be used as the initial fragmented data of the iterative detection model of field 3.

[0085] In the embodiments of the present application, by sharding the battery data set of the battery of the electric device, the amount of data input to the model can be reduced, thereby improving the computational efficiency of the model. Furthermore, by determining the proportion of missing information in each battery parameter field for the sharded data with missing information, the proportion of missing information in the input of the iterative detection model (i.e., the initial sharded data) can be made less than or equal to a preset first proportion threshold. This can improve the accuracy of the supplementary prediction of the iterative detection model.

[0086] In some embodiments, as Figure 3 As shown, the above-mentioned battery data detection method further includes steps S301 to S304:

[0087] Step S301 : determining the fragmented data whose missing information ratio is greater than a preset first ratio threshold as the first fragmented data.

[0088] In an embodiment of the present application, the first shard data of multiple battery parameter fields can be determined in multiple shard data according to the battery parameter field. For example, the multiple battery parameter fields include field 1 and field 2, and the multiple shard data include shard data 1 and shard data 2; for field 1, determine whether the proportion of missing information of shard data 1 under field 1 is greater than a preset first proportion threshold. If it is, shard data 1 is used as the first shard data of field 1; if the proportion of missing information of shard data 2 under field 1 is greater than the preset first proportion threshold, shard data 2 is used as the first shard data of field 1, and the same applies to field 2. In this way, the first shard data of each battery parameter field can be determined by traversing multiple shard data.

[0089] Step S302: Based on the target battery parameter field corresponding to the missing information of the first fragment data and the identifier of the first fragment data, obtain second fragment data that is temporally associated and does not have missing information in the target battery parameter field from the multiple fragment data.

[0090] Here, the identifier of the first shard data is used to represent the time period information of the first shard data. Because the battery data set of the battery of the electrical device is fragmented according to the time period to obtain multiple shard data belonging to different time periods, when two different shard data are associated in time sequence, their respective identifiers are also associated. Therefore, the shard data corresponding to the identifier that is continuous with the identifier can be determined by the identifier of the first shard data, and then it can be determined whether the shard data has missing information in the target battery parameter field. For example, if shard data 1 and shard data 2 are associated in time sequence, the identifier of shard data 1 can be N, and the identifier of shard data 2 can be N+1 or N-1, where N is a positive integer. When the identifier of the first shard data is 2, shard data 1 with identifier 1 and shard data 2 with identifier 3 can be found, and then it can be determined whether shard data 1 and shard data 2 have missing information in the target battery parameter field. If neither of them has missing information, both shard data 1 and shard data 2 can be used as the second shard data.

[0091] In the embodiment of the present application, two shard data are temporally associated, which means that the two shard data are adjacent in time.

[0092] It is understandable that because the proportion of missing information in the first shard data is relatively large, sharding only the first shard data will still result in a relatively large proportion of missing information in the shard data. Therefore, in an embodiment of the present application, second shard data that is temporally associated and does not have missing information in the target battery parameter field is obtained from multiple shard data using the identifier of the first shard data and the target battery parameter field corresponding to the missing information. In this way, because the second shard data is temporally associated with the first shard data and does not have missing information in the target battery parameter field, after combining the first shard data and the second shard data, the proportion of missing information in the first shard data to the combined data of the first and second shard data can be effectively reduced.

[0093] Step S303: sort the first shard data and the second shard data in time sequence to form a first shard data set.

[0094] Because the first and second shard data are temporally related, the sorted first and second shard data, i.e., the first shard data set, are also temporally continuous. For example, if shard data 1 corresponds to the period from March 12th to March 18th, and shard data 2 corresponds to the period from March 19th to March 25th, then the period corresponding to the shard data set obtained by merging shard data 1 and shard data 2 is March 12th to March 25th.

[0095] Step S304: Sharding the first shard data set so that missing information in the first shard data is distributed among different shard data, and a ratio of missing information in the different shard data is less than or equal to a first ratio threshold.

[0096] In the embodiment of the present application, the first sharded data set can be further sharded so that the proportion of missing information in the sharded data after the sharding is less than or equal to a first ratio threshold. The embodiment of the present application does not limit the manner of the further sharding, and can be time-based sharding or random sharding, as long as the proportion of missing information in the sharded data after the sharding is less than or equal to the first ratio threshold.

[0097] In an embodiment of the present application, for the first shard data whose proportion of missing information is greater than a preset first proportion threshold, the second shard data that is associated with the first shard data in time sequence and does not have missing information on the battery parameter field of the missing information of the first shard data can be obtained, and then the first shard data and the second shard data (i.e., the first shard data set) that are sorted in time sequence are re-sharded so that the proportion of missing information on the multiple shard data after sharding is less than or equal to the first proportion threshold. In this way, by obtaining the second shard data, the proportion of missing information in the combined data of the first shard data and the second shard data can be effectively reduced, and the battery information in the first shard data set is continuous in time sequence, and the battery information in the shard data after re-sharding is also continuous in time, so that the iterative detection model can perform supplementary prediction processing on the continuous battery information, thereby improving the accuracy of the iterative detection model.

[0098] In some embodiments, as Figure 4 As shown, the above-mentioned battery data detection method further includes steps S401 to S403:

[0099] Step S401, when the second shard data does not exist in the multiple shard data, based on the target battery parameter field corresponding to the missing information of the first shard data and the identifier of the first shard data, obtain the third shard data associated in time sequence from the target shard data corresponding to the target battery parameter field obtained after supplementary prediction processing of the initial shard data.

[0100] Here, the target slice data corresponding to the target battery parameter field refers to the target slice data output by the iterative detection model corresponding to the target battery parameter field.

[0101] In an embodiment of the present application, when the second slice data does not exist in the multiple slice data, it means that the first slice data cannot be sliced ​​again at the current moment. At this time, the slice processing of the first slice data can be suspended first, and the initial slice data that has been determined is first input into the corresponding iterative detection model for supplementary prediction processing to obtain the corresponding target slice data. Then, after obtaining the target slice data, based on the target battery parameter word corresponding to the missing information of the first slice data, the target slice data output by the iterative detection model corresponding to the target battery parameter word can be determined from the multiple target slice data, that is, the target slice data corresponding to the target battery parameter field. Then, in the target slice data corresponding to the target battery parameter field, the third slice data that is temporally associated with the first slice data is determined. Because the target shard data corresponding to the target battery parameter field is obtained after the iterative detection model performs supplementary prediction processing on the initial shard data, the target shard data corresponding to the target battery parameter field must have no missing information on the target battery parameter field. Therefore, it is only necessary to determine whether there is a third shard data in the target shard data corresponding to the target battery parameter field that is temporally associated with the first shard data.

[0102] Step S402: sort the first shard data and the third shard data in time sequence to form a second shard data set.

[0103] Step S403: Sharding the second shard data set so that the missing information in the first shard data is distributed in different shard data, and the ratio of the missing information in the different shard data is less than or equal to a first ratio threshold.

[0104] In the embodiment of the present application, the implementation process of step S402 and step S403 can refer to step S303 and step S304.

[0105] In an embodiment of the present application, when the second shard data does not exist in the multiple shard data, based on the target battery parameter field corresponding to the missing information of the first shard data and the identifier of the first shard data, the target shard data corresponding to the target battery parameter field obtained after the initial shard data is supplemented with prediction processing is obtained, and then the first shard data and the third shard data (i.e., the second shard data set) sorted in time sequence are re-sharded so that the proportion of missing information on the multiple shard data after sharding is less than or equal to the first proportion threshold. In this way, by obtaining the third shard data, the proportion of missing information in the combined data of the first shard data and the third shard data can be effectively reduced, and the battery information in the second shard data set is continuous in time, and the battery information in the shard data after re-sharding is also continuous in time, so that the iterative detection model can perform supplementary prediction processing on the continuous battery information, thereby improving the accuracy of the iterative detection model.

[0106] In some embodiments, as Figure 5 As shown, the above-mentioned battery data detection method further includes step S501; the above-mentioned step S103 can be implemented through step S502:

[0107] Step S501: Determine the deviation ratio between the target fragment data and the initial fragment data.

[0108] Here, the deviation ratio is the ratio of battery information with deviations in the target slice data to all battery information. Whether the battery information in the target slice data is deviated can be determined by determining whether the battery information in the target slice data is the same as the battery information in the initial slice data. For example, if target slice data 1 is output by the iterative detection model corresponding to field 1, and there are 1,000 pieces of battery information in field 1 for target slice data 1, and 50 pieces of battery information are deviated, then the deviation ratio is 5%.

[0109] Step S502: For the first target slice data in which the deviation ratio in the target slice data is less than or equal to a preset threshold, at least one detection model from a plurality of detection models that match the plurality of battery parameter fields one by one is used to perform accuracy detection on the supplementary information in the first target slice data and the prediction information to obtain a detection result of the first target slice data.

[0110] In an embodiment of the present application, it is possible to first determine whether the deviation ratio of the target slice data is less than or equal to a preset threshold value. If it is less than, the first target slice data whose deviation ratio is less than or equal to the preset threshold value can be input into the detection model, so that the detection model performs accuracy detection on the supplementary information and the prediction information in the first target slice data. In this way, by calculating the deviation ratio of the target slice data, the accuracy of the target slice data can be pre-detected. When the pre-detection passes (that is, the deviation ratio of the target slice data is less than or equal to the preset threshold value), the first target slice data is input into the detection model, and the accuracy detection is performed on the supplementary information and the prediction information in the first target slice data. In this way, the amount of calculation of the detection model can be reduced, and the efficiency of the detection model in performing accuracy detection on the target slice data can be improved.

[0111] In some embodiments, as Figure 6 As shown, the above-mentioned battery data detection method further includes steps S601 to S603:

[0112] Step S601: For second target shard data in the target shard data whose deviation ratio is greater than a preset threshold, fourth shard data associated in time sequence is acquired from the initial shard data based on an identifier of the second target shard data.

[0113] Step S602: sort the initial fragmented data corresponding to the second target fragmented data and the fourth fragmented data in time sequence to form a third fragmented data set.

[0114] In an embodiment of the present application, when the deviation ratio of the second target shard data is greater than a preset threshold, it indicates that the accuracy of the second target shard data is low. The reason may be that the amount of input data of the iterative detection model is small. Therefore, the shard data can be reconstructed based on the identifier of the second target shard data, that is, through the identifier of the second target shard data, the fourth shard data associated in time sequence is obtained from the initial shard data, and then the initial shard data and the fourth shard data corresponding to the second target shard data are sorted in time sequence to form a third shard data set. Wherein, the proportion of missing information in the third shard data set is less than or equal to the first proportion threshold.

[0115] Step S603: Perform supplementary prediction processing on the third shard data set based on the iterative detection model until the deviation ratio between the processed shard data and the third shard data set is less than or equal to a preset threshold, and use the processed shard data as the target shard data of the third shard data set.

[0116] In an embodiment of the present application, the reconstructed third shard data set can be input into the iterative detection model to perform supplementary processing on the missing information and predictive processing on the non-missing information. If the deviation ratio between the processed shard data and the third shard data set is still greater than the preset threshold, the processed shard data is used as the second target shard data, and steps S601 to S603 of "performing supplementary predictive processing on the third shard data set based on the iterative detection model" are repeated until the deviation ratio between the processed shard data and the third shard data set is less than or equal to the preset threshold, and the processed shard data is used as the target shard data of the third shard data set.

[0117] In an embodiment of the present application, for second target slice data whose deviation ratio is greater than a preset threshold, the fourth slice data associated in time sequence can be obtained through the identifier of the second target slice data, thereby expanding the input data of the iterative detection model, thereby making the deviation ratio of the target slice data output by the iterative detection model less than or equal to the preset threshold. In this way, by adjusting the size of the input data of the iterative detection model, the deviation ratio between the output data of the iterative detection model and the input data can be reduced, thereby reducing the number of manual interventions and improving the efficiency of battery information quality testing.

[0118] In some embodiments, as Figure 7 As shown, the above-mentioned battery data detection method further includes steps S701 to S704:

[0119] Step S701: receiving a data processing request sent by a back-end detection module of a battery; the data processing request is used to request detection of battery information of at least one battery parameter field to be processed.

[0120] Here, the back-end detection module is used to detect abnormalities in the battery of the electrical device.

[0121] In practical applications, the backend detection module needs to regularly check the battery of an electrical device for abnormalities. However, the data collected by the data acquisition module may have quality issues for various reasons, making it impossible to directly use this data to detect battery abnormalities. Therefore, the backend detection module can determine at least one battery parameter field to be processed based on the current abnormality detection type and then send a data processing request to the battery data detection device.

[0122] In some embodiments, the data processing request also carries the time period of the battery information, so that the corresponding battery data set can be obtained according to the time period of the battery information.

[0123] Step S702: Determine, based on the data processing request, a battery data set corresponding to the at least one battery parameter field to be processed.

[0124] In the embodiment of the present application, a battery data set corresponding to at least one battery parameter field to be processed may be obtained from a database.

[0125] In some embodiments, if the data processing request also carries a time period of battery information, a battery data set corresponding to at least one battery parameter field to be processed within the time period may be obtained from the database.

[0126] Step S703, based on the at least one battery parameter field to be processed, determine the at least one iterative detection model and the at least one detection model from the multiple iterative detection models that match the multiple battery parameter fields one-to-one and the multiple detection models that match the multiple battery parameter fields one-to-one.

[0127] Exemplarily, if the at least one battery parameter field to be processed includes a battery current, an iterative detection model corresponding to the battery current may be determined from a plurality of iterative detection models, and a detection model corresponding to the battery current may be determined from a plurality of detection models.

[0128] Step S704: When the detection result of the target slice data meets a preset condition, the target slice data is sent to a back-end detection module of the battery.

[0129] In an embodiment of the present application, after the detection model outputs the detection result of the target slice data, it is necessary to determine whether the detection result meets the preset conditions, wherein if the detection result is label information, the label information of the target slice data is label information that characterizes the accuracy of the target slice data, which means that the detection result of the target slice data meets the preset conditions; if the detection result is the accuracy score of the target slice data, the preset condition is that the accuracy score is greater than or equal to the preset score value. Exemplarily, the preset score value can be 80. In other embodiments, the detection result includes label information of whether each supplementary information in the target slice data has passed the accuracy test and label information of whether each predicted information has passed the accuracy test. If the proportion of supplementary information that has passed the accuracy test and predicted information that has passed the accuracy test is greater than or equal to the preset proportion, it is considered that the detection result of the target slice data meets the preset conditions. Exemplarily, the preset proportion can be 90%.

[0130] For each battery parameter field, the detection result of the target slice data output by the detection model of the battery parameter field can be the supplementary information in the target slice data and the detection result under the battery parameter field.

[0131] In an embodiment of the present application, if the detection result of the target slice data meets the preset conditions, indicating that the supplementary information and the predicted information of the target slice data are relatively accurate, the target slice data can be returned to the back-end detection module. After receiving the target slice data, the back-end detection module can perform an abnormality detection on the battery of the electric device based on the target slice data to determine whether the battery of the electric device has an abnormality.

[0132] In an embodiment of the present application, based on a data processing request sent by a back-end detection module for requesting detection of battery information of at least one battery parameter field to be processed, a battery data set corresponding to the at least one battery parameter field to be processed can be determined, and the at least one iterative detection model and the at least one detection model can be determined from a plurality of iterative detection models that match the multiple battery parameter fields one-to-one and a plurality of detection models that match the multiple battery parameter fields one-to-one. In this way, the corresponding battery data set, iterative detection model, and detection model can be determined according to different types of data processing requests. And the target slice data whose detection results meet the preset conditions can be sent to the back-end detection module, so that the battery detection module performs abnormality detection on the battery of the electrical device based on the target slice data. In this way, the battery detection module can obtain more accurate battery detection results through battery information with higher data quality.

[0133] In some embodiments, the battery data detection method further includes steps S801 and S802:

[0134] Step S801: When it is detected through binary logs that a battery source data set is stored in the database of the data warehouse, the battery source data set is extracted into the operational data storage layer of the data warehouse; the battery source data set is a data set obtained by performing attribute detection on the battery of the electrical device.

[0135] Here, the binary log (binlog) is used to record changes in the database table structure. Therefore, the binary log can be used to detect whether the battery source data set is stored in the database. If it is determined that it is stored, the battery source data set can be extracted to the operational data storage layer (ODS) of the data warehouse. For example, attribute detection can be voltage detection, current detection, temperature detection, etc.

[0136] Step S802 : performing logic processing on the battery source data set in the operation data storage layer to obtain the battery data set.

[0137] Here, the logical processing may be at least one of Map processing and Reduce processing.

[0138] In the embodiment of the present application, the binary log can be used to promptly detect whether a battery source data set exists in the database. This allows for timely logical processing of the battery source data set to obtain a battery data set, thereby performing data quality testing on the battery information in the battery data set.

[0139] In some embodiments, the battery data detection method further includes steps S901 and S902:

[0140] Step S901: When the detection result does not meet the preset conditions, determine the abnormal data in the target shard data and store the abnormal data in the database.

[0141] In an embodiment of the present application, when the detection results include label information of whether each supplementary information in the target shard data passes the accuracy detection and label information of whether each predicted information passes the accuracy detection, the abnormal data in the target shard data can be determined through the label information of the supplementary information and the label information of the predicted information, and the abnormal data can be stored in the database.

[0142] Step S902 : generating an alarm tag based on the abnormal data, and generating a data abnormality alarm table based on the abnormal data and the alarm tag; the data abnormality alarm table is used to adjust the logical processing.

[0143] In the embodiment of the present application, the alarm label indicates that the accuracy test of the abnormal data has failed.

[0144] In some embodiments, the abnormality alarm table also includes the location information of the abnormal data in the target slice data and / or the location information of the original battery information corresponding to the abnormal data in the battery data set. In some embodiments, the location information can be represented by (time, battery parameter field). By generating a data abnormality alarm table, an effective basis can be provided for adjusting the logic processing, thereby improving the accuracy of the generated battery data set.

[0145] In this embodiment of the present application, if the test results do not meet the preset conditions, abnormal data in the target slice data is identified and stored in the database. Storing only abnormal data in the database reduces storage space overhead. Furthermore, by generating a data anomaly alarm table, an effective basis for adjusting logical processing is provided, thereby improving the accuracy of the generated battery data set.

[0146] In some embodiments, as Figure 8 As shown, the above-mentioned battery data detection method can also be implemented through steps S1001 to S1008:

[0147] Step S1001 : When target business data is stored in a business system database, the target business data is extracted to an operational data storage layer in a data warehouse.

[0148] In an embodiment of the present application, a binary log (binlog) may be used to detect whether target business data is stored in a business system database.

[0149] Step S1002, first-stage testing.

[0150] In the embodiment of the present application, a first-stage test may be performed on the target service data. The main contents of the first-stage test include:

[0151] a. Whether the target business data is correctly loaded into the Hadoop system;

[0152] b. Whether the source data and target data table structures are consistent;

[0153] c. Test to verify that the correct data is extracted and loaded into the correct location.

[0154] The verification methods for the above verification content include:

[0155] a. Check whether there is data flowing into the table at the specified location;

[0156] b. The field name, field type, and field format of the header field flowing into the data table. The field type includes character type, numeric type, etc., and the field format is the format of the value corresponding to the field.

[0157] Step S1003: Map the tested target business data.

[0158] Step S1004: Perform Reduce processing on the data processed by Map.

[0159] In the embodiment of the present application, Map processing and Reduce processing can be performed in parallel by multiple processing nodes.

[0160] Step S1005, second stage test.

[0161] In this embodiment of the present application, a two-stage test can be performed on the data after Reduce processing. The main contents that need to be verified in the second stage test include:

[0162] a. Whether MapReduce of each node is running normally

[0163] b. Is the output file correct?

[0164] c. Whether the key-value pairs in the MapReduce processing are generated correctly

[0165] d. After the Reduce process is completed, is the data aggregation and merging correct?

[0166] The verification methods for the above verification content include:

[0167] a. Check the node running status and verify the output file format

[0168] b. Compare the keys and values ​​in the single-node and multi-node running results.

[0169] Step S1006: Perform supplementary detection processing on the tested data to obtain a detection result.

[0170] In the embodiment of the present application, step S1006 can be implemented by the following steps:

[0171] Step S1061: The alarm monitoring system performs fragmentation processing on the tested data to obtain fragmented data.

[0172] Here, the processed data is in the Hive table, and the Hive table can be sharded according to the sharding rule. The sharding rule can be sharding by time or by table name.

[0173] Step S1062: Input each shard data into each xgboost model to obtain supplementary prediction data for each field.

[0174] For example, Figure 9 As shown, the shard data includes shard data 1 to shard data x, and shard data 1 to shard data x all include data of field 1, field 2, and field 3. Shard data 1 is input into the xgboost model of field 1 to obtain test 1.1, which represents the supplementary prediction data of shard data 1 for field 1 under the xgboost model; similarly, test 1.2 represents the supplementary prediction data of shard data 1 for field 2 under the xgboost model; test 1.3 represents the supplementary prediction data of shard data 1 for field 3 under the xgboost model; shard data x is input into the xgboost model of field 1 to obtain test x.1, which represents the supplementary prediction data of shard data x for field 1 under the xgboost model; similarly, test x.2 represents the supplementary prediction data of shard data x for field 2 under the xgboost model; test x.3 represents the supplementary prediction data of shard data x for field 3 under the xgboost model.

[0175] In some embodiments, the fields may include at least one of the following: a list of all cell voltages, an average voltage of all cells, a temperature list, an average temperature, a vehicle status, a charging status, an operating status, a total cell voltage, a total cell current, a total cell resistance, a device ID, a vehicle type, and a reporting time.

[0176] In step S1063 , the supplementary prediction data for each field is input into each bagging model to obtain a detection result of the supplementary prediction data for each field.

[0177] For example, Figure 9 As shown, the test 1.1 of the shard data 1 under field 1 can be input into the abnormality judgment module 901, and the abnormality judgment module 901 can use the bagging model corresponding to field 1 to detect test 1.1 and obtain the detection result of test 1.1; similarly, the test 1.2 of the shard data 1 under field 2 is input into the abnormality judgment module 901, and the abnormality judgment module 901 can use the bagging model corresponding to field 2 to detect test 1.2 and obtain the detection result of test 1.2; the test 1.3 of the shard data 1 under field 3 is input into the abnormality judgment module 901, and the abnormality judgment module 901 can use the bagging model corresponding to field 3 to detect test 1.3 and obtain the detection result of test 1.3.

[0178] The test x.1 of the shard data x under field 1 can be input into the abnormality judgment module 901, and the abnormality judgment module 901 can use the bagging model corresponding to field 1 to detect the test x.1 and obtain the detection result of the test x.1; similarly, the test x.2 of the shard data x under field 2 can be input into the abnormality judgment module 901, and the abnormality judgment module 901 can use the bagging model corresponding to field 2 to detect the test x.2 and obtain the detection result of the test x.2; the test x.3 of the shard data x under field 3 can be input into the abnormality judgment module 901, and the abnormality judgment module 901 can use the bagging model corresponding to field 3 to detect the test x.3 and obtain the detection result of the test x.3.

[0179] In some embodiments, the detection results may be manually confirmed, and the xgboost model and the bagging model may be retrained based on the manual confirmation results.

[0180] Step S1007: Perform alarm processing based on the detection result.

[0181] In an embodiment of the present application, alarm processing can be to establish a system monitoring alarm table based on the detection results, and send the system monitoring alarm table to the user. The user can adjust the map processing and reduce processing according to the system monitoring alarm table.

[0182] Step S1008: storing the alarm data corresponding to the detection result in the distributed file system.

[0183] like Figure 9 As shown, after the abnormality judgment module 901 outputs the detection results of the fragment data 1 to the fragment data x under each field, step S1101, i.e., abnormal data processing, can be performed according to each detection result; wherein, the abnormal data processing can be the above-mentioned step S1005 and / or step S1006. In addition, after the abnormal data processing is performed, step S1102, i.e., updating the model, can be performed based on the alarm data obtained after the abnormal data processing after manual confirmation. The model can be an xgboost model and a bagging model. Wherein, updating the model can refer to updating the model according to the manual confirmation result. For example, if it is manually determined that the alarm data does not belong to abnormal data, the xgboost model and the bagging model are retrained based on the alarm data.

[0184] Figure 10 A schematic diagram of the structure of a battery data detection device provided in an embodiment of the present application is shown as follows: Figure 10 As shown, the battery data detection device 1000 includes: a determination unit 1001, a first detection unit 1002 and a second detection unit 1003, wherein:

[0185] A determining unit 1001 is configured to determine initial fragmented data having missing information, wherein the initial fragmented data includes battery information corresponding to a plurality of battery parameter fields;

[0186] A first detection unit 1002 is configured to perform at least supplementary prediction processing on the battery information corresponding to the multiple battery parameter fields in the initial sharded data using at least one iterative detection model from the multiple iterative detection models that are matched one-to-one with the multiple battery parameter fields, to obtain target sharded data corresponding to the initial sharded data; the target sharded data includes supplementary information for the missing information and predicted information for the non-missing information;

[0187] The second detection unit 1003 is used to perform accuracy detection on the target slice data using at least one detection model from multiple detection models that match the multiple battery parameter fields one by one, on the supplementary information in the target slice data and the predicted information, to obtain a detection result of the target slice data.

[0188] In some embodiments, the battery data detection device 1000 also includes: a sharding unit; a sharding unit, used to perform sharding processing on the battery data set of the battery of the electrical device to obtain multiple sharded data belonging to different time periods; each of the sharded data includes battery information corresponding to multiple battery parameter fields; in the case where there is missing information in the sharded data, the proportion information of the missing information in the sharded data is determined according to the battery parameter field; the determination unit 1001 is also used to determine the sharded data whose proportion information of the missing information is less than or equal to a preset first proportion threshold as the initial sharded data.

[0189] In some embodiments, the above-mentioned sharding unit is further used to determine the sharding data whose proportion information of the missing information is greater than a preset first proportion threshold as the first sharding data; based on the target battery parameter field corresponding to the missing information of the first sharding data and the identifier of the first sharding data, obtain the second sharding data that is associated in time series and does not have missing information in the target battery parameter field from the multiple sharding data; sort the first sharding data and the second sharding data in time series to form a first sharding data set; and perform sharding processing on the first sharding data set so that the missing information in the first sharding data is distributed in different sharding data, and the proportion information of the missing information in the different sharding data is less than or equal to the first proportion threshold.

[0190] In some embodiments, the above-mentioned sharding unit is further used to, when the second sharding data does not exist in the multiple sharding data, obtain third sharding data that is associated in time sequence from the target sharding data corresponding to the target battery parameter field obtained after supplementary prediction processing of the initial sharding data based on the target battery parameter field corresponding to the missing information of the first sharding data and the identifier of the first sharding data; sort the first sharding data and the third sharding data in time sequence to form a second sharding data set; and perform sharding processing on the second sharding data set so that the missing information in the first sharding data is distributed in different sharding data, and the proportion of the missing information in the different sharding data is less than or equal to a first proportion threshold.

[0191] In some embodiments, the above-mentioned battery data detection device 1000 also includes a deviation determination unit; the deviation determination unit is used to determine the deviation ratio between the target shard data and the initial shard data; the second detection unit 1003 is also used to, for the first target shard data in which the deviation ratio in the target shard data is less than or equal to a preset threshold, use at least one detection model from multiple detection models that match the multiple battery parameter fields one by one to perform accuracy detection on the supplementary information and the prediction information in the first target shard data to obtain a detection result of the first target shard data.

[0192] In some embodiments, the first detection unit 1002 is further used to obtain, for the second target shard data in the target shard data, fourth shard data associated in time series from the initial shard data based on an identifier of the second target shard data, with the deviation ratio being greater than a preset threshold; sort the initial shard data and the fourth shard data corresponding to the second target shard data in time series to form a third shard data set; perform supplementary prediction processing on the third shard data set based on the iterative detection model until the deviation ratio between the processed shard data and the third shard data set is less than or equal to the preset threshold, and use the processed shard data as the target shard data of the third shard data set.

[0193] In some embodiments, the above-mentioned battery data detection device 1000 also includes a receiving unit, a set determination unit, a model determination unit and a sending unit; the receiving unit is used to receive a data processing request sent by the battery's back-end detection module; the data processing request is used to request detection of battery information of at least one battery parameter field to be processed; the set determination unit is used to determine the battery data set corresponding to the at least one battery parameter field to be processed based on the data processing request; the model determination unit is used to determine the at least one iterative detection model and the at least one detection model from multiple iterative detection models that match the multiple battery parameter fields one-to-one and multiple detection models that match the multiple battery parameter fields one-to-one based on the at least one battery parameter field to be processed; the sending unit is used to send the target shard data to the battery's back-end detection module if the detection result of the target shard data meets a preset condition; the battery detection module is used to perform abnormality detection on the battery of the electric device based on the target shard data to determine whether the battery of the electric device has an abnormality.

[0194] In some embodiments, the first detection unit 1002 is further used to, when the iterative detection model includes a missing supplement model and a prediction model, use the missing supplement model to supplement the missing information in the initial shard data to obtain supplementary information for the missing information; use the prediction model to predict the non-missing information and the supplementary information in the initial shard data to obtain target shard data corresponding to the initial shard data; or use the iterative detection model to supplement the missing information in the initial shard data to obtain supplementary information for the missing information; use the iterative detection model to predict the non-missing information and the supplementary information in the initial shard data to obtain target shard data corresponding to the initial shard data.

[0195] In some embodiments, the above-mentioned battery data detection device 1000 also includes an extraction unit and a logic processing unit; the extraction unit is used to extract the battery source data set into the operation data storage layer of the data warehouse when it is detected through the binary log that a battery source data set is stored in the database of the data warehouse; the battery source data set is a data set obtained by performing attribute detection on the battery of the electrical device; the logic processing unit is used to perform logical processing on the battery source data set in the operation data storage layer to obtain the battery data set; wherein the battery parameter field in the battery data set includes at least one of the following: battery voltage, battery current, battery resistance, battery temperature, battery charging status, battery operating status, battery flag and the type of vehicle on which the battery is installed.

[0196] In some embodiments, the above-mentioned battery data detection device 1000 also includes an abnormality determination unit and a generation unit; the abnormality determination unit is used to determine the abnormal data in the target shard data when the detection result does not meet the preset conditions, and store the abnormal data in the database; the generation unit is used to generate an alarm label based on the abnormal data, and generate a data abnormality alarm table based on the abnormal data and the alarm label; the data abnormality alarm table is used to adjust the logical processing.

[0197] The description of the above device embodiment is similar to the description of the above method embodiment and has similar beneficial effects as the method embodiment. In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to perform the methods described in the above method embodiments. For technical details not disclosed in the device embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0198] It should be noted that, in the embodiment of the present application, if the above-mentioned data processing method is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods of each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM), a magnetic disk or an optical disk. In this way, the embodiment of the present application is not limited to any specific hardware, software or firmware, or any combination of hardware, software and firmware.

[0199] An embodiment of the present application provides a computer device including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the program, some or all of the steps in the above method are implemented.

[0200] The present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements some or all of the steps in the above method. The computer-readable storage medium may be transient or non-transient.

[0201] An embodiment of the present application provides a computer program, including computer-readable code. When the computer-readable code runs in a computer device, a processor in the computer device executes some or all of the steps for implementing the above method.

[0202] The present application provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program. When the computer program is read and executed by a computer, some or all of the steps in the above method are implemented. The computer program product can be implemented in hardware, software, or a combination thereof. In some embodiments, the computer program product is embodied as a computer storage medium. In other embodiments, the computer program product is embodied as a software product, such as a software development kit (SDK).

[0203] It should be noted that the descriptions of the various embodiments above tend to emphasize the differences between the various embodiments, and their similarities or similarities can be referenced to each other. The descriptions of the above device, storage medium, computer program, and computer program product embodiments are similar to the descriptions of the above method embodiments and have similar beneficial effects as the method embodiments. For technical details not disclosed in the embodiments of the device, storage medium, computer program, and computer program product of this application, please refer to the description of the method embodiments of this application for understanding.

[0204] Figure 11 A hardware entity diagram of a battery data detection device in an embodiment of the present application is shown in FIG. Figure 11 As shown, the hardware entity of the battery data detection device 1100 includes: a processor 1101, a communication interface 1102 and a memory 1103, wherein:

[0205] The processor 1101 generally controls the overall operation of the computer device 1100. The overall operation may be to implement the battery data detection method provided in the embodiment of the present application, for example, Figures 1 to 8 The method shown.

[0206] The communication interface 1102 enables the computer device to communicate with other terminals or servers through a network.

[0207] The memory 1103 is configured to store instructions and applications executable by the processor 1101. It can also cache data to be processed or processed by the processor 1101 and various modules in the computer device 1100 (for example, image data, audio data, voice communication data, and video communication data). It can be implemented using flash memory (FLASH) or random access memory (RAM). Data can be transmitted between the processor 1101, the communication interface 1102, and the memory 1103 via the bus 1104.

[0208] An embodiment of the present application provides a computer storage medium storing one or more programs. The one or more programs can be executed by one or more processors to implement the steps of the battery data detection method of any of the above embodiments.

[0209] It should be noted that the description of the above storage medium and device embodiments is similar to the description of the above method embodiments and has similar beneficial effects as the method embodiments. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0210] The processor may be at least one of an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller, and a microprocessor. It is understood that the electronic device that implements the functions of the processor may also be other electronic devices, which are not specifically limited in the embodiments of the present application.

[0211] The above-mentioned computer storage medium / memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory (Flash Memory), a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); it can also be various terminals including one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0212] It should be understood that "one embodiment" or "an embodiment" mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned steps / processes does not mean the order of execution, and the execution order of each step / process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. The above-mentioned serial numbers of the embodiments of the present application are for description only and do not represent the advantages and disadvantages of the embodiments.

[0213] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0214] The above is only an implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A method for detecting battery data, characterized in that: The method comprises: Determining initial fragmented data having missing information, wherein the initial fragmented data includes battery information corresponding to a plurality of battery parameter fields; performing at least supplementary prediction processing on battery information corresponding to each of the plurality of battery parameter fields in the initial sharded data using at least one iterative detection model from a plurality of iterative detection models that are matched one-to-one with the plurality of battery parameter fields, to obtain target sharded data corresponding to the initial sharded data; the target sharded data including supplementary information for the missing information and predicted information for the non-missing information; For the target slice data, at least one detection model from a plurality of detection models that match the plurality of battery parameter fields one by one is used to perform accuracy detection on the supplementary information and the prediction information in the target slice data to obtain a detection result of the target slice data.

2. The battery data detection method according to claim 1, characterized in that: The method further comprises: Slice the battery data set of the battery of the electric device to obtain a plurality of slice data belonging to different time periods; each of the slice data includes battery information corresponding to a plurality of battery parameter fields; If there is missing information in the slice data, determine the proportion of the missing information in the slice data according to the battery parameter field; The determining of the initial shard data having missing information includes: The fragmented data for which the ratio of the missing information is less than or equal to a preset first ratio threshold is determined as the initial fragmented data.

3. The battery data detection method according to claim 2, characterized in that: The method further comprises: Determine the slice data whose missing information ratio is greater than a preset first ratio threshold as the first slice data; Based on the target battery parameter field corresponding to the missing information of the first fragment data and the identifier of the first fragment data, obtaining second fragment data that is associated in time sequence and does not have missing information in the target battery parameter field from the multiple fragment data; Sort the first shard data and the second shard data in time sequence to form a first shard data set; The first shard data set is sharded so that missing information in the first shard data is distributed among different shard data, and a ratio of missing information in the different shard data is less than or equal to a first ratio threshold.

4. The battery data detection method according to claim 3, characterized in that: The method further comprises: If the second slice data does not exist in the plurality of slice data, obtaining, based on the target battery parameter field corresponding to the missing information of the first slice data and the identifier of the first slice data, third slice data associated in time sequence from the target slice data corresponding to the target battery parameter field obtained after performing supplementary prediction processing on the initial slice data; Sort the first shard data and the third shard data in time sequence to form a second shard data set; The second shard data set is sharded so that the missing information in the first shard data is distributed in different shard data, and the proportion of the missing information in the different shard data is less than or equal to a first proportion threshold.

5. The battery data detection method according to claim 1, characterized in that: The method further comprises: Determine the deviation ratio between the target shard data and the initial shard data; The method of using at least one detection model from among the plurality of detection models that are matched one-to-one with the plurality of battery parameter fields to perform accuracy detection on the supplementary information in the target slice data and the prediction information to obtain a detection result of the target slice data includes: For the first target slice data in which the deviation ratio in the target slice data is less than or equal to a preset threshold, at least one detection model from a plurality of detection models that match the plurality of battery parameter fields one by one is used to perform accuracy detection on the supplementary information in the first target slice data and the prediction information to obtain a detection result of the first target slice data.

6. The battery data detection method according to claim 5, characterized in that: The method further comprises: For the second target shard data in the target shard data, the deviation ratio of which is greater than a preset threshold, based on the identifier of the second target shard data, obtaining fourth shard data associated in time sequence from the initial shard data; sorting the initial shard data corresponding to the second target shard data and the fourth shard data in time sequence to form a third shard data set; Based on the iterative detection model, supplementary prediction processing is performed on the third shard data set until the deviation ratio between the processed shard data and the third shard data set is less than or equal to a preset threshold, and the processed shard data is used as the target shard data of the third shard data set.

7. The battery data detection method according to any one of claims 1 to 6, characterized in that: The method further comprises: Receive a data processing request sent by a back-end detection module of the battery; the data processing request is used to request detection of battery information of at least one battery parameter field to be processed; determining, based on the data processing request, a battery data set corresponding to the at least one battery parameter field to be processed; Based on the at least one battery parameter field to be processed, determining the at least one iterative detection model and the at least one detection model from a plurality of iterative detection models that match the plurality of battery parameter fields one-to-one and a plurality of detection models that match the plurality of battery parameter fields one-to-one; When the detection result of the target slice data meets a preset condition, sending the target slice data to the back-end detection module of the battery; The battery detection module is used to perform abnormality detection on the battery of the electric device based on the target shard data to determine whether the battery of the electric device has an abnormality.

8. The battery data detection method according to any one of claims 1 to 6, characterized in that: performing at least a supplementary prediction process on the battery information corresponding to the multiple battery parameter fields in the initial slice data using at least one iterative detection model from the multiple iterative detection models that match the multiple battery parameter fields one-to-one, to obtain target slice data corresponding to the initial slice data; The target fragment data includes supplementary information for the missing information and predicted information for the non-missing information, including: In a case where the iterative detection model includes a missing supplement model and a prediction model, the missing supplement model is used to supplement the missing information in the initial fragmented data to obtain supplementary information for the missing information; and the prediction model is used to predict the non-missing information in the initial fragmented data and the supplementary information to obtain target fragmented data corresponding to the initial fragmented data. or, The iterative detection model is used to supplement the missing information in the initial shard data to obtain supplementary information for the missing information; the iterative detection model is used to predict the non-missing information in the initial shard data and the supplementary information to obtain target shard data corresponding to the initial shard data.

9. The battery data detection method according to any one of claims 1 to 6, characterized in that: The method further comprises: If a battery source data set is stored in a database of the data warehouse through binary log detection, extracting the battery source data set into the operational data storage layer of the data warehouse; the battery source data set is a data set obtained by performing attribute detection on the battery of the electric device; In the operation data storage layer, logically processing the battery source data set to obtain the battery data set; The battery parameter field in the battery data set includes at least one of the following: battery voltage, battery current, battery resistance, battery temperature, battery charge state, battery operating state, battery flag, and the type of vehicle on which the battery is installed.

10. A battery data detection device, characterized in that: include: A determining unit, configured to determine initial fragmented data having missing information, wherein the initial fragmented data includes battery information corresponding to a plurality of battery parameter fields; a first detection unit, configured to perform at least supplementary prediction processing on battery information corresponding to each of the plurality of battery parameter fields in the initial sliced ​​data using at least one iterative detection model from among a plurality of iterative detection models that are matched one-to-one with the plurality of battery parameter fields, to obtain target sliced ​​data corresponding to the initial sliced ​​data; The target fragment data includes supplementary information for the missing information and predicted information for the non-missing information; The second detection unit is used to perform accuracy detection on the target slice data by using at least one detection model from multiple detection models that match the multiple battery parameter fields one by one, on the supplementary information in the target slice data and the predicted information, to obtain a detection result of the target slice data.

11. A battery data detection device, characterized in that: include: a memory for storing executable instructions; The processor is configured to implement the steps of the battery data detection method according to any one of claims 1 to 9 when executing the executable instructions stored in the memory.

12. A computer-readable storage medium, characterized in that The storage medium stores executable instructions, and when the executable instructions are executed by the processor, the steps of the battery data detection method according to any one of claims 1 to 9 are implemented.

13. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of the battery data detection method according to any one of claims 1 to 9 are implemented.

Images