CN116577685A

CN116577685A - Health detection method, data processing method, related device, equipment and medium

Info

Publication number: CN116577685A
Application number: CN202310857199.9A
Authority: CN
Inventors: 黄瑶; 孔思宇; 章东飞; 薛庆瑞; 田达; 翁文辉
Original assignee: Contemporary Amperex Technology Co Ltd
Current assignee: Contemporary Amperex Technology Co Ltd
Priority date: 2023-07-13
Filing date: 2023-07-13
Publication date: 2023-08-11
Anticipated expiration: 2043-07-13
Also published as: CN116577685B

Abstract

The application discloses a health detection method, a data processing method, a related device, equipment and a medium, wherein the health detection method comprises the following steps: acquiring an original test data set generated by testing a battery, wherein the original test data set comprises original test data at a plurality of moments; determining the target data type of the original test data set based on the target change rule of the original test data in time; screening a plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set; and carrying out health detection on the battery based on the target test data set to obtain a health detection result. By means of the scheme, the accuracy of battery health detection can be improved.

Description

Health detection method, data processing method, related device, equipment and medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a health detection method, a data processing method, a related device, equipment, and a medium.

Background

In order to achieve accuracy of test results of a battery or other target objects, the battery or other target objects are often required to be tested through a plurality of test conditions, for example, different test conditions are put into different test flows, and test data under different flows can be obtained after the plurality of tests. Because of the difference between the test conditions in each flow, the difference between the test data of the battery or other target object obtained under different test conditions is caused. If the obtained test data is directly used for performing performance analysis on the battery or other target objects under the condition that the data under different test flows are mixed or cannot be distinguished, the test data is likely to deviate greatly, so that the detection result is inaccurate or other data processing results are likely to be inaccurate.

Disclosure of Invention

The application provides at least one health detection method, a data processing method, a related device, equipment and a medium.

The application provides a health detection method, which comprises the following steps: acquiring an original test data set generated by testing a battery, wherein the original test data set comprises original test data at a plurality of moments; determining the target data type of the original test data set based on the target change rule of the original test data in time; screening a plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set; and carrying out health detection on the battery based on the target test data set to obtain a health detection result.

In the above scheme, the obtained target test data is obtained by determining the data type of the obtained original test data set and processing the obtained test data according to the data processing mode corresponding to the data type, compared with the original test data set, the quality of the target test data is higher, and the health detection result obtained by detecting the health of the battery by using the target test data later is more accurate than the health detection result obtained by using the original test data set.

In some embodiments, the original test data set includes original test data acquired under at least two test flows, and the method includes screening a plurality of original test data in the original test data set by using a target data processing mode corresponding to a target data type to obtain a target test data set, including: distinguishing original test data belonging to different test flows in an original test data group by utilizing a target data processing mode, and determining original test data corresponding to each test flow; and selecting the original test data corresponding to one of the test flows to be combined to obtain a target test data set.

In the above scheme, the test data belonging to different flows in the original test data set is screened by using the target data processing mode, and then the original test data under one of the test flows is used as the target test data set.

In some embodiments, the health detection method further comprises: for each candidate data type, determining the variation rule difference of original test data corresponding to different test flows in the candidate data type, wherein the candidate data type comprises a target data type; and determining a data processing mode of the candidate data type based on the variation rule difference, wherein the data processing mode is used for distinguishing original test data corresponding to different test flows in the candidate data type.

In the scheme, the corresponding data processing method is formulated according to the time change rule difference of the test data under different test flows, so that the original test data under two test flows can be conveniently screened by using the corresponding data processing method, and the target test data set is obtained.

In some embodiments, determining the target data type to which the original test data set belongs based on a target change rule of the original test data in time in the original test data set includes: matching the target change rule with a preset change rule of each candidate data type respectively to obtain the confidence coefficient of the original test data group belonging to each candidate data type; and responding to the confidence coefficient between the original test data set and each candidate data type to meet a preset condition, taking the data type corresponding to the maximum confidence coefficient as the target data type to which the original test data set belongs, wherein the maximum confidence coefficient is the maximum confidence coefficient between the original test data set and each candidate data type.

In the above scheme, the confidence that the original test data set belongs to each candidate data type can be obtained by matching the change rule of the original test data set with the preset change rule of each candidate data type, so that the data type to which the original test data set belongs is determined according to each confidence.

In some embodiments, matching the target change rule with a preset change rule of each candidate data type to obtain a confidence that the original test data set belongs to each candidate data type, including: sampling original test data in the original test data set according to preset sampling parameters to obtain a plurality of groups of data sets to be tested, wherein the preset sampling parameters comprise sampling length and/or sampling times; respectively matching the change rule of each data set to be detected with the preset change rule of each candidate data type to obtain the confidence coefficient of each data set to be detected belonging to each candidate data type; and counting the confidence coefficient of each data set to be tested and the candidate data type for each candidate data type to obtain the confidence coefficient of the original test data set and the candidate data type.

In the above scheme, the original test data set is sampled to obtain a plurality of data sets to be tested, and then the target confidence coefficient of the original test data set belonging to each data type is determined according to the confidence coefficient of the original test data set belonging to each data type.

In some embodiments, the health detection method further comprises: adjusting preset sampling parameters and/or preset conditions in response to the confidence coefficient of the original test data and each candidate data type not meeting the preset conditions; and determining the target data type of the original test data group by utilizing the adjusted preset sampling parameters and/or preset conditions.

In the above scheme, under the condition that each target confidence coefficient meets the adjustment condition, one or more of the preset sampling parameters and the preset conditions are adjusted, and then the data type of the original test data set is redetermined by utilizing the adjusted preset sampling parameters and/or the preset conditions, so that the data type of the original test data set is more accurate.

In some embodiments, the health detection method further comprises: responding to the preset sampling parameters and/or the preset conditions, wherein the total adjustment times reach an adjustment times threshold value, newly adding candidate data types and formulating a data processing mode corresponding to the newly added candidate data types; and taking the newly added candidate data type as the target data type to which the original test data set belongs.

In the above scheme, if the data type to which the original test data set belongs is adjusted for a plurality of times, it is likely that the data type corresponding to the original test data set does not exist in each data type, so that the data type can be supplemented by a mode of adding the data type and a data processing mode corresponding to the newly added data type, and the accuracy of data processing of the original test data set is improved.

In some embodiments, the preset condition includes that the confidence of the original test data with only one of the candidate data types is greater than or equal to a preset confidence.

In the above scheme, if the confidence between two or more data types of the original test data set is greater than or equal to the preset confidence, the erroneous data type is likely to be used as the data type to which the original test data set belongs, or the confidence between the original test data set and each data type is less than the preset confidence, which may be that noise may exist in the sampled data, and the like, so that the confidence is inaccurate, and the accuracy of obtaining the obtained confidence can be improved by adjusting one or more of the sampling times, the sampling length and the preset conditions.

In some embodiments, the original test data is discharge data, the health detection result includes that the electric quantity of the battery has a water jump problem or the electric quantity of the battery does not have a water jump problem, the health detection is performed on the battery based on the target test data set to obtain the health detection result, including: acquiring the slope between at least two original test data in a target test data set; determining that the capacity of the battery has a water jump problem in response to the slope being greater than or equal to a preset slope; or, in response to the slope being less than the preset slope, determining that the capacity of the battery is free of the water jump problem.

In the above scheme, by analyzing the discharge data of the battery, if the discharge data of the battery cell fluctuates greatly, the slope may be larger, if the slope is larger than the preset slope, the electric quantity of the battery is large, the problem of water jump may occur, otherwise, the electric quantity of the battery does not have the problem of water jump.

In some embodiments, the health detection method further comprises: predicting target discharge data of the battery at a plurality of time points in a preset time period in the future based on the change rule of original test data in the target test data set in time; and determining the probability of the battery having a water jump problem in a preset time period in the future based on the target discharge data.

In the scheme, the probability of the water jump problem of the battery in a future period of time can be estimated by predicting the future discharge data of the battery.

The application provides a data processing method, which comprises the following steps: acquiring an original test data set generated by testing a target object, wherein the original test data set comprises original test data at a plurality of moments; determining the target data type of the original test data set based on the target change rule of the original test data in time; and screening the plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set.

The application provides a health detection device, which comprises a first data acquisition module, a first type determination module, a first data processing module and a health detection module; the first data acquisition module is used for acquiring an original test data set generated by testing the battery, wherein the original test data set comprises original test data at a plurality of moments; the first type determining module is used for determining the target data type of the original test data set based on the target change rule of the original test data in time; the first data processing module is used for screening a plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set; and the health detection module is used for carrying out health detection on the battery based on the target test data set to obtain a health detection result.

The application provides a data processing device, which comprises a second data acquisition module, a second type determination module and a second data processing module; the second data acquisition module is used for acquiring an original test data set generated by testing the target object, wherein the original test data set comprises original test data at a plurality of moments; the second type determining module is used for determining the target data type of the original test data set based on the target change rule of the original test data in time; and the second data processing module is used for screening a plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set.

The present application provides an apparatus comprising a memory and a processor for executing program instructions stored in the memory to implement the above data processing method or any of the above health detection methods.

The present application provides a computer readable storage medium having stored thereon program instructions which, when executed by a processor, implement the above-described data processing method or any of the above-described health detection methods.

In the above scheme, the data type of the obtained original test data set is determined, and the obtained test data is processed according to the data processing mode corresponding to the data type, so that the quality of the target test data is higher than that of the original test data set, and the health detection result obtained by detecting the health of the battery by using the target test data is more accurate than that obtained by using the original test data set.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application as claimed.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application.

FIG. 1 is a flow chart of a health detection method according to some embodiments of the present application;

FIG. 2 is another flow chart of a health detection method according to some embodiments of the present application;

FIG. 3 is a schematic diagram of an original test data set and a target test data set under a 1EH type;

FIG. 4 is a schematic diagram of an original test data set and a target test data set under a 1TP type;

FIG. 5 is a schematic diagram of an original test data set and a target test data set under a 1TG type;

FIG. 6 is a schematic diagram of an original test data set and a target test data set for a 01X type;

FIG. 7 is a schematic diagram of an original test data set and a target test data set under 1SP type;

FIG. 8 is a schematic diagram of a primary test data set and a target test data set for a 1RR type;

FIG. 9 is a schematic diagram of an original test data set and a target test data set for a 1TR type;

FIG. 10 is a schematic diagram of a raw test data set and a target test data set under a 1MJ type;

FIG. 11 is a schematic of the raw test data set and the fitting data;

FIG. 12 is a schematic of a target test dataset and fitting data;

FIG. 13 is a flow chart of a data processing method according to some embodiments of the present application;

FIG. 14 is a schematic diagram of a health detection device according to some embodiments of the present application;

FIG. 15 is a schematic diagram of a data processing apparatus according to some embodiments of the present application;

FIG. 16 is a schematic diagram of an apparatus provided by some embodiments of the application;

fig. 17 is a schematic diagram of the structure of a computer-readable storage medium according to some embodiments of the application.

Detailed Description

The following describes embodiments of the present application in detail with reference to the drawings.

In the following description, for purposes of explanation and not limitation, specific details are set forth such as the particular sub-system architecture, interfaces, techniques, etc., in order to provide a thorough understanding of the present application.

The term "and/or" is herein merely an association relationship describing an associated object, meaning that there may be three relationships, e.g., a and/or B, may represent: a exists alone, A and B exist together, and B exists alone. In addition, the character "/" herein generally indicates that the front and rear associated objects are an "or" relationship. Further, "a plurality" herein means two or more than two. In addition, the term "at least one" herein means any one of a plurality or any combination of at least two of a plurality, for example, including at least one of A, B, C, and may mean including any one or more elements selected from the group consisting of A, B and C.

In consideration of the accuracy of the test result of the target object, the battery needs to be tested through a plurality of test conditions in many cases, for example, different test conditions are put into different test flows, and test data under different flows can be obtained after a plurality of tests. Because of the difference between the test conditions in each flow, the test data of the battery acquired under different test conditions is different. If the obtained test data is directly used for performing performance analysis on the target object under the condition that the data under different test flows are mixed or cannot be distinguished, inaccurate health detection results are likely to be caused.

In this embodiment, the battery may be a battery cell, a battery module, or a battery module. Among them, the battery cell may be regarded as the smallest unit constituting the battery. Each battery cell may be a secondary battery or a primary battery; but not limited to, lithium sulfur batteries, sodium ion batteries, or magnesium ion batteries. The battery module may be considered as an integral body obtained by connecting a plurality of battery cells in series, parallel or series-parallel, and accommodated in a case. The series-parallel connection refers to that a plurality of battery monomers are connected in series or in parallel. For example, the plurality of battery cells can be directly connected in series, in parallel or in series-parallel, and then the whole formed by the plurality of battery cells is accommodated in the box body. The battery module can also be in a form that a plurality of battery monomers are connected in series or in parallel or in series-parallel to form a battery module, and a plurality of battery modules are connected in series or in parallel or in series-parallel to form a whole and are accommodated in the box body.

The test conditions may be different between different test flows, for example, the test condition in test flow a is quick charge and quick discharge of the battery, the test condition in test flow b is slow charge and slow discharge of the battery, the test condition in test flow c is the first ambient temperature where the battery is located, the test condition in test flow d is the second ambient temperature where the battery is located, the first ambient temperature and the second ambient temperature are different, the test condition in test flow e is the full charge of the battery is the first full charge, the test condition in test flow f is the full charge of the battery is the second full charge, and the first full charge and the second full charge are different … …

Therefore, the application provides a method for performing the health detection by using the target test data set after performing the data processing on the obtained original test data set to obtain the target test data set, and compared with the method for directly using the original test data set, the method has the advantage that the obtained health detection result is more accurate.

Referring to fig. 1, the data processing method provided by the present application may include the following steps S11 to S13. Step S11: an original test data set generated by testing the battery is obtained. The original test data set comprises original test data at a plurality of moments. Step S12: and determining the target data type of the original test data set based on the target change rule of the original test data in time. Step S13: and screening the plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set. Step S14: and carrying out health detection on the battery based on the target test data set to obtain a health detection result.

The raw test data set generated by testing the battery may be charge data and/or discharge data of the battery cells. For example, if the battery has a water jump problem during charging and discharging, the original test data generated by the test may be discharging data, such as capacity data of the battery at a plurality of moments during discharging. For example, the raw test data generated by testing the charging rate of the battery may be, for example, charged data at different times during the charging process. Of course, the battery may be tested for a wide variety of different performances before shipment, and the original test data in the present application is not limited to the above-listed data, and any original test data generated by testing the battery process may be used. The original test data set may be time sequence data in a period of time, the time sequence may take the start of the test as a start point and the end of the test as an end point, that is, even if the data of two processes are not obtained in a natural period of time, if the two sets of original test data are mixed together, the test start point processes of the two processes are overlapped. Data within the same period of time may also be composed. The change rule includes, but is not limited to, a change rate of data in time or a range in which the data is located. The time-varying rule of the obtained test data generated by the battery may be different due to different test procedures, and different time-varying rules may be presented between different original test data sets respectively formed by the original test data under different test procedures, for example, the test data generated by the battery under a certain test procedure has a low general value, or the test data generated by the battery under another test procedure generally presents a specific fluctuation or a value interval, which is just an example, so that the time-varying rule of the original test data sets formed by the original test data under the two procedures is that the two rules are fused. The target data type may be a data type of the original test data set, or a data type corresponding to a certain flow in the original test data set. The data type may be a type to which the original test data in the two flows together form, and the other original test data set includes the original test data obtained in the flows b and c, or the data type may be a type to which the original test data set formed by the two flows together form. The data processing mode corresponding to the data type can be determined in advance, so that after the target data type of the original test data set is determined, the original test data in the original test data set is conveniently screened according to the data processing mode corresponding to the target data type, so that the original test data under different processes are distinguished, and the target test data set is obtained. The method for detecting the health of the battery by using the target test data set can be used for judging whether the original test data of the abnormality exists in the target test data set or whether the change rule of the original test data in time is abnormal, so that the health detection of the battery is realized.

In the above scheme, the quality of the target test data set obtained by determining the data type to which the obtained original test data set belongs and processing the obtained test data according to the data processing mode corresponding to the data type is higher than that of the original test data set, and the health detection result obtained by detecting the health of the battery by using the target test data set is more accurate than that obtained by using the original test data set.

In some embodiments, the raw test data set includes raw test data acquired under at least two test flows, respectively. The step S13 may include the steps of: and distinguishing the original test data belonging to different test flows in the original test data group by utilizing a target data processing mode, and determining the original test data corresponding to each test flow. And selecting the original test data corresponding to one of the test flows to be combined to obtain a target test data set.

In other words, the target data processing manner corresponding to the target data type can distinguish the original test data belonging to different processes under the target data type, so as to obtain the original test data corresponding to different processes. The original test data corresponding to one of the flows may then be selected as the target test data set. The method for combining the original test data corresponding to one of the flows may be to combine the original test data under the flow according to the time sequence generated by each of the original test data to obtain the target test data set. The selection mode can be random selection or selection of a specific process under the target data type, or can be to take the original test data of the two processes as a target test data set, sequentially detect the health of the battery according to the original test data of the two processes, and then combine the two health detection results to serve as a final health detection result of the battery.

In some embodiments, the health detection method further comprises: and for each candidate data type, determining the variation rule difference of the original test data corresponding to different test flows in the candidate data type. Wherein the candidate data type includes a target data type. And determining a data processing mode of the candidate data type based on the variation rule difference. The data processing mode is used for distinguishing original test data corresponding to different test flows in the candidate data types.

As described above, the data type may be a data type corresponding to an original test data group constituted by original test data under two or more flows. Because the change rules of the original test data in different flows in time are different, for example, the time change rate of the data is different, or the change range of the data is different, etc. The data processing mode for distinguishing the original test data corresponding to each flow can be determined by determining the change rule difference between the original test data of different flows under the candidate data type. By way of example, the data processing method may be regression, random forest, classification, feature screening, etc., and is not limited herein.

In some embodiments, the step S12 may include the following steps: and matching the target change rule with a preset change rule of each candidate data type respectively to obtain the confidence coefficient of the original test data group belonging to each candidate data type. And responding to the confidence coefficient between the original test data set and each candidate data type to meet a preset condition, and taking the data type corresponding to the maximum confidence coefficient as the target data type to which the original test data set belongs. Wherein the maximum confidence is the maximum confidence between the original test data set and each candidate data type.

Alternatively, the flows and the number of flows included under different candidate data types may be different. The preset change rules of the candidate data types are different. The matching mode includes, but is not limited to, one or more of clustering algorithm, neural network algorithm, classification algorithm, regression algorithm, fitting algorithm, feature screening and the like. The preset condition can be set according to the actual requirement, for example, the confidence coefficient is greater than or equal to a certain confidence coefficient threshold value, etc. The original test data set and each candidate data type have a confidence coefficient respectively, and the largest confidence coefficient in each confidence coefficient is the largest confidence coefficient in the embodiment.

In some embodiments, the foregoing manner of matching the target change rule with the preset change rule of each candidate data type to obtain the confidence that the original test data set belongs to each candidate data type may be: and sampling the original test data in the original test data set according to preset sampling parameters to obtain a plurality of groups of data sets to be tested. The preset sampling parameters comprise sampling length and/or sampling times. And then, respectively matching the change rule of each data set to be detected with the preset change rule of each candidate data type to obtain the confidence that each data set to be detected respectively belongs to each candidate data type. And counting the confidence coefficient of each data set to be tested and the candidate data type for each candidate data type to obtain the confidence coefficient of the original test data set and the candidate data type.

The sampling length refers to the number of original test data contained in the data set to be tested obtained by each sampling. The sampling times refer to the number of data sets to be measured obtained by sampling. The number of samples and the sampling length may be adjusted as desired, of course, the sampling length is less than or equal to the length of the original test data set, e.g., the sampling length may be between 70% and 90% of the length of the original test data set. And for each data set to be tested, acquiring the confidence that the data set to be tested belongs to each data type. The sampling length may be multiple, that is, the number of original test data included in the sampled data set to be tested may be different. The matching manner between the change rule of each data set to be tested and the preset change rule of each candidate data type can be referred to above, and will not be described herein. Alternatively, the manner in which the confidence level is calculated for different data types may be different. For example, the way in which the data types calculate the confidence between each data group type to be measured and the data group type to be measured can be a classification algorithm, and the way in which the confidence corresponding to other data types is calculated can also be one or more of a regression algorithm, a fitting algorithm, a feature screening method and the like. The preset statistics may be an average value, a mode value, a median value, and the like, and the statistical average value is taken as an example in this embodiment.

In some embodiments, the health detection method further comprises: and adjusting preset sampling parameters and/or preset conditions in response to the confidence coefficient of the original test data and each candidate data type not meeting the preset conditions. And then, determining the target data type of the original test data group by utilizing the adjusted preset sampling parameters and/or preset conditions.

Adjusting the preset sampling parameter may be one or both of the number of samples, the sampling length. In response to the confidence level of the original test data and each candidate data type not meeting the preset condition, adjusting the preset sampling parameter and/or the preset condition may be adjusting one of the sampling number, the sampling length, the preset condition, or adjusting two or more of the sampling number, the sampling length, the preset condition. The adjustment of the number of samples may be an increase or decrease of the number of samples within a range of allowable numbers of samples. The sampling frequency range may be preset. The sample length may be increased or decreased by the sample length within an allowable length range. In other embodiments, a plurality of sampling lengths may be set in the sampling times, for example, the sampling times are 3 times, the sampling length of the first sampling is a1, the sampling length of the first sampling is a2, the sampling length of the third sampling is a3, and a1, a2 and a3 may be equal or different. The preset conditions may be either relaxed or stricter. The determining the data type to which the original test data set belongs may specifically be to sample the original test data set according to the latest sampling times and sampling lengths and determine the data type to which the latest preset conditions belong by using the adjusted sampling times, sampling lengths and/or preset conditions.

In some embodiments, the health detection method further comprises: and responding to the preset sampling parameters and/or the preset conditions, wherein the total adjustment times reach an adjustment times threshold value, newly adding candidate data types and formulating a data processing mode corresponding to the newly added candidate data types. And taking the newly added candidate data type as the target data type to which the original test data set belongs.

As described above, in the case where the target confidence of the original test data set and each data type does not satisfy the preset condition, the sampling number, the sampling length, and/or the preset condition need to be adjusted, and if the sampling length is adjusted simultaneously when the sampling number is adjusted for the original test data set, the adjustment number is increased by 1 and not increased by 2. The threshold value of the adjustment number is assumed to be 3, if the target confidence coefficient of the original test data set and each data type does not meet the preset condition, the sampling number, the sampling length and/or the preset condition are adjusted for the first time, then the original test data set is sampled and whether the target confidence coefficient between the original test data set and each data type meets the preset condition is judged again by using the sampling number, the sampling length and/or the preset condition after the first time adjustment, if the judgment result is that the target confidence coefficient between the original test data set and each data type does not meet the preset condition, the second time adjustment is performed on the sampling number, the sampling length and/or the preset condition, the original test data set is sampled again by using the sampling number, the sampling length and/or the preset condition after the second time adjustment, if the judgment result is that the target confidence coefficient between the original test data set and each data type does not meet the preset condition, the sampling number, the sampling length and/or the preset condition is adjusted for the third time, and the new data type is further processed by using the sampling number, and if the new confidence coefficient between the original test data set and each data type is not met, the threshold value is further increased, and the threshold value is adjusted. That is, when the total adjustment times reach the adjustment times threshold, and the data type of the original test data set still cannot be determined according to the latest sampling times, sampling lengths and preset conditions, the newly added data type is used as the data type of the original test data set. The data processing mode corresponding to the newly added data type can be determined according to the variation rule difference between the original test data of different processes contained in the newly added data type.

If the target confidence coefficient between the original test data set and two or more data types is greater than or equal to the preset confidence coefficient, or the target confidence coefficient between the original test data set and each data type is smaller than the preset confidence coefficient, determining that the confidence coefficient between the original test data set and each data type does not meet the preset condition. The magnitude of the preset confidence level can be adjusted according to the requirement.

In some embodiments, the raw test data is discharge data, and the health detection result includes that the battery has a water jump problem or that the battery has no water jump problem. The step S14 may include the steps of: the slope between at least two original test data in the target test data set is obtained. And determining that the capacity of the battery has a water jump problem in response to the slope being greater than or equal to a preset slope. Or, in response to the slope being less than the preset slope, determining that the capacity of the battery is free of the water jump problem.

Active lithium loss is the most common and dominant source of battery cell cycling capacity decay. The lithium is separated by charging, and the lithium simple substance can not be completely discharged, so that the discharge capacity is lower than the charge capacity, and the discharge can more represent the real capacity of the battery monomer due to the loss of the battery monomer in the charge and discharge process. In particular the slope between two temporally adjacent raw test data. In some application scenarios, the slopes between multiple sets of adjacent original test data may be obtained, and if one set of slopes is greater than or equal to a preset slope, it is determined that the capacity of the battery has a water jump problem. In some application scenarios, the slope between at least two test data includes, but is not limited to, the slope between each adjacent data in the fit data, the slope between a maximum and a minimum, or the slope between a value intermediate the timing and an end of the timing. If the slope change is large, this means that the battery capacity is abnormal in terms of time variation, and there is a problem of water jump. If the slope is small, this means that the battery capacity is changed normally over time, and no problem of water jump occurs.

In some embodiments, the health detection method further comprises: and predicting target discharge data of the battery at a plurality of time points in a future preset time period based on the change rule of the original test data in the target test data set in time. And determining the probability of the battery having a water jump problem in a preset time period in the future based on the target discharge data.

The change rule of the battery in the time of the original test data is considered to be continuous, so that the change rule of the battery in the future preset time period can be predicted through the change rule of the battery in the current time period, and the discharge data of a plurality of time points in the future preset time period can be determined. Specifically, the target test data set can be fitted to obtain fitting data, and specifically, discharge data of a plurality of time points in a preset time period in the future can be fitted. A slope between at least two of the fit data is obtained. Then, in response to the slope being greater than or equal to the preset slope, determining that the probability of the occurrence of the diving problem within the future preset time period is greater than the first preset probability. Or in response to the slope being less than the preset slope, determining that the probability of the battery showing a water jump problem within a future preset time period is less than a second preset probability. The first preset probabilities may be the same or different.

In some embodiments, feature extraction is performed based on a plurality of known types of test data to obtain each known type of data feature. And performing feature matching on the original test data set to be identified and processed according to the data characteristics of the known types, for example, judging the type probability of the test data according to the data types corresponding to different features, determining the data type of the original feature data according to the threshold k of the data of different types, and performing data processing in a targeted manner, thereby greatly improving the data quality and improving the early warning accuracy of the single battery jump.

As described above, the original test data set may be sampled, the sampling times may be a total randomly selected times or a predetermined total times, the total times Σc of the random number and the predetermined total times D may have different processing methods corresponding to different test data, the random times may default to a value 3, at this time, the calculated amount is smaller and the probability of only acquiring the abnormal data area is not large, and of course, the value of the random times may also be adjusted accordingly according to the actual situation.

The scheme can further clear chaotic and mixed data, can be applied to data preprocessing, online early warning of the battery cells (judging the risk of the battery cells) or other data identification with characteristic rules, can identify data with periodic fluctuation caused by different test flows according to types, can realize different data processing according to different types of data characteristics, and can be used for other industries to provide a specific data identification and a demand scene of a data processing method based on data characteristic extraction.

Referring to fig. 2, the data processing method provided in the embodiment shown in fig. 2 may further include the following steps:

step S21: a raw test data set is obtained.

The raw test data set may be discharge data of an unknown type of battery cell.

Step S22: setting a sampling length, sampling times and preset confidence.

The specific setting mode may be that a sampling length is randomly selected in a preset sampling length interval as the sampling length of the original test data set, and the sampling times and the preset confidence are the same, which is not described herein.

Step S23: and sampling the original test data set to obtain a plurality of data sets to be tested.

The initial position of sampling may be in a range corresponding to the difference between the length of the original test data set and the sampling length, that is, the initial position of sampling needs to ensure that the sampling length of the data set to be tested obtained by sampling reaches the set sampling length.

Step S24: and respectively acquiring the confidence coefficient of each data group to be tested belonging to each candidate data type.

The manner of obtaining the confidence that each data set to be tested belongs to each candidate data type is as described above, and is not repeated here. The confidence may be a probability. And writing corresponding matching algorithms for N candidate data types according to the characteristics of each candidate data type, and calculating the probability of the data set to be tested corresponding to each candidate data type through each characteristic algorithm when the data set to be tested is input, and calculating the probability value of the matching once for each matching as confidence. The algorithm for calculating the probability of each candidate data type can use classification, regression, fitting, feature screening and other methods.

Step S25: and determining the target confidence that the original test data set belongs to each candidate data type based on the confidence that each data set to be tested belongs to each candidate data type.

And for each candidate data type, taking a preset statistical value of the confidence coefficient of each data set to be tested and the candidate data type as a target confidence coefficient of the original test data set and the candidate data type. The preset statistic may be an average value.

Step S26: and judging whether the confidence coefficient between the original test data set and each candidate data type meets a preset condition.

The manner of determining whether the confidence between the original test data set and each candidate data type meets the preset condition is as described above, and will not be described herein. And if the confidence degree between the original test data set and each candidate data type meets the preset condition, executing the step S27, otherwise executing the step S34. If the preset condition is not met, the matching length, the matching times and the preset confidence coefficient can be adjusted until the preset condition is met, and if the matching times are not met, new candidate data types need to be supplemented.

Step S27: and taking the candidate data type corresponding to the maximum target confidence as the target data type to which the original test data set belongs.

Step S28: and processing the original test data set by utilizing a data processing mode corresponding to the target data type to obtain a target test data set.

And carrying out corresponding data processing on the original test data set to obtain data with better quality. The algorithm of data processing corresponding to different data types can use regression, random forest, classification, feature screening and other methods. Referring to fig. 3 to 10, fig. 3 to 10 show data processing methods corresponding to 8 data types. As shown in fig. 3, the original test data set (original data 1 EH) belongs to a 1EH type, in this embodiment, the data type of the original test data with the characteristics of three large and one small data is defined as a 1EH type, and the data processing mode corresponding to the 1EH data type is to be the smallest. That is, the time series data acquired in the 1EH data type is three consecutive larger data in time series, and then the fourth data is smaller data. The difference between the larger data and the smaller data is smaller than the difference between the larger data and the smaller data, specifically, the 1EH type data can be considered as a cycle of every four data, the data in each cycle basically shows the rule, and the data processing mode is to remove the fourth data in each cycle to obtain the target test data set (expected to achieve 1 EH). As shown in fig. 4, the original test data set is of 1TP type, the data type of the original test data with the characteristic of one large data and four small data is defined as 1TP type, the data processing mode corresponding to 1TP type is to be the large data, and as in fig. 3, it can be considered that the data of 1TP type is one cycle of every five data, the data in each cycle is basically the first data and the larger data, and then the four data are smaller data, it is obvious that the first data in fig. 4 is between 101.5% and 102.00%, and then the four data are below 100.50%, the data processing mode of this type is to remove the larger data in each cycle to obtain the target test data set (expected to achieve 1 TP), wherein the relevant description of the data characteristics of one large data and nine small data, thirty four small data or thirty two small data, one small data and one large data and nine large data or thirty four small data in fig. 5 to 10 below can refer to one large data and four data and twelve small data in fig. 3. As shown in fig. 5, the original test data set is 1TG type, in this embodiment, the data type of the original test data with data characteristics of one large and nine small is defined as 1TG type, the data processing mode corresponding to the 1TG type is to be the enlargement, so as to obtain the target test data set (expected to achieve 1 TG), and the data of the 1TG type can be considered as ten data as one cycle, and the data in each cycle presents a characteristic of one large and nine small. As shown in fig. 6, the original test data set is 01X type, in this embodiment, the data type of the original test data with data characteristics of thirty-four or thirty-twelve is defined as 01X type, the data processing mode corresponding to 01X type is to be the smallest, so as to obtain the target test data set (01X is expected to be achieved), that is, each fifteen or seventeen data of 01X type can be considered as one period, the data characteristic presented by each period is thirty-four or thirty-twelve, that is, three larger data in each period will be followed by fourteen or twelve smaller data, the difference between the larger data and the smaller data is larger, and the difference between the larger data or the difference between the smaller data is smaller than the difference between the larger data and the smaller data. As shown in fig. 7, the type of the original test data set is 1SP type, in this embodiment, the data type of the original test data with data characteristics of three small and one large is defined as 1SP type, and the original test data under 1SP type can be considered as one cycle of every four original test data, wherein the larger data is larger than the smaller data, and the difference between the larger data and the difference between the smaller data are smaller than the difference between the larger data and the smaller data. The data processing mode of the 1SP type is to get the target test data set (expected to be achieved), namely to remove smaller data in each cycle. As shown in fig. 8, the original test data set is of 1RR type, in this embodiment, the data type of the original test data with data characteristics of one small and three large is defined as 1RR type, that is, three large data exist continuously after one small data, wherein the large data is larger than the small data, the difference between the large data and the difference between the small data are smaller than the difference between the large data and the small data, and the data processing mode corresponding to 1RR type is to remove the small data in each period. As shown in fig. 9, the type of the original test data set is 1TR, and the data type of the original test data with data characteristics of one large data and nine small data is defined as 1TR in this embodiment, that is, every ten original test data is one period, one large data in each period is followed by nine continuous smaller data, where the large data is larger than the small data, and the difference between the large data and the difference between the small data are smaller than the difference between the large data and the small data. The data processing mode corresponding to the 1TR type is to remove small data, namely smaller data in each period. As shown in fig. 10, the original test data set is 1MJ, in this embodiment, the data type of the original test data with data characteristics of thirty-four or thirty-twelve is defined as 1MJ, that is, the 1MJ uses every seventeen or fifteen original test data as one period, and after three consecutive large data, there are fourteen consecutive or twelve consecutive small data in the data characteristic of one period, and the data processing manner corresponding to the 1MJ is to remove the small data in each period.

Step S29: fitting the target test data set to obtain fitting data.

The data fitting method may be solved by using different fitting functions, and is not particularly limited herein. Fitting data obtained by fitting directly from the original test data set is shown in fig. 11, and fitting data obtained from the target test data set is shown in fig. 12. It is apparent that the noise exists in the original test data set in fig. 11, and the noise of the target test data set obtained through the data processing is significantly reduced.

Step S30: a slope between at least two of the fit data is obtained.

Step S31: judging whether the slope is larger than or equal to a preset slope.

If the slope is greater than or equal to the preset slope, step S32 is executed, otherwise step S33 is executed.

Step S32: and determining that the electric quantity of the battery has the problem of water jump.

Step S33: and determining that the electric quantity of the battery has no water jump problem.

As shown in fig. 11 and fig. 12, the fitting data is analyzed or early-warning is judged based on the fitting data, the better the fitting effect is, the more accurate the pre-warning is, the more reliable the early-warning is, the data which is not subjected to data preprocessing and the fitting are shown in fig. 11 (the accuracy is only 27.5%), the data after data processing and the fitting are shown in fig. 12 (the accuracy for the processed data is 99.1%).

Step S34: and judging whether the sampling times, the sampling length and/or the total adjustment times of the preset conditions reach the adjustment times threshold.

And when the judgment result is that the sampling times, the sampling length and/or the total adjustment times of the preset condition reach the threshold value of the adjustment times, executing the step S35, otherwise executing the step S22.

Step S35: and adding the data types and making a data processing mode corresponding to the added data types.

In some embodiments, before executing step S35, it may be further determined by manual methods, etc. whether the original test data set belongs to a known data type, if the manual determination result is not a known data type, step S35 may be executed, and if the manual determination result is that the test data belongs to a certain data type, step S22 may be executed. After the data types are newly added and the data processing modes corresponding to the newly added data types are formulated, the confidence between each data set to be tested and each data type can be redetermined, and the newly added data types can also be directly used as the data types of the original test data sets.

The data processing method provided by the embodiment can be applied to pre-processing of the pre-warning battery monomer diving data, and improves the data quality so as to evaluate the risk of the battery monomer more accurately, stop testing in time and conduct failure analysis. The method can also be used for identifying the type of the data with fixed characteristics, in the actual business, certain cyclic alternation of the data is caused by the alternate operation of the cyclic test flow, but when the data is analyzed, the data is often required to be analyzed separately, the actual data is often mixed together due to the alternate flow, the data which is not separated is inconvenient to analyze in a single flow, the mixed data processing caused by different flow is particularly important, and other data with obvious periodic characteristics are also suitable for other situations.

Referring to fig. 13, the data processing method provided in this embodiment may include the following steps: step S101: and acquiring an original test data set generated by testing the target object. The original test data set comprises original test data at a plurality of moments; step S102: and determining the target data type of the original test data set based on the target change rule of the original test data in time. Step S103: and screening the plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set.

The target object can be any object requiring performance testing, such as a battery, an air conditioner, a workpiece, and the like. In some application scenarios, the target object is a battery, and the original test data set generated by performing different tests on the battery is discharge or charge data. In some application scenarios, the target object is an air conditioner, and the original test data generated by testing the air conditioner may be the time variation of the actual refrigeration temperature or heating temperature of the air conditioner under different test conditions. In some application scenarios, the target object is a workpiece, and the original test data is the deformation recovery condition of the workpiece in time under the condition that the workpiece is deformed under the action of different external forces. The set of raw test data generated by testing the target object may include raw test data obtained under different test flows. The change rule includes, but is not limited to, a change rate of data in time or a range in which the data is located. Different test flows may cause that the time-varying rule of the test data generated by the acquired target object may be different, and different time-varying rules may also be presented between different original test data sets respectively composed of the original test data under the different test flows. The time-varying rule of the obtained test data generated by the battery may be different due to different test procedures, and different time-varying rules may be presented between different original test data sets respectively formed by the original test data under different test procedures, for example, the test data generated by the battery under a certain test procedure has a low general value, or the test data generated by the battery under another test procedure generally presents a specific fluctuation or a value interval, which is just an example, so that the time-varying rule of the original test data sets formed by the original test data under the two procedures is that the two rules are fused. The target data type may be a data type of the original test data set, or a data type corresponding to a certain flow in the original test data set. The data type may be a type to which the original test data in the two flows together form, and the other original test data set includes the original test data obtained in the flows b and c, or the data type may be a type to which the original test data set formed by the two flows together form. The data processing mode corresponding to the data type can be determined in advance, so that after the target data type of the original test data set is determined, the original test data in the original test data set is conveniently screened according to the data processing mode corresponding to the target data type, so that the original test data under different processes are distinguished, and the target test data set is obtained. And the target object is conveniently and subsequently subjected to health detection by utilizing the target test data set, so that health detection results or other data processing results are obtained, for example, whether abnormal original test data exist in the target test data set or whether the change rule of the original test data in time is abnormal is judged, and the health detection of the target object is realized. The result of the health detection of the battery may be whether the battery has a water jump problem. The result of the health detection of the air conditioner may be whether the air conditioner has a problem of poor heating or cooling. The result of the health detection of the workpiece may be whether the elastic capability of the workpiece meets the standard or not, etc.

In the above scheme, the quality of the target test data obtained by determining the data type of the obtained original test data set and processing the obtained test data according to the data processing mode corresponding to the data type is higher than that of the original test data set, and the health detection result obtained by detecting the health of the target object by using the target test data is more accurate than that obtained by using the original test data set.

In some embodiments, the data processing method further comprises: for each candidate data type, determining the variation rule difference of original test data corresponding to different test flows in the candidate data type, wherein the candidate data type comprises a target data type; and determining a data processing mode of the candidate data type based on the variation rule difference, wherein the data processing mode is used for distinguishing original test data corresponding to different test flows in the candidate data type.

In some embodiments, the data processing method further comprises: adjusting preset sampling parameters and/or preset conditions in response to the confidence coefficient of the original test data and each candidate data type not meeting the preset conditions; and determining the target data type of the original test data group by utilizing the adjusted preset sampling parameters and/or preset conditions.

In some embodiments, the data processing method further comprises: responding to the preset sampling parameters and/or the preset conditions, wherein the total adjustment times reach an adjustment times threshold value, newly adding candidate data types and formulating a data processing mode corresponding to the newly added candidate data types; and taking the newly added candidate data type as the target data type to which the original test data set belongs.

In order to better understand the data processing method provided in fig. 13, the target object is used as a battery in this embodiment, the data types include 8 data types as shown in fig. 3 to 10, and the data processing method corresponding to the data may be the data processing method corresponding to the 8 data types shown in fig. 3 to 10. Referring to fig. 3 to 10, as shown in fig. 3, the original test data set (original data 1 EH) belongs to a 1EH type, in this embodiment, the data type of the original test data with data characteristics of three large and one small is defined as a 1EH type, and the data processing mode corresponding to the 1EH data type is to be small. That is, the time series data acquired in the 1EH data type is three consecutive larger data in time series, and then the fourth data is smaller data. The difference between the larger data and the smaller data is smaller than the difference between the larger data and the smaller data, specifically, the 1EH type data can be considered as a cycle of every four data, the data in each cycle basically shows the rule, and the data processing mode is to remove the fourth data in each cycle to obtain the target test data set (expected to achieve 1 EH). As shown in fig. 4, the original test data set is of 1TP type, the data type of the original test data with the characteristic of one large data and four small data is defined as 1TP type, the data processing mode corresponding to 1TP type is to be the large data, and as in fig. 3, it can be considered that the data of 1TP type is one cycle of every five data, the data in each cycle is basically the first data and the larger data, and then the four data are smaller data, it is obvious that the first data in fig. 4 is between 101.5% and 102.00%, and then the four data are below 100.50%, the data processing mode of this type is to remove the larger data in each cycle to obtain the target test data set (expected to achieve 1 TP), wherein the relevant description of the data characteristics of one large data and nine small data, thirty four small data or thirty two small data, one small data and one large data and nine large data or thirty four small data in fig. 5 to 10 below can refer to one large data and four data and twelve small data in fig. 3. As shown in fig. 5, the original test data set is 1TG type, in this embodiment, the data type of the original test data with data characteristics of one large and nine small is defined as 1TG type, the data processing mode corresponding to the 1TG type is to be the enlargement, so as to obtain the target test data set (expected to achieve 1 TG), and the data of the 1TG type can be considered as ten data as one cycle, and the data in each cycle presents a characteristic of one large and nine small. As shown in fig. 6, the original test data set is 01X type, in this embodiment, the data type of the original test data with data characteristics of thirty-four or thirty-twelve is defined as 01X type, the data processing mode corresponding to 01X type is to be the smallest, so as to obtain the target test data set (01X is expected to be achieved), that is, each fifteen or seventeen data of 01X type can be considered as one period, the data characteristic presented by each period is thirty-four or thirty-twelve, that is, three larger data in each period will be followed by fourteen or twelve smaller data, the difference between the larger data and the smaller data is larger, and the difference between the larger data or the difference between the smaller data is smaller than the difference between the larger data and the smaller data. As shown in fig. 7, the type of the original test data set is 1SP type, in this embodiment, the data type of the original test data with data characteristics of three small and one large is defined as 1SP type, and the original test data under 1SP type can be considered as one cycle of every four original test data, wherein the larger data is larger than the smaller data, and the difference between the larger data and the difference between the smaller data are smaller than the difference between the larger data and the smaller data. The data processing mode of the 1SP type is to get the target test data set (expected to be achieved), namely to remove smaller data in each cycle. As shown in fig. 8, the original test data set is of 1RR type, in this embodiment, the data type of the original test data with data characteristics of one small and three large is defined as 1RR type, that is, three large data exist continuously after one small data, wherein the large data is larger than the small data, the difference between the large data and the difference between the small data are smaller than the difference between the large data and the small data, and the data processing mode corresponding to 1RR type is to remove the small data in each period. As shown in fig. 9, the type of the original test data set is 1TR, and the data type of the original test data with data characteristics of one large data and nine small data is defined as 1TR in this embodiment, that is, every ten original test data is one period, one large data in each period is followed by nine continuous smaller data, where the large data is larger than the small data, and the difference between the large data and the difference between the small data are smaller than the difference between the large data and the small data. The data processing mode corresponding to the 1TR type is to remove small data, namely smaller data in each period. As shown in fig. 10, the original test data set is 1MJ, in this embodiment, the data type of the original test data with data characteristics of thirty-four or thirty-twelve is defined as 1MJ, that is, the 1MJ uses every seventeen or fifteen original test data as one period, and after three consecutive large data, there are fourteen consecutive or twelve consecutive small data in the data characteristic of one period, and the data processing manner corresponding to the 1MJ is to remove the small data in each period.

Referring to fig. 14, the health detection device 40 provided in the present embodiment may include: a first data acquisition module 41, a first type determination module 42, a first data processing module 43, and a health detection module 44. A first data obtaining module 41, configured to obtain an original test data set generated by testing the battery, where the original test data set includes original test data at a plurality of moments; a first type determining module 42, configured to determine, based on a target change rule of the original test data in time in the original test data set, a target data type to which the original test data set belongs; the first data processing module 43 is configured to screen a plurality of original test data in the original test data set by using a target data processing manner corresponding to the target data type, so as to obtain a target test data set; the health detection module 44 is configured to perform health detection on the battery based on the target test data set, so as to obtain a health detection result.

In some embodiments, the original test data set includes original test data acquired under at least two test flows, and the first data processing module 43 screens the plurality of original test data in the original test data set by using a target data processing manner corresponding to a target data type to obtain a target test data set, including: distinguishing original test data belonging to different test flows in an original test data group by utilizing a target data processing mode, and determining original test data corresponding to each test flow; and selecting the original test data corresponding to one of the test flows to be combined to obtain a target test data set.

In some embodiments, the first data processing module 43 is further configured to: for each candidate data type, determining the variation rule difference of original test data corresponding to different test flows in the candidate data type, wherein the candidate data type comprises a target data type; and determining a data processing mode of the candidate data type based on the variation rule difference, wherein the data processing mode is used for distinguishing original test data corresponding to different test flows in the candidate data type.

In some embodiments, the first type determination module 42 determines, based on a target law of change in the original test data over time, a target data type to which the original test data set belongs, including: matching the target change rule with a preset change rule of each candidate data type respectively to obtain the confidence coefficient of the original test data group belonging to each candidate data type; and responding to the confidence coefficient between the original test data set and each candidate data type to meet a preset condition, taking the data type corresponding to the maximum confidence coefficient as the target data type to which the original test data set belongs, wherein the maximum confidence coefficient is the maximum confidence coefficient between the original test data set and each candidate data type.

In some embodiments, the first type determining module 42 matches the target change rule with a preset change rule of each candidate data type, to obtain a confidence that the original test data set belongs to each candidate data type, including: sampling original test data in the original test data set according to preset sampling parameters to obtain a plurality of groups of data sets to be tested, wherein the preset sampling parameters comprise sampling length and/or sampling times; respectively matching the change rule of each data set to be detected with the preset change rule of each candidate data type to obtain the confidence coefficient of each data set to be detected belonging to each candidate data type; and counting the confidence coefficient of each data set to be tested and the candidate data type for each candidate data type to obtain the confidence coefficient of the original test data set and the candidate data type.

In some embodiments, the first type determination module 42 is further to: adjusting preset sampling parameters and/or preset conditions in response to the confidence coefficient of the original test data and each candidate data type not meeting the preset conditions; and determining the target data type of the original test data group by utilizing the adjusted preset sampling parameters and/or preset conditions.

In some embodiments, the first type determination module 42 is further to: responding to the preset sampling parameters and/or the preset conditions, wherein the total adjustment times reach an adjustment times threshold value, newly adding candidate data types and formulating a data processing mode corresponding to the newly added candidate data types; and taking the newly added candidate data type as the target data type to which the original test data set belongs.

In some embodiments, the raw test data is discharge data, the health detection result includes that the battery has a water jump problem or the battery has no water jump problem, and the health detection module 44 performs health detection on the battery based on the target test data set to obtain the health detection result, including: acquiring the slope between at least two original test data in a target test data set; determining that the capacity of the battery has a water jump problem in response to the slope being greater than or equal to a preset slope; or, in response to the slope being less than the preset slope, determining that the capacity of the battery is free of the water jump problem.

In some embodiments, health detection module 44 is further configured to: predicting target discharge data of the battery at a plurality of time points in a preset time period in the future based on the change rule of original test data in the target test data set in time; and determining the probability of the battery having a water jump problem in a preset time period in the future based on the target discharge data.

Referring to fig. 15, the data processing apparatus 50 provided in this embodiment includes: a second data acquisition module 51, a second type determination module 52, and a second data processing module 53; the second data obtaining module 51 is configured to obtain an original test data set generated by testing the target object, where the original test data set includes original test data under multiple moments; a second type determining module 52, configured to determine, based on a target change rule of the original test data in time in the original test data set, a target data type to which the original test data set belongs; the second data processing module 53 is configured to screen the plurality of original test data in the original test data set by using a target data processing manner corresponding to the target data type, so as to obtain a target test data set.

The data processing device can implement the data processing method provided by the data processing method embodiment, and the functions of each module refer to the data processing method embodiment and are not described herein again.

Referring to fig. 16, the apparatus 60 provided by the present application includes a memory 61 and a processor 62, where the processor 62 is configured to execute program instructions stored in the memory 61 to implement steps in any of the above-described embodiments of the data processing method or implement steps in any of the above-described embodiments of the health detection method. In one particular implementation scenario, the device 60 may include, but is not limited to: the device 60 may also include, but is not limited to, a mobile device such as a laptop, a tablet, a computer device, a consumer, a microcomputer, a desktop, a server, etc.

In particular, the processor 62 is adapted to control itself and the memory 61 to implement the steps of any of the data processing method embodiments described above. The processor 62 may also be referred to as a CPU (Central Processing Unit ). The processor 62 may be an integrated circuit chip having signal processing capabilities. The processor 62 may also be a general purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a Field programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. In addition, the processor 62 may be commonly implemented by an integrated circuit chip.

Referring to fig. 17, a computer readable storage medium 70 provided in this embodiment stores a program instruction 71 that can be executed by a processor, where the program instruction 71 is used to implement the steps in any of the above-described data processing method embodiments or implement the steps in any of the above-described health detection method embodiments when executed by the processor.

The foregoing description of various embodiments is intended to highlight differences between the various embodiments, which may be the same or similar to each other by reference, and is not repeated herein for the sake of brevity.

In the several embodiments provided in the present application, it should be understood that the disclosed method and apparatus may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of modules or units is merely a logical functional division, and there may be additional divisions when actually implemented, e.g., the units or components may be combined or integrated into another subsystem, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical, or other forms.

In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units. The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor (processor) to execute all or part of the steps of the methods of the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

Claims

1. A health detection method is characterized in that,

acquiring an original test data set generated by testing a battery, wherein the original test data set comprises original test data at a plurality of moments;

determining a target data type to which the original test data set belongs based on a target change rule of the original test data in the original test data set in time;

screening a plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set;

and carrying out health detection on the battery based on the target test data set to obtain a health detection result.

2. The method for detecting health according to claim 1, wherein the raw test data set includes raw test data obtained under at least two test flows, the screening the plurality of raw test data in the raw test data set by using a target data processing manner corresponding to the target data type to obtain a target test data set includes:

distinguishing original test data belonging to different test flows in the original test data set by utilizing the target data processing mode, and determining original test data corresponding to each test flow;

And selecting the original test data corresponding to one of the test flows to be combined to obtain the target test data set.

3. The method of claim 2, further comprising:

for each candidate data type, determining the variation rule difference of original test data corresponding to different test flows in the candidate data type, wherein the candidate data type comprises the target data type;

and determining a data processing mode of the candidate data type based on the variation rule difference, wherein the data processing mode is used for distinguishing original test data corresponding to different test flows in the candidate data type.

4. A method of health detection according to any one of claims 1-3, wherein determining, based on a target law of change over time of the raw test data in the raw test data set, a target data type to which the raw test data set belongs comprises:

matching the target change rule with a preset change rule of each candidate data type respectively to obtain the confidence coefficient of the original test data group belonging to each candidate data type;

And responding to the confidence coefficient between the original test data set and each candidate data type to meet a preset condition, and taking the data type corresponding to the maximum confidence coefficient as the target data type to which the original test data set belongs, wherein the maximum confidence coefficient is the maximum confidence coefficient between the original test data set and each candidate data type.

5. The method for health detection according to claim 4, wherein the matching the target change rule with the preset change rule of each candidate data type to obtain the confidence that the original test data set belongs to each candidate data type includes:

sampling the original test data in the original test data set according to preset sampling parameters to obtain a plurality of groups of data sets to be tested, wherein the preset sampling parameters comprise sampling length and/or sampling times;

respectively matching the change rule of each data set to be detected with the preset change rule of each candidate data type to obtain the confidence coefficient of each data set to be detected belonging to each candidate data type;

and counting the confidence coefficient of each data set to be tested and the candidate data type for each candidate data type to obtain the confidence coefficient of the original test data set and the candidate data type.

6. The method of claim 5, further comprising:

adjusting the preset sampling parameters and/or the preset conditions in response to the confidence coefficient of the original test data and each candidate data type not meeting the preset conditions;

and determining the target data type of the original test data set by utilizing the adjusted preset sampling parameters and/or the preset conditions.

7. The method of claim 6, further comprising:

responding to the preset sampling parameters and/or the adjustment total times of the preset conditions to reach an adjustment times threshold value, newly adding candidate data types and formulating a data processing mode corresponding to the newly added candidate data types;

and taking the newly added candidate data type as the target data type to which the original test data set belongs.

8. The method of claim 4, wherein the predetermined condition comprises a confidence level of the raw test data with only one of the candidate data types being greater than or equal to a predetermined confidence level.

9. The method for detecting health of any one of claims 1 to 3, wherein the raw test data is discharge data, the health detection result includes that a jump problem exists in the electric quantity of the battery or that the jump problem does not exist in the electric quantity of the battery, the performing health detection on the battery based on the target test data set to obtain the health detection result includes:

Acquiring the slope between at least two original test data in the target test data set;

determining that the capacity of the battery has a water jump problem in response to the slope being greater than or equal to a preset slope;

or, in response to the slope being less than the preset slope, determining that the capacity of the battery is free of a water jump problem.

10. The method of claim 9, further comprising:

predicting target discharge data of the battery at a plurality of time points in a future preset time period based on a change rule of original test data in the target test data set in time;

and determining the probability of the battery having a water jump problem in a preset time period in the future based on the target discharge data.

11. A method of data processing, comprising:

acquiring an original test data set generated by testing a target object, wherein the original test data set comprises original test data at a plurality of moments;

And screening the plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set.

12. A health detection device, comprising:

the first data acquisition module is used for acquiring an original test data set generated by testing the battery, wherein the original test data set comprises original test data at a plurality of moments;

the first type determining module is used for determining a target data type to which the original test data set belongs based on a target change rule of the original test data in the original test data set in time;

the first data processing module is used for screening the plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set;

and the health detection module is used for carrying out health detection on the battery based on the target test data set to obtain a health detection result.

13. A data processing apparatus, comprising:

the second data acquisition module is used for acquiring an original test data set generated by testing the target object, wherein the original test data set comprises original test data at a plurality of moments;

The second type determining module is used for determining the target data type of the original test data set based on the target change rule of the original test data in the original test data set in time;

and the second data processing module is used for screening the plurality of original test data in the original test data set by utilizing a target data processing mode corresponding to the target data type to obtain a target test data set.

14. An apparatus comprising a memory and a processor for executing program instructions stored in the memory to implement the health detection method of any one of the preceding claims 1-10 or to implement the data processing method of claim 11.

15. A computer readable storage medium having stored thereon program instructions, which when executed by a processor, implement the health detection method of any of claims 1 to 10, or implement the data processing method of claim 11.