Data analysis method applied to lithium battery performance detection

Abstract

The invention discloses a data analysis method applied to lithium battery performance detection, relates to the technical field of lithium battery performance detection, and effectively improves the accuracy of lithium battery performance detection. According to the method, a standard state data change interval and a standard test state change tree are generated according to test state data of each lithium battery under the same cycle charging and discharging times, the test state change tree of each lithium battery is established, and a lithium battery performance detection network is established. Further obtaining the state fragment distance between the state data sub-nodes, obtaining the relative state deviation probability between the state data sub-nodes according to the state fragment distance, and obtaining the real-time test state data of the tested lithium battery in each cycle charging and discharging. And inputting the real-time test state data into the lithium battery performance detection network to match related state data sub-nodes so as to obtain the performance score of the tested lithium battery.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery performance detection, and in particular to a data analysis method applied to lithium battery performance detection. Background Art

[0002] In the field of lithium batteries, data analysis methods play a vital role in performance testing. By analyzing battery test data, we can gain a deep understanding of the battery's performance characteristics, health status, and predict its lifespan. Currently, commonly used data analysis methods include traditional analysis based on statistical principles, machine learning, and artificial intelligence technologies.

[0003] The existing lithium battery performance testing technology has the following defects:

[0004] Insufficient consideration of individual differences: Traditional testing methods are often based on average performance parameters and fail to fully consider the differences between individual batteries. This may lead to inaccurate performance evaluation of some batteries, making it difficult to achieve efficient classification and management.

[0005] Unable to handle complex nonlinear relationships: Existing statistical analysis methods have limited capabilities in handling complex nonlinear relationships and find it difficult to detect potential influencing factors and their interactions.

[0006] Therefore, how to improve the accuracy of lithium battery performance detection while fully demonstrating the potential correlation between various abnormal factors is a difficulty in the existing technology. For this purpose, a data analysis method applied to lithium battery performance detection is provided. Summary of the invention

[0007] In order to solve the above technical problems, the purpose of the present invention is to provide a data analysis method for lithium battery performance testing.

[0008] In order to achieve the above object, the present invention provides the following technical solutions:

[0009] A data analysis method for lithium battery performance detection includes the following steps:

[0010] Step S1, setting a plurality of target model lithium batteries to perform cyclic charge and discharge operations simultaneously, and collecting test status data of each lithium battery during each cyclic charge and discharge process until the number of cyclic charge and discharge times reaches the rated number of times of use of the target model lithium battery;

[0011] Step S2, based on the test state data of each lithium battery under the same number of charge and discharge cycles, generate a standard state data change interval and a standard test state change tree of the target model of lithium battery under each charge and discharge cycle, and establish a test state change tree of each lithium battery under each charge and discharge cycle operation, wherein the standard test state change tree and the test state change tree are provided with a plurality of state data nodes;

[0012] Step S3, establishing a lithium battery performance detection network, inputting all test state change trees and standard test state change trees into the lithium battery performance detection network, and then obtaining the state fragment distance between each state data sub-node, establishing a number of local state relationship clusters according to the state fragment distance between each state data sub-node, and obtaining the relative correlation value between each local state relationship cluster, and then obtaining the relative state offset probability between each state data sub-node according to the relative correlation value;

[0013] Step S4, perform multiple cycles of charge and discharge operations on the test lithium battery of the target model, and obtain real-time test status data of each cycle of charge and discharge, input the real-time test status data into the lithium battery performance detection network to match the relevant status data sub-node, and then obtain the performance score of the test lithium battery according to the relative status offset probability between the status data sub-node and other status data sub-nodes.

[0014] Furthermore, the process of acquiring the test status data includes:

[0015] Set n target models of lithium batteries, set a number for each lithium battery, obtain the rated number of uses k of the target model of lithium batteries through the Internet, and then perform k cycles of charge and discharge operations on each lithium battery;

[0016] During the process of cyclic charge and discharge operation of each lithium battery, the historical temperature change curve, historical voltage change curve and historical internal resistance change curve of the lithium battery are collected. At the end of each cyclic charge and discharge operation, the historical voltage change curve, historical internal resistance change curve and historical temperature change curve generated by the corresponding cyclic charge and discharge operation of each lithium battery are integrated to obtain the test status data.

[0017] Furthermore, the process of generating the standard state data change interval includes:

[0018] A two-dimensional coordinate system is established, and then the historical change curves corresponding to the same data type in the test state data with the same second digit of the serial number subscript are mapped to the same two-dimensional coordinate system, and a number of time points are set in the two-dimensional coordinate system, and then each historical change curve is divided into h historical change segments according to the time points, where h is a natural number greater than 0;

[0019] Obtain the spatial distances between the corresponding time points on each two-dimensional coordinate system but between the historical change segments corresponding to different historical change segments, and set the spatial distance threshold;

[0020] Compare the spatial distance between each historical change segment with the spatial distance threshold. If the spatial distance is greater than the spatial distance threshold, it is determined that there is no correlation between the corresponding historical change segments.

[0021] If the spatial distance is less than or equal to the spatial distance threshold, the corresponding historical change segments are associated, and the number of associations is recorded once for each of the two historical change segments;

[0022] Select the historical change segments associated with the historical change segments with the largest number of associations to form the standard state data change interval segments of the target model of lithium battery between the corresponding time points;

[0023] Then, the standard state data change interval segments of various types of historical change curves of the lithium battery between various time points are obtained respectively, and the standard state data change interval segments between various time points are connected in chronological order, so as to obtain the standard state data change interval of the lithium battery in each cycle charge and discharge operation.

[0024] Furthermore, the standard test state change tree and the process of establishing the test state change tree include:

[0025] Establish k*(h-1) state event main nodes for each lithium battery respectively, and establish three state data sub-nodes according to the type of historical change curve, and connect them with the state event main node, and then connect each state event main node according to the cycle charge and discharge sequence to obtain the test state change tree of each lithium battery under each cycle charge and discharge operation, and store the historical change fragments between each time point in the corresponding state data sub-node;

[0026] At the same time, according to the process of establishing the test state change tree, a standard test state change tree is established, and the standard state data change interval fragments between each time point are stored in each standard state data sub-node.

[0027] Furthermore, the process of obtaining the state fragment distance between the state data sub-nodes includes:

[0028] Establishing a lithium battery performance detection network, and then inputting all test state change trees and standard test state change trees into the lithium battery performance detection network;

[0029] Obtain the state data sub-node of each state event main node, and the calibration segment distance between the historical change segment and the standard state data change interval contained in the standard state data sub-node of the same type as the standard state event main node in the corresponding time sequence;

[0030] At the same time, the state fragment distances between the state data subnodes of each state event main node in the same time sequence are obtained, and the state fragment distance threshold is set.

[0031] Furthermore, the process of establishing the local state relationship cluster includes:

[0032] Each state data sub-node is recorded as a central state data sub-node in sequence, and then the relationship between the state fragment distance between the central state data sub-node and the state data sub-nodes of the same type and the state fragment distance threshold value in the same time sequence is determined;

[0033] According to the judgment result, a local state relationship cluster is established with the central state data sub-node as the center and the corresponding state data sub-node as the associated state data sub-node, and then it is judged whether the state fragment distance between the central state data sub-nodes of adjacent local state relationship clusters is less than the state fragment distance threshold value, and the corresponding two local state relationship clusters are locally merged according to the judgment result;

[0034] With the central state data sub-node as the center and the state fragment distance threshold as the radius, the association range area is set for each local state relationship cluster, and the relative association value between the corresponding two central state data sub-nodes is obtained according to the value of the ratio of the association range area occupied by the relative local merged part between each local state relationship cluster;

[0035] Obtain the relative correlation values ​​between the state data sub-nodes associated with each state event master node at the same time point, and then obtain the relative state offset probability between each state event master node according to the relative correlation values;

[0036] Furthermore, a state abnormality distance is set to determine whether the calibration fragment distance between each state data sub-node and the corresponding standard state data sub-node is less than or equal to the state abnormality distance. If so, the corresponding state data sub-node is marked as a normal state data sub-node, otherwise the corresponding state data sub-node is marked as an abnormal state data sub-node.

[0037] Furthermore, the process of obtaining the performance score of the tested lithium battery includes:

[0038] The real-time change curves in the real-time test status data of each cycle of charge and discharge are divided into a number of real-time change segments in the order of time points, and then each real-time change segment is input into the lithium battery performance detection network, and the real-time detection distance between each real-time change segment and each status data sub-node at the same time point is obtained;

[0039] If it is determined that the real-time detection distance is less than or equal to the state segment distance threshold, then it is determined that the real-time change segment is associated with the corresponding state data sub-node, otherwise it is determined that there is no association;

[0040] The state data sub-node corresponding to the real-time change segment with the smallest real-time detection distance is selected as the detection state node, and then the performance score of the test lithium battery of the target model is obtained according to the detection state node at each time point.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] The present invention establishes a lithium battery performance detection network according to the test status data of each lithium battery under the same number of charge and discharge cycles, and then obtains the state fragment distance between each status data sub-node, obtains the relative state offset probability between each status data sub-node according to the state fragment distance, and obtains the real-time test status data of the tested lithium battery in each charge and discharge cycle, inputs the real-time test status data into the lithium battery performance detection network to match the relevant status data sub-node, and then obtains the performance score of the tested lithium battery, thereby achieving the goal of fully displaying the potential correlation between various abnormal factors while improving the accuracy of lithium battery performance detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention.

[0044] Figure 1 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0045] To make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0046] Example 1

[0047] like Figure 1 As shown, the data analysis method applied to lithium battery performance detection includes the following steps:

[0048] Step S1, setting a plurality of target model lithium batteries to perform cyclic charge and discharge operations simultaneously, and collecting test status data of each lithium battery during each cyclic charge and discharge process until the number of cyclic charge and discharge times reaches the rated number of times of use of the target model lithium battery;

[0049] Step S2, based on the test state data of each lithium battery under the same number of charge and discharge cycles, generate a standard state data change interval and a standard test state change tree of the target model of lithium battery under each charge and discharge cycle, and establish a test state change tree of each lithium battery under each charge and discharge cycle operation, wherein the standard test state change tree and the test state change tree are provided with a plurality of state data nodes;

[0050] Step S3, establishing a lithium battery performance detection network, inputting all test state change trees and standard test state change trees into the lithium battery performance detection network, and then obtaining the state fragment distance between each state data sub-node, establishing a number of local state relationship clusters according to the state fragment distance between each state data sub-node, and obtaining the relative correlation value between each local state relationship cluster, and then obtaining the relative state offset probability between each state data sub-node according to the relative correlation value;

[0051] Step S4, perform multiple cycles of charge and discharge operations on the test lithium battery of the target model, and obtain real-time test status data of each cycle of charge and discharge, input the real-time test status data into the lithium battery performance detection network to match the relevant status data sub-node, and then obtain the performance score of the test lithium battery according to the relative status offset probability between the status data sub-node and other status data sub-nodes.

[0052] Example 2

[0053] This embodiment is a further limitation of Embodiment 1, and Step S1 is implemented by the following process:

[0054] Set n target lithium batteries and set numbers a1, a2, a3, ..., a for each lithium battery. n , where n is a natural number greater than 0;

[0055] Obtain the rated number of times k of use of the target model of lithium battery through the Internet, then perform k cycles of charge and discharge operations on each lithium battery, and install a voltage sensor, an internal resistance sensor, and a temperature sensor on each lithium battery;

[0056] In the process of cyclic charge and discharge operation of each lithium battery, the historical temperature change curve of each lithium battery is generated by the temperature sensor, and the historical voltage change curve and the historical internal resistance change curve of each lithium battery in the charging or discharging process are generated by the voltage sensor and the internal resistance sensor. It should be noted that in the process of the voltage sensor collecting the historical voltage change curve of the lithium battery, the slope of the historical voltage change curve at each time point is obtained in real time. If the slope of the historical voltage change curve at three consecutive time points is equal to 0, the currently generated historical voltage change curve is output, and a new historical voltage change curve is regenerated;

[0057] At the end of each cycle charge and discharge operation, the historical voltage change curve, historical internal resistance change curve and historical temperature change curve generated by the corresponding cycle charge and discharge operation of each lithium battery are integrated to obtain the test status data, and each test status data is numbered according to the number of cycles and the lithium battery number. 1，m 、a 2，m 、a 3，m ,……,a n，m , where a n，i The number indicates the number is a n The test status data of the lithium battery at the mth cycle charge and discharge operation, where m is a natural number greater than 0 and less than or equal to k;

[0058] The above operation of obtaining the test status data of each cycle charge and discharge operation is repeated until the number of cycle charge and discharge operations reaches k.

[0059] Example 3

[0060] This embodiment is a further limitation of Embodiment 1, and Step S2 is implemented by the following process:

[0061] A two-dimensional coordinate system is established, and then the historical change curves corresponding to the same data type in the test state data with the same second digit of the serial number subscript are mapped to the same two-dimensional coordinate system, and a number of time points are set on the x-axis of the two-dimensional coordinate system, and then each historical change curve is divided into h historical change segments according to the time points, where h is a natural number greater than 0;

[0062] Obtain the spatial distances between the corresponding time points on each two-dimensional coordinate system but between the historical change segments corresponding to different historical change segments, and set the spatial distance threshold;

[0063] Compare the spatial distance between each historical change segment with the spatial distance threshold. If the spatial distance is greater than the spatial distance threshold, it is determined that there is no correlation between the corresponding historical change segments.

[0064] If the spatial distance is less than or equal to the spatial distance threshold, the corresponding historical change segments are associated, and the number of associations is recorded once for each of the two historical change segments;

[0065] Select the historical change segments associated with the historical change segments with the largest number of associations to form the standard state data change interval segments of the target model of lithium battery between the corresponding time points;

[0066] Then, the standard state data change interval segments of various types of historical change curves of the lithium battery between various time points are obtained respectively, and the standard state data change interval segments between various time points are connected in chronological order, so as to obtain the standard state data change interval of the lithium battery in each cycle charge and discharge operation.

[0067] Furthermore, k*(h-1) state event main nodes are established for each lithium battery, and three state data sub-nodes are established according to the type of historical change curve, and connected to the state event main node, and then each state event main node is connected according to the cycle charge and discharge sequence to obtain the test state change tree of each lithium battery under each cycle charge and discharge operation, and the historical change fragments between each time point are stored in the corresponding state data sub-node;

[0068] At the same time, according to the process of establishing the test state change tree, a standard test state change tree is established, and the standard state data change interval fragments between each time point are stored in each standard state data sub-node.

[0069] Example 4

[0070] This embodiment is a further limitation of Embodiment 1, and step S3 is implemented by the following process:

[0071] Establishing a lithium battery performance detection network, and setting a three-dimensional coordinate system in the lithium battery performance detection network, and then inputting all test state change trees and standard test state change trees into the lithium battery performance detection network;

[0072] First, the standard test state change trees corresponding to each cycle charge and discharge operation are arranged in chronological order in the center of the lithium battery performance detection network, and at the same time, the state change trees of each lithium battery in each cycle test are arranged around the standard test state change tree in chronological order;

[0073] Obtain the state data sub-node of each state event main node, and the calibration segment distance between the historical change segment and the standard state data change interval contained in the standard state data sub-node of the same type as the standard state event main node in the corresponding time sequence;

[0074] It should be noted that if the historical change segment is within the corresponding standard state data change interval, the calibration segment distance between the corresponding historical change segment and the standard state data change interval is recorded as 0;

[0075] At the same time, obtain the state fragment distance between the state data sub-nodes of each state event main node in the same time sequence, and set the state fragment distance threshold;

[0076] Each state data sub-node is recorded as a central state data sub-node in sequence, and then the relationship between the state fragment distance between the central state data sub-node and the state data sub-nodes of the same type and the state fragment distance threshold value in the same time sequence is determined;

[0077] If the state fragment distance between the central state data subnode and the state data subnode is greater than or equal to the state fragment distance threshold, no operation is performed;

[0078] If the state fragment distance between the central state data subnode and the state data subnode is less than the state fragment distance threshold, a local state relationship cluster is established with the central state data subnode as the center and the corresponding state data subnode as the associated state data subnode;

[0079] When all the associated state data sub-nodes of each central state data sub-node are judged, it is judged whether the state fragment distance between the central state data sub-nodes of adjacent local state relationship clusters is less than the state fragment distance threshold. If the judgment is not established, no operation is performed;

[0080] If the judgment is established, the corresponding two local state relationship clusters are locally merged, where the merging range is the central state data sub-node of any local state relationship cluster, and the distance extending to the other local state relationship cluster is covered by the state fragment distance threshold as the radius;

[0081] It is judged in turn whether each local state relationship cluster can be merged, and each local state relationship cluster is merged in turn according to the judgment result, and at the same time, the association range area is set for each local state relationship cluster with the central state data sub-node as the center and the state fragment distance threshold as the radius;

[0082] Then, according to the value of the ratio of the area of ​​the association range occupied by the relatively local merged parts between each local state relationship cluster, the relative association value between the corresponding two central state data sub-nodes is obtained;

[0083] For example, there are two local state relationship clusters, where the ratio of the associated range area of ​​the first local state relationship cluster to the relatively local merged part of the second local state relationship cluster is 0.45, and the ratio of the associated range area of ​​the second local state relationship cluster to the relatively local merged part of the first local state relationship cluster is 0.30. Then the relative association value of the first local state relationship cluster to the second local state relationship cluster is 4.5, and the relative association value of the first local state relationship cluster to the second local state relationship cluster is 3.

[0084] Furthermore, the relative correlation values ​​between the state data sub-nodes associated with each state event main node at the same time point are obtained, and then the relative state offset probability between each state event main node is obtained according to the relative correlation value, wherein the calculation formula of the relative state offset probability is:

[0085] where μ i，j，time , α i，j，time , β i，j，time , γ i，j，time Respectively represent the number a i The lithium battery corresponding status event master node at time point is time and number is a j The relative state offset probability between the main nodes of the lithium battery corresponding state event, as well as the relative correlation value between each state data sub-node, ω is a constant, i and j are not the same number, and are both natural numbers greater than 0 and less than or equal to n;

[0086] The state abnormal distance is set to determine whether the calibration fragment distance between each state data sub-node and the corresponding standard state data sub-node is less than or equal to the state abnormal distance. If so, the corresponding state data sub-node is marked as a normal state data sub-node, otherwise the corresponding state data sub-node is marked as an abnormal state data sub-node.

[0087] Example 5

[0088] This embodiment is a further limitation of Embodiment 1, and step S4 is implemented by the following process:

[0089] Perform multiple cycles of charge and discharge on the target model of the test lithium battery, and collect real-time test status data of each cycle of charge and discharge through voltage sensors, internal resistance sensors and temperature sensors;

[0090] The real-time change curves in the real-time test status data of each cycle of charge and discharge are divided into a number of real-time change segments in the order of time points, and then each real-time change segment is input into the lithium battery performance detection network, and the real-time detection distance between each real-time change segment and each status data sub-node at the same time point is obtained;

[0091] If it is judged that the real-time detection distance is less than or equal to the state segment distance threshold, then it is judged that the real-time change segment is associated with the corresponding state data sub-node, otherwise it is judged that there is no association. It should be noted that if it is judged that the real-time detection distances between the real-time change segment and each state data sub-node at the same time point are greater than the state segment distance threshold, then the lithium battery performance detection network is updated according to the real-time test state data of the test lithium battery of the target model;

[0092] The state data subnode with the smallest real-time detection distance corresponding to the real-time change segment is selected as the detection state node, and then the performance score A of the test lithium battery of the target model is obtained according to the detection state node at each time point. The calculation formula of the performance score A is:

[0093] Where L a，内阻 , L a，电压 , L a，温度 Respectively represent the abnormal state distance between the detection state node corresponding to the internal resistance, voltage and temperature and the corresponding standard state data sub-node at the a-th time point, ∈ a，内阻 ,∈ a，温度 ,∈ a，电压 Indicates the total number of offsets of the detection status nodes of the internal resistance, voltage and temperature at the ath time point, Δ a，内阻 , Δ a，电压 , Δ a，温度 Indicates the correlation distance difference of the detection status nodes of internal resistance, voltage and temperature at the a-th time point;

[0094] The calculation formula for the total number of offsets ∈ is:

[0095] The calculation formula of the correlation distance difference Δ is: Δ=L a -L 状态异常距离 ;

[0096] Set the performance score threshold. If the performance score A is less than or equal to the performance score threshold, the corresponding tested lithium battery is judged to be defective.

[0097] If the performance score A is greater than the performance score threshold, the corresponding tested lithium battery is judged to be a normal product.

[0098] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A data analysis method for lithium battery performance testing, characterized in that: The following steps are involved: Step S1, setting a plurality of target model lithium batteries to perform cyclic charge and discharge operations simultaneously, and collecting test status data of each lithium battery during each cyclic charge and discharge process until the number of cyclic charge and discharge times reaches the rated number of times of use of the target model lithium battery; Step S2, based on the test state data of each lithium battery under the same number of charge and discharge cycles, generate a standard state data change interval and a standard test state change tree of the target model of lithium battery under each charge and discharge cycle, and establish a test state change tree of each lithium battery under each charge and discharge cycle operation, wherein the standard test state change tree and the test state change tree are provided with a plurality of state data nodes; Step S3, establishing a lithium battery performance detection network, inputting all test state change trees and standard test state change trees into the lithium battery performance detection network, and then obtaining the state fragment distance between each state data sub-node, establishing a number of local state relationship clusters according to the state fragment distance between each state data sub-node, and obtaining the relative correlation value between each local state relationship cluster, and then obtaining the relative state offset probability between each state data sub-node according to the relative correlation value; Step S4, perform multiple cycles of charge and discharge operations on the test lithium battery of the target model, and obtain real-time test status data of each cycle of charge and discharge, input the real-time test status data into the lithium battery performance detection network to match the relevant status data sub-node, and then obtain the performance score of the test lithium battery according to the relative status offset probability between the status data sub-node and other status data sub-nodes.

2. A data analysis method for lithium battery performance testing according to claim 1, characterized in that: The process of acquiring the test status data includes: Set n target models of lithium batteries, set a number for each lithium battery, obtain the rated number of uses k of the target model of lithium batteries through the Internet, and then perform k cycles of charge and discharge operations on each lithium battery; During the process of cyclic charge and discharge operation of each lithium battery, the historical temperature change curve, historical voltage change curve and historical internal resistance change curve of the lithium battery are collected. At the end of each cyclic charge and discharge operation, the historical voltage change curve, historical internal resistance change curve and historical temperature change curve generated by each lithium battery in the corresponding cyclic charge and discharge operation are integrated to obtain the test status data.

3. A data analysis method for lithium battery performance testing according to claim 2, characterized in that: The generation process of the standard state data change interval includes: Establish a two-dimensional coordinate system, map the historical change curves with the same data type to the same two-dimensional coordinate system, set a number of time points, and then divide each historical change curve into h historical change segments according to the time points, where h is a natural number greater than 0; Obtain the spatial distances between the historical change segments corresponding to the time points on each two-dimensional coordinate system but corresponding to different historical change segments, set a spatial distance threshold, compare the spatial distance between each historical change segment with the spatial distance threshold, and if the spatial distance is greater than the spatial distance threshold, it is determined that there is no correlation between the corresponding historical change segments; If the spatial distance is less than or equal to the spatial distance threshold, the corresponding historical change segments are associated, and the number of associations is recorded once for each of the two historical change segments. The historical change segments associated with the historical change segment with the largest number of associations are selected to form the standard state data change interval segments of the target model lithium battery between the corresponding time points; The standard state data change interval segments between various time points are connected in chronological order, thereby obtaining the standard state data change interval of the lithium battery in each cycle charge and discharge operation.

4. A data analysis method for lithium battery performance testing according to claim 3, characterized in that: The standard test state change tree and the test state change tree establishment process include: Establish k*(h-1) state event main nodes for each lithium battery respectively, establish three state data sub-nodes according to the type of historical change curve, and connect them with the state event main node, then connect each state event main node according to the cyclic charge and discharge sequence, and obtain the test state change tree of each lithium battery under each cyclic charge and discharge operation, and store the historical change fragments between each time point in the corresponding state data sub-node. At the same time, according to the process of establishing the test state change tree, establish a standard test state change tree, and store the standard state data change interval fragments between each time point in each standard state data sub-node.

5. A data analysis method for lithium battery performance testing according to claim 4, characterized in that: The process of obtaining the state fragment distance between state data subnodes includes: Establish a lithium battery performance detection network, and obtain the state data sub-nodes of each state event main node, and the calibration segment distance between the historical change segment and the standard state data change interval contained in the standard state data sub-node of the same type carried by the standard state event main node of the corresponding time sequence; At the same time, the state fragment distances between the state data subnodes of each state event main node in the same time sequence are obtained, and the state fragment distance threshold is set.

6. A data analysis method for lithium battery performance testing according to claim 5, characterized in that: The process of establishing the local state relation cluster includes: Each state data sub-node is recorded as a central state data sub-node in sequence, and then the relationship between the state fragment distance between the central state data sub-node and the state data sub-nodes of the same type and the state fragment distance threshold value in the same time sequence is determined; According to the judgment result, a local state relationship cluster is established with the central state data sub-node as the center and the corresponding state data sub-node as the associated state data sub-node, and then it is judged whether the state fragment distance between the central state data sub-nodes of adjacent local state relationship clusters is less than the state fragment distance threshold value, and the corresponding two local state relationship clusters are locally merged according to the judgment result; With the central state data sub-node as the center and the state fragment distance threshold as the radius, the association range area is set for each local state relationship cluster, and the relative association value between the corresponding two central state data sub-nodes is obtained according to the value of the ratio of the association range area occupied by the relative local merged part between each local state relationship cluster; The relative correlation values ​​between the state data sub-nodes associated with each state event main node at the same time point are obtained, and then the relative state offset probability between each state event main node is obtained according to the relative correlation values.

7. A data analysis method for lithium battery performance testing according to claim 6, characterized in that: The state abnormal distance is set to determine whether the calibration fragment distance between each state data sub-node and the corresponding standard state data sub-node is less than or equal to the state abnormal distance. If so, the corresponding state data sub-node is marked as a normal state data sub-node, otherwise the corresponding state data sub-node is marked as an abnormal state data sub-node.

8. A data analysis method for lithium battery performance testing according to claim 7, characterized in that: The process of obtaining the performance score of the tested lithium battery includes: The real-time change curves in the real-time test status data of each cycle of charge and discharge are divided into a number of real-time change segments in the order of time points, and then each real-time change segment is input into the lithium battery performance detection network, and the real-time detection distance between each real-time change segment and each status data sub-node at the same time point is obtained; If it is determined that the real-time detection distance is less than or equal to the state segment distance threshold, then it is determined that the real-time change segment is associated with the corresponding state data sub-node, otherwise it is determined that there is no association; The state data sub-node corresponding to the real-time change segment with the smallest real-time detection distance is selected as the detection state node, and then the performance score of the test lithium battery of the target model is obtained according to the detection state node at each time point.

Images