A method and system for detecting performance of a lithium battery starting power supply

Abstract

The present application belongs to the technical field of automobile starting power supply, and particularly relates to a lithium battery starting power supply performance detection method and system, which comprises the following steps: clustering power supply usage information according to the differences in the power sequence and voltage sequence of the power usage information, and obtaining a plurality of power supply usage information clusters; obtaining the performance of the power usage information based on the data differences in the voltage sequence of the power usage information; establishing a power supply performance change model for each power supply usage information cluster according to the performance change differences between each power supply usage information and the power supply usage information in the same cluster; obtaining the relative performance position of the power usage information; establishing a power supply performance change model for each power supply usage information according to the change rule of the relative performance position of all power usage information of the power supply usage information; and establishing a power supply performance prediction model based on a machine learning model to complete the performance detection of the starting power supply. The present application improves the accuracy and timeliness of starting power supply performance detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automobile starting power supply. More particularly, the present application relates to a lithium battery starting power supply performance detection method and system. BACKGROUND

[0002] The starting power supply is mainly a portable mobile power supply for emergency starting of a vehicle. When the vehicle battery cannot be started due to low temperature, aging and insufficient energy, the starting power supply can provide a portable starting solution for the vehicle. Compared with the traditional lead-acid battery, the lithium ion battery has higher energy density, longer cycle life and lighter weight, and can provide stronger acceleration performance and endurance for the vehicle. Therefore, the lithium battery starting power supply is currently used as an emergency starting power supply for a vehicle.

[0003] In order to ensure the availability and safety of the lithium battery starting power supply, the lithium battery starting power supply needs to be subjected to charge-discharge cycle performance detection and starting performance test under different temperature environments. Compared with the ordinary lithium battery power supply, the starting power supply has higher performance requirements, and therefore the lithium battery starting power supply needs to be subjected to additional performance detection. In the related art, for example, a system and method for detecting and identifying positive and negative connection of an emergency starting power supply under ultra-low voltage are disclosed in a Chinese patent document with the authorization announcement number CN113433486B. The system and method detect whether the clip of the emergency starting power supply is connected through a clip on-off information acquisition module, and detect whether the clip connection polarity is correct through a positive and negative sampling detection module, so as to avoid damage to the vehicle and the emergency starting power supply caused by ultra-low voltage or short circuit fault of the vehicle battery, and improve the reliability and safety of the emergency starting power supply.

[0004] However, the detection of the emergency starting power supply is usually performed when the user is ready to use the emergency starting power supply. If the detection finds that the emergency starting power supply cannot be used, the user has no alternative solution and can only wait for rescue, resulting in additional time cost. In addition, the use scenario of the emergency starting power supply is special and the use frequency is low. With the increase of the placement time of the power supply, the parameters such as the remaining power may deviate, so that the power supply can be used when the power supply is detected, but the vehicle cannot be started, resulting in poor detection effect. SUMMARY

[0005] To solve the above technical problems of poor performance detection effect and poor emergency effect of the emergency starting power supply, the present application provides solutions in the following aspects.

[0006] In a first aspect, the present application provides a lithium battery starting power supply performance detection method, comprising:

[0007] The system collects power usage information from several startup power supplies. The power usage information includes several power consumption records for each startup power supply, including the time of use, power consumption sequence, and voltage sequence. Based on the differences in the power consumption sequence and voltage sequence of the power consumption information from different power supplies, the system clusters the power usage information to obtain several power usage information clusters. Based on the data differences in the voltage sequence of the power consumption information, the system obtains the performance of the power consumption information.

[0008] Based on the performance change differences between each power supply usage information and the power supply usage information of the same cluster, a power supply performance change model for each power supply usage information cluster is established: the relative performance position of the power consumption information is obtained, including: plotting the performance of the power consumption information at the same performance position in the power supply performance change model of the corresponding power supply usage information cluster, and recording it as the relative performance position of each power consumption information.

[0009] Based on the changing patterns of the relative performance positions of all power consumption information, a power performance change model for each power consumption information is established. A machine learning model is then used to learn the power performance change models for all power consumption information to establish a power performance prediction model and complete the performance testing of the starting power supply.

[0010] This invention establishes a power supply performance variation model for startup power supplies by analyzing their performance changes. This provides a basis for performance testing of startup power supplies, enabling the determination of testing effectiveness and improving the accuracy of startup power supply performance testing. Furthermore, this invention obtains the relative position of power supply usage information performance within the power supply performance variation model of power supply usage information in the same cluster. This provides a reference for the performance of power supply usage information, avoiding excessive deviations in startup power supply performance testing that could affect the use of the startup power supply.

[0011] Preferably, the acquisition of several power usage information clusters includes:

[0012] Based on the relative power consumption differences and interval power consumption rate differences of different power usage information, the power consumption information differences of different power usage information are obtained; the power consumption information sets of the i-th and m-th power usage information are denoted as . , ,right and Perform matching and obtain China regarding Matching a subset of electricity information; calculating China regarding The matching subset of electricity information, and The mean of the differences in electricity consumption information among all corresponding matching electricity consumption information is denoted as . right The unilateral difference in electricity consumption; right the single-sided power consumption difference of the minimum value of the single-sided power consumption difference of , recorded as the distance of the i-th and m-th power usage information; using a distance-based clustering algorithm, all power usage information is clustered to obtain a plurality of power usage information clusters.

[0013] The application classifies power usage information, and refines the analysis of power usage information to similar use environments and vehicles, so as to avoid that the difference of power usage information is too large to cause poor model establishment effect.

[0014] Preferably, the relative power consumption is obtained by:

[0015] The c-th and c+1-th power consumption information of the i-th power usage information is recorded as , ; the maximum value and the minimum value of the power sequence of are subtracted and positively correlated normalized, and recorded as the relative power consumption of .

[0016] Preferably, the interval power consumption speed is obtained by:

[0017] The use time of and is subtracted, recorded as the interval time of , and the interval time is measured in minutes; the first power value of the power sequence of is subtracted from the last power value of the power sequence of , and then divided by the interval time of , and positively correlated normalized, recorded as the interval power consumption speed of .

[0018] The application measures the difference of the application vehicle of different starting power based on the interval power consumption speed, so as to avoid that the large power consumption difference causes data dispersion and affects the analysis of the rule of power usage information.

[0019] Preferably, the power consumption information difference satisfies the expression:

[0020] ;

[0021] In the formula, represents the c-th power consumption information of the i-th power usage information, represents the d-th power consumption information of the m-th power usage information; represents the power consumption information difference of and . represents the power consumption information difference of relative power consumption, indicates relative power consumption; indicates interval power consumption speed; indicates interval power consumption speed; indicates absolute value function; indicates normalization function.

[0022] Preferably, the acquisition the matching power information subset of includes:

[0023] The first power information of is recorded as the matching power information of the first power information of The power information with the minimum difference from the second power information of is acquired from the remaining power information of , recorded as the matching power information of the second power information of , and the matching power information of all power information of is sequentially acquired, obtaining the matching power information subset of about ; the remaining power information of only contains power information with a serial number greater than the matched power information in , when there is still unmatched power information and the remaining power information of is 0, or all power information of

[0024] has matching power information, the matching is completed.

[0025] Preferably, the data difference of the voltage sequence based on the power information is used to acquire the performance of the power information, including:

[0026] The time point corresponding to the minimum value of the voltage sequence of is recorded as the characteristic time point of , the difference between the last time point of the voltage sequence of and the characteristic time point of is recorded as the first difference value; the difference between the last voltage value of the voltage sequence of and the minimum value is recorded as the second difference value; the second difference value is compared with the first difference value, and positive correlation normalization is performed, and the performance of is recorded as .

[0027] The application quantifies the performance of the starting power supply corresponding to different power consumption information of power supply use information, improves the accuracy of analysis of performance changes of the starting power supply, and makes performance detection of the starting power supply more accurate.

[0028] Preferably, the power supply performance change model of each power supply use information cluster is established by:

[0029] The performance of all power consumption information of the i-th power supply use information is obtained, a time-performance coordinate system is established with the use time as the horizontal axis and the performance as the vertical axis, a scatter plot of the i-th power supply use information is drawn, and linear fitting is performed to obtain a performance fitting straight line of the i-th power supply use information; the Euclidean distance mean of the scatter plot of the i-th power supply use information and the performance fitting straight line of the i-th power supply use information is negatively correlated and normalized, and is recorded as the fitting effect of the i-th power supply use information; the performance fitting straight line of the power supply use information with the largest fitting effect in each power supply use information cluster is recorded as the power supply performance change model corresponding to the power supply use information cluster.

[0030] Preferably, the power supply performance change model of each power supply use information is established by:

[0031] The distance mean of the relative positions of all adjacent power consumption information of the i-th power supply use information is obtained, and is recorded as The region growing algorithm is used to cluster all power consumption information of the i-th power supply use information; the region growing condition is that the clustering of adjacent power consumption information is greater than A plurality of sparse power consumption information clusters of the i-th power supply use information are obtained; the region growing condition is that the clustering of adjacent power consumption information is less than A plurality of close power consumption information clusters of the i-th power supply use information are obtained; the time range contained by any power consumption information cluster is recorded as a performance stage, a plurality of performance stages of the i-th power supply use information are obtained, and all the performance stages are arranged in time sequence to form a power supply performance change model of the i-th power supply use information.

[0032] In a second aspect, the application provides a lithium battery starting power supply performance detection system, which comprises a processor and a memory, and the memory stores computer program instructions.

[0033] By using the above technical solution, the above-mentioned lithium battery starting power supply performance detection method is generated into a computer program and stored in the memory to be loaded and executed by the processor, so that a terminal device is manufactured according to the memory and the processor, and the use is convenient.

[0034] The application has the following advantages:

[0035] (1) The power supply performance change model of the starting power supply can provide reference and prediction for performance detection of the starting power supply, and the performance detection accuracy of the starting power supply can be improved;

[0036] (2) The power supply performance change model of the starting power supply can provide reference and prediction for performance detection of the starting power supply, and the performance detection accuracy of the starting power supply can be improved;

[0037] (3) The power supply performance change model of the starting power supply can provide reference and prediction for performance detection of the starting power supply, and the performance detection accuracy of the starting power supply can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is a flow chart schematically showing a lithium battery starting power supply performance detection method in the present application;

[0039] Figure 2 is a schematic diagram showing two battery health state change diagrams;

[0040] Figure 3 is a schematic diagram showing two battery health state change diagrams;

[0041] Figure 4 is a schematic diagram showing the relative position of power supply information. DETAILED DESCRIPTION

[0042] The embodiment of the present application discloses a lithium battery starting power supply performance detection method, referring to Figure 1 , comprising steps S1-S3:

[0043] S1: Collecting the use information of a plurality of starting power supplies.

[0044] It needs to be explained that when the battery of the car is insufficient to start the car, by connecting the starting power supply and the battery, using the large current provided by the built-in battery, the car battery is charged instantaneously, so that the car can start. During the charging process, the voltage will drop and then rise, for example, the voltage 13.6V drops to 10V and then rises to 11V, and after stopping discharging, the voltage returns to 13V; the power will gradually decrease, for example, the power consumption of a small gasoline car is about 1Ah. With the increase of use frequency, the performance of the starting power supply gradually weakens, the change of voltage will appear deviation, for example, the voltage drops to 9V, which leads to ignition failure, and the amplitude of voltage recovery will decrease; the consumption of power will also change, with the decrease of battery health status, the actual capacity of the battery will decrease, so that the starting car needs to consume a higher percentage of power. Therefore, the present application first acquires the voltage use information of the starting power supply, so as to analyze the voltage and power data of the use process of the starting power supply.

[0045] Specifically, the power use information of a plurality of starting power supplies is collected, the power use information includes a plurality of power consumption information corresponding to the starting power supply, the power consumption information includes a use time, and a power sequence and a voltage sequence, the power sequence and the voltage sequence are power data and voltage data sorted in time order at T minutes after the use time. The sampling unit of the power sequence and the voltage sequence is 0.1 seconds each time, and the T is a preset time length, which is set by the implementer according to the actual implementation situation, because the voltage is recovered within five minutes, so the value of T can be set to 5.

[0046] At this point, the power use information of a plurality of starting power supplies is obtained.

[0047] S2: according to the difference of the power sequence and the voltage sequence of the power consumption information of different power use information, the power use information is clustered to obtain a plurality of power use information clusters; based on the data difference of the voltage sequence of the power consumption information, the performance of the power consumption information is obtained; according to the performance change difference of each power use information and the same cluster power use information, the power performance change model of each power use information cluster is established; the relative performance position of the power consumption information is obtained; according to the change rule of the relative performance position of all power consumption information of the power use information, the power performance change model of each power use information is established.

[0048] It should be noted that the use of starting power supply is widely used, involving many types of vehicles, regions, for example, 12V car voltage suitable for cars, 24V truck and large engineering vehicles, the climate difference in different regions may be different, and the starting power supply is affected by temperature, the same performance of the same lithium battery starting power supply can be used less times in low temperature environment than in normal temperature environment, so it is impossible to detect all lithium battery starting power supplies in all use scenarios by a unified performance detection method. Considering that the type of vehicle mainly affects the single consumption of the starting power supply during use, and the region mainly affects the consumption of the starting power supply in the placement state and the single consumption of the starting power supply during use, therefore, the power supply use information is clustered according to the power difference of the starting power supply.

[0049] It should be noted that for each power supply use information cluster, as the number of uses of the starting power supply accumulates, the performance of the starting power supply gradually decreases, so the performance decline rule of each power supply use information cluster is analyzed to obtain the power supply performance change model of each power supply use information cluster. However, users still have different habits of using starting power supply, for example, users who are used to maintaining the vehicle at high power and high oil quantity are less likely to use the starting power supply, and users who are not sensitive to the remaining energy use the starting power supply more frequently, so the time change and the performance change of the starting power supply are not synchronous, and the starting power supply has different power performance changes under different charging habits, so it is difficult to represent a power supply use information cluster by a power performance change model. Considering that time, charging habit and power habit are different, but the same is that the starting power supply changes from optimal performance to worst performance until it cannot be used, therefore, the relative performance index is established, for example, the relative performance of a starting power supply that is used only once but idle for a year is equivalent to a starting power supply that is used many times in a year, so that all power supply use information of a cluster can be placed in the same performance reference system, improving the scope of application of the power supply performance change model.

[0050] It should be further noted that the difference in power before and after the use of the starting power supply reflects the single consumption of the starting power supply, so the more similar the power difference before and after the use of the starting power supply, the more similar the single consumption of the starting power supply; the power consumption speed between two uses of the starting power supply reflects the consumption of the starting power supply in the placement state, so the more similar the power consumption speed between two uses of the starting power supply, the more similar the consumption of the starting power supply in the placement state. Therefore, the power supply use information is clustered according to the power information difference of the power supply use information.

[0051] Specifically, the power usage information is clustered according to the differences between the power consumption sequences and the voltage sequences of the power usage information, and a plurality of power usage information clusters are obtained, including:

[0052] The cth and (c+1)th power usage information of the ith power usage information is denoted as 、 The maximum value and the minimum value of the power consumption sequence of are subtracted and positively correlated normalized, and denoted as the relative power consumption of and The usage time of is subtracted, and denoted as the interval time of The first power value of the power consumption sequence of is subtracted from the last power value of the power consumption sequence of , and then divided by the interval time of , and positively correlated normalized, and denoted as

[0053] the interval power consumption speed of .

[0054] ;

[0055] In the formula, denotes the cth power usage information of the ith power usage information, denotes the dth power usage information of the mth power usage information; denotes the power usage information difference between and ; denotes the relative power consumption of , denotes the relative power consumption of ; denotes the interval power consumption speed of ; denotes the interval power consumption speed of ; denotes the absolute value function; denotes the normalization function.

[0056] In the formula, denotes the relative power consumption difference between and , denotes the interval power consumption speed difference between and , the interval power consumption speed difference between and the relative power consumption difference and the interval power consumption speed difference of the i-th power usage information and the m-th power usage information, the greater the value, the greater the power information difference between the i-th power usage information and the m-th power usage information. and the power information difference between the i-th power usage information and the m-th power usage information. and the power information difference between the i-th power usage information and the m-th power usage information.

[0057] It should be noted that when the i-th power usage information is similar to the m-th power usage information, there is a part of the power information with a smaller power information difference between the power information set of the i-th power usage information and the power information set of the m-th power usage information. Therefore, the power information set of the i-th power usage information and the power information set of the m-th power usage information are matched, and the sum of the matching power information differences can be used to describe the similarity of the two power usage information.

[0058] The power information set of the i-th power usage information and the m-th power usage information is denoted as , The power information set of the i-th power usage information and the m-th power usage information is matched to obtain the matching power information subset of the i-th power usage information and the m-th power usage information: the first power information of the i-th power usage information is denoted as the matching power information of the first power information of the m-th power usage information, and the power information with the smallest power information difference with the second power information of the i-th power usage information is obtained from the remaining power information of the i-th power usage information, denoted as the matching power information of the second power information of the m-th power usage information. The matching power information of all power information of the i-th power usage information is obtained in turn to obtain the matching power information subset of the i-th power usage information and the m-th power usage information; the remaining power information of the i-th power usage information only contains power information with a serial number greater than the matched power information in the i-th power usage information, and when the i-th power usage information still has unmatched power information and the remaining power information of the i-th power usage information is 0, or all power information of the i-th power usage information has matching power information, the matching is completed. and

[0059] ​​​​​​​​​​​​​​​​​​​​​​​​​the minimum value of the single-sided power consumption difference of the i-th and m-th power supply usage information is recorded as the distance between the i-th and m-th power supply usage information.

[0060] It should be noted that the greater the distance between the i-th power supply usage information and the m-th power supply usage information, the less likely it is that the i-th power supply usage information and the m-th power supply usage information belong to the same cluster. Since the distance between different power supply usage information can be used as a clustering basis, a plurality of power supply usage information clusters are obtained.

[0061] The distance-based clustering algorithm is used to cluster all power supply usage information and obtain a plurality of power supply usage information clusters. It should be noted that, for example, using the DBSCAN clustering algorithm, the parameters used include the neighborhood radius and the minimum number of points, which are set by the implementer according to the actual implementation situation, for example, the neighborhood radius can be set to 0.1 and the minimum number of points can be set to 10.

[0062] At this point, a plurality of power supply usage information clusters are obtained.

[0063] It should be noted that in one power supply usage information cluster, due to the different charging and discharging habits of the driver to the starting power supply, the performance of the starting power supply varies, for example, in the case of high-frequency use of shallow charging and shallow discharging, the battery health state decays linearly, and in the case of low-frequency use and full-power storage, the battery health state decays in a stepwise jump. Figure 2 is a schematic diagram of two battery health state changes, wherein represents a stepwise jump decay of the battery health state, represents a linear decay of the battery health state. For example, Figure 3 is a schematic diagram of corresponding data points with the same performance of the two battery health state changes, , is two data points, , is , the corresponding data points with the same performance on . Since the battery health state is a strictly monotonic decreasing function, any data point of has and only has a unique corresponding data point with the same performance in , therefore, establishing a power supply performance change model to obtain the relative performance of each power supply usage information in the power supply performance change model can improve the scope of application of the power supply performance change model and improve the accuracy of starting power supply performance detection.

[0064] Preferably, the performance of the power consumption information is obtained based on the data difference of the voltage sequence of the power consumption information, comprising: obtaining the time corresponding to the minimum value of the voltage sequence of .The characteristic moment will The last moment of the voltage sequence and The difference at the characteristic time points is recorded as the first difference; The difference between the last voltage value of the voltage sequence and the minimum value is denoted as the second difference; the second difference is compared with the first difference and normalized by positive correlation, denoted as... The performance of a starting power supply is measured by the speed at which its voltage recovers after a sudden drop. It should be noted that a degraded battery cannot restore its voltage and recovers slowly.

[0065] Preferably, based on the performance change differences between each power supply usage information cluster and the power supply usage information of the same cluster, a power supply performance change model is established for each power supply usage information cluster, including:

[0066] To obtain the performance data of all power consumption information for the i-th power source, establish a time-performance coordinate system with usage time as the horizontal axis and performance as the vertical axis. Plot a scatter plot of the i-th power source's usage information and perform linear fitting to obtain a performance fitting line for the i-th power source's usage information. It should be noted that the linear fitting method can employ the least squares method.

[0067] The mean Euclidean distance between the scatter plot of the i-th power usage information and the performance fitting line of the i-th power usage information is negatively correlated and normalized, denoted as the fitting effect of the i-th power usage information. It should be noted that the better the fitting effect, the more linearly the performance change of the i-th power usage information follows, and thus the more suitable it is as a standard model to characterize the performance change of a cluster of power usage information.

[0068] The performance fitting line of the power usage information with the best fitting effect among each power usage information cluster is denoted as the power performance change model of the corresponding power usage information cluster.

[0069] At this point, the power performance change model for each power supply usage information cluster has been obtained.

[0070] It should be noted that the performance changes of all power usage information in each power usage information cluster can be represented by the power performance change model of the corresponding power usage information cluster, only the corresponding time is different. Therefore, this invention obtains the relative position of the performance of all power usage information of each power usage information in the power performance change model, and analyzes the change law of the relative position, thereby obtaining the performance change model of each power usage information.

[0071] Preferably, the relative performance position of electricity consumption information is obtained, and based on the changing patterns of the relative performance positions of all electricity consumption information, a power performance change model for each power consumption information is established, including:

[0072] The relative position of the power consumption information to the power consumption information is recorded as the relative position of each power consumption information in the power performance change model corresponding to the power consumption information cluster. Figure 4 The relative position of the power consumption information to the power consumption information is recorded as the relative position of each power consumption information in the power performance change model corresponding to the power consumption information cluster.

[0073] It should be noted that the slower the power performance decreases, the closer the distance between the power consumption information, and the more concentrated in the power performance change model, so the power consumption information can be divided into several stages, and each stage represents a performance change rule.

[0074] The distance average of the relative positions of all adjacent power consumption information of the i-th power consumption information is recorded as The region growing algorithm is used to cluster all power consumption information of the i-th power consumption information; the region growing condition is that the clustering of adjacent power consumption information is greater than A number of sparse power consumption information clusters of the i-th power consumption information are obtained; the region growing condition is that the clustering of adjacent power consumption information is less than The time range contained in any power consumption information cluster is recorded as a performance stage, and a number of performance stages of the i-th power consumption information are obtained, and all the performance stages are arranged in time sequence to form a power performance change model of the i-th power consumption information.

[0075] So far, the power performance change model of each power consumption information has been obtained.

[0076] S3: Use a machine learning model to learn the power performance change model of all power consumption information, establish a power performance prediction model, and complete the performance detection of the starting power supply.

[0077] It should be noted that by learning the power performance change model of all power consumption information, the subsequent performance of each power consumption information can be predicted, so the present application adopts an unsupervised machine learning model to establish a power performance prediction model. Considering that the LSTM model has flexible memory ability and can effectively propagate long-term information, the present application adopts an LSTM model to establish a power performance prediction model.

[0078] Specifically, the power performance change model of all power supply usage information is input into the LSTM model for training, including the relative performance position of each power usage information and the performance stage corresponding to the power performance change model, and the loss function used is the mean square error function, and the activation function includes the tanh function and the softmax function. The trained LSTM model is recorded as a power performance prediction model. When detecting the performance of the starting power supply, the relative performance position of all power usage information of the starting power supply and the performance stage corresponding to the power performance change model are obtained through the method in steps S1-S2, and are input into the power performance prediction model, and the voltage performance of the starting power supply is output.

[0079] At this point, the performance detection of the lithium battery starting power supply is completed.

[0080] The embodiment of the application further discloses a lithium battery starting power supply performance detection system, comprising a processor and a memory, and the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the lithium battery starting power supply performance detection method is realized.

[0081] The system further comprises a communication bus and a communication interface and other components well known to those skilled in the art, and the settings and functions thereof are known in the art, and thus will not be described here.

[0082] Although the present specification has shown and described several embodiments of the present application, it will be apparent to those skilled in the art that many modifications, changes and substitutions can be made without departing from the spirit and scope of the present application. It should be understood that various alternatives to the embodiments of the present application described herein can be employed in practicing the present application.

Claims

1. A method for testing the performance of a lithium-ion battery starting power supply, characterized in that, include: The system collects power usage information from several startup power supplies. This information includes several instances of power consumption by the corresponding startup power supply, along with the time of use, power consumption sequence, and voltage sequence. Based on the differences in the power consumption sequence and voltage sequence of different power supply usage information, the system clusters the power usage information to obtain several power usage information clusters. Based on the data differences in the voltage sequence of the power consumption information, the system obtains the performance of the power usage information. Based on the performance change differences between each power supply usage information and the power supply usage information of the same cluster, a power supply performance change model for each power supply usage information cluster is established. This includes: obtaining the performance of all power consumption information of the i-th power supply usage information, establishing a time-performance coordinate system with usage time as the horizontal axis and performance as the vertical axis, plotting a scatter plot of the i-th power supply usage information, and performing linear fitting to obtain the performance fitting line of the i-th power supply usage information; performing negative correlation normalization on the mean Euclidean distance between the scatter plot of the i-th power supply usage information and the performance fitting line of the i-th power supply usage information, and recording this as the fitting effect of the i-th power supply usage information; and recording the performance fitting line of the power supply usage information with the largest fitting effect in each power supply usage information cluster as the power supply performance change model of the corresponding power supply usage information cluster. Obtaining the relative performance position of electricity consumption information includes: plotting the performance of electricity consumption information at the same performance position in the power performance change model of the corresponding power consumption information cluster class, and recording it as the relative performance position of each piece of electricity consumption information; Based on the changing patterns of the relative performance positions of all power consumption information, a power performance change model for each power consumption information is established, including: obtaining the average distance between the relative positions of all adjacent power consumption information of the i-th power consumption information, denoted as . The region growing algorithm is used to cluster all power consumption information of the i-th power consumption information; the condition for region growing is that the cluster size of adjacent power consumption information is greater than 1. This yields several sparse power consumption clusters for the i-th power consumption information; the region growing condition is that the clustering of adjacent power consumption information is less than 1. We obtain several closely packed power consumption information clusters for the i-th power consumption information; the time range contained in any power consumption information cluster is denoted as a performance stage, and we obtain several performance stages for the i-th power consumption information. We then construct the power performance change model of the i-th power consumption information in chronological order by arranging all the performance stages. A machine learning model is used to learn the power performance variation model of all power usage information, and a power performance prediction model is established to complete the performance testing of the startup power supply.

2. The method for testing the performance of a lithium-ion battery starting power supply according to claim 1, characterized in that, The acquisition of several power usage information clusters includes: Based on the relative power consumption differences and interval power consumption rate differences of different power usage information, the power consumption information differences of different power usage information are obtained; the power consumption information sets of the i-th and m-th power usage information are denoted as . , ,right and Perform matching and obtain China regarding Matching a subset of electricity information; calculating China regarding The matching subset of electricity information, and The mean of the differences in electricity consumption information among all corresponding matching electricity consumption information is denoted as . right The unilateral difference in electricity consumption; right Unilateral electricity consumption differences and right The minimum value of the unilateral electricity consumption difference is denoted as the distance between the i-th and m-th power consumption information. A distance-based clustering algorithm is used to cluster all power consumption information to obtain several power consumption information clusters.

3. The method for testing the performance of a lithium-ion battery starting power supply according to claim 2, characterized in that, The acquisition of the relative electricity consumption includes: Let the c-th and (c+1)-th power consumption information of the i-th power consumption information be denoted as... , ;Will The difference between the maximum and minimum values ​​of the electricity sequence is calculated and then normalized by positive correlation, denoted as . The relative electricity consumption.

4. The method for testing the performance of a lithium-ion battery starting power supply according to claim 3, characterized in that, The acquisition of the interval power consumption rate includes: Will and The difference between the usage time and the time of use is recorded as follows: The interval time, which is measured in minutes; The first energy value of the energy sequence and Subtract the last energy value from the energy sequence, then multiply by... The intervals are compared and positive correlation normalization is performed, denoted as . The rate at which the battery is consumed at intervals.

5. The method for testing the performance of a lithium-ion battery starting power supply according to claim 2, characterized in that, The differences in electricity consumption information satisfy the expression: ； In the formula, This represents the c-th power consumption information of the i-th power consumption information. This represents the d-th power consumption information of the m-th power consumption information; express and Differences in electricity consumption information; express The relative power consumption express The relative electricity consumption; express The rate at which power is consumed at intervals; express The rate at which power is consumed at intervals; Represents the absolute value function; This represents the normalization function.

6. The method for testing the performance of a lithium-ion battery starting power supply according to claim 2, characterized in that, The acquisition China regarding The subset of matching electricity consumption information includes: Will The first electricity consumption information is recorded as The first electricity consumption information is matched with the electricity consumption information, in Obtain from the remaining electricity consumption information The electricity consumption information with the smallest difference in the second electricity consumption information is denoted as... The second electricity consumption information is matched with the electricity consumption information, and then obtained sequentially. Matching all electricity consumption information with the electricity consumption information to obtain China regarding The matching power information subset; The remaining electricity consumption information only includes When the serial number of the electricity consumption information is greater than that of the matched electricity consumption information, There is still unmatched electricity usage information. When the remaining electricity consumption information is 0, or Matching is complete when all electricity consumption information has a matching electricity consumption information.

7. The method for testing the performance of a lithium-ion battery starting power supply according to claim 3, characterized in that, The performance of obtaining electricity consumption information based on the data difference of the voltage sequence of electricity consumption information includes: Get The time corresponding to the minimum value of the voltage sequence is denoted as . The characteristic moment will The last moment of the voltage sequence and The difference at the characteristic time points is recorded as the first difference; The difference between the last voltage value of the voltage sequence and the minimum value is denoted as the second difference; the second difference is compared with the first difference and normalized by positive correlation, denoted as... Performance.

8. A lithium battery starting power supply performance testing system, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a lithium-ion battery starting power supply performance testing method according to any one of claims 1-7.

Images