Sorting method and device for retired power batteries, electronic equipment and storage medium

Abstract

The invention provides a sorting method and device for decommissioned power batteries, electronic equipment and a storage medium. The sorting method comprises the following steps: acquiring a first battery characteristic parameter of each decommissioned power battery in a target batch of decommissioned power batteries; on the basis of the first battery characteristic parameters and a battery residual capacity prediction model, predicting the battery residual capacity of each retired power battery; based on the first battery characteristic parameter of each sample battery, the battery residual capacity and an entropy weight method, determining a sorting factor of the target batch of decommissioned power batteries; and based on the sorting factor and a K-means clustering algorithm, clustering the target batch of decommissioned power batteries to obtain a sorting result of each decommissioned power battery in the target batch of decommissioned power batteries. Thus, according to the dispersion degree of the battery characteristic parameters in the target batch of decommissioned power batteries, the battery characteristic parameters with the large dispersion degree are selected as the sorting factors of the K-means clustering to sort the target batch of decommissioned power batteries, and the sorting accuracy of the decommissioned power batteries is improved.

Description

Technical Field

[0001] The present application relates to the technical field of cascade utilization of retired power batteries, and in particular to a method, device, electronic equipment and storage medium for sorting retired power batteries. Background Art

[0002] With the rapid popularization of electric vehicles, the cascade utilization of retired power batteries has been a hot issue in the development of energy storage technology in recent years. Due to the differences in material size, appearance structure, power performance, ohmic internal resistance, residual capacity, process manufacturing, etc. of retired power batteries, and the lack of complete operation data records, the performance of retired power batteries is seriously restricted by the batteries with poor performance. If such batteries with large differences are connected in series and parallel to form a battery pack and then put into the cascade utilization scenario, the batteries will be affected by the different current shunting, voltage division, and input and output power, which will affect the operating status and service life of the batteries, and the reliability and safety of the battery cascade utilization cannot be guaranteed.

[0003] Therefore, as an important part of cascade utilization, it is very important to sort and process retired batteries. The existing sorting method of retired power batteries is to sort retired power batteries according to the relationship between coulomb efficiency and battery capacity attenuation. Since the sorting standard is single and the error for retired power batteries in different batches is large, the accuracy of retired power battery sorting is low. Summary of the invention

[0004] In view of this, the embodiments of the present application at least provide a method, device, electronic device and storage medium for sorting retired power batteries. According to the degree of dispersion of battery characteristic parameters in a target batch of retired power batteries, battery characteristic parameters with a larger degree of dispersion are selected as sorting factors for K-means clustering to sort the target batch of retired power batteries, thereby improving the accuracy of sorting retired power batteries.

[0005] This application mainly includes the following aspects:

[0006] In a first aspect, an embodiment of the present application provides a method for sorting retired power batteries, the method comprising:

[0007] Acquire a first battery characteristic parameter of each retired power battery in a target batch of retired power batteries; the first battery characteristic parameter is a plurality of preset battery characteristic parameters that affect aging characteristics of the retired power batteries;

[0008] Based on the first battery characteristic parameter and a battery remaining capacity prediction model, predicting the battery remaining capacity of each of the retired power batteries; the battery remaining capacity prediction model is a model constructed based on a support vector machine;

[0009] Determine the sorting factor of the target batch of retired power batteries based on the first battery characteristic parameter, the remaining battery capacity and the entropy weight method of each sample battery; the sample battery is a preset second number of retired power batteries selected from the target batch of retired power batteries; the sorting factor is a preset third number of battery characteristic parameters with the highest weight value between the first battery characteristic parameter and the remaining battery capacity;

[0010] Based on the sorting factors and the K-means clustering algorithm, the target batch of retired power batteries is clustered to obtain a sorting result of each retired power battery in the target batch of retired power batteries.

[0011] In a second aspect, an embodiment of the present application further provides a sorting device for retired power batteries, the sorting device for retired power batteries comprising:

[0012] An acquisition module, used for acquiring a first battery characteristic parameter of each retired power battery in a target batch of retired power batteries; the first battery characteristic parameter is a plurality of preset battery characteristic parameters that affect the aging characteristics of the retired power batteries;

[0013] A prediction module, configured to predict the remaining battery capacity of each of the retired power batteries based on the first battery characteristic parameters and a battery remaining capacity prediction model; the battery remaining capacity prediction model is a model constructed based on a support vector machine;

[0014] A determination module, configured to determine a sorting factor of the target batch of retired power batteries based on a first battery characteristic parameter of each sample battery, a battery remaining capacity, and an entropy weight method; the sample battery is a preset second number of retired power batteries selected from the target batch of retired power batteries; the sorting factor is a preset third number of battery characteristic parameters having the highest weight value among the first battery characteristic parameter and the battery remaining capacity;

[0015] The sorting module is used to cluster the target batch of retired power batteries based on the sorting factor and the K-means clustering algorithm to obtain a sorting result of each retired power battery in the target batch of retired power batteries.

[0016] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the method for sorting retired power batteries as described above.

[0017] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for sorting retired power batteries as described above are executed.

[0018] The embodiments of the present application provide a method, device, electronic device and storage medium for sorting retired power batteries, wherein the method comprises: obtaining a first battery characteristic parameter of each retired power battery in a target batch of retired power batteries; the first battery characteristic parameter is a plurality of preset battery characteristic parameters that affect the aging characteristics of retired power batteries; based on the first battery characteristic parameter and a battery remaining capacity prediction model, predicting the battery remaining capacity of each retired power battery; the battery remaining capacity prediction model is a model constructed based on a support vector machine; based on the first battery characteristic parameter, the battery remaining capacity and the entropy weight method of each sample battery, determining the sorting factor of the target batch of retired power batteries; the sample battery is a preset second number of retired power batteries selected from the target batch of retired power batteries; the sorting factor is a preset third number of battery characteristic parameters with the highest weight value between the first battery characteristic parameter and the battery remaining capacity; based on the sorting factor and the K-means clustering algorithm, clustering the target batch of retired power batteries to obtain the sorting results of each retired power battery in the target batch of retired power batteries. In this way, according to the degree of dispersion of battery characteristic parameters in the target batch of retired power batteries, battery characteristic parameters with a larger degree of dispersion are selected as sorting factors for K-means clustering to sort the target batch of retired power batteries, thereby improving the accuracy of sorting retired power batteries.

[0019] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 A flow chart of a method for sorting retired power batteries provided in an embodiment of the present application is shown;

[0022] Figure 2 A schematic diagram showing the capacity prediction result of retired power batteries based on a support vector machine in an embodiment of the present application is shown;

[0023] Figure 3A schematic diagram showing a variation trend curve of the sample average error and SSE relative to the number of cluster centers in an embodiment of the present application is shown;

[0024] Figure 4 A functional module diagram of a retired power battery sorting device provided in an embodiment of the present application is shown;

[0025] Figure 5 A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0027] To facilitate the understanding of the present application, the technical solution provided by the present application is described in detail below in conjunction with specific embodiments.

[0028] The embodiments of the present application provide a method, device, electronic device and storage medium for sorting retired power batteries. According to the degree of dispersion of battery characteristic parameters in a target batch of retired power batteries, battery characteristic parameters with a larger degree of dispersion are selected as sorting factors for K-means clustering to sort the target batch of retired power batteries, thereby improving the accuracy of sorting retired power batteries.

[0029] See also Figure 1 , Figure 1 This is a flow chart of a method for sorting retired power batteries provided in an embodiment of the present application. Figure 1 As shown, the method for sorting retired power batteries provided in the embodiment of the present application comprises the following steps:

[0030] S101, obtaining a first battery characteristic parameter of each retired power battery in a target batch of retired power batteries; the first battery characteristic parameter is a plurality of preset battery characteristic parameters that affect aging characteristics of the retired power batteries.

[0031] In the embodiment of the present application, the battery parameter characteristics, as an important performance indicator to measure the battery performance, have an important influence on the sorting of retired power batteries. In the process of cascade utilization of retired power batteries, the performance of retired power batteries is affected by many factors such as temperature, capacity, power, internal resistance, etc. Based on the evaluation of the performance and parameters of retired power batteries, 10 important battery characteristic parameters that affect the aging characteristics of retired power batteries are determined, namely open circuit voltage, battery remaining capacity (actual capacity), ohmic internal resistance, voltage change, polarization performance (polarization voltage), energy density, self-discharge rate, charge and discharge rate, charging temperature rise and discharge temperature rise. Among them, except for the battery remaining capacity that cannot be directly obtained, the remaining 9 battery characteristic parameters are determined as the first characteristic parameters of retired power batteries.

[0032] Specifically, the data information of each first characteristic parameter of each retired power battery in the monitored target batch of retired power batteries is acquired from the acquisition system in the battery management system (BMS) in the Supervisory Control And Data Acquisition (SCADA) of the power battery, and then the data are preprocessed, extracted, and other operations are performed to obtain the first characteristic parameters of the retired power batteries.

[0033] S102, predicting the remaining battery capacity of each of the retired power batteries based on the first battery characteristic parameters and a battery remaining capacity prediction model; the battery remaining capacity prediction model is a model constructed based on a support vector machine.

[0034] In the embodiment of the present application, since the battery characteristic parameter of the remaining battery capacity cannot be directly obtained, it is necessary to predict the remaining battery capacity of each retired power battery based on the first battery characteristic parameter and the battery remaining capacity model trained in advance. Among them, the battery remaining capacity prediction model is a model built based on a support vector machine. Specifically, the support vector machine (SVM) algorithm has better effects in nonlinear, small sample and high-dimensional pattern recognition processing, and can also be applied to other machine learning problems such as function fitting. In machine learning, support vector machine is a supervised learning model related to the learning algorithm, which can analyze data, identify patterns, and be used for classification and regression prediction analysis. It has the advantages of high prediction accuracy, high prediction efficiency, and short prediction time. The embodiment of the present application uses the support vector machine algorithm to estimate the actual remaining battery capacity of retired power batteries, and can more reliably regulate the operation of the entire battery pack according to the state of the retired power batteries.

[0035] S103, determining the sorting factor of the target batch of retired power batteries based on the first battery characteristic parameter of each sample battery, the remaining battery capacity and the entropy weight method; the sample batteries are a preset second number of retired power batteries selected from the target batch of retired power batteries; the sorting factor is a preset third number of battery characteristic parameters with the highest weight value among the first battery characteristic parameter and the remaining battery capacity.

[0036] In the embodiment of the present application, when sorting the retired power batteries in batches, if a large number of battery characteristic parameters are used to sort the retired power batteries, the data storage cost will be too high and the classification will be poor. Therefore, before sorting the retired power batteries, the 10 important battery characteristic parameters that affect the performance of the retired power batteries should be screened. Since the higher the data discreteness of the battery characteristic parameters, the easier and more accurate the classification is, the embodiment of the present application selects a preset second number of retired power batteries from the target batch of retired power batteries as sample batteries, and discards the battery characteristic parameters with low data discreteness, i.e., low weight, in the sample batteries through the entropy weight method (EWM), and retains the battery characteristic parameters with the highest data discreteness, i.e., the highest weight, which is the preset third number of battery characteristic parameters. That is to say, the battery characteristic parameters with the highest data discreteness in the target batch of retired power batteries should be selected from the open circuit voltage, battery remaining capacity, ohmic internal resistance and many other battery characteristic parameters as the sorting factors for cluster sorting of retired power batteries. This can effectively reduce the calculation amount of the battery sorting process and improve the accuracy of sorting.

[0037] S104, clustering the target batch of retired power batteries based on the sorting factors and the K-means clustering algorithm to obtain a sorting result of each retired power battery in the target batch of retired power batteries.

[0038] In the embodiment of the present application, the K-means clustering algorithm is used to divide the target batch of retired power batteries into K categories with similar battery characteristic parameters according to the sorting factors of each retired power battery in the target batch of retired power batteries. Furthermore, the retired power batteries are classified into grades according to the battery characteristic parameters corresponding to each classification, for example, the retired power batteries are classified into four grades of excellent, good, medium and poor according to clustering, and the retired power batteries of different grades are applied to the application scenarios of the corresponding grades, or the retired power batteries are scrapped due to their low performance.

[0039] Furthermore, the battery remaining capacity prediction model is constructed according to the following steps:

[0040] Step a1: Select a preset fourth number and a preset fifth number of retired power batteries from the target batch of retired power batteries as training set sample batteries and test set sample batteries, respectively.

[0041] In this step, before constructing the battery remaining capacity prediction model, a preset fourth number and a preset fifth number of retired power batteries are respectively selected from the target batch of retired power batteries as training set sample batteries and test set sample batteries. For example, 10 retired power batteries are selected as training set sample batteries, and 1 retired power battery is selected as test set sample battery.

[0042] Step a2, performing charge and discharge cycles on the training set sample batteries, using the first battery characteristic parameter and the measured value of the battery remaining capacity of the training set sample batteries at different cycle times as input variables and output variables respectively, and training using a support vector machine algorithm to obtain an initial prediction model.

[0043] In this step, the number of training samples is expanded by repeatedly charging and discharging the above 10 training set sample batteries, and the first characteristic parameters and the actual measured values ​​of the remaining battery capacity of the 10 training set sample batteries at different cycle numbers are obtained. The first battery characteristic parameters and the actual measured values ​​of the remaining battery capacity of the 10 training set sample batteries at different cycle numbers are used as input variables and output variables, respectively, and training is performed using the support vector machine algorithm to obtain an initial prediction model.

[0044] Specifically, for a set of data sets {(x1, y1),…,(x i ,y i )}, where the input variable x i and the output variable y i Together as the input of the support vector machine, x i ∈R n ,y i ∈R n , R n Represents the dimension of the input space, then the support vector machine is defined as:

[0045] f(x)=ω T φ(x)+b;

[0046] Where: T is the transpose of the weight vector; b is the bias constant; φ(x) is the nonlinear mapping function.

[0047] Furthermore, the structural risk function R is defined as:

[0048]

[0049] Where, L ε (y i,f(x i )) is an insensitive loss function;

[0050] Introducing the slack variable ξ i and According to the minimization principle, the structural risk function R is transformed into:

[0051]

[0052] By optimizing and solving the quadratic form with inequality constraints, we get the weight vector:

[0053]

[0054] Where: β i and It is obtained by solving quadratic programming and Lagrange multiplier method. Then the support vector machine regression function is as follows:

[0055] That is the initial prediction model.

[0056] Step a3, performing charge and discharge cycles on the test set sample battery, inputting the first battery characteristic parameters of the test set sample battery at different cycle times into the initial prediction model, and outputting the battery remaining capacity prediction value of the test set sample battery at different cycle times.

[0057] In this step, the test set sample batteries are charged and discharged in cycles, the first battery characteristic parameters of the test set sample batteries at different cycle times are input into the initial prediction model, and the battery remaining capacity prediction values ​​of the test set sample batteries at different cycle times are output.

[0058] Step a4: if the curve of the battery remaining capacity prediction value and the charge and discharge cycle number is similar to the curve of the battery remaining capacity actual measurement value and the charge and discharge cycle number of the test set sample battery, the initial prediction model is determined as the battery remaining capacity prediction model.

[0059] In this step, see Figure 2 , Figure 2 Schematic diagram of the prediction result of retired power battery capacity based on support vector machine in the embodiment of the present application. Figure 2 As shown, Figure 2 The original data of the sample battery of the test set in the embodiment of the present application, that is, the curve of the measured value of the remaining battery capacity of the retired power battery and the number of charge and discharge cycles; and the regression data of the sample battery of the test set, that is, the curve of the predicted value of the remaining battery capacity of the retired power battery and the number of charge and discharge cycles, are shown. Figure 2It can be seen that the two curves are similar, and with the increase of the number of cycles, the remaining capacity of the retired power battery decreases linearly, which is basically in line with the degradation trend of the remaining capacity of the power battery. It can be seen that the overall prediction result of the remaining capacity of the battery based on the support vector machine is stable, and the prediction effect is good. The initial prediction model can be determined as the remaining capacity prediction model of the battery.

[0060] Furthermore, the method further comprises:

[0061] If the curve of the predicted value of the remaining battery capacity and the number of charge and discharge cycles deviates from the curve of the measured value of the remaining battery capacity and the number of charge and discharge cycles of the test set sample batteries, the training set sample batteries and the test set sample batteries are re-selected from the target batch of retired power batteries to construct the battery remaining capacity prediction model.

[0062] In this step, if the above two curves deviate, it means that the prediction effect is not good, and it is necessary to re-select training set sample batteries and test set sample batteries from the target batch of retired power batteries to build a battery remaining capacity prediction model.

[0063] Furthermore, S103 specifically includes:

[0064] Step b1, constructing a battery original data matrix according to the first battery characteristic parameter and the battery remaining capacity of each sample battery.

[0065] In the embodiment of the present application, since the entropy weight method can determine the degree of discreteness of the battery characteristic parameters and weight the battery characteristic parameters, and has the characteristics of objective weighting, intuitive output results, and obvious quantization, compared with other methods, it has fewer steps, fast operation, and is easy to understand and implement. Therefore, the embodiment of the present application uses the entropy weight method to calculate the weights of the battery characteristic parameters. The entropy weight method determines the weights by the amount of information carried by the data. It is an objective comprehensive evaluation method and can obtain a more objective indicator weight calculation result. Entropy is a measure of the degree of disorder of the system. The smaller the entropy value of the battery characteristic parameter, the greater the degree of variation of the data corresponding to the battery characteristic parameter, the more information it contains, the greater the role of the battery characteristic parameter in the sorting of retired power batteries, and the greater the weight; vice versa. In the calculation process, according to the degree of variation of each battery characteristic parameter of the retired power battery, the information entropy is calculated to determine the weight of each battery characteristic parameter.

[0066] Specifically, first, assuming that there are m sample batteries and n battery characteristic parameters, the battery original data matrix R = (r ij ) m×n ;

[0067]

[0068] In the formula, r ij is the value of the j-th battery characteristic parameter in the ith sample battery.

[0069] Step b2, standardizing the battery original data matrix according to the battery characteristic parameters and battery characteristic parameter classification of the sample battery corresponding to each element in the battery original data matrix to obtain a battery standardized data matrix; the battery characteristic parameter classification includes positive parameters and negative parameters.

[0070] In the embodiment of the present application, since the measurement units of various battery characteristic parameters of retired power batteries are not unified, the battery original data matrix is ​​first standardized before calculating the weight. The standardization process includes dimensionless processing. The dimensionless processing can solve the problem that the units are inconsistent and cannot be combined for calculation. In addition, since the positive and negative parameter values ​​represent different meanings, battery characteristic parameters with large differences in order of magnitude need to be classified according to different algorithms for data standardization. Specifically, the battery characteristic parameter classification includes positive parameters and negative parameters.

[0071] Specifically, since the higher the forward parameter value, the better, the forward parameter normalization can be expressed as:

[0072]

[0073] Since the lower the negative parameter value, the better, the negative parameter normalization can be expressed as:

[0074]

[0075] Then we get the battery standardized data matrix R ′ =(r i ′ j ) m×n .

[0076] Step b3, determining the information entropy value of each battery characteristic parameter based on the battery standardized data matrix.

[0077] In this step, firstly, the battery standardized data matrix R ′ Calculate the proportion p of the jth battery characteristic parameter in the i-th sample battery ij ; As shown below:

[0078]

[0079] Then, the information entropy value e of the jth battery characteristic parameter is calculated according to the proportion of the jth battery characteristic parameter in each sample battery. j ; As shown below:

[0080] In the formula,

[0081] Step b4, determining the weight value of each battery characteristic parameter according to the information entropy value of each battery characteristic parameter, and determining the third number of battery characteristic parameters with the highest weight value as the sorting factor of the target batch of retired power batteries.

[0082] In this step, firstly, according to the information entropy value e of the jth battery characteristic parameter j Calculate the information utility value d of the jth battery characteristic parameter j =1-e j ; Then calculate the weight value ω of the jth battery characteristic parameter j ;

[0083] Specifically,

[0084] Through calculation and analysis, the battery characteristic parameters with lower weight values ​​are discarded, and the preset third number of battery characteristic parameters with the highest weight values ​​are retained to determine the sorting factors for the target batch of retired power batteries.

[0085] For example, the entropy weight method is used to perform weight calculation and analysis on 10 important battery characteristic parameters, including open circuit voltage, battery remaining capacity (actual capacity), ohmic internal resistance, voltage change, polarization performance (polarization voltage), energy density, self-discharge rate, charge and discharge rate, charging temperature rise, and discharging temperature rise, to obtain the weight value of each battery characteristic parameter, as shown in Table 1 below:

[0086] Table 1 Calculation results of battery characteristic parameter weights

[0087]

[0088] According to the results shown in Table 1, it can be seen that the ohmic internal resistance, battery remaining capacity and open circuit voltage of retired power batteries have the largest weights compared with other battery characteristic parameters. These three battery characteristic parameters will be selected as sorting factors for subsequent cluster analysis of retired power batteries.

[0089] Furthermore, in addition to the weight values ​​of the characteristic parameters of each battery, when determining the sorting factor, the difficulty of measuring and obtaining the characteristic parameters of the battery can also be considered to determine the sorting factor, as shown in Table 2 below:

[0090] Table 2 Statistics of characteristic parameter measurement and acquisition information of retired power batteries

[0091]

[0092] As shown in Table 2, the open circuit voltage, ohmic internal resistance, voltage change, energy density, self-discharge rate, charge and discharge rate, and charge and discharge temperature rise are relatively easy to measure; the open circuit voltage, ohmic internal resistance, voltage change, and energy density test time is relatively short; some of these parameters can also be read from the battery technical parameter table to reduce the time for laboratory measurements. The above indicators can be used as objective reference factors in actual engineering applications and should be comprehensively considered when selecting sorting factors.

[0093] Furthermore, S104 specifically includes:

[0094] Step c1, determining the number of clustering centers of the K-means clustering algorithm based on the sorting factors of each of the sample batteries.

[0095] In the embodiment of the present application, after starting K-means clustering, the battery characteristic parameter data corresponding to the sorting factor of each retired power battery in the target batch of retired power batteries is first obtained, and the battery characteristic parameter data corresponding to the sorting factor is preprocessed, such as dimensionless processing and other standardization processing. Then, the number of clustering centers of the K-means clustering algorithm is determined.

[0096] Specifically, in the K-means clustering algorithm, the number of cluster centers K is manually defined, and the selection and rationality of the K value are difficult to estimate. It is necessary to continuously adjust the classification to achieve the optimal effect in order to determine the appropriate number of cluster centers K. In the embodiment of the present application, the number of cluster centers of the K-means clustering algorithm is determined based on the sorting factors of each sample battery.

[0097] Step c2: randomly determining the initialization cluster center of each retired power battery classification according to the number of cluster centers.

[0098] In this step, the sorting factor data set based on the retired power batteries is randomly initialized according to the number of cluster centers, and any different groupings are attempted to be identified based on the initialized cluster centers. Each initialized cluster center is the center point of the cluster, and the initialized cluster center point of each retired power battery classification is randomly determined.

[0099] Step c3, assigning a retired power battery classification corresponding to the nearest initialization cluster center to each of the retired power batteries according to the sorting factor of each of the retired power batteries and the Euclidean distance between each of the initialization cluster centers.

[0100] In this step, according to the Euclidean distance between the sorting factor of each retired power battery and each initialized cluster center, each retired power battery is assigned a retired power battery classification corresponding to the nearest initialized cluster center. Specifically, the Euclidean distance formula can be expressed as:

[0101]

[0102] Among them, d(x i , μ j ) represents the data point x i and the initial cluster center point μ of retired power battery category j j The Euclidean distance between ik Represents data point x i The kth battery characteristic parameter; μ jk Represents the initial cluster center point μ of retired power battery category j j The kth battery characteristic parameter is obtained by using the above formula to calculate the Euclidean distance between the sorting factor and the initialization cluster center of any retired power battery, and find the nearest initialization cluster center, and classify the retired power battery into the retired power battery classification corresponding to the nearest initialization cluster center.

[0103] Step c4: updating the cluster center of each retired power battery classification according to the mean value of the sorting factor in each retired power battery classification.

[0104] In this step, the cluster center of each retired power battery classification is updated according to the mean value of the sorting factor in each retired power battery classification, as shown in the following formula:

[0105] Where μ′ j C is a new cluster center for retired power batteries; j is the set of all data points in retired power battery category j; |C j | represents the number of data points in retired power battery category j.

[0106] Step c5, repeatedly assigning the retired power battery classification corresponding to the nearest cluster center to each of the retired power batteries, and updating the cluster center of each of the retired power battery classifications, until the average error criterion function of each of the retired power battery classifications converges to a preset minimum value.

[0107] In this step, steps c3 and c4 are repeated until the average error criterion function E of each retired power battery classification converges to a preset minimum value. In the embodiment of the present application, in the process of sorting retired power batteries, the measurement standard of clustering results is to minimize the average error criterion function W, which can effectively ensure the reliability, scientificity and effectiveness of the algorithm operation. The method of continuous iterative updating is adopted until all cluster center points no longer change, and the final battery classification is obtained and output. Among them, each iterative calculation process is accompanied by a decrease in the average error criterion function E. When the average error criterion function E reaches a minimum value, it means that the clustering effect is better at this time.

[0108] Step c6, classifying the retired power batteries corresponding to the cluster centers allocated to each of the retired power batteries for the last time, and determining them as the sorting results of each of the retired power batteries.

[0109] In this step, the retired power batteries corresponding to the cluster centers last allocated to the retired power batteries are classified to determine the sorting results of the retired power batteries.

[0110] Furthermore, step c1 specifically includes:

[0111] Step d1, selecting a range according to the number of preset cluster centers, and performing K-means clustering on the sorting factors of each of the sample batteries according to different numbers of cluster centers.

[0112] In this step, by inputting different K values ​​in the range of 0 to 6, K-means clustering is performed on the sorting factors of each sample battery.

[0113] Step d2, respectively calculating the sum of squares of the sorting factors of the sample batteries clustered according to different numbers of cluster centers.

[0114] In this step, the sum of squared errors (SSE) of the sorting factors of all sample batteries clustered with different numbers of cluster centers are calculated respectively. The sum of squared errors SSE is used to represent the clustering error of the battery data sample and can characterize the quality of the clustering effect. The specific calculation results are shown in Table 3 below:

[0115] Table 3. The sum of squares of battery data errors under different numbers of cluster centers

[0116]

[0117] Step d3, according to the changing trend of the sum of squared errors corresponding to different numbers of cluster centers relative to the number of cluster centers and the elbow rule, the number of cluster centers corresponding to the elbow point of the changing trend is determined as the number of cluster centers of the K-means clustering algorithm.

[0118] Specifically, see Figure 3 , Figure 3 Schematic diagram of the variation trend curve of the sample average error and SSE relative to the number of cluster centers in the embodiment of the present application. Figure 3As shown in the figure, as the number of cluster centers K increases, the classification of the battery data set becomes more refined, the degree of aggregation of each cluster increases, and the error square sum SSE decreases. According to the Elbow Method, the elbow method is a technique for determining the optimal number of clusters in the K-means clustering algorithm. The core idea is to find the point where the SSE decreases fastest, that is, the elbow point in the figure, as the optimal K value by drawing a relationship graph between SSE and K value under different K values. Specifically, as Figure 3 As shown in the figure, when the actual number of cluster centers K is lower than the optimal number of clusters, increasing the K value causes a sudden increase in the degree of aggregation, and the error sum square SSE decreases significantly; when the actual number of cluster centers K is close to the optimal number of clusters, the effect of increasing the K value on the clustering results gradually decreases, and the error sum square SSE decreases slowly; when the actual number of cluster centers K is greater than the optimal number of clusters, the continued increase of the K value has little effect on the clustering results, and the error sum square SSE eventually tends to be flat. The general trend of the error sum square SSE and the number of cluster centers K obtained by actual measurement is basically consistent with the theoretical effect. Figure 3 It can be found that when K=4 is located at the elbow point, the clustering effect is optimal, that is, K=4 is the optimal number of clusters in the battery data, and the elbow point K=4 is determined as the number of cluster centers of the K-means clustering algorithm.

[0119] Furthermore, the method further comprises:

[0120] If after a preset number of repetitions of assigning the retired power battery classification corresponding to the nearest cluster center to each of the retired power batteries and updating the cluster centers of each of the retired power battery classifications, the average error criterion function of each of the retired power battery classifications has not converged, the initialization cluster center of each of the retired power battery classifications is randomly determined based on the number of cluster centers.

[0121] In this step, if the average error criterion function E of each retired power battery classification has not converged after repeating step c3 and step c4 for a preset number of times, it means that the initialization cluster center randomly selected in step c2 is improperly selected, causing the algorithm to fall into a local optimal solution and fail to converge to the global optimal solution. Therefore, the initialization cluster center of each retired power battery classification is randomly determined based on the number of cluster centers. Furthermore, you can choose to try to randomly initialize the cluster center points multiple times, and then select the sorting result with the best clustering effect for output.

[0122] An embodiment of the present application provides a method for sorting retired power batteries, comprising: obtaining a first battery characteristic parameter of each retired power battery in a target batch of retired power batteries; the first battery characteristic parameter is a plurality of preset battery characteristic parameters that affect the aging characteristics of the retired power batteries; based on the first battery characteristic parameter and a battery remaining capacity prediction model, predicting the battery remaining capacity of each retired power battery; the battery remaining capacity prediction model is a model constructed based on a support vector machine; based on the first battery characteristic parameter of each sample battery, the battery remaining capacity and an entropy weight method, determining a sorting factor of the target batch of retired power batteries; the sample battery is a preset second number of retired power batteries selected from the target batch of retired power batteries; the sorting factor is a preset third number of battery characteristic parameters with the highest weight value between the first battery characteristic parameter and the battery remaining capacity; based on the sorting factor and a K-means clustering algorithm, clustering the target batch of retired power batteries to obtain a sorting result of each retired power battery in the target batch of retired power batteries. In this way, according to the degree of dispersion of battery characteristic parameters in the target batch of retired power batteries, battery characteristic parameters with a larger degree of dispersion are selected as sorting factors for K-means clustering to sort the target batch of retired power batteries, thereby improving the accuracy of sorting retired power batteries.

[0123] Based on the same application concept, the embodiments of the present application also provide a device for sorting retired power batteries corresponding to the method for sorting retired power batteries provided in the above embodiments. Since the principle of solving the problem by the device in the embodiments of the present application is similar to the method for sorting retired power batteries in the above embodiments of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0124] See also Figure 4 , Figure 4 This is a functional module diagram of a retired power battery sorting device provided in an embodiment of the present application. As shown in Figure 4, the retired power battery sorting device 400 includes:

[0125] The acquisition module 410 is used to acquire the first battery characteristic parameter of each retired power battery in the target batch of retired power batteries; the first battery characteristic parameter is a plurality of preset battery characteristic parameters that affect the aging characteristics of the retired power batteries.

[0126] The prediction module 420 is used to predict the remaining battery capacity of each of the retired power batteries based on the first battery characteristic parameters and a battery remaining capacity prediction model; the battery remaining capacity prediction model is a model constructed based on a support vector machine.

[0127] The determination module 430 is used to determine the sorting factor of the target batch of retired power batteries based on the first battery characteristic parameter of each sample battery, the remaining battery capacity and the entropy weight method; the sample batteries are a preset second number of retired power batteries selected from the target batch of retired power batteries; the sorting factor is a preset third number of battery characteristic parameters with the highest weight value among the first battery characteristic parameter and the remaining battery capacity.

[0128] The sorting module 440 is used to cluster the target batch of retired power batteries based on the sorting factors and the K-means clustering algorithm to obtain a sorting result of each retired power battery in the target batch of retired power batteries.

[0129] Furthermore, the prediction module 420 is used to construct the battery remaining capacity prediction model according to the following steps:

[0130] Selecting a preset fourth number and a preset fifth number of retired power batteries from the target batch of retired power batteries as training set sample batteries and test set sample batteries respectively;

[0131] The training set sample batteries are charged and discharged cycled, and the first battery characteristic parameter and the measured value of the battery remaining capacity of the training set sample batteries at different cycle times are used as input variables and output variables respectively, and training is performed using a support vector machine algorithm to obtain an initial prediction model;

[0132] Performing charge and discharge cycles on the test set sample battery, inputting the first battery characteristic parameters of the test set sample battery at different cycle times into the initial prediction model, and outputting the battery remaining capacity prediction value of the test set sample battery at different cycle times;

[0133] If the curve of the battery remaining capacity prediction value and the number of charge and discharge cycles is similar to the curve of the battery remaining capacity measured value and the number of charge and discharge cycles of the test set sample battery, the initial prediction model is determined as the battery remaining capacity prediction model.

[0134] Furthermore, the prediction module 420 is also used for:

[0135] If the curve of the predicted value of the remaining battery capacity and the number of charge and discharge cycles deviates from the curve of the measured value of the remaining battery capacity and the number of charge and discharge cycles of the test set sample batteries, the training set sample batteries and the test set sample batteries are re-selected from the target batch of retired power batteries to construct the battery remaining capacity prediction model.

[0136] Further, when the determination module 430 is used to determine the sorting factor of the target batch of retired power batteries based on the first battery characteristic parameter of each sample battery, the battery remaining capacity and the entropy weight method, the determination module 430 is specifically used to:

[0137] Constructing a battery raw data matrix according to the first battery characteristic parameter and the battery remaining capacity of each sample battery;

[0138] According to the battery characteristic parameters and battery characteristic parameter classification of the sample battery corresponding to each element in the battery original data matrix, the battery original data matrix is ​​standardized to obtain a battery standardized data matrix; the battery characteristic parameter classification includes positive parameters and negative parameters;

[0139] Determining information entropy values ​​of characteristic parameters of each battery according to the battery standardized data matrix;

[0140] According to the information entropy value of each battery characteristic parameter, the weight value of each battery characteristic parameter is determined, and a preset third number of battery characteristic parameters with the highest weight value are determined as the sorting factors of the target batch of retired power batteries.

[0141] Further, when the sorting module 440 is used to cluster the target batch of retired power batteries based on the sorting factor and the K-means clustering algorithm to obtain the sorting result of each retired power battery in the target batch of retired power batteries, the sorting module 440 is specifically used to:

[0142] Determining the number of cluster centers of the K-means clustering algorithm based on the sorting factor of each of the sample batteries;

[0143] According to the number of cluster centers, randomly determining the initialization cluster center of each retired power battery classification;

[0144] According to the Euclidean distance between the sorting factor of each of the retired power batteries and each of the initialized cluster centers, assigning a retired power battery classification corresponding to the nearest initialized cluster center to each of the retired power batteries;

[0145] updating the cluster center of each retired power battery classification according to the mean value of the sorting factor in each retired power battery classification;

[0146] Repeatingly assigning a retired power battery classification corresponding to the nearest cluster center to each of the retired power batteries, and updating the cluster center of each of the retired power battery classifications, until the average error criterion function of each of the retired power battery classifications converges to a preset minimum value;

[0147] The retired power batteries corresponding to the cluster centers allocated to the retired power batteries for the last time are classified to determine the sorting results of the retired power batteries.

[0148] Furthermore, when the sorting module 440 is used to determine the number of clustering centers of the K-means clustering algorithm based on the sorting factors of each of the sample batteries, the sorting module 440 is specifically used to:

[0149] According to a preset number of cluster centers, a K-means clustering is performed on the sorting factors of each sample battery according to different numbers of cluster centers;

[0150] Calculating the sum of squares of the sorting factors of each of the sample batteries clustered according to different numbers of cluster centers;

[0151] According to the changing trend of the sum of squared errors corresponding to different numbers of cluster centers relative to the number of cluster centers and the elbow rule, the number of cluster centers corresponding to the elbow point of the changing trend is determined as the number of cluster centers of the K-means clustering algorithm.

[0152] Furthermore, the sorting module 440 is also used for:

[0153] If after a preset number of repetitions of assigning the retired power battery classification corresponding to the nearest cluster center to each of the retired power batteries and updating the cluster centers of each of the retired power battery classifications, the average error criterion function of each of the retired power battery classifications has not converged, the initialization cluster center of each of the retired power battery classifications is randomly determined based on the number of cluster centers.

[0154] The embodiment of the present application provides a sorting device for retired power batteries, comprising: an acquisition module, used to acquire a first battery characteristic parameter of each retired power battery in a target batch of retired power batteries; the first battery characteristic parameter is a plurality of preset battery characteristic parameters that affect the aging characteristics of the retired power batteries; a prediction module, used to predict the battery remaining capacity of each retired power battery based on the first battery characteristic parameter and a battery remaining capacity prediction model; the battery remaining capacity prediction model is a model constructed based on a support vector machine; a determination module, used to determine a sorting factor of the target batch of retired power batteries based on the first battery characteristic parameter, the battery remaining capacity and an entropy weight method of each sample battery; the sample battery is a preset second number of retired power batteries selected from the target batch of retired power batteries; the sorting factor is a preset third number of battery characteristic parameters with the highest weight value between the first battery characteristic parameter and the battery remaining capacity; a sorting module, used to cluster the target batch of retired power batteries based on the sorting factor and a K-means clustering algorithm, to obtain a sorting result of each retired power battery in the target batch of retired power batteries. In this way, according to the degree of dispersion of battery characteristic parameters in the target batch of retired power batteries, battery characteristic parameters with a larger degree of dispersion are selected as sorting factors for K-means clustering to sort the target batch of retired power batteries, thereby improving the accuracy of sorting retired power batteries.

[0155] Based on the same application idea, please refer to Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 5 As shown, the electronic device 500 includes: a processor 510 , a memory 520 and a bus 530 .

[0156] The memory 520 stores machine-readable instructions executable by the processor 510. When the electronic device 500 is running, the processor 510 communicates with the memory 520 through the bus 530. When the processor 510 is running, the machine-readable instructions execute the steps of the method for sorting retired power batteries provided in the above embodiment. The specific implementation method can be found in the method embodiment, which will not be repeated here.

[0157] Based on the same application concept, an embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for sorting retired power batteries provided in the above embodiment are executed. The specific implementation method can be found in the method embodiment, which will not be repeated here.

[0158] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0159] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0160] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0161] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0162] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0163] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0164] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for sorting retired power batteries, characterized in that: The method comprises: Acquire a first battery characteristic parameter of each retired power battery in a target batch of retired power batteries; the first battery characteristic parameter is a plurality of preset battery characteristic parameters that affect aging characteristics of the retired power batteries; Based on the first battery characteristic parameter and a battery remaining capacity prediction model, predicting the battery remaining capacity of each of the retired power batteries; the battery remaining capacity prediction model is a model constructed based on a support vector machine; Determine the sorting factor of the target batch of retired power batteries based on the first battery characteristic parameter, the remaining battery capacity and the entropy weight method of each sample battery; the sample battery is a preset second number of retired power batteries selected from the target batch of retired power batteries; the sorting factor is a preset third number of battery characteristic parameters with the highest weight value between the first battery characteristic parameter and the remaining battery capacity; Based on the sorting factors and the K-means clustering algorithm, the target batch of retired power batteries is clustered to obtain a sorting result of each retired power battery in the target batch of retired power batteries.

2. The method for sorting retired power batteries according to claim 1, characterized in that: The battery remaining capacity prediction model is constructed according to the following steps: Selecting a preset fourth number and a preset fifth number of retired power batteries from the target batch of retired power batteries as training set sample batteries and test set sample batteries respectively; The training set sample batteries are charged and discharged cycled, and the first battery characteristic parameter and the measured value of the battery remaining capacity of the training set sample batteries at different cycle times are used as input variables and output variables respectively, and training is performed using a support vector machine algorithm to obtain an initial prediction model; Performing charge and discharge cycles on the test set sample battery, inputting the first battery characteristic parameters of the test set sample battery at different cycle times into the initial prediction model, and outputting the battery remaining capacity prediction value of the test set sample battery at different cycle times; If the curve of the battery remaining capacity prediction value and the number of charge and discharge cycles is similar to the curve of the battery remaining capacity measured value and the number of charge and discharge cycles of the test set sample battery, the initial prediction model is determined as the battery remaining capacity prediction model.

3. The method for sorting retired power batteries according to claim 2, characterized in that: The method further comprises: If the curve of the predicted value of the remaining battery capacity and the number of charge and discharge cycles deviates from the curve of the measured value of the remaining battery capacity and the number of charge and discharge cycles of the test set sample batteries, the training set sample batteries and the test set sample batteries are re-selected from the target batch of retired power batteries to construct the battery remaining capacity prediction model.

4. The method for sorting retired power batteries according to claim 1, characterized in that: The determining of the sorting factors of the target batch of retired power batteries based on the first battery characteristic parameters of each sample battery, the remaining battery capacity and the entropy weight method includes: Constructing a battery raw data matrix according to the first battery characteristic parameter and the battery remaining capacity of each sample battery; According to the battery characteristic parameters and battery characteristic parameter classification of the sample battery corresponding to each element in the battery original data matrix, the battery original data matrix is ​​standardized to obtain a battery standardized data matrix; the battery characteristic parameter classification includes positive parameters and negative parameters; Determining information entropy values ​​of characteristic parameters of each battery according to the battery standardized data matrix; According to the information entropy value of each battery characteristic parameter, the weight value of each battery characteristic parameter is determined, and a preset third number of battery characteristic parameters with the highest weight value are determined as the sorting factors of the target batch of retired power batteries.

5. The method for sorting retired power batteries according to claim 1, characterized in that: The clustering of the target batch of retired power batteries based on the sorting factor and the K-means clustering algorithm to obtain the sorting results of each retired power battery in the target batch of retired power batteries includes: Determining the number of cluster centers of the K-means clustering algorithm based on the sorting factor of each of the sample batteries; According to the number of cluster centers, randomly determining the initialization cluster center of each retired power battery classification; According to the Euclidean distance between the sorting factor of each of the retired power batteries and each of the initialized cluster centers, assigning a retired power battery classification corresponding to the nearest initialized cluster center to each of the retired power batteries; updating the cluster center of each retired power battery classification according to the mean value of the sorting factor in each retired power battery classification; Repeatingly assigning a retired power battery classification corresponding to the nearest cluster center to each of the retired power batteries, and updating the cluster center of each of the retired power battery classifications, until the average error criterion function of each of the retired power battery classifications converges to a preset minimum value; The retired power batteries corresponding to the cluster centers allocated to the retired power batteries for the last time are classified to determine the sorting results of the retired power batteries.

6. The method for sorting retired power batteries according to claim 5, characterized in that: The step of determining the number of clustering centers of the K-means clustering algorithm based on the sorting factors of the sample batteries comprises: According to a preset number of cluster centers, a K-means clustering is performed on the sorting factors of each sample battery according to different numbers of cluster centers; Calculating the sum of squares of the sorting factors of each of the sample batteries clustered according to different numbers of cluster centers; According to the changing trend of the sum of squared errors corresponding to different numbers of cluster centers relative to the number of cluster centers and the elbow rule, the number of cluster centers corresponding to the elbow point of the changing trend is determined as the number of cluster centers of the K-means clustering algorithm.

7. The method for sorting retired power batteries according to claim 5, characterized in that: The method further comprises: If after a preset number of repetitions of assigning the retired power battery classification corresponding to the nearest cluster center to each of the retired power batteries and updating the cluster centers of each of the retired power battery classifications, the average error criterion function of each of the retired power battery classifications has not converged, the initialization cluster center of each of the retired power battery classifications is randomly determined based on the number of cluster centers.

8. A sorting device for retired power batteries, characterized in that: The sorting device for retired power batteries comprises: An acquisition module, used for acquiring a first battery characteristic parameter of each retired power battery in a target batch of retired power batteries; the first battery characteristic parameter is a plurality of preset battery characteristic parameters that affect the aging characteristics of the retired power batteries; A prediction module, configured to predict the remaining battery capacity of each of the retired power batteries based on the first battery characteristic parameters and a battery remaining capacity prediction model; the battery remaining capacity prediction model is a model constructed based on a support vector machine; A determination module, configured to determine a sorting factor of the target batch of retired power batteries based on a first battery characteristic parameter of each sample battery, a battery remaining capacity, and an entropy weight method; the sample battery is a preset second number of retired power batteries selected from the target batch of retired power batteries; the sorting factor is a preset third number of battery characteristic parameters having the highest weight value among the first battery characteristic parameter and the battery remaining capacity; The sorting module is used to cluster the target batch of retired power batteries based on the sorting factor and the K-means clustering algorithm to obtain a sorting result of each retired power battery in the target batch of retired power batteries.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate through the bus. When the machine-readable instructions are run by the processor, the steps of the method for sorting retired power batteries as described in any one of claims 1 to 7 are executed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for sorting retired power batteries as described in any one of claims 1 to 7 are executed.

Images