Battery Pack Consistency Evaluation and Rating Method and Device

Abstract

The present invention discloses a method and device for evaluating the consistency and grading of battery packs. The method includes: obtaining the original battery characteristics of multiple battery packs to be tested; performing data cleaning on the original battery characteristics, extracting features from the original battery characteristics after data cleaning, and using the extracted features to construct a feature sequence corresponding to each battery pack to be tested; using the feature sequence of each battery pack to be tested as a detection sample, and calculating the dominance degree of each detection sample according to the feature sequence; arranging all detection samples in descending or ascending order according to the dominance degree, and dividing all detection samples into multiple grades according to a predetermined ratio; wherein, the worse the consistency of the battery pack to be tested corresponding to the detection sample with a larger dominance degree. Constructing multiple feature sequences to enrich the feature space to improve the accuracy of consistency evaluation, without the need to solve the index weights and provide grading evaluation thresholds, and improving the data processing efficiency.

Description

Technical Field

[0001] The present invention relates to the field of performance evaluation of energy storage battery packs, and particularly to a method and device for evaluating the consistency and grading of battery packs. Background Art

[0002] In recent years, the proportion of renewable power sources such as wind power and photovoltaic power in the power system has gradually increased. At the same time, the demand for energy storage applications in all links of the power system has gradually increased, and energy storage technology has developed rapidly. Among various energy storage technology routes, electrochemical energy storage has developed relatively fast. As the most common energy storage method, lithium-ion batteries have a very wide range of application fields due to their advantages such as longer life cycle, faster response speed, lower self-discharge rate, and higher energy conversion efficiency. Lithium-ion battery energy storage has become one of the most important energy storage technology routes.

[0003] Due to the continuous increase in the demand for energy storage scale and the limitation of the capacity of a single battery cell, in energy storage application scenarios, multiple single battery cells usually need to be designed with a certain series-parallel topology structure to form battery packs at different levels, such as battery modules, battery packs, and battery clusters, etc., to meet the actual required usage scale and performance parameter requirements. However, due to factors such as the inability to completely unify the battery processing and manufacturing processes and the inability to ensure exactly the same battery usage environment, the single battery cells in the battery pack will show certain differences during use, that is, there is a certain degree of inconsistency. And with the continuous increase in the scale of energy storage power stations, the number of integrated single battery cells increases sharply, and the consistency differences between the batteries are more prominent. The consistency differences of the battery pack will lead to a significant decline in the overall service life and performance of the battery pack, which is a key issue affecting the performance of the battery pack and even the power station.

[0004] Evaluating the overall inconsistency performance of the battery pack and making a comprehensive grading evaluation of the overall health state of the battery pack based on the inconsistency is an important means for the operation optimization of energy storage power stations. In the prior art, the calculation of battery pack consistency is divided into methods such as single static parameter, multi-static parameter, and dynamic parameter. Among them, the single static parameter method uses a single static parameter to evaluate the consistency of the battery pack, and it is necessary to obtain the evaluation threshold in advance according to professional knowledge or production experience. The implementation process is not flexible enough and not comprehensive and accurate enough; the multi-static parameter method uses multiple static parameter characteristics within the same time to evaluate the consistency. It often needs to calculate the weights of multiple indicators based on various methods such as subjective weighting and objective weighting, and design a comprehensive evaluation function. The implementation process is relatively complex; the dynamic parameter method makes full use of the dynamic charge and discharge characteristics of the battery, including more comprehensive battery state information. However, it often needs to design test conditions and obtain specific charge and discharge curves, and still needs to design and extract one or more consistency evaluation indicators and design relevant thresholds and weights.

[0005] In view of the problems in the prior art that the evaluation means are not flexible and accurate enough and the evaluation means are relatively complex, there is currently no effective solution. Summary of the Invention

[0006] The embodiments of the present invention provide a method and device for evaluating the consistency and grading of battery packs, which uniformly extract the original parameters of the same type of batteries from multiple battery packs, characterize the battery packs more comprehensively, and do not depend on specific experimental conditions and laboratory test environments, so as to solve the problems of inflexible and inaccurate evaluation means; by calculating the degree of domination and then performing consistency and grading evaluation, it is not necessary to determine the threshold or solve the index weight in real time according to different batches of battery packs, so as to solve the problem of complex evaluation means.

[0007] To achieve the above object, the embodiments of the present invention provide a method for evaluating the consistency and grading of battery packs, including: obtaining the original battery characteristics of multiple battery packs to be tested; performing data cleaning on the original battery characteristics, extracting features from the original battery characteristics after data cleaning, and using the extracted features to construct a feature sequence corresponding to the battery packs to be tested; using the feature sequence of each battery pack to be tested as a detection sample, and calculating the degree of domination of each detection sample according to the feature sequence; arranging all detection samples in descending or ascending order according to the degree of domination, and dividing all detection samples into multiple grades according to a predetermined ratio; wherein, the worse the consistency of the battery pack to be tested corresponding to the detection sample with a larger degree of domination.

[0008] Further optionally, the extracting features from the original battery characteristics after data cleaning and using the extracted features to construct a feature sequence corresponding to the battery packs to be tested includes: calculating corresponding static feature indexes and dynamic feature indexes according to the original battery characteristics of each battery pack to be tested; using the static characteristic indexes and the dynamic characteristic indexes as items in the feature sequence corresponding to the battery packs to be tested, and constructing the corresponding feature sequence for each battery pack to be tested.

[0009] Further optionally, calculating the degree of domination of each detection sample according to the feature sequence includes: comparing the corresponding items of the feature sequence of any detection sample with the feature sequences of the remaining all detection samples; if all items in the feature sequence of the current detection sample are not greater than the corresponding items in the feature sequence of the comparison detection sample, and at least one item in the feature sequence of the current detection sample is less than the corresponding item in the feature sequence of the comparison detection sample, it is considered that the comparison detection sample is dominated by the current detection sample, and the degree of domination of the comparison detection sample is incremented by 1.

[0010] Further optionally, the static characteristic indexes include: voltage range, temperature range, voltage standard deviation coefficient, and temperature standard deviation coefficient; the dynamic characteristic indexes include: characteristic parameters representing the discharge ohmic internal resistance, characteristic parameters representing the dynamic polarization characteristic, characteristic parameters representing the ohmic internal resistance, and parameters representing the overall discrete characteristic of the voltage or temperature curve during the charge and discharge process.

[0011] Further optionally, the data cleaning of the original battery characteristics includes: for any target parameter value in the original battery parameters, if it exceeds the reasonable threshold of the parameter or is a missing value, then use the parameter value at the previous moment or the parameter value at the next moment of the target parameter value to replace the target parameter value; or, calculate the average value of the parameter values in a preset period through a sliding window, and use the average value as the target parameter value.

[0012] On the other hand, an embodiment of the present invention further provides a battery pack consistency evaluation and grading device, including: a data acquisition module for acquiring the original battery characteristics of a plurality of battery packs to be tested; a feature sequence construction module for performing data cleaning on the original battery characteristics, extracting features from the data-cleaned original battery characteristics, and using the extracted features to construct a feature sequence corresponding to each battery pack to be tested; a dominance calculation module for using the feature sequence of each battery pack to be tested as a test sample and calculating the dominance of each test sample according to the feature sequence; an evaluation module for arranging all test samples in descending or ascending order according to the dominance and dividing all test samples into multiple grades according to a predetermined ratio; wherein, the greater the dominance of a test sample, the worse the consistency of the corresponding battery pack to be tested.

[0013] Further optionally, the feature sequence construction module includes: a feature extraction sub-module for calculating corresponding static characteristic indexes and dynamic characteristic indexes according to the original battery characteristics of each battery pack to be tested; a data connection sub-module for using the static characteristic indexes and the dynamic characteristic indexes as items in the feature sequence corresponding to each battery pack to be tested to construct the feature sequence corresponding to each battery pack to be tested.

[0014] Further optionally, the dominance calculation module includes: an item comparison sub-module for comparing the corresponding items of the feature sequence of any test sample with the feature sequences of all the other test samples; a dominance determination sub-module for, if all items in the feature sequence of the current test sample are not greater than the corresponding items in the feature sequence of the comparison test sample, and at least one item in the feature sequence of the current test sample is less than the corresponding item in the feature sequence of the comparison test sample, then it is considered that the comparison test sample is dominated by the current test sample, and the dominance of the comparison test sample is incremented by 1.

[0015] Further optionally, the static characteristic indexes include: voltage range, temperature range, voltage standard deviation coefficient, and temperature standard deviation coefficient; the dynamic characteristic indexes include: characteristic parameters representing the discharge ohmic internal resistance, characteristic parameters representing the dynamic polarization characteristics, characteristic parameters representing the ohmic internal resistance, and parameters representing the overall discrete characteristics of the voltage or temperature curve during the charge and discharge process.

[0016] Further optionally, the feature sequence construction module includes: a first correction sub-module, which is used for any target parameter value in the original battery parameters. If it exceeds the parameter reasonable threshold or is a missing value, the parameter value at the previous moment or the next moment of the target parameter value is used to replace the target parameter value; a second correction sub-module, which is used to calculate the average value of the parameter values in a preset period through a sliding window, and use the average value as the target parameter value.

[0017] Meanwhile, the present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above method is implemented.

[0018] Meanwhile, the present invention also provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program for executing the above method.

[0019] The above technical solutions have the following beneficial effects:

[0020] 1. Extract the original battery parameters to construct a feature sequence, more comprehensively characterize the consistency of the battery pack, design reasonable consistency evaluation indexes, and avoid relying on specific experimental conditions and laboratory test environments;

[0021] 2. Propose a grade evaluation method under multiple indexes. Different from the existing grade evaluation methods, it does not require weight calculation and feature fusion of multiple indexes, nor does it require determining grade thresholds based on expert knowledge and operation experience. Therefore, it is not targeted at batteries of specific manufacturers or models, and has flexibility and universality;

[0022] 3. Based on the Pareto optimal idea, through the calculation of the dominance degree, the multi-dimensional indexes are reduced to one dimension for evaluation. The dimension reduction method is objective, effective, does not lose the original information, and does not require solving the index weights. It can be carried out at any use stage of the battery pack, does not require designing and adjusting weights, and has better flexibility and promotion;

[0023] 4. Divide the battery pack into multiple grades according to the degree of being dominated according to the actual situation, and conduct grade evaluation, which is simple and feasible and does not require providing equivalent evaluation thresholds.

[0024] To make the above and other purposes, features and advantages of the present invention more obvious and understandable, the following provides preferred embodiments and, in conjunction with the accompanying drawings, detailed descriptions are as follows. Description of the Drawings

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0026] Figure 1 It is a flowchart of the battery pack consistency evaluation and grading method provided in this embodiment;

[0027] Figure 2 It is a flowchart of the feature sequence construction method provided in this embodiment;

[0028] Figure 3 It is a flowchart of the domination degree calculation method provided in this embodiment;

[0029] Figure 4 It is a flowchart of the data cleaning method provided in this embodiment;

[0030] Figure 5 It is a structural schematic diagram of the battery pack consistency evaluation and grading device provided in this embodiment;

[0031] Figure 6 It is a structural schematic diagram of the feature sequence construction module provided in this embodiment;

[0032] Figure 7 It is a structural schematic diagram of the domination degree calculation module provided in this embodiment;

[0033] Figure 8 It is another structural schematic diagram of the feature sequence construction module provided in this embodiment. Detailed Embodiments

[0034] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0035] The single static parameter method in the prior art is as follows:

[0036] Step 1: Collect the original battery features;

[0037] Step 2: Data cleaning;

[0038] Step 3: Extracting single indicators of battery pack consistency;

[0039] Step 4: Specifying the grading evaluation threshold to complete the grading evaluation of battery pack consistency.

[0040] The multi-static parameter method in the prior art is as follows:

[0041] Steps 1 to 3 are the same as above;

[0042] Step 4: Normalize multiple single indicators, and the conversion formula is as follows:

[0043] x′(i,j) = [x max (j) - x * (i,j)] / [x max (j) - x min (j)]

[0044] where x max (j0, x min (j) are the maximum and minimum values of the extracted indicators respectively; x * (i,j), x′(i,j0 are the indicators of the battery pack before and after data processing respectively.

[0045] Step 5: Weight design. Assign certain weights to each indicator according to the subjective weighting method, such as expert knowledge; or according to the objective weighting method, use principal component analysis or entropy weight method, etc., and obtain certain weight values through algorithm calculation.

[0046] Step 6: Feature fusion. Weight multiple indicators into a comprehensive indicator according to the designed weights.

[0047] Step 7: Set the grading evaluation threshold, similar to Step 4 in the single-static parameter method.

[0048] The dynamic parameter method in the prior art is as follows:

[0049] Step 1: Design test conditions, such as constant current - constant voltage charge and discharge. And conduct charge and discharge experiments.

[0050] Step 2: Extract characteristic parameters, such as extracting characteristic parameters representing discharge ohmic resistance, characteristic parameters representing dynamic polarization characteristics, characteristic parameters representing ohmic resistance, characteristic parameters representing battery capacity and deterioration degree, and characteristic parameters of charge and discharge characteristics in the constant current or constant voltage stage, etc. from the charge and discharge curves.

[0051] Steps 3 - 6 are the same as Steps 4 - 7 in the multi-static parameter method.

[0052] The prior art has the following disadvantages:

[0053] 1. It is necessary to provide the grading evaluation threshold according to the battery manufacturer or long-term operation experience, and the implementation process is inconvenient.

[0054] 2. It is necessary to set different grading evaluation threshold intervals for different battery types, battery models, and even different usage scenarios, and the real-time process is not flexible.

[0055] 3. When it comes to the grading evaluation of multiple indicators, it is necessary to design weights and fuse multiple indicators. The weight design process is not objective enough or accurate enough.

[0056] 4. It is necessary to design specific test conditions or experimental environments, and the test process is not convenient.

[0057] 5. It is impossible to globally and dynamically adjust the weights of multiple indicators according to the usage status of the battery over time, and it does not have self-adaptability.

[0058] To solve the problems of the existing technology, an embodiment of the present invention provides a method for evaluating the consistency and grading of a battery pack. Figure 1 For the flowchart of the method for evaluating the consistency and grading of the battery pack provided in this embodiment, as Figure 1 shown, the method includes:

[0059] S1. Obtain the original battery characteristics of multiple battery packs to be tested;

[0060] Measure the original battery parameters of the battery packs to be tested. Each item in the original battery parameters can be used as a feature value in the subsequent feature sequence. Therefore, in order to enrich the feature space and improve the final evaluation accuracy, as many parameters in the original battery parameters as possible can be measured.

[0061] As an optional implementation method, collect the time (timestamp), battery voltage (V), battery current (I), battery temperature (t), etc. of each battery pack to be tested.

[0062] S2. Clean the original battery characteristics, extract features from the original battery characteristics after data cleaning, and use the extracted features to construct a feature sequence corresponding to the battery packs to be tested;

[0063] In order to avoid the influence of incorrect data on the evaluation result, it is necessary to clean the data in the original battery parameters to replace the incorrect data or redundant data therein, and improve the reliability of the evaluation result.

[0064] Construct a feature sequence x using the feature values of each item in the original battery parameters after cleaning i =[x i,1 ,x i,2 ,…,x i,j, where x i,j is the j-th consistency index of the i-th battery pack extracted, and each battery pack to be tested corresponds to a feature sequence, which represents the overall consistency degree of the corresponding battery pack.

[0065] S3. Using the feature sequence of each battery pack to be tested as a detection sample, calculate the dominance degree of each detection sample according to the feature sequence;

[0066] Take the feature sequence as the consistency index sequence of the corresponding battery pack to be tested, and the consistency index sequences of all battery packs to be tested constitute the battery pack consistency sample set X. And each feature sequence is used as a detection sample, and the dominance degree of each detection sample needs to be calculated, that is, the number of samples that dominate the current sample.

[0067] S4. Sort all detection samples in descending or ascending order according to the dominance degree, and divide all detection samples into multiple levels according to a predetermined ratio; among them, the worse the consistency of the battery pack to be tested corresponding to the detection sample with a larger dominance degree.

[0068] Reduce the multi-dimensional features of the consistency of all battery packs to one dimension and evaluate the consistency of the battery packs, that is, sort all detection samples from large to small or from small to large according to the dominance degree. The smaller the dominance degree of the battery pack, the better the consistency; on the contrary, the larger the dominance degree of the battery pack, the worse the consistency.

[0069] In addition, it is also necessary to evaluate the level of the battery pack consistency. For the scenario where the battery pack consistency needs to be divided into L levels, after sorting the battery packs according to the size of the dominance degree, they can be divided into L levels in turn according to a predetermined ratio. It should be noted that during the division process, it is necessary to ensure that the battery packs with the same dominance degree should be divided into the same level.

[0070] For example, if it is necessary to divide into 3 levels, all battery packs can be sorted from small to large according to the dominance degree, and then all battery packs can be divided into 3 levels according to a predetermined ratio of 3:3:4.

[0071] As an alternative implementation manner, Figure 2 is the flowchart of the feature sequence construction method provided in this embodiment. As Figure 2 shown, perform feature extraction on the original battery features after data cleaning, and use the extracted features to construct the feature sequence of the corresponding battery pack to be tested, including:

[0072] S201. Calculate the corresponding static feature index and dynamic feature index according to the original battery features of each battery pack to be tested;

[0073] S202. Take the static characteristic indexes and dynamic characteristic indexes as items in the characteristic sequence corresponding to the battery pack to be tested, and construct the characteristic sequence corresponding to each battery pack to be tested.

[0074] Extract characteristics according to the original battery characteristics to obtain a variety of characteristic indexes, including: static characteristic indexes and dynamic characteristic indexes.

[0075] Each static characteristic index and dynamic characteristic index can be used as a parameter in the characteristic sequence. A variety of static characteristic indexes and a variety of dynamic characteristic indexes can be randomly arranged in the characteristic sequence, or arranged in the order of static characteristic index - dynamic parameter index or dynamic parameter index - static parameter index. It should be noted that different battery packs to be tested should adopt a unified arrangement standard so that the corresponding items of all battery packs to be tested are of the same type of characteristic index.

[0076] As an optional implementation method, the static characteristic indexes include: voltage range, temperature range, voltage standard deviation coefficient, temperature standard deviation coefficient;

[0077] The voltage range is calculated by the following formula:

[0078] ΔU max =U max -U min

[0079] Where, ΔU max is the voltage range; U max , U min are respectively the maximum value and minimum value of the single-cell voltage in a group of batteries.

[0080] The temperature range is calculated by the following formula:

[0081] ΔT max =T max -T min

[0082] Where, ΔT max is the temperature range, T max , T min are respectively the maximum value and minimum value of the single-cell temperature in a group of batteries.

[0083] The voltage standard deviation coefficient is calculated by the following formula:

[0084]

[0085] Where, δ U is the voltage standard deviation coefficient, N is the total number of battery monomers in a group of batteries; U i is the voltage of the i-th battery monomer in a group of batteries; U mis the average voltage of all battery cells in a group of batteries.

[0086] The temperature standard deviation coefficient is calculated by the following formula:

[0087]

[0088] where δ T is the temperature standard deviation coefficient, and T i is the temperature of the i-th battery cell in a group of batteries; T m is the average temperature of all battery cells in a group of batteries.

[0089] The dynamic characteristic indexes include: characteristic parameters representing the discharge ohmic resistance, characteristic parameters representing the dynamic polarization characteristics, characteristic parameters representing the ohmic resistance, and parameters representing the overall discrete characteristics of the voltage or temperature curve during the charge and discharge process.

[0090] As an optional implementation manner, Figure 3 is the flowchart of the domination degree calculation method provided in this embodiment. As Figure 3 shown, calculating the domination degree of each detection sample according to the feature sequence includes:

[0091] S301. Compare the corresponding items of the feature sequence of any detection sample with those of all the other detection samples;

[0092] S302. If all items in the feature sequence of the current detection sample are not greater than the corresponding items in the feature sequence of the comparison detection sample, and at least one item in the feature sequence of the current detection sample is less than the corresponding item in the feature sequence of the comparison detection sample, it is considered that the comparison detection sample is dominated by the current detection sample, and the domination degree of the comparison detection sample is incremented by 1.

[0093] When calculating the domination degree of each detection sample, all detection samples in the sample set need to be compared with the remaining detection samples one by one. If samples p and q simultaneously meet the following two conditions:

[0094] (1) All indicators x p,i of sample p are not greater than the corresponding indicators x q,i of sample q.

[0095] (2) Sample p has at least one indicator x p,j less than the corresponding indicator x q,j of sample q.

[0096] A more rigorous statement is:

[0097] For the minimization of multi-objective problems, a vector composed of n objective components f i (1,..., n) Arbitrarily given two decision variables

[0098] if and only if, for all have then dominates

[0099] if and only if, for there is and there is at least one j ∈ {1,..., n} such that then weakly dominates

[0100] then it can be said that sample p is superior to sample q (in this embodiment, it is stipulated that the smaller the value, the better the index), denoted as p dominates q or q is dominated by p, that is:

[0101] p < q

[0102] Calculate the number of times each sample is dominated by different samples in the sample set to obtain the degree of being dominated.

[0103] As an alternative implementation manner, Figure 4 is the flowchart of the data cleaning method provided in this embodiment. As Figure 4 shown, the data cleaning of the original battery characteristics includes:

[0104] S203. For any target parameter value in the original battery parameters, if it exceeds the reasonable parameter threshold or is a missing value, then replace the target parameter value with the parameter value at the previous moment or the parameter value at the next moment of the target parameter value;

[0105] Or, S204. Calculate the average value of the parameter values in the preset time period through a sliding window, and use the average value as the target parameter value.

[0106] To ensure the effectiveness of the data and reduce the influence of incorrect data on the judgment result, in this embodiment, for a certain value in the original battery parameters that is a missing value or an unreasonable value significantly exceeding the threshold range, the value at the previous moment of this value, or the value at the next moment, or the average value within a period of time calculated through a sliding window is used for assignment replacement to ensure that each item in the characteristic sequence of each battery group to be tested is accurate and reliable.

[0107] The embodiment of the present invention also provides a battery pack consistency evaluation and grading device. Figure 5 is the structural schematic diagram of the battery pack consistency evaluation and grading device provided in this embodiment. As Figure 5 shown, the device includes:

[0108] A data acquisition module 100, configured to acquire the original battery characteristics of a plurality of battery packs to be tested;

[0109] Measure the original battery parameters of the battery pack to be measured. Each item in the original battery parameters can be used as a feature value in the subsequent feature sequence. Therefore, in order to enrich the feature space and improve the final evaluation accuracy, as many parameters in the original battery parameters as possible should be measured.

[0110] As an alternative implementation, collect the time (timestamp), battery voltage (V), battery current (I), battery temperature (t), etc. of each battery pack to be measured.

[0111] The feature sequence construction module 200 is used to perform data cleaning on the original battery features, extract features from the original battery features after data cleaning, and construct a feature sequence corresponding to the battery pack to be measured using the extracted features;

[0112] To avoid the influence of incorrect data on the evaluation result, it is necessary to clean the data in the original battery parameters to replace the incorrect data or redundant data therein, and improve the reliability of the evaluation result.

[0113] Construct a feature sequence x using each feature value in the original battery parameters after cleaning i =[x i,1 ,x i,2 ,…,x i,j , where x i,j is the j-th consistency index of the i-th battery pack extracted. Each battery pack to be measured corresponds to a feature sequence, and this feature sequence represents the overall consistency degree of the corresponding battery pack.

[0114] The domination degree calculation module 300 is used to use the feature sequence of each battery pack to be measured as a detection sample, and calculate the domination degree of each detection sample according to the feature sequence;

[0115] Use the feature sequence as the consistency index sequence of the corresponding battery pack to be measured. The consistency index sequences of all battery packs to be measured constitute the battery pack consistency sample set X. And each feature sequence is used as a detection sample, and calculate the domination degree of each detection sample, that is, the number of samples that dominate the current sample.

[0116] The evaluation module 400 is used to sort all detection samples in descending or ascending order according to the domination degree, and divide all detection samples into multiple levels according to a predetermined ratio; among them, the battery pack to be measured corresponding to the detection sample with a larger domination degree has worse consistency.

[0117] Reduce the multi-dimensional features of the consistency of all battery packs to one dimension, and evaluate the consistency of the battery packs, that is, sort all detection samples from largest to smallest or from smallest to largest according to the domination degree. The battery pack with a smaller domination degree has better consistency; on the contrary, the battery pack with a larger domination degree has worse consistency.

[0118] In addition, it is also necessary to evaluate the consistency level of the battery pack. For the scenario where the consistency of the battery pack needs to be divided into L levels, the battery packs can be sorted according to the degree of dominance, and then divided into L levels in turn according to a predetermined ratio. It should be noted that during the division process, it is necessary to ensure that the battery packs with the same degree of dominance should be divided into the same level.

[0119] For example, if it is necessary to divide into 3 levels, all the battery packs can be sorted from small to large according to the degree of dominance, and then all the battery packs can be divided into 3 levels according to the predetermined ratio of 3:3:4.

[0120] As an alternative implementation Figure 6 The structural schematic diagram of the feature sequence construction module provided for this embodiment is as Figure 6 shown. The feature sequence construction module 200 includes:

[0121] The feature extraction sub-module 2001 is used to calculate the corresponding static feature index and dynamic feature index according to the original battery features of each battery pack to be tested;

[0122] The data connection sub-module 2002 is used to construct the feature sequence corresponding to each battery pack to be tested by using the static characteristic index and the dynamic characteristic index as items in the feature sequence of the corresponding battery pack to be tested.

[0123] Feature extraction is performed based on the original battery features to obtain various feature indexes, including: static feature indexes and dynamic feature indexes.

[0124] Each static feature index and dynamic feature index can be used as a parameter in the feature sequence. Multiple static feature indexes and multiple dynamic feature indexes can be randomly arranged in the feature sequence, or arranged in the order of static feature index - dynamic parameter index or dynamic parameter index - static parameter index. It should be noted that different battery packs to be tested should adopt a unified arrangement standard so that the corresponding items of all battery packs to be tested are of the same type of feature index.

[0125] As an alternative implementation, the static characteristic indexes include: voltage range, temperature range, voltage standard deviation coefficient, temperature standard deviation coefficient;

[0126] The voltage range is calculated by the following formula:

[0127] ΔU max =U max -U min

[0128] where ΔU max is the voltage range; U max , Umin The maximum and minimum values of the single - cell voltages in a group of batteries, respectively.

[0129] The temperature range difference is calculated by the following formula:

[0130] ΔT max = T max - T min

[0131] Where, ΔT max is the temperature range difference, T max , T min are the maximum and minimum values of the single - cell temperatures in a group of batteries, respectively.

[0132] The voltage standard - deviation coefficient is calculated by the following formula:

[0133]

[0134] Where, δ U is the voltage standard - deviation coefficient, N is the total number of battery cells in a group of batteries; U i is the voltage of the i - th battery cell in a group of batteries; U m is the average value of the voltages of all battery cells in a group of batteries.

[0135] The temperature standard - deviation coefficient is calculated by the following formula:

[0136]

[0137] Where, δ T is the temperature standard - deviation coefficient, T i is the temperature of the i - th battery cell in a group of batteries; T m is the average value of the temperatures of all battery cells in a group of batteries.

[0138] The dynamic characteristic indexes include: characteristic parameters representing the discharge ohmic resistance, characteristic parameters representing the dynamic polarization characteristics, characteristic parameters representing the ohmic resistance, and parameters representing the overall discrete characteristics of the voltage or temperature curves during the charge - discharge process.

[0139] As an optional implementation manner, Figure 7 is a schematic structural diagram of the domination - degree calculation module provided in this embodiment. As shown in Figure 7 , the domination - degree calculation module 300 includes:

[0140] An item - comparison sub - module 3001, configured to perform corresponding - item comparison of the feature sequences between any detection sample and all the remaining detection samples;

[0141] The domination degree determination sub-module 3002 is used to determine that if all items in the feature sequence of the current detection sample are not greater than the corresponding items in the feature sequence of the comparison detection sample, and there is at least one item in the feature sequence of the current detection sample that is less than the corresponding item in the feature sequence of the comparison detection sample, then it is considered that the comparison detection sample is dominated by the current detection sample, and the domination degree of the comparison detection sample is incremented by 1.

[0142] When calculating the domination degree of each detection sample, all detection samples in the sample set need to be compared one by one with the remaining detection samples. If samples p and q simultaneously satisfy the following two conditions:

[0143] (1) All indicators x of sample p p,i are not greater than the corresponding indicators x of sample q q,i .

[0144] (2) There is at least one indicator x in sample p p,j that is less than the corresponding indicator x of sample q q,j .

[0145] A more rigorous statement is:

[0146] For the minimization multi-objective problem, the vector composed of n objective components f i (1,..., n) Given any two decision variables

[0147] If and only if, for all have then dominates

[0148] If and only if, for there is and there is at least one j ∈ {1,..., n} such that then weakly dominates

[0149] Then it can be said that sample p is superior to sample q (in this embodiment, it is stipulated that the smaller the value, the better the indicator), denoted as p dominates q or q is dominated by p, that is:

[0150] p < q

[0151] Calculate the number of times each sample is dominated by different samples in the sample set to obtain the domination degree.

[0152] As an alternative implementation, Figure 8 This is another structural schematic diagram of the feature sequence construction module provided in this embodiment. As shown in Figure 8 , the feature sequence construction module 200 further includes:

[0153] The first correction sub-module 2003 is configured to, for any target parameter value in the original battery parameters, if it exceeds the reasonable parameter threshold or is a missing value, replace the target parameter value with the parameter value at the previous moment or the parameter value at the next moment of the target parameter value;

[0154] The second correction sub-module 2004 is configured to calculate the average value of the parameter values in the preset time period through a sliding window and use the average value as the target parameter value.

[0155] To ensure the validity of the data and reduce the influence of incorrect data on the judgment result, in this embodiment, for an unreasonable value in the original battery parameters where a certain numerical value is a missing value or significantly exceeds the threshold range, the numerical value at the previous moment of this numerical value, or the numerical value at the next moment, or the average value within a period of time calculated through a sliding window is used for assignment replacement to ensure that each item in the characteristic sequence of each battery pack to be tested is accurate and reliable.

[0156] The above technical solutions have the following beneficial effects:

[0157] 1. Extract multiple characteristic indicators such as static characteristics and dynamic characteristics to more comprehensively characterize the consistency of the battery pack, design reasonable consistency evaluation indicators, and avoid relying on specific experimental conditions and laboratory test environments;

[0158] 2. Propose a grading evaluation method under multiple indicators. Different from the existing grading evaluation methods, it does not require weight calculation and feature fusion of multiple indicators, nor does it require giving a definite grading threshold based on expert knowledge and operation experience. Therefore, it is not targeted at batteries of specific manufacturers or models and has flexibility and universality;

[0159] 3. Based on the Pareto optimality idea, through the calculation of the dominance degree, the multi-dimensional indicators are reduced to one dimension for evaluation. The dimension reduction method is objective, effective, does not lose the original information, and does not require solving the indicator weights. It can be carried out at any usage stage of the battery pack, does not require designing and adjusting weights, and has better flexibility and promotion;

[0160] 4. Divide the battery packs into multiple grades according to the degree of being dominated according to the actual scenario for grading evaluation, which is simple and feasible and does not require providing an equivalent evaluation threshold.

[0161] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0162] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general purpose computers, special purpose computers, embedded processors, or other programmable data processing devices to produce a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in one or more flows and / or blocks Figure 1 in one or more flows and / or blocks Figure 1 or in one or more blocks.

[0163] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including instruction means that implement the functions specified in one or more flows and / or blocks Figure 1 in one or more flows and / or blocks Figure 1 or in one or more blocks.

[0164] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are performed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows and / or blocks Figure 1 in one or more flows and / or blocks Figure 1 or in one or more blocks.

[0165] Specific embodiments are applied in the present invention to elaborate on the principles and implementation manners of the present invention. The descriptions of the above embodiments are only used to help understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for evaluating the consistency and grading of a battery pack, characterized in that, Including: Obtaining the original battery characteristics of multiple battery packs to be tested; Performing data cleaning on the original battery characteristics, extracting features from the original battery characteristics after data cleaning, and constructing a feature sequence corresponding to each battery pack to be tested using the extracted features; Using the feature sequence of each battery pack to be tested as a test sample, and calculating the dominance degree of each test sample according to the feature sequence; Sorting all test samples in descending or ascending order according to the dominance degree, and dividing all test samples into multiple levels according to a predetermined ratio; among them, the worse the consistency of the battery pack corresponding to the test sample with a larger dominance degree; The step of extracting features from the original battery characteristics after data cleaning and constructing a feature sequence corresponding to each battery pack to be tested using the extracted features includes: Calculating corresponding static feature indexes and dynamic feature indexes according to the original battery characteristics of each battery pack to be tested; Taking the static feature index and the dynamic feature index as items in the feature sequence corresponding to each battery pack to be tested, and constructing the corresponding feature sequence for each battery pack to be tested; Calculating the dominance degree of each test sample according to the feature sequence includes: Comparing the corresponding items of the feature sequence of any test sample with the corresponding items of the feature sequences of all the other test samples; If all items in the feature sequence of the current test sample are not greater than the corresponding items in the feature sequence of the comparison test sample, and at least one item in the feature sequence of the current test sample is less than the corresponding item in the feature sequence of the comparison test sample, it is considered that the comparison test sample is dominated by the current test sample, and the dominance degree of the comparison test sample is incremented by 1.

2. The battery pack consistency evaluation and grading method according to claim 1, characterized in that: The static feature indexes include: voltage range, temperature range, voltage standard deviation coefficient, temperature standard deviation coefficient; The dynamic feature indexes include: characteristic parameters representing the discharge ohmic resistance, characteristic parameters representing the dynamic polarization characteristics, characteristic parameters representing the ohmic resistance, and parameters representing the overall discrete characteristics of the voltage or temperature curve during the charge and discharge process.

3. The battery pack consistency evaluation and grading method according to claim 1, wherein The data cleaning of the original battery characteristics includes: For any target parameter value in the original battery characteristics, if it exceeds the reasonable parameter threshold or is a missing value, replacing the target parameter value with the parameter value at the previous moment or the parameter value at the next moment of the target parameter value; Or, calculating the average value of the parameter values in a preset time period through a sliding window, and taking the average value as the target parameter value.

4. A battery pack consistency evaluation and grading device, characterized in that Including: A data acquisition module for obtaining the original battery characteristics of multiple battery packs to be tested; A feature sequence construction module for performing data cleaning on the original battery characteristics, extracting features from the original battery characteristics after data cleaning, and constructing a feature sequence corresponding to each battery pack to be tested using the extracted features; A dominance degree calculation module for using the feature sequence of each battery pack to be tested as a test sample and calculating the dominance degree of each test sample according to the feature sequence; An evaluation module, configured to sort all detection samples in descending or ascending order according to the degree of domination, and divide all detection samples into multiple levels according to a predetermined ratio; wherein, the worse the consistency of the battery pack to be tested corresponding to the detection sample with a greater degree of domination. The feature sequence construction module includes: A feature extraction sub-module, configured to calculate corresponding static feature indicators and dynamic feature indicators according to the original battery features of each battery pack to be tested. A data connection sub-module, configured to use the static feature indicators and the dynamic feature indicators as items in the feature sequence corresponding to the battery pack to be tested, and construct the feature sequence corresponding to each battery pack to be tested. The degree of domination calculation module includes: An item comparison sub-module, configured to compare the corresponding items of the feature sequence of any detection sample with the feature sequences of all the other detection samples. A degree of domination determination sub-module, configured to, if all items in the feature sequence of the current detection sample are not greater than the corresponding items in the feature sequence of the comparison detection sample, and at least one item in the feature sequence of the current detection sample is less than the corresponding item in the feature sequence of the comparison detection sample, then consider that the comparison detection sample is dominated by the current detection sample, and add 1 to the degree of domination of the comparison detection sample.

5. The battery pack consistency evaluation and grading device according to claim 4, wherein: The static feature indicators include: voltage range, temperature range, voltage standard deviation coefficient, and temperature standard deviation coefficient. The dynamic feature indicators include: characteristic parameters representing discharge ohmic resistance, characteristic parameters representing dynamic polarization characteristics, characteristic parameters representing ohmic resistance, and parameters representing the overall discrete characteristics of the voltage or temperature curve during the charge and discharge process.

6. The battery pack consistency evaluation and grading device according to claim 4, wherein the feature sequence construction module includes: A first correction sub-module, configured to, for any target parameter value in the original battery features, if it exceeds the parameter reasonable threshold or is a missing value, replace the target parameter value with the parameter value at the previous moment or the parameter value at the next moment of the target parameter value. A second correction sub-module, configured to calculate the average value of the parameter values in a preset period through a sliding window, and use the average value as the target parameter value.

Images