CN112327192B - Battery capacity diving phenomenon identification method based on curve form - Google Patents
Battery capacity diving phenomenon identification method based on curve form Download PDFInfo
- Publication number
- CN112327192B CN112327192B CN202011131864.9A CN202011131864A CN112327192B CN 112327192 B CN112327192 B CN 112327192B CN 202011131864 A CN202011131864 A CN 202011131864A CN 112327192 B CN112327192 B CN 112327192B
- Authority
- CN
- China
- Prior art keywords
- slope
- curve
- battery
- battery capacity
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/392—Determining battery ageing or deterioration, e.g. state of health
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
Abstract
The invention discloses a battery capacity diving identification method based on a curve form, which comprises the following steps: acquiring battery capacity degradation curve data according to the battery capacity degradation data subjected to data smoothing pretreatment; extracting slope characteristics of a battery capacity degradation curve from the battery capacity degradation curve data; performing form recognition on the battery capacity degradation curve according to the extracted slope characteristics; and identifying the battery capacity diving according to the battery capacity degradation curve form identification result.
Description
Technical Field
The invention relates to battery capacity decline detection, in particular to a battery capacity diving phenomenon identification method based on a curve form.
Background
Lithium batteries are widely used in various industries at present, particularly in the field of automobile power energy, and have been developed as an indispensable part due to their characteristics of long cycle life, high energy density and good safety performance. With the continuous expansion of the application range of lithium batteries, the market and customers have paid more attention to and expect the service life and safety performance of lithium batteries.
During normal charge and discharge cycles of a lithium battery, as a Solid Electrolyte Interface (SEI) film increases, its capacity suffers from normal capacity fading. However, as the number of battery charge and discharge cycles increases, polarization within the battery can lead to the occurrence of a particular lithium precipitation phenomenon. Lithium precipitation has a great influence on the life and safety of lithium batteries: on the one hand, rapid degradation of battery capacity is caused by precipitation of lithium metal; on the other hand, the precipitated lithium metal forms dendrites, and the dendrites may pierce the separator due to the continuous growth of the dendrites, so that the positive electrode and the negative electrode inside the battery are short-circuited, thereby causing serious safety accidents. In addition, as the application range of the battery is increased, low-temperature and high-rate charging conditions are more common, and the risk of lithium precipitation of the battery is increased due to the factors. Since the lithium separation of the battery seriously affects the cycle life and the safety performance of the battery, it is important to judge the lithium separation phenomenon of the lithium battery in advance and accurately.
In order to meet the requirements of users and markets, lithium battery designers can carry out a cycle life test aiming at a new developing and allocating party, so as to determine whether the service life of the battery meets the requirements. Aiming at the design and test stages of the lithium battery, the phenomenon of lithium separation of the battery needs to be identified and an alarm is given, and on one hand, the battery with the lithium separation is found and confirmed, so that the test is finished as early as possible, and the cost of an enterprise is saved. Meanwhile, a formula which can cause lithium precipitation is proposed from a design formula matrix, the formula design matrix is simplified, and the design and development efficiency is improved. Furthermore, it is also possible to avoid that the battery with the risk of lithium precipitation enters a subsequent production phase and thus eventually enters the market, leading to greater economic losses and safety risks.
At present, a common method for detecting lithium separation from a battery mainly comprises the steps of manually disassembling the battery, and observing the appearance state of a disassembled battery cell, so as to judge whether lithium separation occurs. The method is mainly judged according to the experience of detection personnel, and errors caused by subjective judgment exist. Meanwhile, the battery disassembling process is complex, a large amount of manpower and material resources are consumed, the efficiency is low, certain potential safety hazards exist, and the battery disassembling has irreversible destructiveness. In addition, since the lithium deposition of the battery is a result exhibited by the accumulation of chemical reactions for a long period of time to some extent, when the occurrence of lithium deposition in the battery is observed manually, it is indicated that the battery has been in a lithium deposition state for a long period of time, resulting in a hysteresis in the identification of lithium deposition. In addition to manual disassembly, common methods for detecting lithium by analysis include voltage analysis and elemental analysis. The voltage analysis method is to judge and analyze lithium by measuring the voltage between the metal shell and the negative electrode of the battery, the method needs additional equipment for testing, the monitoring result has certain error and is influenced by factors such as the battery structure, and the method has certain limitation. The elemental analysis method is to use a scanning electron microscope to perform in-situ synchronous observation on the battery, so as to judge whether lithium is separated in the charge-discharge cycle. The method needs additional equipment to carry out measurement and is high in cost. In summary, the existing lithium analysis detection method has hysteresis or limitation, and cannot meet the requirement of accurately and previously identifying the lithium analysis phenomenon of the battery under a general battery test scene.
The lithium precipitation of the battery is mainly caused by unreasonable formula design, defects (such as process problems and quality deviation) introduced in processing and manufacturing, and the like. The rapid decline of the battery capacity is an important performance of the battery after lithium precipitation, and the rapid decline of the battery capacity is defined as 'water jump'. During the entire life cycle of a lithium battery, the battery will exhibit normal performance degradation, i.e., a degradation law that remains stable over a period of time. When the battery has water jump, the battery can have abnormal decline, namely when the life cycle of the battery exceeds a certain critical point, the capacity decline rate of the battery is rapidly increased, the capacity of the battery is rapidly reduced, and the critical point is defined as a water jump point. Generally, the occurrence of battery water-jumping is caused by lithium precipitation of the battery. Therefore, the battery capacity curve can be analyzed, the diving point of the capacity curve is identified in advance, and the battery diving linearity is early warned, so that the lithium analysis detection of the battery is supported.
The invention realizes the identification and alarm of the battery water-jumping phenomenon by using the battery capacity degradation curve based on the curve form and the geometric shape parameters, supports the development of lithium analysis, thereby providing decision support for testers in a test stage, optimizing a test formula and avoiding safety accidents.
Disclosure of Invention
The invention provides a battery diving phenomenon identification method based on a curve form aiming at a lithium battery diving phenomenon.
The invention discloses a battery capacity diving identification method based on a curve form, which comprises the following steps:
acquiring battery capacity degradation curve data according to the battery capacity degradation data subjected to data smoothing pretreatment;
extracting slope characteristics of a battery capacity degradation curve from the battery capacity degradation curve data;
performing form recognition on the battery capacity degradation curve according to the extracted slope characteristics;
and identifying the battery capacity diving according to the battery capacity degradation curve form identification result.
Preferably, the slope characteristic comprises: performing characteristic extraction of a primary slope on battery capacity degradation curve data of each charge-discharge cycle to obtain a primary slope; and carrying out slope extraction again on the primary slope characteristics to obtain a secondary slope.
Preferably, the morphological identifying of the battery capacity degradation curve according to the extracted slope characteristics includes: setting a reference interval in which all values in the interval are approximate to 0; defining a slope value of the secondary slope that does not exceed a range of values within the reference interval as "zero data"; defining a slope value of the secondary slope greater than a range of values within the reference interval as "positive data"; defining a slope value of the secondary slope smaller than a range of values within the reference interval as "negative data"; and carrying out shape recognition on the battery capacity degradation curve by using the zero data, the positive data and the negative data.
Preferably, using the "zero data", "positive data", and "negative data" to perform morphology recognition on the battery capacity degradation curve includes: when N positive data continuously appear, identifying the battery capacity degradation curve form as a concave curve; when N negative data continuously appear, identifying the battery capacity degradation curve form as a convex curve; when N "zero data" appear consecutively, the battery capacity degradation curve morphology is recognized as a straight curve.
Preferably, it is judged that the battery has not been subjected to the water jump when the battery capacity degradation curve shape is recognized as a straight curve or a concave curve.
Preferably, the primary slope is used for determining whether the water jump occurs for the battery with the curve form identified as a convex curve, and the method specifically comprises the following steps: calculating the difference value between the primary slope value of the battery capacity degradation curve data and a slope reference value to obtain a slope difference value; and comparing the slope difference value with a slope difference threshold value, and if the slope difference value is greater than the slope difference value, judging that the battery has jumped.
Preferably, the slope reference value is a slope value of interval data in which a battery capacity decrease trend is stable and linear in an initial period of a charge-discharge cycle of the lithium battery. Selecting the slope reference value comprises: selecting initial data of a battery life cycle test; setting a plurality of windows in the initial data of the battery life cycle test, calculating the slope difference between each window and the front window and the rear window, and selecting the minimum value as a reference window; and calculating a reference slope value according to the capacity retention rate data of the lithium batteries at two ends of the reference window and the window length.
Preferably, the data smoothing preprocessing is performed by using a local weighted regression method. The battery is a lithium battery.
The invention has the beneficial effects that: 1) only the battery capacity data obtained in the cycle test is used, destructive disassembly and measurement of other parameters are not needed, and the cost investment is low; 2) the invention can assist in carrying out batch analysis after the battery cycle test, identify whether the battery is a diving sample, and provide information support for batch identification test; 3) by adopting a two-stage decision method combining curve form discrimination and slope difference calculation, samples obviously without a diving risk can be eliminated in the curve form discrimination stage, and then the remaining samples with a certain diving risk are converted into slope difference calculation to further judge whether the battery is a diving sample, thereby being beneficial to improving the stability and accuracy of discrimination.
The above technical solution of the present invention will be described in detail with reference to the accompanying drawings.
Drawings
FIG. 1 is a schematic diagram of a method for identifying battery capacity diving based on a curve form according to the present invention;
FIG. 2 is a schematic diagram of a battery capacity diving identification process based on a curve form according to the present invention;
FIG. 3 is a graph showing the relationship between a straight curve and a quadratic slope;
FIG. 4 is a schematic diagram of a concave curve versus a quadratic slope;
FIG. 5 is a schematic diagram of a convex curve versus a quadratic slope;
FIG. 6 is a raw degradation curve for four sample cells;
FIG. 7 is a graph of data smoothing and noise reduction pre-processing results;
FIG. 8 is a result of a primary slope feature calculation;
FIG. 9 is a result of a quadratic slope characteristic calculation;
FIG. 10 is a graph showing the results of curve shape calculations for a normal battery sample;
FIG. 11 is a graph showing the results of curve shape calculations for a sample diving battery;
FIG. 12 is a slope reference build result;
FIG. 13 is a graph showing the results of real-time calculation of slope difference for normal battery samples;
FIG. 14 is a diagram illustrating the real-time calculation result of the slope difference of the sample of the diving battery.
Detailed Description
Fig. 1 shows a battery capacity diving identification method based on a curve form, which comprises the following steps:
acquiring battery capacity degradation curve data according to the battery capacity degradation data subjected to data smoothing pretreatment;
extracting slope characteristics of a battery capacity degradation curve from the battery capacity degradation curve data;
performing form recognition on the battery capacity degradation curve according to the extracted slope characteristics;
and identifying the battery capacity diving according to the battery capacity degradation curve form identification result.
The slope characteristics of the present invention include: performing characteristic extraction of a primary slope on battery capacity degradation curve data of each battery charge-discharge cycle to obtain a primary slope; and carrying out slope extraction again on the primary slope characteristics to obtain a secondary slope.
That is, the present invention extracts a first slope and a second slope per charge-discharge cycle of a battery.
Generally, the morphological recognition of the battery capacity degradation curve according to the extracted slope characteristics includes: setting a reference interval in which all values in the interval are approximate to 0; defining a slope value of the secondary slope that does not exceed a range of values within the reference interval as "zero data"; defining a slope value of the secondary slope greater than a range of values within the reference interval as "positive data"; defining a slope value of the secondary slope smaller than a range of values within the reference interval as "negative data"; and carrying out shape recognition on the battery capacity degradation curve by using the zero data, the positive data and the negative data.
The morphological recognition of the battery capacity degradation curve by using the zero data, the positive data and the negative data comprises the following steps: when N positive data continuously appear, identifying the battery capacity degradation curve form as a concave curve; when N negative data continuously appear, identifying the battery capacity degradation curve form as a convex curve; when N "zero data" appear consecutively, the battery capacity degradation curve morphology is recognized as a straight curve. And when the battery capacity degradation curve form is identified as a straight curve or a concave curve, judging that the battery has no water jump.
The method for judging whether the water jump occurs or not by using the primary slope for the battery with the curve form identified as the convex curve specifically comprises the following steps: calculating the difference value between the primary slope value of the battery capacity degradation curve data and a slope reference value to obtain a slope difference value; and comparing the slope difference value with a slope difference threshold value, and if the slope difference value is greater than the slope difference value, judging that the battery has jumped. The slope reference value is generally a slope value of interval data with a stable and linear battery capacity reduction trend in the initial charge-discharge cycle of the lithium battery. Selecting the slope reference value comprises: selecting initial data of a battery life cycle test; setting a plurality of windows in the initial data of the battery life cycle test, calculating the slope difference between each window and the front window and the rear window, and selecting the minimum value as a reference window; and calculating a reference slope value according to the capacity retention rate data of the lithium batteries at two ends of the reference window and the window length.
In addition, the invention adopts a local weighted regression method to carry out data smoothing pretreatment, of course, the invention can also adopt any existing data smoothing treatment technology to carry out pretreatment on the original battery capacity degradation data.
The method for identifying battery capacity diving based on the curve form of the present invention is further explained with reference to fig. 2-14.
Fig. 2 shows a battery diving identification process of the present invention, and as shown in fig. 2, the battery capacity diving identification process based on the curve form of the present invention mainly includes data preprocessing and feature extraction, capacity degradation curve form identification, and diving identification based on a slope difference.
Firstly, carrying out data preprocessing on original battery capacity degradation data to obtain smoothed battery capacity degradation data. Based on the curve data, primary and secondary slope characteristics are extracted.
And then, carrying out form recognition on the curve by utilizing a constructed curve form standard and combining with a secondary slope characteristic, wherein if the current capacity curve is a convex curve, the current capacity curve is the occurrence of the water jump.
And finally, calculating primary slope difference data by using the primary slope characteristics and the normal slope reference of the data with the diving risk, and combining a threshold value to finish diving identification.
And based on actually measured lithium battery life cycle test data, utilizing curve classification and a primary slope difference value calculation result to identify whether the current battery has the water jump or not. The invention has the following advantages:
1) the real-time monitoring and alarming can be carried out aiming at the water jumping phenomenon of the battery only by utilizing the capacity data of the real-time battery life cycle process, so that the development of the lithium analysis of the battery is supported;
2) an effective data smoothing method is provided for the battery data characteristics, so that the local fluctuation of a battery capacity curve is eliminated, and the accuracy of the identification of the water jump phenomenon is improved;
3) and sensitive diving characteristic parameters are extracted according to a battery capacity curve, and a diving warning can be identified in advance, so that a lithium analysis warning is given out.
In addition, the algorithm model can be transplanted to a user use stage, decision support is provided for task planning of the user, meanwhile, an alarm is given in advance, and safety accidents are avoided.
Data pre-processing
In the process of developing a battery life test, measurement errors and environmental interference are not generated, collected battery capacity data can have noise and random fluctuation, data noise can generate certain interference on subsequent identification, in order to eliminate the data noise, the collected original data needs to be subjected to smoothing processing, and the random noise is removed while the data change trend is kept.
The method adopts a local weighted regression method (LOWESS) to carry out data smoothing, and the main idea of the method is to select a section of data before and after a target point to fit a polynomial regression curve, so that the estimated value of the curve is used for replacing the true value, and the effect of removing the noise and retaining the curve trend is realized.
The principle of the local weighted regression algorithm is that local observation data are fitted through polynomial weighting, and then the fitting result is estimated through the least square method. The method comprises the following specific steps:
for a single parameter sample xi,y i1,2, …, n, establishing a model.
In the formula, betai0,βi1,…,βidIs relative to xiThe unknown parameters of (1); epsiloniI is 1,2, …, n is an independently distributed random error term; d is a value given in advance.
For each xiCalculating all points x aroundj(j ═ 1,2, …, n) at a distance of
dij＝|xi-xj|
hiIs dijThe r-th value in (j ═ 1,2 … n) is small. r is a distance fnThe most recent integer, r ═ fnIs the window width selected when the data is partially returned. f is belonged to (0, 1)]Represents the point of influence yiX ofjAnd (3) a range.
Given a weight function W (x), for each point xiAll x's within the windowkK is 1,2, …, n, and weights are calculated
Parameter betaikThe estimate of (i-1, 2, …, n; k-1, 2, …, n) is given by the following criteria
Thereby obtaining yiFitting value of
Given a weight function w (x), the following four basic conditions are satisfied:
(1) w (x) > 0 vs | x | < 1;
(2)W(-x)＝W(x)
(3) for x ≧ 0, W (x) is an increasing function;
(4) w (x) 0 vs | x | ≧ 1
Currently, the most common is the cubic weight function
And (m, n) type weight function
The local weighted regression algorithm integrates the traditional local polynomial fitting and local weighted regression, and is a fitting algorithm with strong robustness. Different smoothness requirements can be achieved by selecting proper window width, and when the window width is larger, a large-range trend can be obtained; when the window width is small, the smoothing result is again closer to the original data. The characteristics of the algorithm enable parameters in the linear regression model to change along with different values of the independent variable, and the algorithm can ensure that the estimated value of the regression function is given at any point of the independent variable space due to different observed values corresponding to different parameters.
Slope feature calculation
As the number of charge and discharge cycles of the lithium ion battery increases, reactions occur inside the battery, and a Solid Electrolyte Interface (SEI) film grows continuously, thereby causing degradation of the battery capacity. The battery capacity degradation caused by the increase of the SEI film is linear, the degradation rate is stable, and the degradation rate is generally in a linear degradation mode. When the battery has lithium precipitation or other abnormal conditions, the battery capacity can have accelerated degradation, and a capacity curve can have obvious inflection points.
From the view point of curve morphology, when the battery capacity is normally degraded due to the increase of an SEI film, the capacity degradation curve is a straight line, and no water jump condition exists; when the battery capacity is subjected to deceleration degradation due to other reasons, namely the battery degradation rate is reduced compared with the initial rate, the battery capacity curve is a concave curve, and the battery does not have water jumping; when the capacity of the battery is degraded in an accelerating way due to lithium precipitation or other abnormalities, the capacity curve has an obvious inflection point, the battery capacity curve is a convex curve, and water jump occurs at a high probability. Therefore, the method utilizes the curve form to carry out primary classification and screening on the battery capacity curve and identify whether the battery capacity curve has the probability of water diving. The secondary slope is a direct parameter for representing the form of the curve (concave, convex and straight), and secondary slope characteristics are extracted to carry out classification and identification on the battery capacity curve.
From the angle of primary slope, when the battery capacity is normally declined, the linear degradation rule is formed, and the primary slope of the capacity degradation curve is a constant value; when the battery is abnormally accelerated to decline in capacity due to lithium precipitation, the primary slope of the battery is gradually increased. Therefore, after the curve form is utilized to preliminarily identify whether the battery capacity has the diving probability, the decay rate of the battery capacity is further quantified by utilizing the primary slope according to the convex curve. And extracting the primary slope characteristic to finally identify whether the battery has a diving condition.
Battery capacity curve form identification and classification
And (4) carrying out curve form identification by utilizing the secondary slope, and firstly constructing a curve form identification reference. The slope of the straight curve generally fluctuates above and below 0, the slope of the concave curve generally is constantly greater than 0, and the slope of the convex curve generally is constantly less than 0, as shown in fig. 3, 4 and 5.
Therefore, when constructing the morphology recognition reference, the secondary slope normal reference space is defined as [ -a, b ], and in this interval range, the secondary slope is considered to be approximately 0, the battery capacity curve is a straight curve, and the battery degradation condition is normal.
On the basis of the reference interval, data points in the reference interval are defined as 'zero data', data points larger than the reference interval are defined as 'positive data', and data points smaller than the reference interval are defined as 'negative data'.
And (3) setting a curve form recognition threshold value by combining expert experience and partial historical data:
when N "positive data" appear consecutively, classifying the original curve as a concave curve (convex);
when N "zero data" continuously appear, classifying the original curve as a straight curve (zero);
when N "negative data" appear consecutively, the original curve is classified as a convex curve (concave)
The form of the curve (concave curve, convex curve and straight curve) can qualitatively reflect whether the current battery capacity curve has the water jump or not, and the secondary slope is a direct representation of the form of the curve. And classifying the battery capacity curve by utilizing the secondary slope characteristics and through a curve form identification standard, and primarily screening whether the battery capacity data has the probability of water diving.
And calculating statistical data of positive data, zero data and negative data of the battery capacity degradation data obtained by the current test by using the secondary characteristics of the battery capacity degradation data, and carrying out shape recognition and classification on the current battery capacity curve based on the distribution of the three data and the threshold value of shape recognition.
When the battery capacity curve is classified into a straight curve, the current battery capacity data is not subjected to water diving, but the risk of water diving exists in the future, and testers need to further observe the data;
when the battery capacity curve is classified into a concave curve, the current battery has good performance and is in a deceleration decline phenomenon, and the probability of water jump in the future is low;
when the battery capacity curve is classified into a convex curve, the battery is indicated to be accelerated to decline, the probability of occurrence of the water jump is high, and quantitative analysis needs to be carried out by combining the primary slope to identify whether the water jump occurs.
Normal slope reference selection
The primary slope may characterize how fast the battery capacity degrades. Considering the difference of design formulas of different batteries and the difference of conditions of different test platforms, the degradation performance of the batteries is directly represented by the primary slope, and the degradation performance of the batteries is greatly influenced by individual difference.
Therefore, the method provides a slope difference value, the slope value which is relatively stable in the initial degradation stage of the current battery is used as a reference, and the difference value between the current slope and the slope reference is used as a quantization parameter to represent the decline speed of the battery capacity, so that the individual difference of the battery caused by design and test is eliminated.
The slope difference is calculated by first constructing a normal slope reference.
The reference value of the battery capacity degradation under the normal slope reference is generally, the battery capacity degradation is stable at the initial stage of the life cycle test, and the battery capacity degradation is generally in a linear rule. Therefore, the degradation data in the initial stage of the battery test is used, and the obvious interval data with stable capacity descending trend and linear law is judged and selected through the sliding window to serve as the slope reference. And calculating the slope difference between each window and the front window and the rear window, selecting the window with the minimum difference as a reference window, completing the construction of the normal slope reference, and calibrating the normal track of the battery capacity degradation.
Diving identification based on slope difference
On the basis of the curve form, whether the water jump occurs or not is further quantitatively identified by utilizing the primary slope aiming at the battery with the identification result of a convex curve, and reference is provided for a tester.
And calculating real-time primary slope characteristics according to the battery capacity degradation data obtained by the current test, and combining the constructed normal slope reference to obtain the difference value between the real-time primary slope parameter and the normal slope reference, and recording the difference value as a slope difference value.
The normal slope reference is the calibration of the battery stability degradation, therefore, the slope difference is compared with the normal reference to obtain a parameter which can quantitatively represent the battery real-time capacity degradation rate.
When the slope difference data is less than 0, the real-time capacity degradation rate of the battery is smaller than the reference rate, and the battery performance is good;
when the slope difference data is approximate to 0, the real-time capacity degradation rate of the battery is basically the same as the reference rate, and the capacity degradation rate of the battery has no obvious change;
when the slope difference value is larger than 0, the real-time capacity degradation rate of the battery is larger than the reference rate, the capacity degradation rate of the battery is increased, and the capacity of the battery is degraded in an accelerating mode. The larger the value, the more severe its accelerated degradation. And combining expert experience and historical data, setting a threshold value, and considering that the battery has jumped when the slope difference value exceeds the threshold value.
Identifying and screening a convex curve according to the curve form, calculating the current slope difference, and if the slope difference exceeds a threshold value, determining that the battery has an obvious water jump phenomenon, wherein generally speaking, a tester needs to stop testing; if the slope difference value exceeds the threshold value, the battery is not in a remarkable diving phenomenon, but still in a dangerous state, the risk of the battery in a diving condition in the future is high, and testers need to pay attention.
DETAILED DESCRIPTION OF EMBODIMENT (S) OF INVENTION
In case analysis, the actual lithium ion storage battery test data is adopted, and a battery capacity diving phenomenon recognition method based on the curve form and the slope difference is applied to carry out capacity diving recognition and diving early warning. The method comprises the steps of firstly simply introducing data applied in a case, then describing an application process of the battery capacity diving phenomenon identification method based on a curve form and a slope difference value by combining actual data, and finally analyzing test results, effectiveness, sensitivity and the like of the method.
The feasibility and the effectiveness of the provided lithium ion battery service life prediction method are verified by adopting test data of Ningde time New energy science and technology Limited company (note that the battery used in the test is a soft package battery specially used in the product design stage, which is different from the battery used in the real product of the company).
The lithium battery samples under the test condition of 25 ℃ are selected for analysis and verification, and case analysis is carried out on four samples, wherein the case analysis comprises two diving samples and two normal samples. The numbering of the four samples is marked as follows:
normal samples: n1, N2
Diving sample: d1 and D2
The raw degradation curves for the four sample cells are shown in fig. 6.
1. Data pre-processing
The battery capacity diving phenomenon based on the curve form and the slope difference value is identified to carry out smooth noise reduction processing on a battery capacity degradation curve, retain the general trend of capacity degradation, and remove the interference of local noise, self-recovery effect and the like. The capacity degradation curve was smoothed using the LOWESS method, with the smoothing scale parameter uniformly set to 0.4.
The capacity degradation curve results after the smoothing noise reduction processing are shown in fig. 7.
2. Slope feature calculation
After the smoothing noise reduction processing, the data is subjected to the curve feature extraction of the primary slope and the secondary slope to perform the following curve form classification and the diving state identification, and the slope feature calculation results are shown in fig. 8 and 9.
3. Battery capacity curve form identification and classification
When the secondary feature extraction of the degradation curve of the battery capacity is completed, the classification of the curve forms can be performed. Firstly, defining a secondary slope normal reference space as [ -0.000002,0.000002], defining data points in the interval as 'zero data', data points larger than the interval as 'positive data', and data points smaller than the interval as 'negative data'. Then, identifying the curve form, and classifying the curve form into a concave curve when 20 points continuously appear as 'positive data'; when 20 points appear continuously as negative data, classifying the curve form into a convex curve; after identifying as a concave curve or a convex curve, 20 points appear consecutively as "zero data", and the curve morphology is classified back as a straight curve.
In this example, the dynamic propulsion process of the battery charge-discharge cycle test was simulated using all the test data of the sample battery. The specific method comprises the following steps: and supplementing the real capacity retention rate data of the sample battery corresponding to the cycle in the capacity degradation curve sequence of the sample battery every time the charge-discharge cycle of the battery is increased by 1, and classifying the curve form of the current state by using the classification method to obtain a real-time state sequence of the curve form.
The results of the classification of the battery capacity curves based on the curve morphology are shown in fig. 10 and 11:
in fig. 10, the battery capacity degradation curve of sample N2 in the charge-discharge cycle test is always in a straight curve state, which is consistent with the actual degradation curve shape, and this shows that the method has higher accuracy in classification. In the charge/discharge cycle test, in the sample N1, the battery capacity degradation curve was in a straight curve state in the first 800 cycles, and the battery capacity degradation was approximately in a linear degradation state, while after 800 cycles, the curve state was changed to a convex curve, which is consistent with the actual battery capacity degradation curve, but it is obvious from the figure that the battery did not generate the capacity jump phenomenon, so that it is necessary to further identify the jump phenomenon after the curve classification. In fig. 11, before 1000 cycles, the battery capacity degradation is approximately in a phenomenon degradation state, the curve state is also in a straight curve state, and after 1000 cycles, the battery flux degradation is in a water jump phenomenon, at which the capacity degradation curve becomes a convex curve state.
The real-time curve classification and the battery degradation state of the method are high in consistency by combining the actual battery capacity degradation curve and the curve state real-time classification curve, the change of the curve state can be accurately and quickly reflected, and accurate and reliable conditions are provided for further identifying the diving state.
4. Normal slope reference selection
The method replaces the slope with the primary difference to replace the primary slope value, and calculates the slope of each window between 95 percent and 99 percent of the residual life by using the sliding window. The window width is selected in a self-adaptive mode, the 10% of the data length of 95% -99% takes the charge-discharge cycle number 50 as the lower limit and 100 as the upper limit, the step length is 10, the difference value of the slope of each window and the slope of the two windows before and after is calculated, the minimum value is selected as a reference window, and the construction of the normal slope reference is completed.
The slope reference construction results are shown in fig. 12.
5. Diving identification based on slope difference
Since the battery capacity curve classification method based on the curve form identifies that the battery with the convex curve not only contains the battery with the diving but also contains the battery with the normal degradation state, the method further identifies the battery capacity diving based on the slope difference.
In this case, the dynamic propulsion process of the battery charge-discharge cycle test is simulated using all the test data of the sample battery. The specific method comprises the following steps: and calculating the difference value between the real-time slope of the current battery capacity retention rate and the normal slope standard by the method to form a change sequence of the real-time slope difference value along with the increase of the number of charge-discharge cycles.
In the invention, through data verification and expert experience, the threshold value of the slope difference is set to be 0.0001, and when the slope difference is higher than the threshold value, the battery degradation is considered to have the water jump, and a tester is reminded to stop the test immediately.
The results of the battery capacity diving identification method based on the slope difference are shown in fig. 12 and 13:
in the figure, the thin curve is a sample battery actual capacity degradation curve, and the thick curve is a change curve of the slope difference of the sample battery as the number of charge and discharge cycles increases. The black dotted line is a set alarm threshold value, the black dotted line is a characteristic included angle threshold value, when the characteristic included angle exceeds the threshold value, a diving alarm is sent out, and the test is recommended to be stopped.
In fig. 13, the slope difference of sample N1 in the charge-discharge cycle test is kept low, which is consistent with the results of the curve classification method and the actual degradation curve, and also illustrates the correctness of the battery capacity curve classification method based on the curve morphology. Although the degradation curve of the sample N2 is classified as a convex curve, the slope difference increases as the number of charge/discharge cycles increases, but the slope difference is kept below the threshold value, which indicates that the battery capacity degradation is approximately in a linear degradation state, no water jump occurs, and the slope difference is consistent with the actual degradation curve, thus showing the effectiveness and correctness of the water jump sample identification by the water jump identification method based on the slope difference. In fig. 14, the battery capacity degradation of the sample battery before 800 cycles is in a substantially linear degradation state, and the slope difference is maintained substantially at 0 or so and is stabilized in a normal state. After 800 cycles, the battery capacity degradation starts to generate the diving, the slope difference value also rises rapidly and finally reaches over 0.001, and the diving state early warning is sent out to prompt experimenters to stop the test. By combining with the actual battery capacity degradation curve, the method has the advantages of high consistency between the real-time state evaluation and the battery degradation state, and quick, instant and accurate early warning on the diving state.
Therefore, the method can effectively and stably perform real-time risk monitoring and evaluation on battery capacity degradation and timely discovery and early warning of battery capacity diving, so that the purposes of real-time risk evaluation, early test termination, test cost saving, safety improvement and the like are achieved.
Although the present invention has been described in detail hereinabove, the present invention is not limited thereto, and various modifications can be made by those skilled in the art in light of the principle of the present invention. Thus, modifications made in accordance with the principles of the present invention should be understood to fall within the scope of the present invention.
Claims (9)
1. A battery capacity diving identification method based on curve form is characterized by comprising the following steps:
acquiring battery capacity degradation curve data according to the battery capacity degradation data subjected to data smoothing pretreatment;
extracting slope characteristics of a battery capacity degradation curve from the battery capacity degradation curve data, wherein the slope characteristics comprise a primary slope and a secondary slope;
for a reference interval in which all values are approximately 0, defining the slope value of the secondary slope of the slope characteristic which does not exceed the value range in the reference interval as 'zero data';
defining a slope value of a secondary slope of the slope characteristic greater than the range of values within the reference interval as "positive data";
defining a slope value of a secondary slope of the slope characteristic smaller than the range of values within the reference interval as "negative data";
carrying out shape recognition on a battery capacity degradation curve by using the zero data, the positive data and the negative data;
and identifying the battery capacity diving according to the battery capacity degradation curve form identification result.
2. The method for identifying battery capacity diving according to the curve form of claim 1, wherein a primary slope of the slope feature is obtained by performing feature extraction of the primary slope on battery capacity degradation curve data of each battery charge-discharge cycle; and the secondary slope of the slope characteristic is obtained by performing slope extraction on the primary slope characteristic again.
3. The method for identifying battery capacity diving according to claim 2, wherein the step of performing morphology identification on the battery capacity degradation curve by using the zero data, the positive data and the negative data comprises the steps of:
when N positive data continuously appear, identifying the battery capacity degradation curve form as a concave curve;
when N negative data continuously appear, identifying the battery capacity degradation curve form as a convex curve;
when N "zero data" appear consecutively, the battery capacity degradation curve morphology is recognized as a straight curve.
4. The method of claim 3, wherein it is determined that the battery has not been subjected to the diving when the battery capacity degradation curve shape is identified as a straight curve or a concave curve.
5. The method of claim 3, wherein the primary slope is used to determine whether or not a water jump has occurred for a battery whose curve form is identified as a convex curve.
6. The method of claim 4, wherein the determining whether the diving has occurred using the primary slope for the battery whose curve form is identified as a convex curve comprises:
calculating the difference value between the primary slope value of the battery capacity degradation curve data and a slope reference value to obtain a slope difference value;
and comparing the slope difference value with a slope difference threshold value, and if the slope difference value is greater than the slope difference value, judging that the battery has jumped.
7. The method for identifying battery capacity diving according to the curve form of claim 6, wherein the slope reference value is a slope value of interval data with a steady and linear battery capacity descending trend in the initial period of the charge-discharge cycle of the lithium battery.
8. The method for battery capacity diving identification based on curve morphology as claimed in claim 7, wherein selecting said slope reference value comprises:
selecting initial data of a battery life cycle test;
setting a plurality of windows in the initial data of the battery life cycle test, calculating the slope difference between each window and the front window and the rear window, and selecting the minimum value as a reference window;
and calculating a reference slope value according to the capacity retention rate data of the lithium batteries at two ends of the reference window and the window length.
9. The curve morphology-based battery capacity diving identification method according to any one of claims 1-8, characterized in that a local weighted regression method is adopted for the data smoothing preprocessing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011131864.9A CN112327192B (en) | 2020-10-21 | 2020-10-21 | Battery capacity diving phenomenon identification method based on curve form |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011131864.9A CN112327192B (en) | 2020-10-21 | 2020-10-21 | Battery capacity diving phenomenon identification method based on curve form |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112327192A CN112327192A (en) | 2021-02-05 |
| CN112327192B true CN112327192B (en) | 2021-11-30 |
Family
ID=74311934
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011131864.9A Active CN112327192B (en) | 2020-10-21 | 2020-10-21 | Battery capacity diving phenomenon identification method based on curve form |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112327192B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113359042A (en) * | 2021-06-07 | 2021-09-07 | 芜湖天弋能源科技有限公司 | Method for nondestructive testing of carbon coating missing layer of anode of lithium ion battery |
| CN113655398B (en) * | 2021-08-11 | 2024-04-05 | 远景动力技术(江苏)有限公司 | Method for predicting circulating water jump of lithium battery |
| CN113777495A (en) * | 2021-08-25 | 2021-12-10 | 同济大学 | Lithium battery capacity diving online multi-stage early warning method and system based on characteristic area |
| JP2023180752A (en) * | 2022-06-10 | 2023-12-21 | 株式会社トヨタシステムズ | Secondary battery deterioration determination device and secondary battery deterioration determination method |
| CN116413609B (en) * | 2023-06-08 | 2023-08-29 | 江苏正力新能电池技术有限公司 | Battery diving identification method and device, electronic equipment and storage medium |
| CN116541718B (en) * | 2023-07-07 | 2023-09-22 | 江苏正力新能电池技术有限公司 | Training method, device, equipment and medium for battery capacity jump recognition model |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247237A (en) * | 2017-05-26 | 2017-10-13 | 联想(北京)有限公司 | A kind of detection method of battery, electronic equipment and computer-readable recording medium |
| SE1851061A1 (en) * | 2018-09-07 | 2018-09-07 | Abb Schweiz Ag | Voltage source converter generates an output voltage having at least three voltage levels |
| CN109596983A (en) * | 2018-11-19 | 2019-04-09 | 清华大学 | The prediction technique that capacity is dived during a kind of cell degradation |
| CN111175653A (en) * | 2020-01-06 | 2020-05-19 | 国网内蒙古东部电力有限公司电力科学研究院 | Method for identifying and prejudging capacity 'water-jumping' fault of ternary battery |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SG115592A1 (en) * | 2002-12-19 | 2005-10-28 | Eta Sa Mft Horlogere Suisse | Energy source mounting device for a portable electronic apparatus |
| CN109143078A (en) * | 2018-08-28 | 2019-01-04 | 中航锂电技术研究院有限公司 | A kind of identification pre-judging method of lithium iron phosphate dynamic battery " diving " failure |
-
2020
- 2020-10-21 CN CN202011131864.9A patent/CN112327192B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247237A (en) * | 2017-05-26 | 2017-10-13 | 联想(北京)有限公司 | A kind of detection method of battery, electronic equipment and computer-readable recording medium |
| SE1851061A1 (en) * | 2018-09-07 | 2018-09-07 | Abb Schweiz Ag | Voltage source converter generates an output voltage having at least three voltage levels |
| CN109596983A (en) * | 2018-11-19 | 2019-04-09 | 清华大学 | The prediction technique that capacity is dived during a kind of cell degradation |
| CN111175653A (en) * | 2020-01-06 | 2020-05-19 | 国网内蒙古东部电力有限公司电力科学研究院 | Method for identifying and prejudging capacity 'water-jumping' fault of ternary battery |
Non-Patent Citations (2)
| Title |
|---|
| Tadiran HLCs extend battery life for oceanographic deployments;Lee Gordon 等;《OCEANS 2016 MTS/IEEE Monterey》;20161201;1-4页 * |
| 三元材料锂离子电池老化诊断、评估与建模方法;高洋;《中国博士学位论文全文数据库工程科技Ⅱ辑》;20200115(第01期);第C042-274页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112327192A (en) | 2021-02-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112327192B (en) | Battery capacity diving phenomenon identification method based on curve form | |
| CN110224192B (en) | Method for predicting service life of power battery by gradient utilization | |
| CN111707951B (en) | Battery pack consistency evaluation method and system | |
| CN112327167B (en) | Battery capacity diving risk assessment method and system | |
| CN112858919B (en) | Battery system online fault diagnosis method and system based on cluster analysis | |
| CN112327191B (en) | Battery diving probability evaluation method based on geometric feature fusion decision | |
| CN116502112B (en) | New energy power supply test data management method and system | |
| CN112327193B (en) | Lithium battery capacity diving early warning method | |
| CN113030761B (en) | Method and system for evaluating battery health state of ultra-large-scale energy storage power station | |
| CN112327194B (en) | Lithium battery capacity diving identification method and device | |
| CN113052464A (en) | Method and system for evaluating reliability of battery energy storage system | |
| CN112287980B (en) | Power battery screening method based on typical feature vector | |
| CN111983464B (en) | Lithium battery lithium separation testing method based on pure electric vehicle | |
| CN114280479A (en) | Electrochemical impedance spectrum-based rapid sorting method for retired batteries | |
| CN111983474A (en) | Lithium ion battery life prediction method and system based on capacity decline model | |
| CN115061049B (en) | Method and system for rapidly detecting UPS battery fault of data center | |
| CN114462259B (en) | SOC-based battery remaining life prediction method and system, automobile and medium | |
| CN110927609B (en) | Decline evaluation method and device for battery energy storage system by utilizing battery in echelon | |
| CN112327189A (en) | KNN algorithm-based energy storage battery health state comprehensive judgment method | |
| Wang et al. | Dynamic early recognition of abnormal lithium-ion batteries before capacity drops using self-adaptive quantum clustering | |
| CN112937369A (en) | Active equalization control method for power battery pack based on Mahalanobis process | |
| CN114974477B (en) | Nickel-iron electrode material optimization method and system for double-flow-direction lightning arrester | |
| CN111178383A (en) | Battery monomer quality classification method based on multi-core support vector machine | |
| CN115937195A (en) | Surface defect detection method and device and computer equipment | |
| CN112485694B (en) | Battery pack detection method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |