Battery characteristics analysis method, device, equipment and storage medium

Abstract

The present invention discloses a battery characteristic analysis method, device, equipment and storage medium, the method comprising: obtaining target ultrasonic characteristics of a target battery; obtaining target charging characteristics of a target battery; inputting the target ultrasonic characteristics and the target charging characteristics into a preset deep LSTM model to obtain the battery characteristics of the target battery. The present invention proposes a lithium-ion battery characteristic analysis method based on ultrasonic energy characteristic fusion, that is, the present invention can obtain rich battery degradation information of the target battery based on the target ultrasonic characteristics and the target charging characteristics, thereby improving the accuracy of subsequent battery characteristic analysis.

Description

Technical Field

[0001] The present invention relates to the field of battery technology, and in particular to a battery characteristic analysis method, device, equipment and storage medium. Background Art

[0002] Lithium-ion batteries are widely used in new energy vehicles, energy storage power stations and 3C consumer fields due to their advantages such as high energy density, long cycle life and low self-discharge rate.

[0003] However, during the use of lithium batteries, with repeated charging and discharging, the positive and negative electrode materials of the battery will gradually age, and the internal electrolyte will continue to be consumed, resulting in a gradual decrease in battery capacity. Usually, the battery capacity needs to be stopped when it declines to a certain stage, and its remaining useful life (Remaining Useful Life, RUL) is the difference between the current number of cycles of the battery and the number of cycles when the capacity declines to 80% of the rated capacity. Accurate battery capacity and RUL analysis facilitates users to maintain and replace batteries in a timely manner, which plays an important role in ensuring the safety and reliability of batteries. Therefore, how to provide a high-precision battery characteristic analysis method has become a problem that needs to be solved urgently.

[0004] The above contents are only used to assist in understanding the technical solution of the present invention and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] The main purpose of the present invention is to provide a battery characteristic analysis method, device, equipment and storage medium, aiming to solve the technical problem of how to provide a high-precision battery characteristic analysis method.

[0006] To achieve the above object, the present invention provides a battery characteristics analysis method, the method comprising:

[0007] Acquire target ultrasonic characteristics of a target battery;

[0008] Acquiring a target charging characteristic of the target battery;

[0009] The target ultrasonic feature and the target charging feature are input into a preset deep LSTM model to obtain the battery characteristics of the target battery.

[0010] Optionally, the step of acquiring the ultrasonic characteristics of the target battery includes:

[0011] Obtaining an initial ultrasonic characteristic of a target battery based on the initial ultrasonic pulse signal;

[0012] Performing correlation matching on the initial ultrasonic features and preset ultrasonic data to obtain sample ultrasonic data;

[0013] The sample ultrasonic data is fitted by the least square method to obtain the target ultrasonic features corresponding to the target battery.

[0014] Optionally, the step of acquiring the target charging characteristic of the target battery includes:

[0015] Performing constant current charging on the target battery to obtain initial charging characteristics of the target battery;

[0016] Performing correlation matching on the initial charging characteristics and preset charging data to obtain sample charging data;

[0017] The sample charging data is fitted by the least square method to obtain the target charging characteristics corresponding to the target battery.

[0018] Optionally, the step of obtaining the initial ultrasonic characteristics of the target battery based on the initial ultrasonic pulse signal includes:

[0019] Sending an initial ultrasonic pulse signal to a first end of a target battery, and acquiring a changing ultrasonic pulse signal outputted from a second end of the target battery;

[0020] The peak-to-valley values ​​and signal waveform areas corresponding to the changing ultrasonic pulse signal are used as initial ultrasonic features of the target battery.

[0021] Optionally, the step of performing constant current charging on the target battery to obtain the initial charging characteristics of the target battery includes:

[0022] Performing constant current charging on the target battery to obtain target charging information corresponding to the target battery during the constant current charging process;

[0023] Determine the target sampling point according to the preset loss information;

[0024] An initial charging characteristic of the target battery is acquired based on the target charging information and the target sampling point.

[0025] Optionally, before inputting the ultrasonic feature and the charging feature into a preset deep LSTM model, the method further includes:

[0026] Construct an initial deep LSTM model through preset ultrasound data and preset charging data;

[0027] The initial deep LSTM model is optimized by an artificial bee colony algorithm to obtain a preset deep LSTM model.

[0028] Optionally, the battery characteristics include: battery capacity and remaining service life; the step of inputting the target ultrasonic characteristics and the target charging characteristics into a preset deep LSTM model to obtain the battery characteristics of the target battery includes:

[0029] Inputting the target ultrasonic feature and the target charging feature into a preset deep LSTM model to obtain the battery capacity of the target battery;

[0030] The remaining service life of the target battery is determined based on the battery capacity and a preset capacity.

[0031] In addition, to achieve the above-mentioned purpose, the present invention also provides a battery characteristic analysis device, the battery characteristic analysis device comprising:

[0032] A first feature acquisition module, used to acquire target ultrasonic features of a target battery;

[0033] A second feature acquisition module, used to acquire a target charging feature of the target battery;

[0034] The characteristic analysis module is used to input the target ultrasonic characteristics and the target charging characteristics into a preset deep LSTM model to obtain the battery characteristics of the target battery.

[0035] In addition, to achieve the above-mentioned purpose, the present invention also proposes a battery characteristic analysis device, which includes: a memory, a processor, and a battery characteristic analysis program stored in the memory and executable on the processor, wherein the battery characteristic analysis program is configured to implement the steps of the battery characteristic analysis method described above.

[0036] In addition, to achieve the above-mentioned purpose, the present invention further proposes a storage medium, on which a battery characteristic analysis program is stored, and when the battery characteristic analysis program is executed by a processor, the steps of the battery characteristic analysis method described above are implemented.

[0037] The present invention discloses a battery characteristic analysis method, device, equipment and storage medium, the method comprising: obtaining a target ultrasonic feature of a target battery; obtaining a target charging feature of a target battery; constructing an initial deep LSTM model by preset ultrasonic data and preset charging data; optimizing the initial deep LSTM model by an artificial bee colony algorithm to obtain a preset deep LSTM model. In addition, the battery characteristics include: battery capacity and remaining service life, inputting the target ultrasonic feature and the target charging feature into a preset deep LSTM model to obtain the battery capacity of the target battery; and determining the remaining service life of the target battery based on the battery capacity and the preset capacity. The present invention can propose a lithium-ion battery capacity and remaining service life analysis method based on ultrasonic energy characteristic fusion for the problem of insufficient analysis accuracy of lithium-ion battery capacity and RUL, that is, the present invention can obtain rich battery degradation information of the target battery based on the target ultrasonic feature and the target charging feature, thereby significantly improving the subsequent battery characteristic analysis accuracy. At the same time, for the problem of insufficient accuracy of short-term capacity analysis of lithium batteries by traditional LSTM models, the present invention can introduce a bidirectional deep LSTM model for battery characteristic analysis, and optimize the bidirectional deep LSTM model by an artificial bee colony algorithm to effectively improve the short-term capacity analysis accuracy of the preset deep LSTM model. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a structural schematic diagram of a battery characteristic analysis device in a hardware operating environment involved in an embodiment of the present invention;

[0039] Figure 2 A first flow chart of a first embodiment of a battery characteristic analysis method according to the present invention;

[0040] Figure 3 A second flow chart of the first embodiment of the battery characteristic analysis method of the present invention;

[0041] Figure 4 A schematic diagram of a battery characteristic analysis process of a first embodiment of a battery characteristic analysis method of the present invention;

[0042] Figure 5 A schematic diagram of a first process flow of a second embodiment of a battery characteristic analysis method of the present invention;

[0043] Figure 6 A schematic diagram of a response to a changing ultrasonic pulse signal according to a second embodiment of a battery characteristic analysis method of the present invention;

[0044] Figure 7 A second flow chart of a second embodiment of the battery characteristic analysis method of the present invention;

[0045] Figure 8 FIG. 1 is a structural block diagram of a first embodiment of a battery characteristic analysis device according to the present invention.

[0046] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0047] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0048] Reference Figure 1 , Figure 1 The figure is a schematic diagram of the structure of a battery characteristic analysis device in the hardware operating environment involved in the embodiment of the present invention.

[0049] like Figure 1 As shown, the battery characteristic analysis device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0050] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the battery characteristic analysis device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0051] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a data storage module, a network communication module, a user interface module, and a battery characteristic analysis program.

[0052] exist Figure 1In the battery characteristic analysis device shown, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the battery characteristic analysis device of the present invention can be set in the battery characteristic analysis device, and the battery characteristic analysis device calls the battery characteristic analysis program stored in the memory 1005 through the processor 1001, and executes the battery characteristic analysis method provided in an embodiment of the present invention.

[0053] The embodiment of the present invention provides a battery characteristic analysis method, referring to Figure 2 , Figure 2 FIG. 1 is a schematic diagram of a first process flow of a first embodiment of a battery characteristic analysis method according to the present invention.

[0054] It is easy to understand that, since the battery historical degradation data contains rich information about the internal material properties and operating conditions of the battery, the data-based method performs performance analysis on the battery historical degradation data without considering the internal electrochemical degradation mechanism of the battery, has strong robustness and a simpler model, and the present embodiment can perform battery characteristic analysis based on this principle. In addition, since the Long Short Term Memory (LSTM) network adds the concept of gates to the ordinary Recurrent Neural Network (RNN) unit to control the RNN unit, the problem of gradient vanishing and gradient exploding in the RNN long sequence training process is solved, so the present embodiment can analyze the battery characteristics based on the battery historical degradation data and LSTM.

[0055] It is understandable that simply analyzing the capacity and RUL of lithium-ion batteries during the recycling process based on historical degradation data and LSTM has the problem of insufficient accuracy and cannot meet application requirements. To solve this problem, this embodiment can propose a lithium-ion battery characteristic method based on ultrasonic energy characteristic fusion. Therefore, in this embodiment, the battery characteristic analysis method includes the following steps:

[0056] Step S10: Acquire target ultrasonic characteristics of the target battery;

[0057] It should be noted that the execution subject of the method of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer or a personal computer, etc., or other electronic devices that can achieve the same or similar functions. Here, the battery characteristic analysis method provided in this embodiment and the following embodiments is specifically described using the above-mentioned battery characteristic analysis device (hereinafter referred to as the analysis device).

[0058] It should be understood that the target battery can be a lithium battery to be analyzed for characteristics. By constructing a lithium-ion battery characteristics analysis model based on LSTM, it is necessary to effectively select the input features of the LSTM model to accurately analyze the battery characteristics of the lithium battery. Therefore, the present embodiment can obtain the target ultrasonic characteristics of the target battery to obtain the battery degradation information of the target battery. Specifically, the present embodiment can use an ultrasonic probe to send a pulse signal to the target battery, and then obtain the target ultrasonic characteristics based on the ultrasonic signal output by the target battery. The target ultrasonic characteristics contain rich battery degradation information, and inputting them into the battery characteristics analysis model can effectively improve the accuracy of the battery characteristics analysis.

[0059] Step S20: Acquire the target charging characteristics of the target battery;

[0060] It is understandable that the charging process of lithium batteries requires constant current followed by constant voltage, and effective battery degradation information can also be contained in the constant current and constant voltage charging process of lithium batteries. Therefore, this embodiment can charge the target battery through a constant current signal or a constant voltage signal, and obtain the target charging characteristics corresponding to the target battery during the constant current charging or constant voltage charging process, so as to subsequently input it into the battery characteristic analysis model to improve the accuracy of the battery characteristic analysis.

[0061] Step S30: Input the target ultrasonic feature and the target charging feature into a preset deep LSTM model to obtain the battery characteristics of the target battery.

[0062] It should be noted that in this embodiment, the above-mentioned preset depth LSTM model can adopt a bidirectional long short-term memory network (Bi-directional Long Short Term Memory, Bi-LSTM) model, which can be divided into two independent LSTM neural networks. The data input to the Bi-LSTM can be input into the two LSTM neural networks in forward order and reverse order for feature extraction, and then the two extracted feature quantities are spliced ​​to form the final feature. The Bi-LSTM model is superior to a single LSTM model in extraction efficiency and performance, and has obvious advantages in the nonlinear prediction of lithium-ion battery capacity and RUL, thereby further improving the accuracy of battery characteristic analysis.

[0063] Further, as an implementation method, the battery characteristics include: battery capacity and remaining service life, such as Figure 3 As shown, Figure 3 This is a second flow chart of the first embodiment of the battery characteristic analysis method of the present invention. In this embodiment, step S30 includes:

[0064] Step S301: inputting the target ultrasonic feature and the target charging feature into a preset deep LSTM model to obtain the battery capacity of the target battery;

[0065] Step S302: determining the remaining service life of the target battery based on the battery capacity and a preset capacity.

[0066] It should be noted that, in this embodiment, the battery characteristics to be analyzed of the target battery may include battery capacity and remaining useful life (Remaining Useful Life, RUL). Therefore, in this embodiment, the ultrasonic characteristics and charging characteristics of the lithium battery can be used as the input sequence, and the battery capacity of the lithium battery can be used as the output to construct a preset deep LSTM model. Therefore, when the target ultrasonic characteristics and target charging characteristics of the target battery are input into the preset deep LSTM model, this embodiment can output the battery capacity of the target battery based on the preset deep LSTM model.

[0067] It is easy to understand that the input information and output information of the preset deep LSTM model can both be sequence data containing time series information. The preset deep LSTM model can perform self-learning (self-computation) based on cyclic calculations to find the relationship between input and output. Therefore, when calculating sequence tasks with long time series dependencies, it is possible to better capture the key information in the input sequence. Accordingly, in this embodiment, the battery capacity of the target battery obtained by the preset deep LSTM model based on the target ultrasonic features and target charging features at the current moment is available as Q d (k+i), i=1,2,...,N1 represents, k represents the time.

[0068] It should be understood that lithium batteries usually need to be retired when their battery capacity declines to 80% of the rated capacity. Therefore, in this embodiment, the above-mentioned preset capacity can be 80% of the initial capacity, but the preset capacity can also be greater than 80% of the initial capacity. The specific value can be adjusted according to the actual situation, and this embodiment does not limit this. If it is assumed that the above-mentioned preset capacity in this embodiment is 80% of the initial capacity, then this embodiment can screen the battery capacity of the target battery output by the preset deep LSTM model based on 80% of the initial capacity. Specifically, if there is a comparison with Q d (k+i) is greater than or equal to 80% of the rated capacity, and Q d When (k+i+1) is less than 80% of the rated capacity, the value of i at this time is the estimated remaining service life of the target battery.

[0069] It is easy to understand that the above-mentioned preset depth LSTM model can be a Bi-LSTM model that has been pre-trained and iterated, that is, training is required before using the preset depth LSTM model to perform battery characteristic analysis of the target battery. Therefore, further, as an implementable method, in this embodiment, before step S30, it also includes;

[0070] Step S21: constructing an initial deep LSTM model through preset ultrasound data and preset charging data;

[0071] It can be understood that the preset ultrasonic data and preset charging data can be lithium battery data collected in advance for different models and different working conditions. Specifically, ultrasonic signals of multiple time nodes can be collected according to a preset sampling period during the charging process of the lithium battery, and the training ultrasonic features corresponding to each sampling time point can be extracted therefrom, that is, the preset ultrasonic data; and then the training charging features corresponding to the lithium battery and the charging process are extracted, that is, the preset charging data. Thus, this embodiment can construct a bidirectional deep LSTM capacity analysis model with preset ultrasonic data and preset charging data as input and battery capacity as output, that is, the initial deep LSTM model.

[0072] Step S22: Optimize the initial deep LSTM model through an artificial bee colony algorithm to obtain a preset deep LSTM model.

[0073] It should be understood that in order to improve the analysis efficiency of the above-mentioned initial deep LSTM model, the present embodiment can optimize the constructed initial deep LSTM model through an artificial bee colony algorithm (ABC algorithm). Specifically, the present embodiment can optimize the number of hidden layers, the number of neurons and the learning rate of the neural network of the initial deep LSTM model through an artificial bee colony algorithm, and then obtain a preset deep LSTM model to effectively improve the short-term capacity analysis accuracy of the preset deep LSTM model.

[0074] In the specific implementation, Figure 4 As shown, Figure 4 This is a schematic diagram of the battery characteristic analysis process of the first embodiment of the battery characteristic analysis method of the present invention. In this embodiment, the target ultrasonic characteristics and target charging characteristics of the target battery can be first obtained, and then the obtained characteristics are input into the bidirectional deep LSTM capacity analysis model optimized by the artificial bee colony algorithm. The bidirectional deep LSTM capacity analysis model outputs the capacity of the target battery accordingly. Finally, the output capacity at different times is analyzed to obtain the corresponding remaining service life of the target battery.

[0075] In summary, this embodiment can propose a lithium-ion battery capacity and remaining service life analysis method based on ultrasonic energy characteristic fusion to address the problem of insufficient accuracy of lithium-ion battery capacity and RUL analysis. That is, this embodiment can obtain rich battery degradation information of the target battery based on the target ultrasonic characteristics and target charging characteristics, thereby significantly improving the accuracy of subsequent battery characteristic analysis. At the same time, in order to address the problem of insufficient accuracy of lithium battery short-term capacity analysis using the traditional LSTM model, this embodiment can introduce a bidirectional deep LSTM model for battery characteristic analysis, and optimize the bidirectional deep LSTM model through an artificial bee colony algorithm to effectively improve the short-term capacity analysis accuracy of the preset deep LSTM model.

[0076] The present embodiment discloses a battery characteristic analysis method, which includes: obtaining a target ultrasonic feature of a target battery; obtaining a target charging feature of a target battery; constructing an initial deep LSTM model through preset ultrasonic data and preset charging data; optimizing the initial deep LSTM model through an artificial bee colony algorithm to obtain a preset deep LSTM model. In addition, the battery characteristics include: battery capacity and remaining service life, inputting the target ultrasonic feature and the target charging feature into a preset deep LSTM model to obtain the battery capacity of the target battery; and determining the remaining service life of the target battery based on the battery capacity and the preset capacity. In view of the problem of insufficient analysis accuracy of lithium-ion battery capacity and RUL, the present embodiment can propose a lithium-ion battery capacity and remaining service life analysis method based on ultrasonic energy characteristic fusion, that is, the present embodiment can obtain rich battery degradation information of the target battery based on the target ultrasonic feature and the target charging feature, thereby significantly improving the subsequent battery characteristic analysis accuracy. At the same time, in view of the problem of insufficient accuracy of short-term capacity analysis of lithium batteries by traditional LSTM models, the present embodiment can introduce a bidirectional deep LSTM model for battery characteristic analysis, and optimize the bidirectional deep LSTM model through an artificial bee colony algorithm to effectively improve the short-term capacity analysis accuracy of the preset deep LSTM model.

[0077] Reference Figure 5 , Figure 5 This is a schematic diagram of the first process of the second embodiment of the battery characteristic analysis method of the present invention. Figure 2 The illustrated embodiment provides a second embodiment of the battery characteristic analysis method of the present invention.

[0078] As an implementable method, in this embodiment, step S10 includes:

[0079] Step S101: obtaining an initial ultrasonic feature of a target battery based on an initial ultrasonic pulse signal;

[0080] It is easy to understand that the initial ultrasonic pulse signal may be a pulse signal sent to one end of the target battery by an ultrasonic probe. After the initial ultrasonic pulse signal propagates inside the target battery, it may be output from the other end of the target battery, so that the degradation information inside the target battery may be obtained based on the ultrasonic pulse signal output from the other end of the target battery. Therefore, further, as an implementable method, in this embodiment, step S101 includes:

[0081] Step S101a: sending an initial ultrasonic pulse signal to the first end of the target battery, and acquiring a changing ultrasonic pulse signal output by the second end of the target battery;

[0082] It is understandable that in this embodiment, ultrasonic probes can also be placed at both ends of the target battery. The first ultrasonic probe is used to send a pulse signal to one end of the target battery, and the second ultrasonic probe can be used to receive the pulse signal output by the target battery, that is, the above-mentioned changing ultrasonic pulse signal.

[0083] Step S101b: taking the peak-to-valley values ​​and signal waveform areas corresponding to the changed ultrasonic pulse signal as the initial ultrasonic features of the target battery.

[0084] It is important to understand that the remaining capacity of a lithium-ion battery reflects the aging of the lithium-ion battery, along with lithium deposition, relaxation of the electrode material, etc. All of these conditions will cause the Young's modulus of the electrode material to change, and the acoustic impedance calculation formula is as follows:

[0085]

[0086] In the formula, Z ultra is the acoustic impedance, ρ is the material density, and E is the material Young's modulus.

[0087] It can be seen from the above formula that the material density and Young's modulus determine the size of the acoustic impedance, which in turn affects the energy attenuation of ultrasound passing through the medium, that is, ultrasound is very sensitive to changes in material properties. Therefore, this embodiment can obtain rich degradation information inside the target battery by sending an ultrasonic pulse signal to the target battery.

[0088] Specifically, in this embodiment, the sampling period of the variable ultrasonic pulse signal can be set to 6us, and 1200 sampling points can be obtained (the sampling period and the number of sampling points can be set according to actual conditions, and this embodiment does not impose any restrictions on this). At this time, the received complete response signal can be f(t), t = 1, 2, ..., 1200. For ease of understanding, Figure 6 Take this as an example to illustrate. Figure 6 This is a schematic diagram of the response of the ultrasonic pulse signal according to the second embodiment of the battery characteristic analysis method of the present invention. Figure 6The horizontal axis is the sampling period, and the vertical axis is the signal amplitude. The overall trend of the response waveform of the changing pulse signal is to rise first, reach a peak, and then fall. In addition, Figure 6 In the figure, the solid line curve corresponds to the ultrasonic pulse response of the 10th cycle, and the dotted line curve corresponds to the ultrasonic pulse response of the 200th cycle. Therefore, it can be seen that with the increase in the number of charge and discharge cycles, the ultrasonic peak-to-valley value of the changing ultrasonic pulse signal gradually decreases, and the area of ​​the ultrasonic response signal waveform also gradually decreases, where the area of ​​the ultrasonic waveform is the sum of the areas of 1200 sampling points. That is, the signal waveform area and peak-to-valley value of the changing ultrasonic pulse signal are highly correlated with the number of battery cycles, that is, the signal waveform area and peak-to-valley value of the changing ultrasonic pulse signal are closely related to the degree of battery aging. Therefore, this embodiment can select the peak-to-valley value as the battery health feature 1, and express it as FV, ​​which is calculated as follows:

[0089] FV=|maxf(t)-min f(t)|,t=1,2,....,1200;

[0090] At the same time, the waveform area of ​​the ultrasonic response signal is used as the battery health feature 2 and is expressed as Area, which is calculated as follows:

[0091]

[0092] The battery health feature 1 and the battery health feature 2 are the initial ultrasonic features.

[0093] Step S102: performing correlation matching on the initial ultrasonic features and preset ultrasonic data to obtain sample ultrasonic data;

[0094] It should be understood that the above-mentioned preset ultrasonic data can be the ultrasonic characteristics of lithium batteries collected in advance for different models and different working conditions. Specifically, this embodiment can respectively collect the corresponding ultrasonic pulse signals from k=1 to the cycle capacity to 70%, and collect the changing ultrasonic pulse signals corresponding to the output of lithium batteries of different models and different working conditions based on the preset sampling period (such as 10s) during charging, and then determine the ultrasonic characteristics corresponding to the jth time slice at the kth moment corresponding to different models and different working conditions based on the collected changing pulse ultrasonic signals. The pre-collected ultrasonic characteristics can be expressed as: [FV(k,j), Area(k,j)], j=1,2,...,N2; k=1,2,...,N3, and finally they are summarized as preset ultrasonic data, and the preset ultrasonic data is stored in the database for subsequent calls.

[0095] It is understandable that in order to ensure the accuracy of the analysis of battery characteristics, the above ultrasonic characteristic analysis needs to be based on the entire life cycle of the battery. However, the lithium-ion battery ultrasonic signal corresponding to the target battery to be measured before detection is usually not collected. Therefore, this embodiment can first perform constant voltage and constant current charging on the target battery, and determine the peak and valley values ​​and waveform area corresponding to the target battery during the charging process, and obtain the initial ultrasonic characteristics corresponding to the jth time slice at the kth moment corresponding to the target battery, and express the initial ultrasonic characteristics as: [FV(k,j),Area(k,j)], j=1,2,...,N4.

[0096] Then, this embodiment can match the initial ultrasonic features corresponding to the target battery and the preset ultrasonic data in the database according to the correlation, and use [FV(k,j), Area(k,j)], j = 1, 2, ..., N2; k = 1, 2, ..., N3 in the preset ultrasonic data with the highest correlation with the initial ultrasonic features as the above-mentioned sample ultrasonic data, and obtain the sample peak and valley values ​​FV(i), i = 1, 2, ..., k and the sample waveform area Area(i), i = 1, 2, ..., k at and before time k corresponding to the target battery.

[0097] Step S103: fitting the sample ultrasonic data by the least square method to obtain the target ultrasonic features corresponding to the target battery.

[0098] It should be noted that, in this embodiment, according to the peak-valley values ​​FV(i) of the sample in the historical data set at time k and before in the sample ultrasound data, i=1,2,...,k, a sample set {i,FV(i)} is constructed, where i is the number of cycles, and a quadratic polynomial of peak-valley values ​​is constructed for the data set:

[0099] FV(i)=b0+b1(i)+b2(i) 2 ;

[0100] Then, this embodiment can solve the above [b0, b1, b2] by the least square method. Then, this embodiment can substitute the number of cycles i into the peak-to-valley value quadratic polynomial to analyze the subsequent sound wave peak-to-valley value FV(i) at time k, i=k, k+1.

[0101] As above, this embodiment can construct a sample set {i, Area(i)} based on the historical data set Area(i) before and at time k according to the sample ultrasound data, i=1,2,...,k, where i is the number of cycles, and construct a waveform area quadratic polynomial for the data set:

[0102] Area(i)=r0+r1(i)+r2(i) 2 ;

[0103] Then, this embodiment can also solve [r0, r1, r2] by the least square method. Substituting the number of cycles i into the quadratic polynomial of the waveform area can analyze the subsequent ultrasonic area characteristic value Area(i) at time k, i=k, k+1.

[0104] It is understandable that, in addition to the above ultrasonic pulse signal, this embodiment can also perform battery characteristic analysis based on the charging information of the target battery. Therefore, further, as an implementable method, Figure 8 As shown, Figure 7 This is a second flow chart of the second embodiment of the battery characteristic analysis method of the present invention. In this embodiment, step S20 includes:

[0105] Step S201: performing constant current charging on the target battery to obtain initial charging characteristics of the target battery;

[0106] It is easy to understand that the battery charging process includes two stages: constant current (Constant Current) and constant voltage (Constant Voltage), that is, the battery is first charged with a constant current to the charging cut-off voltage, and then charged with a constant voltage until the current drops to the charging cut-off current. The constant current and constant voltage charging process of the lithium battery can also contain effective battery degradation information. During the experiment, it was found that as the number of battery cycles increases, the time required for the battery voltage to rise to the cut-off voltage gradually shortens. Therefore, this embodiment preferably obtains the initial charging characteristics of the target battery based on the constant current charging signal of the lithium battery. Therefore, further, as an implementable method, in this embodiment, step S201 includes:

[0107] Step S201a: performing constant current charging on the target battery, and obtaining target charging information corresponding to the target battery during the constant current charging process;

[0108] Step S201b: determining a target sampling point according to preset loss information;

[0109] Step S201c: Acquire the initial charging characteristics of the target battery based on the target charging information and the target sampling point.

[0110] It is easy to understand that the charging information of the above-mentioned target battery can be the constant current charging voltage curve corresponding to each cycle number of the target battery. In addition, it was found during the experiment that the time required for the battery voltage of the lithium battery to rise to the cut-off voltage will gradually shorten with the increase in the number of battery cycles, especially in the range of 3.40 to 3.75V, that is, the above-mentioned preset loss information. Therefore, this embodiment can use the points corresponding to 3.40V and 3.75V as the above-mentioned target sampling points. Finally, this embodiment can use the time required for the voltage of the corresponding target battery in the constant current charging voltage curve to rise from 3.40V to 3.75V as the initial charging characteristics corresponding to the target battery.

[0111] Specifically, this embodiment can collect and obtain the voltage signal g(x) of the target battery during the constant current charging process, x=1, 2, ..., N5, the sampling frequency can be set to 100 Hz, and the number of sampling points required for the battery voltage of the target battery to rise from 3.40V to 3.75V is taken as the characteristic value, and the initial charging characteristic can be expressed as:

[0112] T cr =x2-x1;

[0113] Wherein, x1 is the time corresponding to the number of sampling points when g(x), x=1, 2, ..., N5 is 3.40V; x2 is the time corresponding to the number of sampling points when g(x), x=1, 2, ..., N5 is 3.75V.

[0114] Step S202: performing correlation matching on the initial charging characteristics and preset charging data to obtain sample charging data;

[0115] Step S203: fitting the sample charging data by the least square method to obtain the target charging characteristics corresponding to the target battery.

[0116] It is easy to understand that, similar to the acquisition process of the target ultrasonic characteristics of the target battery, in this embodiment, the charging characteristics of lithium batteries of different models and working conditions can be collected in advance, and the pre-collected charging characteristics can be summarized as the preset charging data and stored in the database. After the initial charging characteristics corresponding to the target battery are obtained, the present embodiment can perform correlation matching on the initial charging characteristics and the preset charging data, and according to the matching result, the charging data with the highest correlation with the target battery in the preset charging data is used as the sample charging data, which can be expressed as T cr (i),i=kd,k-d+1,...,k.

[0117] Then, this embodiment can construct a sample set {i,T cr(i)}, i = kd, k-d+1, ..., k, i is the number of cycles, and a quadratic polynomial of charging time is constructed for the sample set:

[0118] T cr (i) = a0 + a1 (i) + a2 (i) 2 ;

[0119] Similarly, in this embodiment, [a0, a1, a2] can be solved by the least square method, and the rise time characteristic value of the subsequent constant current charging voltage curve segment after time k can be analyzed by substituting the cycle number i into the quadratic polynomial of the charging time.

[0120] In summary, in view of the fact that the degradation characteristics of lithium-ion batteries are not obvious, this embodiment collects the battery ultrasonic signals of the target battery at multiple time nodes, and extracts the peak value FV and waveform area value Area of ​​each time node; and then extracts the specific rise time characteristic value T of the constant current charging voltage curve segment. cr , fully excavated the three characteristic quantities that are highly correlated with battery degradation at the target battery k moment, and obtained rich battery degradation information of the target battery. In addition, after extracting the three typical features of the target battery at the k moment, this embodiment can screen out the sample ultrasonic data and sample charging data corresponding to the target battery by matching the preset ultrasonic data and preset charging data in the database, and optimize and fit the subsequent ultrasonic peak characteristics, subsequent waveform area characteristics and subsequent rise time characteristics corresponding to the target battery by the least squares method, that is, the target ultrasonic characteristics and target charging characteristics corresponding to the target battery, thereby improving the subsequent ultrasonic peak-to-peak value FV, ultrasonic area Area and rise time T of the kth cycle. cr As input, the current capacity Q(k) of the kth cycle is used as output to build a preset deep LSTM model to analyze the capacity and RUL analysis accuracy of the target battery.

[0121] In this embodiment, an initial ultrasonic pulse signal is sent to the first end of the target battery to obtain a change ultrasonic pulse signal output from the second end of the target battery; the peak-to-valley value and signal waveform area corresponding to the change ultrasonic pulse signal are used as the initial ultrasonic feature of the target battery. The initial ultrasonic feature and the preset ultrasonic data are matched for correlation to obtain sample ultrasonic data; the sample ultrasonic data are fitted by the least squares method to obtain the target ultrasonic feature corresponding to the target battery. The target battery is charged with a constant current to obtain the target charging information corresponding to the target battery during the constant current charging process; the target sampling point is determined according to the preset loss information; the initial charging feature of the target battery is obtained based on the target charging information and the target sampling point. The initial charging feature and the preset charging data are matched for correlation to obtain sample charging data; the sample charging data are fitted by the least squares method to obtain the target charging feature corresponding to the target battery. In view of the fact that the degradation feature of lithium-ion batteries is not obvious, this embodiment collects the battery ultrasonic signals of multiple time nodes of the target battery, and extracts the peak value and waveform area value of each time node; then extracts the specific rise time feature value of the constant current charging voltage curve segment, fully excavates the feature quantity highly correlated with the battery degradation of the target battery, and obtains rich battery degradation information of the target battery. In addition, this embodiment can further filter out sample ultrasonic data and sample charging data corresponding to the target battery by matching the preset ultrasonic data and preset charging data in the database, and optimize and fit the target ultrasonic features and target charging features corresponding to the target battery through the least squares method to improve the analysis accuracy of the preset deep LSTM model.

[0122] In addition, an embodiment of the present invention further provides a storage medium, on which a battery characteristic analysis program is stored. When the battery characteristic analysis program is executed by a processor, the steps of the battery characteristic analysis method described above are implemented.

[0123] refer to Figure 8 , Figure 8 FIG. 1 is a structural block diagram of a first embodiment of a battery characteristic analysis device according to the present invention.

[0124] like Figure 8 As shown, the battery characteristic analysis device proposed in the embodiment of the present invention includes:

[0125] A first feature acquisition module 801 is used to acquire target ultrasonic features of a target battery;

[0126] A second feature acquisition module 802 is used to acquire a target charging feature of the target battery;

[0127] The characteristic analysis module 803 is used to input the target ultrasonic characteristics and the target charging characteristics into a preset deep LSTM model to obtain the battery characteristics of the target battery.

[0128] Further, as an implementable method, in this embodiment, the characteristic analysis module 803 is also used to construct an initial deep LSTM model through preset ultrasound data and preset charging data;

[0129] The characteristic analysis module 803 is also used to optimize the initial deep LSTM model through an artificial bee colony algorithm to obtain a preset deep LSTM model.

[0130] Further, as an implementable embodiment, the battery characteristics include: battery capacity and remaining service life. In this embodiment, the characteristic analysis module 803 is also used to input the target ultrasonic characteristics and the target charging characteristics into a preset deep LSTM model to obtain the battery capacity of the target battery;

[0131] The characteristic analysis module 803 is further configured to determine the remaining service life of the target battery based on the battery capacity and a preset capacity.

[0132] This embodiment obtains the target ultrasonic characteristics of the target battery; obtains the target charging characteristics of the target battery; constructs an initial deep LSTM model through preset ultrasonic data and preset charging data; optimizes the initial deep LSTM model through the artificial bee colony algorithm to obtain a preset deep LSTM model. In addition, the battery characteristics include: battery capacity and remaining service life, input the target ultrasonic characteristics and the target charging characteristics into the preset deep LSTM model to obtain the battery capacity of the target battery; and determine the remaining service life of the target battery based on the battery capacity and the preset capacity. This embodiment can propose a lithium-ion battery capacity and remaining service life analysis method based on ultrasonic energy characteristics fusion to address the problem of insufficient analysis accuracy of lithium-ion battery capacity and RUL, that is, this embodiment can obtain rich battery degradation information of the target battery based on the target ultrasonic characteristics and the target charging characteristics, thereby significantly improving the subsequent battery characteristics analysis accuracy. At the same time, in view of the problem of insufficient accuracy of short-term capacity analysis of lithium batteries by traditional LSTM models, this embodiment can introduce a bidirectional deep LSTM model for battery characteristics analysis, and optimize the bidirectional deep LSTM model through the artificial bee colony algorithm to effectively improve the short-term capacity analysis accuracy of the preset deep LSTM model.

[0133] Based on the first embodiment of the battery characteristic analysis device of the present invention, a second embodiment of the battery characteristic analysis device of the present invention is proposed.

[0134] As an implementable method, in this embodiment, the first feature acquisition module 801 is further used to obtain the initial ultrasonic feature of the target battery based on the initial ultrasonic pulse signal;

[0135] The first feature acquisition module 801 is further used to perform correlation matching between the initial ultrasound feature and the preset ultrasound data to obtain sample ultrasound data;

[0136] The first feature acquisition module 801 is further used to fit the sample ultrasonic data by the least square method to obtain the target ultrasonic feature corresponding to the target battery.

[0137] Further, as an implementable embodiment, in this embodiment, the first feature acquisition module 801 is also used to send an initial ultrasonic pulse signal to the first end of the target battery to obtain a change ultrasonic pulse signal output by the second end of the target battery;

[0138] The first feature acquisition module 801 is further configured to use the peak-to-valley value and the signal waveform area corresponding to the changed ultrasonic pulse signal as the initial ultrasonic feature of the target battery.

[0139] Further, as an implementable embodiment, in this embodiment, the second feature acquisition module 802 is also used to perform constant current charging on the target battery to obtain the initial charging feature of the target battery;

[0140] The second feature acquisition module 802 is further used to perform correlation matching on the initial charging feature and the preset charging data to obtain sample charging data;

[0141] The second feature acquisition module 802 is further used to fit the sample charging data by the least square method to obtain the target charging feature corresponding to the target battery.

[0142] Further, as an implementable embodiment, in this embodiment, the second feature acquisition module 802 is also used to perform constant current charging on the target battery to obtain target charging information corresponding to the target battery during the constant current charging process;

[0143] The second feature acquisition module 802 is further used to determine the target sampling point according to the preset loss information;

[0144] The second feature acquisition module 802 is further configured to acquire the initial charging feature of the target battery based on the target charging information and the target sampling point.

[0145] In this embodiment, an initial ultrasonic pulse signal is sent to the first end of the target battery to obtain a change ultrasonic pulse signal output from the second end of the target battery; the peak-to-valley value and signal waveform area corresponding to the change ultrasonic pulse signal are used as the initial ultrasonic feature of the target battery. The initial ultrasonic feature and the preset ultrasonic data are matched for correlation to obtain sample ultrasonic data; the sample ultrasonic data are fitted by the least squares method to obtain the target ultrasonic feature corresponding to the target battery. The target battery is charged with a constant current to obtain the target charging information corresponding to the target battery during the constant current charging process; the target sampling point is determined according to the preset loss information; the initial charging feature of the target battery is obtained based on the target charging information and the target sampling point. The initial charging feature and the preset charging data are matched for correlation to obtain sample charging data; the sample charging data are fitted by the least squares method to obtain the target charging feature corresponding to the target battery. In view of the fact that the degradation feature of lithium-ion batteries is not obvious, this embodiment collects the battery ultrasonic signals of multiple time nodes of the target battery, and extracts the peak value and waveform area value of each time node; then extracts the specific rise time feature value of the constant current charging voltage curve segment, fully excavates the feature quantity highly correlated with the battery degradation of the target battery, and obtains rich battery degradation information of the target battery. In addition, this embodiment can further filter out sample ultrasonic data and sample charging data corresponding to the target battery by matching the preset ultrasonic data and preset charging data in the database, and optimize and fit the target ultrasonic features and target charging features corresponding to the target battery through the least squares method to improve the analysis accuracy of the preset deep LSTM model.

[0146] Other embodiments or specific implementations of the battery characteristic analysis device of the present invention can refer to the above-mentioned method embodiments, which will not be described in detail here.

[0147] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0148] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0149] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0150] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A battery characteristics analysis method, characterized in that: The battery characteristic analysis method comprises: S1, obtaining the target ultrasonic characteristics of the target battery; the steps include: S11, obtaining the initial ultrasonic characteristics of the target battery based on the initial ultrasonic pulse signal; the steps include: S111, sending an initial ultrasonic pulse signal to a first end of a target battery, and acquiring a changed ultrasonic pulse signal output by a second end of the target battery; S112, using the peak-to-valley value and signal waveform area corresponding to the changed ultrasonic pulse signal as the initial ultrasonic feature of the target battery; S12, performing correlation matching on the initial ultrasonic features and preset ultrasonic data to obtain sample ultrasonic data; S13, fitting the sample ultrasonic data by the least square method to obtain the target ultrasonic features corresponding to the target battery; S2, obtaining the target charging characteristics of the target battery; the steps include: S21, performing constant current charging on the target battery to obtain initial charging characteristics of the target battery; the steps include: S211, performing constant current charging on the target battery, and obtaining target charging information corresponding to the target battery during the constant current charging process; S212, determining a target sampling point according to preset loss information; S213, acquiring an initial charging characteristic of the target battery based on the target charging information and the target sampling point; S22, performing correlation matching on the initial charging characteristics and preset charging data to obtain sample charging data; S23, fitting the sample charging data by the least square method to obtain the target charging characteristics corresponding to the target battery; S3, inputting the target ultrasonic feature and the target charging feature into a preset deep LSTM model to obtain the battery characteristics of the target battery; Before inputting the ultrasonic feature and the charging feature into the preset deep LSTM model, the method further includes: constructing an initial deep LSTM model through preset ultrasonic data and preset charging data; optimizing the initial deep LSTM model through an artificial bee colony algorithm to obtain a preset deep LSTM model; The battery characteristics include: battery capacity and remaining service life; the step of inputting the target ultrasonic characteristics and the target charging characteristics into a preset deep LSTM model to obtain the battery characteristics of the target battery includes: inputting the target ultrasonic characteristics and the target charging characteristics into a preset deep LSTM model to obtain the battery capacity of the target battery; determining the remaining service life of the target battery based on the battery capacity and the preset capacity.

2. A battery characteristic analysis device based on the battery characteristic analysis method according to claim 1, characterized in that: The battery characteristic analysis device comprises: A first feature acquisition module, used to acquire target ultrasonic features of a target battery; A second feature acquisition module, used to acquire a target charging feature of the target battery; The characteristic analysis module is used to input the target ultrasonic characteristics and the target charging characteristics into a preset deep LSTM model to obtain the battery characteristics of the target battery.

3. A battery characteristic analysis device, characterized in that: The device comprises: a memory, a processor, and a battery characteristic analysis program stored in the memory and executable on the processor, wherein the battery characteristic analysis program is configured to implement the steps of the battery characteristic analysis method according to claim 1.

4. A storage medium, characterized in that: The storage medium stores a battery characteristic analysis program. When the battery characteristic analysis program is executed by a processor, the steps of the battery characteristic analysis method according to claim 1 are implemented.

Images