Lithium battery life detection method, device and medium based on SDP aging profile and CGAN

Abstract

The present invention discloses a lithium battery life detection method, device and storage medium based on SDP aging profile and CGAN; the detection method comprises: generating a symmetrical point pattern SDP aging profile sample set based on acquired random walk charge and discharge data; building an initial condition generative adversarial network model CGAN; using the SDP aging profile sample set to iteratively train the initial condition generative adversarial network model CGAN to obtain a target condition generative adversarial network model CGAN; based on the target condition generative adversarial network model CGAN, identifying the corresponding life state and type of satellite lithium batteries.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of lithium battery operating status measurement, and in particular to a lithium battery life detection method, device and storage medium based on a symmetric dot pattern (SDP) aging profile and a conditional generative adversarial network (CGAN). Background Art

[0002] Lithium-ion batteries (hereinafter referred to as "lithium batteries") have the advantages of long cycle life, light weight, and high reliability. They have been widely used in aerospace equipment such as satellites and space stations. The lithium battery system is the only energy source for satellites during the shadow period. If the lithium battery ages, the satellite will not work properly, which will lead to serious safety accidents and huge economic losses. Understandably, as the use time increases, the charge capacity of the lithium battery will continue to decrease and the internal resistance will continue to increase, affecting the performance of the lithium battery. Therefore, the state of life (SOL) detection of lithium batteries is an important task in spacecraft fault diagnosis. The traditional detection method of the service life state of lithium batteries is mainly based on charge capacity measurement. Specifically, the lithium battery needs to complete a complete charge and discharge process regularly and detect the degree of charge capacity attenuation to determine the service life state of the lithium battery. The detection time is relatively long. In addition, due to the complexity of the spacecraft working environment, the working state of the lithium battery often needs to be switched quickly, and the working intensity varies with time. Therefore, there may not be enough time to complete a standard and complete charge and discharge operation process. Summary of the invention

[0003] In view of this, the embodiments of the present invention hope to provide a lithium battery life detection method, device and storage medium based on SDP aging profile and CGAN; it is capable of detecting the service life status of satellite lithium batteries in real time and providing a reference for the maintenance and replacement of satellite lithium batteries.

[0004] The technical solution of the embodiment of the present invention is achieved as follows:

[0005] In a first aspect, an embodiment of the present invention provides a lithium battery life detection method based on SDP aging profile and CGAN, the detection method comprising:

[0006] Generate a symmetric point mode SDP aging face spectrum sample set based on the acquired random walk charging and discharging data;

[0007] Build the initial condition generative adversarial network model CGAN;

[0008] Iteratively training the initial conditional generative adversarial network model CGAN using the SDP aging face spectrum sample set to obtain a target conditional generative adversarial network model CGAN;

[0009] Based on the target conditions, a generative adversarial network model CGAN is used to identify the life status and type of satellite lithium batteries.

[0010] In a second aspect, an embodiment of the present invention provides a lithium battery life detection device based on SDP aging facial profile and CGAN, the detection device comprising: a generation part, a construction part, a training part and a recognition part; wherein,

[0011] The generating part is configured to generate a symmetrical point mode SDP aging facial spectrum sample set based on the acquired random walk charging and discharging data;

[0012] The building part is configured to build an initial condition generative adversarial network model CGAN;

[0013] The training part is configured to iteratively train the initial conditional generative adversarial network model CGAN using the SDP aging face spectrum sample set to obtain a target conditional generative adversarial network model CGAN;

[0014] The identification part is configured to generate a confrontation network model CGAN based on the target condition to identify the life status and type corresponding to the satellite lithium battery.

[0015] In a third aspect, an embodiment of the present invention provides a computing device, the computing device comprising: a communication interface, a memory and a processor; each component is coupled together through a bus system; wherein,

[0016] The communication interface is used to receive and send signals during the process of sending and receiving information with other external network elements;

[0017] The memory is used to store a computer program that can be run on the processor;

[0018] The processor is used to execute the steps of the lithium battery life detection method based on SDP aging profile and CGAN described in the first aspect when running the computer program.

[0019] In a fourth aspect, an embodiment of the present invention provides a storage medium, wherein the storage medium stores a lithium battery life detection program based on the SDP aging profile and CGAN, and when the lithium battery life detection program based on the SDP aging profile and CGAN is executed by at least one processor, the steps of the lithium battery life detection method based on the SDP aging profile and CGAN described in the first aspect are implemented.

[0020] The embodiments of the present invention provide a lithium battery life detection method, device and storage medium based on SDP aging spectrum and CGAN; by generating a lithium battery SDP aging spectrum from lithium battery voltage and current signal data, and using CGAN to perform data enhancement and classification on the lithium battery SDP aging spectrum, real-time detection of the service life status of satellite lithium batteries is achieved, providing a reference for the maintenance and replacement of satellite lithium batteries. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flow chart of a lithium battery life detection method based on SDP aging profile and CGAN provided in an embodiment of the present invention;

[0022] Figure 2 A statistical diagram of the number of working time sections under each working state provided by an embodiment of the present invention;

[0023] Figure 3 A schematic diagram of a satellite lithium battery charge capacity variation trend under a reference charge and discharge working state provided by an embodiment of the present invention;

[0024] FIG4 (a) is a schematic diagram of a voltage variation trend of a satellite lithium battery in the first 50 working time segments in a random walk charge-discharge working state provided by an embodiment of the present invention;

[0025] FIG4 (b) is a schematic diagram of a satellite lithium battery current variation trend in the first 50 working time segments in a random walk charge-discharge working state provided by an embodiment of the present invention;

[0026] FIG5 (a) is a schematic diagram of a voltage variation trend of a satellite lithium battery in the last 50 working time segments in a random walk charge-discharge working state provided by an embodiment of the present invention;

[0027] FIG5( b ) is a schematic diagram of a satellite lithium battery current variation trend in the last 50 working time segments in a random walk charge-discharge working state provided by an embodiment of the present invention;

[0028] Figure 6 A schematic diagram of SDP for generating data of a satellite lithium battery constant current random walk charge and discharge working time segment provided by an embodiment of the present invention;

[0029] Figure 7 A schematic diagram of the SDP aging spectrum of random walk charging and discharging in each time period of group B provided by an embodiment of the present invention;

[0030] Figure 8 A schematic diagram of iterative training of the CGAN model provided in an embodiment of the present invention;

[0031] Fig. 9 The SDP effect diagram randomly generated by the target generator provided in the embodiment of the present invention (training cycle = 50,000 times);

[0032] Fig.10 A schematic diagram of the test results of a simulation experiment of Group A provided in an embodiment of the present invention;

[0033] Fig.11 A schematic diagram of test results of a simulation experiment of group B provided in an embodiment of the present invention;

[0034] Fig.12 A schematic diagram of a lithium battery life detection device based on SDP aging profile and CGAN provided in an embodiment of the present invention;

[0035] Fig.13 A schematic diagram of a specific hardware structure of a computing device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0037] At present, the traditional detection methods for the detection of the service life status of satellite lithium batteries are mainly based on charge capacity measurement and internal resistance measurement methods. For example, the charge capacity measurement method requires the lithium battery to perform a complete charge and discharge operation. Taking an 18650 lithium battery with a charge capacity of 2Ah as an example, when the discharge current of the lithium battery is 0.05A, it can be continuously discharged for nearly 40 hours; when the reference discharge (RD) current is 1A, it needs to be continuously discharged for about 2 hours. The detection time is long. Under complex working conditions, the lithium battery may not have enough time to perform a complete charge and discharge operation process. On the other hand, the detection method of lithium battery internal resistance measurement is greatly affected by the working state of the lithium battery and requires offline measurement. This is because the lithium battery can be regarded as a resistor-capacitance circuit (RC) model. When performing rapid charge and discharge operations, the internal resistance of the battery measured in the loop will change due to the dynamic characteristics of the lithium battery. In addition, the lithium battery on the orbiting satellite can only send voltage and current telemetry signals to the ground base station, and cannot perform offline measurement of circuit components.

[0038] The decrease in charge capacity and increase in internal resistance of lithium batteries can both provide feedback on the aging status of satellite lithium batteries, but the traditional life detection method for satellite lithium batteries cannot comprehensively evaluate these two detection indicators. Therefore, the embodiment of the present invention is expected to provide a satellite lithium battery life detection technology solution that can quickly and conveniently detect the service life status of satellite lithium batteries and provide a reference for the maintenance and replacement of satellite lithium batteries.

[0039] See also Figure 1, which shows a lithium battery life detection method based on SDP aging profile and CGAN provided by an embodiment of the present invention, which specifically includes:

[0040] S101, generating a symmetrical point mode SDP aging face spectrum sample set based on the acquired random walk charge and discharge data;

[0041] S102. Build the initial condition generative adversarial network model CGAN;

[0042] S103, using the SDP aging face spectrum sample set to iteratively train the initial conditional generative adversarial network model CGAN to obtain a target conditional generative adversarial network model CGAN;

[0043] S104. Generate a confrontation network model CGAN based on the target condition to identify the life status and type of the satellite lithium battery.

[0044] for Figure 1 In some possible implementations of the technical solution shown, the random walk charging and discharging data includes:

[0045] When the satellite lithium battery does not have enough time to complete a complete charge and discharge operation, a constant current random walk charge and discharge operation is performed on the satellite lithium battery, and the obtained voltage signal is used as random walk charge and discharge data;

[0046] When the satellite lithium battery has enough time to complete a complete charge and discharge operation, the voltage and current signals obtained are used as the random walk charge and discharge data.

[0047] It should be noted that in the specific implementation process of the present invention, the random walk charge and discharge experimental data of the satellite lithium battery of NASA Ames Experiment Center is used as the random walk charge and discharge data for the satellite lithium battery life detection in the embodiment of the present invention. It should be noted here that when obtaining the random walk charge and discharge data of the satellite lithium battery, multiple lithium batteries of the same model, such as four 18650 lithium batteries, can be used, and a series of continuous charge and discharge operations are performed using a current between -4.5A and 4.5A. This type of continuous charge and discharge operation is called a random walk charge and discharge operation. In order to obtain more accurate random walk charge and discharge data of satellite lithium batteries, the above-mentioned random walk charge and discharge experiment needs to perform charge and discharge operations according to the set charging cycle and discharge cycle within the set time period. For example, the set time period is from January to June, a total of 6 months, each working time segment lasts about 5 minutes, and a series of reference charge and discharge cycles are performed after 1500 cycles (about 5 days), so as to provide a reference benchmark for the service life status of the satellite lithium battery based on the acquired experimental data.

[0048] Specifically, the charge and discharge operations of satellite lithium batteries are divided into four working states: low-current discharge, pulse charge and discharge, random walk charge and discharge, and reference charge and discharge. The number of working time segments under each working state obtained through the above experiments is as follows: Figure 2 As shown by Figure 2 It can be seen that the satellite lithium battery is mainly in the random walk charge and discharge working state, and this working state is distributed between other working states; however, it should be noted that the charge capacity of the current working time segment can be given under both the low current discharge and reference charge and discharge working states. For example, the change trend of the reference charge and discharge charge capacity is as follows: Figure 3 As shown, the charge capacity of the current working time segment cannot be directly measured in the random walk charge and discharge working state. Figures 4 (a) and 4 (b) show the change trends of the voltage and current signals of the satellite lithium battery in the first 50 working time segments in the random walk charge and discharge working state. With the decrease of charge capacity and the increase of internal resistance, it can be seen from Figure 4 (a) that in the first 50 time segments, the voltage of the satellite lithium battery only reaches the voltage boundaries of 3.2V and 4.2V in individual working time segments. And whenever the voltage of the satellite lithium battery reaches the voltage boundary, the random walk charge and discharge operation of the above working time segment will be terminated, and a new charge or discharge set point will be selected. Figures 5 (a) and 5 (b) show the change trends of the voltage and current signals of the satellite lithium battery in the last 50 working time segments in the random walk charge and discharge working state. It can be seen from Figure 5(a) and Figure 5(b) that in the last 50 working time segments, the satellite lithium battery reaches the upper and lower voltage boundaries in almost every step that requires a current of 2A or higher. This is mainly due to the degradation of the health of the satellite lithium battery, which leads to a decrease in charge capacity and an increase in internal resistance.

[0049] for Figure 1 In some possible implementations of the technical solution shown, the generation of a symmetrical point pattern SDP aging facial spectrum sample set based on the acquired random walk charge and discharge data includes:

[0050] Performing data preprocessing on the acquired random walk charge and discharge data;

[0051] performing grouping processing on the preprocessed random walk charge and discharge data to obtain multiple groups of target random walk charge and discharge data;

[0052] For each group of the target random walk charge and discharge data, based on the set SDP graph production method, a rule of producing an SDP graph according to the voltage and current signals of a set number of time segments is used to produce multiple SDP graphs to form an SDP aging facial spectrum sample set, and the graphs are randomly divided into a training set and a test set; wherein the training set is divided into multiple groups of target training sets in chronological order.

[0053] For the above implementation, in some examples, the SDP graph preparation method includes:

[0054] For a one-dimensional voltage and current signal, the voltage and current signal are projected multiple times in a polar coordinate system at different set angles; wherein a projected set of voltage and current signal data is marked as a petal;

[0055] The size of the petals is calculated according to formula (1):

[0056] (1)

[0057] in, represents the number of times the projection is made, ; Represents the total number of projection groups, used to represent the number of petals in each floret; represents the signal time corresponding to the voltage and current signals, Indicates the signal amplitude of the voltage and current signals, and sets the maximum signal amplitude to , the minimum signal amplitude is ; represents the polar radius, is the polar coefficient, used to express the magnification, which determines the size of the petals;

[0058] The density of petals is calculated according to formula (2):

[0059] (2)

[0060] (3)

[0061] Among them, For The positive and negative rotation angle at the beginning, that is, each signal amplitude must be projected Second-rate; represents the time lag coefficient; Represents the polar angle coefficient, which determines the density of the distribution of a petal.

[0062] Understandably, the SDP diagram is a snowflake-like polar coordinate image conversion method, which uses a simple polar coordinate transformation to convert a one-dimensional signal into a two-dimensional petal-symmetrical petal diagram. Specifically, all voltage and current signal amplitudes are converted into polar angles and polar radii and covered on symmetrically distributed petals at one time.

[0063] Based on the above settings and definitions, , , , sampling frequency The SDP diagram generated by the satellite lithium battery constant current random walk charge and discharge working time segment data is as follows Figure 6 shown.

[0064] For the above implementation, after obtaining the voltage and current signal data of the satellite lithium battery under the random walk charge and discharge working state, the wild points are removed to obtain the random walk charge and discharge data that can be used to make the SDP graph; then the above SDP graph method is used to make the corresponding SDP graph sample set. In the embodiment of the present invention, the above SDP graph sample set is referred to as the SDP aging spectrum sample set under the random walk charge and discharge working state of the satellite lithium battery. This is mainly because as the use time of the satellite lithium battery increases, its performance decays, the charge capacity decreases and the internal resistance increases, and the voltage reaches the upper and lower voltage boundaries more and more frequently, so there are more and more "burrs" on the edge of the SDP graph, so this phenomenon is called SDP spectrum aging.

[0065] With respect to the above-mentioned implementation mode, in order to verify the influence of the number of working time segments on the identification result of the service life of the satellite lithium battery, the embodiment of the present invention divides the random walk charge and discharge data obtained above into two groups of target random walk charge and discharge data, A and B. For the target random walk charge and discharge data of group A, a SDP graph is generated by using the voltage and current signals of every 5 working time segments, and a total of 10,800 SDP graphs are generated, of which 1,080 SDP graphs are used as test sets, and the remaining SDP graphs are used as training sets, and the data in the training set are divided into 12 groups in chronological order; at the same time, for the target random walk charge and discharge data of group B, a SDP graph is generated by using the voltage and current signals of every 10 working time segments, and a total of 5,400 SDP graphs are generated, of which 540 are used as test sets, and the remaining SDP graphs are used as training sets, and the data in the training set are also divided into 12 groups in chronological order. Among them, Table 1 shows the charge capacity and internal resistance data of the lithium battery of group B in each time period of random walk charge and discharge. Figure 7 A schematic diagram of the SDP aging spectrum in each time period of random walk charging and discharging of group B is shown.

[0066] Table 1

[0067]

[0068] for Figure 1 In some possible implementations of the technical solution shown, the initial condition generates a confrontation network model CGAN, including:

[0069] A generator, the generator is used to receive the target random walk charge and discharge data, a first condition vector and a second condition vector, so as to generate an SDP aging pseudo-face spectrum based on the input target random walk charge and discharge data, the first condition vector and the second condition vector; wherein the first condition vector represents a first one-hot code vector corresponding to each group of the target training set; the second condition vector represents a second one-hot code vector corresponding to each group of the target random walk charge and discharge data;

[0070] A discriminator is used to discriminate the SDP aging false face spectrum generated by the generator with the corresponding SDP aging face spectrum samples in the target training set to output the confidence probability and type of the SDP aging false face spectrum.

[0071] It should be noted that the conditional generative adversarial network model CGAN is a derivative of the generative adversarial network (GAN). The original GAN ​​uses the generator (hereinafter referred to as "G") and the discriminator (hereinafter referred to as "D") to learn the distribution of sample data in an adversarial manner. It is a typical unsupervised learning network. In order to generate samples of a specified type or batch generate samples of different types, the conditional generative adversarial network CGAN uses random noise and conditional vectors as the input of the generator G. The conditional vector can be a label or other information. Its objective function is shown in formula (4). CGAN can limit the overly free modeling ability of GAN, which is conducive to the convergence of the loss function.

[0072] (4)

[0073] Among them, CGAN has no fixed form. It can add conditional vectors based on the traditional GAN ​​network, or make improvements based on DCGAN to batch generate multiple images.

[0074] For the CGAN provided in the embodiment of the present invention, the input part of the generator G includes three parts, namely, a target random walk charge and discharge data, a first conditional vector and a second conditional vector; wherein, in the specific implementation process, the target random walk charge and discharge data is 100-dimensional, the first conditional vector is a 12-dimensional vector, and the second conditional vector is a 2-dimensional vector, wherein the target random walk charge and discharge data, the first conditional vector and the second conditional vector all generate different one-hot code vectors according to the time sequence and the difference between samples of groups A and B. At the same time, the fully connected layer of the generator G is set as a transposed convolution layer; the fully connected layer of the discriminator D is set as a convolution layer, and a batch normalization layer is added after the convolution layer, and the activation function adopts ReLU; based on the discriminator D provided by the embodiment of the present invention, it can not only output the true and false probability of the SDP aging face spectrum, but also output the recognized type. Specifically, the recognized type is the time period in which the above-mentioned SDP aging face spectrum is located, that is, the specific time period in which the voltage and current signals of the above-mentioned SDP aging face spectrum are generated, and the approximate range of the charge capacity and ohmic internal resistance of the satellite lithium battery is obtained according to the time period in which the SDP aging face spectrum is located, and then the life state of the satellite lithium battery is evaluated. The recognition result is presented using the normalized exponential function (softmax) function. The specific softmax function is shown in formula (5):

[0075] (5)

[0076] in, Indicates the number of SDP aging mask recognition types.

[0077] The comparison table of the first conditional vector and the second conditional vector input into the generator G is shown in Table 2.

[0078] Table 2

[0079]

[0080] for Figure 1 In some embodiments of the technical solution shown, the use of the SDP aging face sample set to iteratively train the initial conditional generative adversarial network model CGAN to obtain a target conditional generative adversarial network model CGAN includes:

[0081] Based on the SDP aging face spectrum in the target training set and the SDP aging fake face spectrum, iteratively training the discriminator in the initial condition generative adversarial network model CGAN to obtain a target discriminator, so that the target discriminator determines the SDP aging fake face spectrum as false, and determines the SDP aging face spectrum in the training set as true;

[0082] Based on the target random walk charge and discharge data in the target training set, the first condition vector and the second condition vector, the SDP aging pseudo-face spectrum is generated to obtain a target condition generator, so that the SDP aging pseudo-face spectrum generated by the target condition generator is determined to be true by the target discriminator;

[0083] Based on the target discriminator and the target condition generator, the target condition generative adversarial network model CGAN is obtained.

[0084] For the above implementation, in some examples, the generating the adversarial network model CGAN based on the target condition to identify the life state and type of the satellite lithium battery includes:

[0085] The target conditional generative adversarial network model CGAN is tested using the SDP aging profile in the test set to identify the time period, charge capacity and ohmic internal resistance range of the SDP aging profile according to the output results of the target conditional generative adversarial network model CGAN.

[0086] like Figure 8 As shown, it shows a schematic diagram of the training process of the CGAN provided by the implementation of the present invention. Specifically, an initial CGAN is built, and two groups of target random walk charge and discharge data A and B are used to iteratively train the initial CGAN, and the trained target discriminator is used to identify the time period, charge capacity and approximate range of ohmic internal resistance of the SDP aging spectrum.

[0087] It should be noted that in the embodiment of the present invention, the CGAN optimization algorithm uses the Adam algorithm, the generator G loss function uses binary cross entropy, and the discriminator D uses a linear combination of binary cross entropy and two classification cross entropies, with a combination ratio of 1:1:0.2. In the embodiment of the present invention, the training cycle is 50,000 times, and the batch processing batch_size=64. The target discriminator after training discriminates the above test set, and the target random walk charge and discharge data set corresponding to the test set is input into the trained target generator. The randomly generated SDP effect diagram is as follows Fig. 9 As shown by Fig. 9 It can be seen that the target generator can generate very realistic SDP aging faces. The trained target discriminator is tested using the test sets of groups A and B. The confusion matrix of the test results is as follows: Fig.10 and Fig.11 As shown by Fig.10 and Fig.11 It can be seen that the confidence probabilities of Group A and Group B are 98.52% and 98.89% respectively. At this time, the target discriminator also has the ability to better discriminate the authenticity and type of the input SDP aging facial spectrum.

[0088] Compared with the conventional charge capacity measurement method for battery life, the technical solution provided by the embodiment of the present invention has the characteristics of convenience, speed, strong compatibility, etc., as shown in Table 3.

[0089] Table 3

[0090]

[0091] Based on the above content, the technical solution proposed in the embodiment of the present invention generates a lithium battery SDP aging spectrum from the satellite lithium battery voltage and current signal data, and uses CGAN to perform data enhancement and classification on the satellite lithium battery SDP aging spectrum, thereby realizing real-time detection of the satellite lithium battery service life status and providing a reference for the maintenance and replacement of satellite lithium batteries.

[0092] Based on the same inventive concept as the above technical solution, see Fig.12 , which shows a lithium battery life detection device 120 based on SDP aging profile and CGAN provided by an embodiment of the present invention, the detection device 120 includes: a generation part 1201, a construction part 1202, a training part 1203 and a recognition part 1204; wherein,

[0093] The generating part 1201 is configured to generate a symmetrical point mode SDP aging facial spectrum sample set based on the acquired random walk charge and discharge data;

[0094] The building part 1202 is configured to build an initial condition generative adversarial network model CGAN;

[0095] The training part 1203 is configured to iteratively train the initial conditional generative adversarial network model CGAN using the SDP aging face spectrum sample set to obtain a target conditional generative adversarial network model CGAN;

[0096] The identification part 1204 is configured to generate a confrontation network model CGAN based on the target condition to identify the life status and type corresponding to the satellite lithium battery.

[0097] It should be noted that for the specific implementation methods or implementation examples of the functions configured by the above-mentioned components, please refer to the corresponding steps, implementation methods and examples of the aforementioned technical solution, and the embodiments of the present invention will not be described in detail here.

[0098] It can be understood that in this embodiment, "part" can be part of a circuit, part of a processor, part of a program or software, etc., and of course it can also be a unit, a module, or a non-modular one.

[0099] In addition, each component in this embodiment may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of software function modules.

[0100] If the integrated unit is implemented in the form of a software function module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) or a processor to perform all or part of the steps of the method described in this embodiment. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0101] Therefore, this embodiment provides a computer storage medium, which stores a lithium battery life detection program based on the SDP aging profile and CGAN. When the lithium battery life detection program based on the SDP aging profile and CGAN is executed by at least one processor, it implements the steps of the lithium battery life detection method based on the SDP aging profile and CGAN described in the above technical solution.

[0102] According to the above-mentioned lithium battery life detection device 120 based on SDP aging profile and CGAN and computer storage medium, see Fig.13, which shows the specific hardware structure of a computing device 130 provided by an embodiment of the present invention that can implement the above-mentioned lithium battery life detection device 120 based on SDP aging profile and CGAN. The computing device 130 can be a wireless device, a mobile or cellular phone (including a so-called smart phone), a personal digital assistant (PDA), a video game console (including a video display, a mobile video game device, a mobile video conferencing unit), a laptop computer, a desktop computer, a TV set-top box, a tablet computing device, an e-book reader, a fixed or mobile media player, etc. The computing device 130 includes: a communication interface 1301, a memory 1302 and a processor 1303; the various components are coupled together through a bus system 1304. It can be understood that the bus system 1304 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 1304 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, in Fig.13 In FIG. 1 , various buses are labeled as bus system 1304. Among them,

[0103] The communication interface 1301 is used to receive and send signals during the process of sending and receiving information with other external network elements;

[0104] The memory 1302 is used to store a computer program that can be run on the processor 1303;

[0105] The processor 1303 is used to execute the steps of the satellite lithium battery life detection method based on the SDP aging profile and CGAN described in the aforementioned technical solution when running the computer program, which will not be repeated here.

[0106] It can be understood that the memory 1302 in the embodiment of the present invention can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus random access memory (DRRAM). The memory 1302 of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0107] The processor 1303 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 1303. The above processor 1303 may be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present invention can be directly embodied as a hardware decoding processor to be executed, or a combination of hardware and software modules in the decoding processor to be executed. The software module may be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 1302, and the processor 1303 reads the information in the memory 1302 and completes the steps of the above method in combination with its hardware.

[0108] It is understood that the embodiments described herein may be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit may be implemented in one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), general purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in the present application, or a combination thereof.

[0109] For software implementation, the techniques described herein can be implemented by modules (e.g., procedures, functions, etc.) that perform the functions described herein. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0110] It can be understood that the exemplary technical solutions of the lithium battery life detection device 120 and the computing device 130 based on the SDP aging spectrum and CGAN are the same concept as the technical solution of the lithium battery life detection method based on the SDP aging spectrum and CGAN. Therefore, the details not described in detail in the technical solution of the lithium battery life detection device 120 and the computing device 130 based on the SDP aging spectrum and CGAN can be found in the description of the technical solution of the lithium battery life detection method based on the SDP aging spectrum and CGAN. The embodiment of the present invention will not elaborate on this.

[0111] It should be noted that the technical solutions described in the embodiments of the present invention can be combined arbitrarily without conflict.

[0112] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A lithium battery life detection method based on SDP aging profile and CGAN, characterized in that: The detection method comprises: Generate a symmetrical point mode SDP aging facial spectrum sample set based on the acquired random walk charging and discharging data; wherein, generating a symmetrical point mode SDP aging facial spectrum sample set based on the acquired random walk charging and discharging data includes: Performing data preprocessing on the acquired random walk charge and discharge data; performing grouping processing on the preprocessed random walk charge and discharge data to obtain multiple groups of target random walk charge and discharge data; For each group of the target random walk charge and discharge data, based on the set SDP graph making method, a rule of making an SDP graph according to the voltage and current signals of a set number of time segments is used to make a plurality of SDP graphs to form an SDP aging face spectrum sample set, and the SDP graphs are randomly divided into a training set and a test set; wherein the training set is divided into a plurality of target training sets in chronological order; The SDP diagram making method includes: For a one-dimensional voltage and current signal, the voltage and current signal are projected multiple times in a polar coordinate system at different set angles; wherein a projected set of voltage and current signal data is marked as a petal; The size of the petals is calculated according to formula (1): （1） in, represents the number of times the projection is made, ; Represents the total number of projection groups, used to represent the number of petals in each flower; represents the signal time corresponding to the voltage and current signals, Indicates the signal amplitude of the voltage and current signals, and sets the maximum signal amplitude to , the minimum signal amplitude is ; represents the polar radius, is the polar coefficient, used to express the magnification, which determines the size of the petals; According to formula (2) and formula (3), the density of petals is calculated as: （2） （3） in, and For The positive and negative rotation angle at the beginning, that is, each signal amplitude must be projected Second-rate; represents the time lag coefficient; Represents the polar angle coefficient, which determines the density of a petal distribution; Constructing an initial condition generative adversarial network model CGAN; wherein the initial condition generative adversarial network model CGAN comprises: A generator, the generator is used to receive the target random walk charge and discharge data, a first condition vector and a second condition vector, so as to generate an SDP aging pseudo-face spectrum based on the input target random walk charge and discharge data, the first condition vector and the second condition vector; wherein the first condition vector represents a first one-hot code vector corresponding to each group of the target training set; the second condition vector represents a second one-hot code vector corresponding to each group of the target random walk charge and discharge data; A discriminator, the discriminator is used to discriminate the SDP aging false face spectrum generated by the generator with the corresponding SDP aging face spectrum samples in the target training set to output the confidence probability and type of the SDP aging false face spectrum; The initial conditional generative adversarial network model CGAN is iteratively trained using the SDP aging face spectrum sample set to obtain a target conditional generative adversarial network model CGAN; wherein, the initial conditional generative adversarial network model CGAN is iteratively trained using the SDP aging face spectrum sample set to obtain a target conditional generative adversarial network model CGAN, including: Based on the SDP aging face spectrum in the target training set and the SDP aging fake face spectrum, iteratively training the discriminator in the initial condition generative adversarial network model CGAN to obtain a target discriminator, so that the target discriminator determines the SDP aging fake face spectrum as false, and determines the SDP aging face spectrum in the training set as true; Based on the target random walk charge and discharge data in the target training set, the first condition vector and the second condition vector, the SDP aging pseudo-face spectrum is generated to obtain a target condition generator, so that the SDP aging pseudo-face spectrum generated by the target condition generator is determined to be true by the target discriminator; Based on the target discriminator and the target condition generator, the target condition generative adversarial network model CGAN is obtained; Based on the target conditions, a generative adversarial network model CGAN is used to identify the life status and type of satellite lithium batteries.

2. The detection method according to claim 1, characterized in that: The random walk charging and discharging data includes: When the satellite lithium battery does not have enough time to complete a complete charge and discharge operation, a constant current random walk charge and discharge operation is performed on the satellite lithium battery, and the obtained voltage signal is used as random walk charge and discharge data; When the satellite lithium battery has enough time to complete a complete charge and discharge operation, the voltage and current signals obtained are used as the random walk charge and discharge data.

3. The detection method according to claim 1, characterized in that: The generating adversarial network model CGAN based on the target condition identifies the life state and type of the satellite lithium battery, including: The target conditional generative adversarial network model CGAN is tested using the SDP aging profile in the test set to identify the time period, charge capacity and ohmic internal resistance range of the SDP aging profile according to the output results of the target conditional generative adversarial network model CGAN.

4. A lithium battery life detection device based on SDP aging profile and CGAN, characterized in that: The detection device comprises: a generation part, a construction part, a training part and a recognition part; wherein, The generating part is configured to generate a symmetrical point mode SDP aging facial spectrum sample set based on the acquired random walk charging and discharging data; Wherein, the generating part is configured as follows: Performing data preprocessing on the acquired random walk charge and discharge data; performing grouping processing on the preprocessed random walk charge and discharge data to obtain multiple groups of target random walk charge and discharge data; For each group of the target random walk charge and discharge data, based on the set SDP graph making method, a rule of making an SDP graph according to the voltage and current signals of a set number of time segments is used to make a plurality of SDP graphs to form an SDP aging face spectrum sample set, and the SDP graphs are randomly divided into a training set and a test set; wherein the training set is divided into a plurality of target training sets in chronological order; The generating part is further configured as follows: For a one-dimensional voltage and current signal, the voltage and current signal are projected multiple times in a polar coordinate system at different set angles; wherein a projected set of voltage and current signal data is marked as a petal; The size of the petals is calculated according to formula (1): （1） in, represents the number of times the projection is made, ; Represents the total number of projection groups, used to represent the number of petals in each flower; represents the signal time corresponding to the voltage and current signals, Indicates the signal amplitude of the voltage and current signals, and sets the maximum signal amplitude to , the minimum signal amplitude is ; represents the polar radius, is the polar coefficient, used to express the magnification, which determines the size of the petals; According to formula (2) and formula (3), the density of petals is calculated as: （2） （3） in, and For The positive and negative rotation angle at the beginning, that is, each signal amplitude must be projected Second-rate; represents the time lag coefficient; Represents the polar angle coefficient, which determines the density of a petal distribution; The building part is configured to build an initial condition generative adversarial network model CGAN; Wherein, the construction part is configured as follows: A generator, the generator is used to receive the target random walk charge and discharge data, a first condition vector and a second condition vector, so as to generate an SDP aging pseudo-face spectrum based on the input target random walk charge and discharge data, the first condition vector and the second condition vector; wherein the first condition vector represents a first one-hot code vector corresponding to each group of the target training set; the second condition vector represents a second one-hot code vector corresponding to each group of the target random walk charge and discharge data; A discriminator, the discriminator is used to discriminate the SDP aging false face spectrum generated by the generator with the corresponding SDP aging face spectrum samples in the target training set to output the confidence probability and type of the SDP aging false face spectrum; The training part is configured to iteratively train the initial conditional generative adversarial network model CGAN using the SDP aging face spectrum sample set to obtain a target conditional generative adversarial network model CGAN; Wherein, the training part is configured as follows: Based on the SDP aging face spectrum in the target training set and the SDP aging fake face spectrum, iteratively training the discriminator in the initial condition generative adversarial network model CGAN to obtain a target discriminator, so that the target discriminator determines the SDP aging fake face spectrum as false, and determines the SDP aging face spectrum in the training set as true; Based on the target random walk charge and discharge data in the target training set, the first condition vector and the second condition vector, the SDP aging pseudo-face spectrum is generated to obtain a target condition generator, so that the SDP aging pseudo-face spectrum generated by the target condition generator is determined to be true by the target discriminator; Based on the target discriminator and the target condition generator, the target condition generative adversarial network model CGAN is obtained; The identification part is configured to generate a confrontation network model CGAN based on the target condition to identify the life status and type corresponding to the satellite lithium battery.

5. A computing device, characterized in that: The computing device includes: a communication interface, a memory and a processor; each component is coupled together through a bus system; wherein, The communication interface is used to receive and send signals during the process of sending and receiving information with other external network elements; The memory is used to store a computer program that can be run on the processor; The processor is used to execute the steps of the lithium battery life detection method based on SDP aging profile and CGAN as described in any one of claims 1 to 3 when running the computer program.

6. A storage medium, characterized in that: The storage medium stores a lithium battery life detection program based on the SDP aging profile and CGAN. When the lithium battery life detection program based on the SDP aging profile and CGAN is executed by at least one processor, the steps of the lithium battery life detection method based on the SDP aging profile and CGAN according to any one of claims 1 to 3 are implemented.

Images