Impedance spectrum acquisition method and device, electronic equipment and storage medium

Abstract

The invention discloses an impedance spectrum acquisition method and device, electronic equipment and a storage medium, and belongs to the technical field of batteries. The method is applied to the electronic equipment and comprises the steps that target charging data of a battery in the current charging process are obtained, and the target charging data comprise multi-dimensional charging parameters collected based on a target time sequence; the target charging data is input into a preset impedance model, a predicted impedance spectrum of the battery is obtained, the preset impedance model is obtained through training based on a training sample data set, and the training sample data set comprises historical multi-dimensional charging parameters, collected in the historical charging process, of the battery and a corresponding historical actual impedance spectrum. According to the embodiment of the invention, a complex integrated circuit does not need to be designed, the time consumed by the process is reduced, the multi-dimensional charging parameters provide diversified characteristics, the accuracy of the preset impedance model is improved, and the efficiency of obtaining the electrochemical impedance spectrum of the battery is improved.

Description

Technical Field

[0001] The present application relates to the field of battery technology, and in particular to an impedance spectrum acquisition method, device, electronic device and storage medium. Background Art

[0002] With the development of science and technology, various electronic devices appear in people's daily lives. People can use electronic devices for entertainment, work, etc.

[0003] At present, hardware devices such as batteries have become an indispensable part of electronic devices. During the charging process or use of electronic devices, the battery's power, temperature, current, impedance and other parameters will change over time. Therefore, it is necessary to monitor the performance of the battery at all times. For example, by predicting the electrochemical impedance spectrum of the battery, the internal changes and health status of the battery during aging can be reflected, which is of great significance to the safe operation of the battery. However, most of the existing methods for obtaining electrochemical impedance spectra are obtained through integrated circuit measurement, material and electrochemical analysis and calculation, or mathematical model parameter fitting. This process requires traditional means to establish an integrated circuit measurement system for measurement, or analysis and calculation based on electrochemistry, materials and traditional mathematical methods, which requires a lot of time, manpower and material resources, and the process is cumbersome, which affects the efficiency of obtaining the electrochemical impedance spectrum of the battery. Summary of the invention

[0004] In order to improve the efficiency of obtaining the electrochemical impedance spectrum of a battery, the present application provides an impedance spectrum acquisition method, device, electronic device and storage medium. The technical solution is as follows:

[0005] In one aspect, the present application provides an impedance spectrum acquisition method, which is applied to electronic equipment, and the method comprises:

[0006] Acquire target charging data of the battery during the current charging process, wherein the target charging data includes multi-dimensional charging parameters collected based on a target time series;

[0007] The target charging data is input into a preset impedance model to obtain a predicted impedance spectrum of the battery, wherein the preset impedance model is trained based on a training sample data set, and the training sample data set includes historical multi-dimensional charging parameters of the battery collected during historical charging processes and corresponding historical actual impedance spectra.

[0008] In one aspect, the present application provides an impedance spectrum acquisition device, which is applied to electronic equipment, and the device comprises:

[0009] A first acquisition module is used to acquire target charging data of the battery in the current charging process, wherein the target charging data includes multi-dimensional charging parameters collected based on a target time series;

[0010] A second acquisition module is used to input the target charging data into a preset impedance model to obtain a predicted impedance spectrum of the battery, wherein the preset impedance model is trained based on a training sample data set, and the training sample data set includes historical multi-dimensional charging parameters of the battery collected during historical charging processes and corresponding historical actual impedance spectra.

[0011] In another aspect, the present application provides an electronic device, comprising a processor and a memory, wherein the memory stores a computer program, and the computer program is executed by the processor to implement the impedance spectrum acquisition method as described in one aspect.

[0012] In another aspect, the present application provides a computer-readable storage medium, wherein the storage medium stores a computer program, and the computer program is executed by a processor to implement the impedance spectrum acquisition method as described in one aspect.

[0013] On the other hand, an embodiment of the present application provides a computer program product. When the computer program product is run on a computer, the computer is enabled to execute the impedance spectrum acquisition method as described in one aspect above.

[0014] On the other hand, an embodiment of the present application provides an application publishing platform, which is used to publish a computer program product, wherein when the computer program product is run on a computer, the computer executes the impedance spectrum acquisition method as described in one aspect above.

[0015] The beneficial effects brought by the technical solution provided by the embodiment of the present application include at least:

[0016] Obtain the target charging data of the battery in the current charging process, the target charging data includes multi-dimensional charging parameters collected based on the target time series; input the target charging data into the preset impedance model to obtain the predicted impedance spectrum of the battery, the preset impedance model is trained based on the training sample data set, and the training sample data set includes the historical multi-dimensional charging parameters of the battery collected during the historical charging process and the corresponding historical actual impedance spectrum. In an embodiment of the present application, the collected target charging data includes multi-dimensional charging parameters, and the target charging data including the multi-dimensional charging parameters are predicted by the preset impedance model to obtain the predicted impedance spectrum of the battery. This process does not require the design of complex integrated circuits, which reduces the time spent on the process, and the multi-dimensional charging parameters provide diversified features, which increases the accuracy of the preset impedance model and improves the efficiency of obtaining the electrochemical impedance spectrum of the battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 is a schematic structural diagram of an example of an electronic device provided in an embodiment of the present application;

[0019] Figure 2 is a method flow chart of a method for obtaining an impedance spectrum provided by an exemplary embodiment of the present application;

[0020] Figure 3 is a training flow chart of a preset impedance model provided by an exemplary embodiment of the present application;

[0021] Figure 4 is a structural schematic diagram of a preset impedance model involved in an exemplary embodiment of the present application;

[0022] Figure 5 It is a structural diagram of an LSTM model unit involved in an exemplary embodiment of the present application;

[0023] Figure 6 is a method flow chart of a method for obtaining an impedance spectrum provided by an exemplary embodiment of the present application;

[0024] Figure 7 is a structural block diagram of an impedance spectrum acquisition device provided by an exemplary embodiment of the present application;

[0025] Figure 8 It is a structural schematic diagram of another example of the impedance spectrum acquisition device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0027] The term "multiple" as used herein refers to two or more than two. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.

[0028] The solution provided by the present application can be used in real-life scenarios where people monitor the battery performance of electronic devices during their daily use of electronic devices. To facilitate understanding, some terms and application scenarios involved in the embodiments of the present application are briefly introduced below.

[0029] Electrochemical impedance spectroscopy: It is one of the important methods of electrochemical measurement. It uses a small-amplitude sinusoidal potential or current as a disturbance signal to make the electrode system produce an approximately linear response, and measures the impedance spectrum of the electrode system in a wide frequency range, so as to study the electrode system and measure the electrochemical impedance spectrum.

[0030] AC impedance is also called electrochemical impedance spectroscopy (EIS), and was called AC impedance (AC Impedance) in early electrochemical literature. Impedance measurement was originally a method for studying the frequency response characteristics of linear circuit networks in electricity. It was applied to the study of electrode processes and became an experimental method in electrochemical research. The measurement principle is that when the electrode system is disturbed by an AC signal of a sinusoidal voltage (current), a corresponding current (voltage) response signal will be generated, and the impedance or admittance of the electrode can be obtained from these signals. The impedance spectrum generated by a series of sinusoidal wave signals is called electrochemical impedance spectroscopy.

[0031] Battery health (State of health, SOH): SOH = current capacity of the battery / initial capacity of the battery * 100%).

[0032] Long Short-Term Memory (LSTM) is a time-recurrent neural network designed to solve the long-term dependency problem of general recurrent neural networks (RNNs). All RNNs have a chain form of repeated neural network modules. In standard RNNs, this repeated structural module has only a very simple structure, such as a tanh layer.

[0033] With the development of science and technology, batteries, as mobile power supply devices, have been widely used in various electronic devices. Among them, lithium batteries are currently widely used in many fields such as electrified transportation and electronic equipment due to their own characteristics of light weight and large capacity. However, the aging of lithium-ion batteries (referred to as lithium batteries) can easily bring great safety hazards. During the operation of lithium batteries, the parameters of lithium batteries such as power, temperature, current, impedance, etc. usually change with time. Timely acquisition of the current performance indicators of lithium batteries will help improve the charging and discharging strategies.

[0034] Please refer to Figure 1 , is a schematic diagram of an example structure of an electronic device provided in an embodiment of the present application. Figure 1 The electronic device shown includes components such as a processor 110 , a memory 120 , a power module 130 , and a charging management module 140 .

[0035] The processor 110 is the control center of the electronic device, and uses various interfaces and lines to connect various parts of the entire electronic device. By running or executing software programs and / or modules stored in the memory 120, and calling data stored in the memory 120, the processor 110 performs various functions of the electronic device and processes data, thereby monitoring the electronic device as a whole. Optionally, the processor 110 may include one or more processing units; optionally, the processor 110 may integrate an application processor, which mainly processes operating devices, user interfaces, and application programs, etc. Of course, other processors may also be included, which are not listed here one by one.

[0036] The memory 120 can be used to store software programs and modules. The processor 110 executes various functional applications and data processing of the electronic device by running the software programs and modules stored in the memory 120. The memory 120 may mainly include a program storage area and a data storage area, wherein the program storage area may store operating devices, applications required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device (such as audio data, phone book, etc.), etc. In addition, the memory 120 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0037] The electronic device also includes a power module 130 for supplying power to various components. Optionally, the power module 130 can be logically connected to the processor 110 via a charging management module 140, so that functions such as charging, discharging, and power consumption management can be implemented through the charging management module 140.

[0038] Although not shown, the electronic device may further include a camera. Optionally, the camera may be located at the front or rear of the electronic device, which is not limited in the present embodiment.

[0039] It is to be understood that the structure illustrated in the embodiments of the present application does not constitute a specific limitation on the electronic device. In other embodiments of the present application, the electronic device may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0040] Optionally, the above-mentioned electronic devices may include but are not limited to wearable devices (such as smart bracelets, smart watches, smart glasses, etc.), mobile phones, tablet computers, laptops, smart glasses, smart watches, MP3 players (Moving Picture Experts Group Audio Layer III, Moving Picture Experts Compression Standard Audio Layer 3), MP4 (Moving Picture Experts Group Audio Layer IV, Moving Picture Experts Compression Standard Audio Layer 4) players, desktop computers, laptops, etc.

[0041] Usually, the above Figure 1 The electronic device shown needs to be equipped with a corresponding operating system and run based on the operating system. For example, the operating system of the electronic device can be an Android system, an iOS system, a Linux system, etc.

[0042] For the above-mentioned electronic devices, predicting the electrochemical impedance spectrum of the battery can reflect the internal changes and health status of the battery during aging, which is of great significance to the safe operation of the battery. Therefore, in the operation of electronic equipment, it is crucial to measure the electrochemical impedance spectrum of the battery in real time and accurately. At present, most of the existing methods of obtaining electrochemical impedance spectra are obtained through integrated circuit measurement, material and electrochemical analysis and calculation, or mathematical model parameter fitting. This process requires traditional means to establish an integrated circuit measurement system for measurement, or analysis and calculation based on electrochemistry, materials and traditional mathematical methods.

[0043] Although some solutions use machine learning methods to predict impedance spectra, most of them are based on traditional machine learning models, using single battery voltage and other features to predict EIS under a certain state of charge (SOC), and have certain limitations in the selection of models and data. For example, based on the offline battery test, the offline relaxation voltage curve data and the corresponding electrochemical impedance spectrum data are obtained to train the machine learning model; the online monitoring relaxation voltage curve data of the battery to be estimated is obtained, and the trained machine learning model is called; according to the online monitoring relaxation voltage curve data, the battery to be estimated and the EIS data are estimated online through the machine learning model to obtain the online EIS curve of the battery to be estimated. This method does not require the integration of a dedicated measurement circuit, and compared with traditional measurement methods, it has the advantages of low cost and ease of use; at the same time, it eliminates the interference problem caused by SOC changes in traditional online estimation methods, greatly improves the estimation accuracy, and is of great significance to the development of advanced battery diagnostic technology based on EIS.

[0044] For the contents mentioned in the above-mentioned related technologies, the process requires the establishment of an integrated circuit measurement system for measurement by traditional means, or analysis and calculation based on electrochemistry, materials and traditional mathematical methods, which requires a lot of time, manpower and material resources, and the process is cumbersome, affecting the efficiency of obtaining the electrochemical impedance spectrum of the battery. In other machine learning model methods, the collected features are too single, and the adaptability and accuracy of the model are not high enough.

[0045] In order to solve the problems existing in the above-mentioned technology, the present application provides an impedance spectrum acquisition method, which can be applied to electronic devices. By acquiring multi-dimensional charging parameters based on target time series collection and inputting them into a preset impedance model, an accurate predicted impedance spectrum can be obtained. There is no need to design complicated hardware circuits. Multi-dimensional charging parameters are also combined for prediction, which improves the efficiency and accuracy of obtaining the electrochemical impedance spectrum of the battery.

[0046] Please refer to Figure 2 , which shows a method flow chart of an impedance spectrum acquisition method provided by an exemplary embodiment of the present application, the impedance spectrum acquisition method can be applied to electronic devices and executed by a processor or a charging management module in the electronic device. Figure 2 As shown, the impedance spectrum acquisition method may include the following steps:

[0047] Step 201 : acquiring target charging data of the battery in the current charging process, where the target charging data includes multi-dimensional charging parameters collected based on a target time series.

[0048] Among them, users can charge electronic devices at any time during the use of electronic devices. For example, the electronic device can be charged by connecting the charging cable to the interface on the electronic device. After connecting the charger, the battery of the electronic device can be charged. When obtaining the target charging data, the electronic device can monitor the charging process and measure the charging parameters such as voltage, current, temperature, capacitance, and capacitance increment in real time. Most of these charging parameters are curves that change with time. The charging parameters during the battery charging process are collected and obtained during the charging time.

[0049] Optionally, the electronic device collects all the above parameters during the collection process to obtain multi-dimensional charging parameters, and obtains target charging data by sorting them in time series. That is to say, the target charging data includes charging parameters of multiple dimensions collected within a period of time. The above charging process can be carried out during constant current charging. Taking the charging parameters of voltage, current, temperature, and capacitance as an example, the electronic device can collect voltage, current, temperature, and capacitance during the constant current charging process of the battery to obtain four corresponding charging parameter curves. From a time perspective, each collection moment will correspond to parameter values ​​of the four dimensions of voltage, current, temperature, and capacitance. The charging parameters of the four dimensions are divided based on the time series to obtain the target charging data U. t =[U t，1 , U t，2 , …, U t，k ], where t represents the duration of this collection, and each collection moment corresponds to a U t , k represents the dimension. In the above example, k=4, and the target charging data includes 4-dimensional features.

[0050] Step 202: input the target charging data into a preset impedance model to obtain a predicted impedance spectrum of the battery. The preset impedance model is trained based on a training sample data set. The training sample data set includes historical multi-dimensional charging parameters of the battery collected during historical charging processes and corresponding historical actual impedance spectra.

[0051] Optionally, the electronic device inputs the target charging data into a preset impedance model, and calculates the predicted impedance spectrum of the battery through the preset impedance model. The preset impedance model is trained in advance and designed in the electronic device, and the training sample data set used by the preset impedance model includes the historical multi-dimensional charging parameters of the battery collected during the historical charging process and the corresponding historical actual impedance spectrum. That is, during the historical charging process, the multi-dimensional charging parameters are collected, and the historical actual impedance spectrum is detected as a training sample data set and trained to obtain a preset impedance model, which is preset in the electronic device.

[0052] In summary, the target charging data of the battery in the current charging process is obtained, and the target charging data includes multi-dimensional charging parameters collected based on the target time series; the target charging data is input into the preset impedance model to obtain the predicted impedance spectrum of the battery, and the preset impedance model is obtained by training based on the training sample data set, and the training sample data set includes the historical multi-dimensional charging parameters of the battery collected during the historical charging process and the corresponding historical actual impedance spectrum. In an embodiment of the present application, the collected target charging data includes multi-dimensional charging parameters, and the target charging data including the multi-dimensional charging parameters is predicted by the preset impedance model to obtain the predicted impedance spectrum of the battery. This process does not require the design of complex integrated circuits, which reduces the time spent on the process, and the multi-dimensional charging parameters provide diversified features, which increases the accuracy of the preset impedance model and improves the efficiency of obtaining the electrochemical impedance spectrum of the battery.

[0053] Before executing the above-mentioned method for acquiring battery performance, this solution needs to train the preset impedance model in advance to obtain the trained preset impedance model. In one possible implementation, whether it is the same battery or different batteries, the duration and battery status required for each charging are different. Therefore, for the target charging data in the collected training sample data set, the length of each sequence in the target charging data is inconsistent. During training, in order to ensure the uniformity of the model input length, the original sequence needs to be padded or truncated to the same length.

[0054] Please refer to Figure 3 , which shows a flow chart of training a preset impedance model provided by an exemplary embodiment of the present application, and the training of the preset impedance model can be performed by an electronic device. Figure 3 As shown, the training process of the preset impedance model may include the following steps:

[0055] Step 301 : obtaining a training sample data set, where the training sample data set includes historical multi-dimensional charging parameters of the battery collected during a historical charging process and a corresponding historical actual impedance spectrum.

[0056] The process of obtaining the training sample data set can be accumulated over a long period of time, for example, various charging parameters are collected by monitoring the battery during the charging time of the electronic device during one or two months. The time of one or two months can be pre-set in the electronic device by the developer.

[0057] Taking the collected charging parameters as voltage, current, battery capacity, and battery surface temperature as an example, during the constant current charging process of the battery, the electronic device obtains the curves of the voltage, current, battery capacity, and battery surface temperature as they change over time through real-time measurement. Considering that in actual scenarios, the time required for each charging is different, the collection time of the charging parameters collected each time will usually be different during the data collection process. For these charging data collected at different times, their specific parameter values ​​will also be inconsistent in time. For example, taking voltage as an example, the length of each charging voltage in the time series of these voltages collected at different times is inconsistent. Therefore, this solution can unify the collected voltage data so that the length of the voltage collected at different times is consistent in the time series. In actual collection, it is also not limited to the several charging parameters involved in this embodiment, and other types of charging parameters can be added or reduced, which is not limited here.

[0058] For example, the voltage data collected during a charging process is as follows: V = [v 0 , v 1 , …, v t The electronic device can intercept or fill in the voltage data according to the required time series length to obtain the following processed voltage data: V = [v 0 , v 1 , …, v i ], where the required time series length can be pre-set by the developer in the electronic device. For example, the required time series length is i. If t is greater than i, the redundant voltage data is truncated and discarded, and the remaining voltage data is the processed voltage data. If t is less than i, the voltage data is padded so that the time series length reaches i, which is the processed voltage data. For charging parameters of other dimensions (current, battery capacity, battery surface temperature, etc.), unified processing can be performed in a similar manner to obtain historical multi-dimensional charging parameters, which will not be repeated here.

[0059] The actual historical impedance spectrum can be obtained through actual measurement. For example, when obtaining the charging parameters of the above-mentioned dimensions, the electrochemical impedance spectrum under different battery states corresponding to the charging parameters is measured by the way. The charging parameter curve of each dimension can correspond to the EIS under p SOCs. Among them, p can be pre-set in the electronic device by the developer. The real part and imaginary part of the EIS under different SOCs are respectively used as one dimension in the output sequence, and the dimension of a data point in the output sequence is 2p. The EIS measured at different SOCs will also have a certain length difference. At the same time, in the measurement and calculation, there may be some data points less than 0 in the high-frequency segment of the EIS. Therefore, for the preprocessing of EIS, the first step is to remove the data points less than 0, and intercept the first M data points of each EIS as the true value / target value output of the model training. Among them, M is an integer and can also be pre-set by the developer.

[0060] That is to say, for the historical actual impedance spectrum, the charging parameters of the above dimensions are collected together during measurement. The curve corresponding to each charging parameter will measure the EIS under p SOCs, and through data processing, the first M data points are intercepted as the true values ​​during model training. The impedance spectra obtained after these interceptions are the historical actual impedance spectra.

[0061] Step 302: train the preset impedance model according to the training sample data set, and obtain the trained preset impedance model when the loss function of the preset impedance model converges, wherein the loss function includes a root mean square error function.

[0062] Optionally, the preset impedance model is trained according to the obtained training sample data set to obtain a trained preset impedance model. The training sample data set includes the historical multi-dimensional charging parameters of the battery collected during the historical charging process and the corresponding historical actual impedance spectrum. During training, the historical multi-dimensional charging parameters of the battery collected during the historical charging process are used as the input sequence, and the historical actual impedance spectrum is used as the output sequence for verification, and finally a trained preset impedance model is obtained.

[0063] In one possible implementation, the preset impedance model includes a feature extraction module and a prediction module. The feature extraction module is used to extract hidden features corresponding to the input sequence and input the hidden features into the prediction module; the prediction module is used to operate on the hidden features to obtain a predicted impedance spectrum of the battery. The training process uses the root mean square error (RMSE) as a loss function to calculate the average error between the predicted value and the true value of each impedance point in the output sequence, and the trained preset impedance model is obtained when the loss function converges.

[0064] Optionally, the feature extraction module can be implemented by an encoder, and the encoder includes multiple first model subunits connected in series. Among them, the number of first model subunits in the encoder can be flexibly set by the developer, and its length is related to the length of the time series of the charging parameters that the developer wants to select. For example, the developer wants the charging parameters of a time series with a length of N. When designing the encoder, the number of first model subunits inside the encoder is also set to N. When the historical multi-dimensional charging parameters of the battery collected during the above-mentioned processing of the historical charging process are padded or truncated to a uniform length N based on the time series, the uniformity of the model input length is achieved.

[0065] Optionally, the prediction module can be implemented by a decoder, which includes a plurality of second model subunits connected in series. The number of second model subunits in the decoder can also be flexibly set by the developer, and its length is related to the length of the predicted impedance spectrum that the developer wants to select. For example, if the developer wants to preset the output length of the impedance model to be an impedance spectrum of M, when designing the decoder, the number of second model subunits in the decoder is also set to M.

[0066] For example, the first model subunit and the second model subunit can both be long short-term memory (LSTM) model units. Figure 4 , which shows a schematic diagram of the structure of a preset impedance model involved in an exemplary embodiment of the present application. Figure 4 As shown, it includes an encoder 401 and a decoder 402. The structure of the encoder 401 is composed of N LSTM model units, and the structure of the decoder 402 is composed of M LSTM model units. Figure 4 The structures shown are connected in series.

[0067] For the historical multi-dimensional charging parameters of the battery collected in the historical charging process in the training sample data set, from the perspective of time series, each collection moment corresponds to the charging parameters of multiple dimensions, that is, the time series of multi-dimensional features. The data point U corresponding to each collection moment in the input sequence U is converted into t Input to the encoder. For example, for each data point U corresponding to the acquisition time in the input sequence U t , contains k-dimensional features. Taking the charging parameters of voltage, current, temperature, and capacity as an example, the voltage, current, temperature, and capacity are collected during the historical charging process to obtain the corresponding four charging parameter curves. From the perspective of time, each collection moment will correspond to the parameter values ​​of the four dimensions of voltage, current, temperature, and capacity. The charging parameters of these four dimensions are divided based on the time series. The data point U corresponding to each collection moment in the input sequence Ut =[U t，1 , U t，2 , …, U t，k ], where t represents the duration of the collection, and each collection moment corresponds to a U t , k represents the dimension. In the above example, k=4, and the target charging data includes 4-dimensional features.

[0068] Optionally, after the historical multi-dimensional charging parameters of the battery collected in the historical charging process of the training sample data set are input into the encoder, the model training can be started. During the training process, the decoder receives the results output by the encoder, and performs impedance prediction in combination with its own initial parameters to obtain an output result of each of the M LSTM model units. These output results can form a predicted impedance spectrum, which is verified in combination with the actual historical impedance spectrum. The loss function is trained until convergence to obtain a trained predicted impedance model, which is set in an electronic device.

[0069] In the above training process, when the collected charging data needs to be input into the preset impedance model for training, the multi-dimensional charging parameters corresponding to each collection moment in each time series are input into each LSTM model unit. For example, the multi-dimensional charging parameters corresponding to the first collection moment are input into the first LSTM model unit, the multi-dimensional charging parameters corresponding to the second collection moment are input into the second LSTM model unit, and so on, and the multi-dimensional charging parameters corresponding to the N collection moments are respectively input into the N LSTM model units in the encoder.

[0070] When training the encoder and decoder, each LSTM model unit also needs to be input by its own cell state C i and the hidden state H i , according to the above Figure 4 The connection relationship in each cell state C i and the hidden state H i You can follow Figure 4 The cell state C of the first LSTM model unit in the encoder 0 and the hidden state H 0 It can be automatically and randomly generated by electronic devices, so that according to the series relationship, each LSTM model unit will get its own cell state and hidden state.

[0071] The initial impedance S of the first LSTM model unit in the decoder 0 It can also be automatically and randomly generated by the electronic device, and the calculation starts according to the data transmitted from the encoder to obtain the output S of the first LSTM model unit in each decoder 1 To S M , is the predicted result.

[0072] Please refer to Figure 5 , which shows a schematic diagram of the structure of an LSTM model unit involved in an exemplary embodiment of the present application. Figure 5 As shown, the LSTM model unit includes three input ports and three output ports, namely, a first input port 501, a second input port 502, a third input port 503, a first output port 504, a second output port 505, and a third output port 506. Among them, the first input port 501 is used to input the cell state C t-1 The second input port 502 is used to input the hidden state H t-1 , the third input port 503 is used for input. The first output port 504 is used to output the new cell state C calculated by the LSTM model unit. t The second output port 505 is used to output the new hidden state H calculated by the LSTM model unit. t-1 .

[0073] Optionally, when the above LSTM model unit is arranged in the encoder, the third input port 503 is used to input the historical multi-dimensional charging parameters of the battery collected during the historical charging process, and the output of the third output port 506 may not be collected. When the above LSTM model unit is arranged in the decoder, if it is the first LSTM model unit of the encoder, then the third input port 503 is used to input the initial impedance value. If it is any LSTM model unit in the middle and the last of the encoder, then the third input port 503 is used to input the result value output by the third output port of the previous LSTM, and the output of the third output port 506 needs to be collected, and the result obtained by the collection is the predicted impedance spectrum.

[0074] exist Figure 5 The structure of the LSTM model unit shown is combined into Figure 4 For each LSTM unit in the encoder, there are three sets of inputs (the tth data point U in the input sequence t , the cell state C output by the previous unit t-1 and the hidden state H t-1 ) and two sets of outputs (cell states C t and the hidden state H t ), its internal structure can be divided into three parts according to the functional principle: forget gate, input gate and output gate.

[0075] a) Forget gate: used to selectively "forget" the cell state information transmitted by the previous unit and output a value f between 0 and 1 t , 1 represents “completely retain”, and 0 represents “completely abandon”;

[0076] f t =σ(W f ·[H,U t ]+b f );

[0077] Among them, W f , b f is the learnable weight matrix and bias term, and σ is the Sigmoid function.

[0078] b) Input gate: also called memory gate, used to "remember" the newly added input information and output an alternative update information vector At the same time, output a value i between 0 and 1 t , decide what new information to keep;

[0079] i t =σ(W i ·[H t-1 , U t ]+b i );

[0080]

[0081] Among them, W i , W C and b i , b C are the learnable weight matrix and bias term respectively, and σ is the Sigmoid function.

[0082] c) Output gate: used to generate the output value (hidden state) Ht, and based on the output of the forget gate and the memory gate, update the cell state C that needs to be passed to the next unit t .

[0083]

[0084] o t =σ(W o •[H t-1 , U t ]+b o );

[0085] H t =o t *tanh(C t );

[0086] Among them, W o , b o is the learnable weight matrix and bias term, and σ is the Sigmoid function.

[0087] When training starts, initialize a cell state C 0 ∈RH and a hidden state H 0 ∈R H , where H is the hidden dimension. t ∈R k , input the LSTM units in the encoder one by one in sequence, and obtain the corresponding encoded output cell state C t ∈R H and the hidden state H t ∈R H . Where R represents the set of real numbers.

[0088] For the historical multi-dimensional charging parameters with a long original time series in the training sample data set, the encoding output C corresponding to the truncated last dimensional data point (i.e., the data point at the last acquisition moment) is used. N , H N ; For the historical multi-dimensional charging parameters with shorter original time series in the training sample data set, the encoding output C corresponding to the last dimensional data point in the original length before the padding operation is used T , H T ; The obtained output is used as the final output C′ of the encoder 0 , H′ 0 .

[0089] The decoder contains M LSTM units. Initialize the starting impedance input S of a decoder 0 ∈R 2p , and the final output of the encoder C′ 0 , H′ 0 ∈R H Together, input the first LSTM unit of the decoder to obtain the first set of output cell states C′ 1 ∈R H and hidden state H′ 1 ∈R H .

[0090] The decoder output (hidden state) H′ 1 Through a fully connected layer, the decoded output is obtained.

[0091] s 1 =ReLU(H′ 1 W 1 +b 1 )∈R 2p ;

[0092] Among them, W 1 ∈R H×2p and b 1 ∈R 2p are the learnable weight matrix and bias vector respectively.

[0093] S 1 Input the next LSTM unit and repeat the above steps until the output S of the Mth LSTM unit is obtained. M , that is, the predicted value of the output sequence S = [S 1 , S 2 , ... S M ].

[0094] After the preset impedance model is trained through the above process, the trained preset impedance model is set in the electronic device. After the electronic device collects the target charging data during the battery charging process, it inputs it into the preset impedance model to obtain the desired predicted impedance spectrum of the battery.

[0095] Please refer to Figure 6 , which shows a method flow chart of an impedance spectrum acquisition method provided by an exemplary embodiment of the present application, and the impedance spectrum acquisition method can be applied to electronic devices. Figure 6 As shown, the impedance spectrum acquisition method may include the following steps:

[0096] Step 601, obtaining target charging data of the battery in the current charging process, where the target charging data includes multi-dimensional charging parameters collected based on a target time series.

[0097] In practical applications, the electronic device is already equipped with the above-mentioned trained preset impedance model. During the process of charging the battery, the data can be collected to obtain the charging parameters collected during the corresponding charging process. Among them, multiple charging durations can also include the time spent collecting charging parameters during the previous charging processes. For example, the collection time within a charging duration is short. The electronic device can combine the charging parameters collected when connecting the charger for charging several times to obtain the charging parameters collected within multiple charging durations, and divide them according to time to obtain multi-dimensional charging parameters collected based on the target time series, which can be used as target charging data.

[0098] In one possible implementation, the electronic device can first obtain the curves of various charging parameters changing over time during the collection, and intercept these parameter curves according to the time series to obtain the multi-dimensional charging parameters corresponding to each collection moment. For example, through the collection, one or more types of data such as the voltage curve, current curve, temperature curve, capacitance curve, and incremental capacitance curve (IC curve) of the battery during the charging process can be obtained. By intercepting these collected curves according to the time series, the multi-dimensional charging parameters on the length of the time series can be obtained.

[0099] In one possible implementation, the preset impedance model in the electronic device is as described above. Figure 4 The structure shown is obtained by training, then, the time series that may be obtained during the acquisition in this step is different from the number of the first model sub-units in the encoder, so it can be intercepted or supplemented again according to the target time series, that is, the number of acquisition moments included in the target time series is the same as the number of multiple first model sub-units. When obtaining the target charging data of the battery in the current charging process, the initial charging data in the current charging process can be obtained first, and the initial charging data is a multi-dimensional charging parameter collected based on the initial time series; when the number of acquisition moments included in the initial charging data is greater than the number of multiple first model sub-units, the initial charging data is intercepted to obtain the target charging data corresponding to the target time series; when the number of acquisition moments included in the initial charging data is less than the number of multiple first model sub-units, the initial charging data is supplemented to obtain the target charging data corresponding to the target time series.

[0100] For example, the initial charging data is as follows: U 初始 =[U 1 , U 2 , …, U t ], and the target time series contains N acquisition moments. If t is greater than N, U is obtained based on the initial charging data by truncation. 目标 =[U 1 , U 2 , …, U N ], if t is less than N, the initial charging data is adjusted to U by padding. 目标 =[U 1 , U 2 , …, U N ]. When padding, you can choose a negative value as a placeholder, pad the length of the sequence to N at the end, and record the original length of the sequence.

[0101] Step 602: input the target charging data into a feature extraction module, extract hidden features corresponding to the target charging data through the feature extraction module, and input the hidden features into a prediction module.

[0102] Optionally, after obtaining the above-mentioned target charging data, the target charging data is input into a preset impedance model. In one possible implementation, the preset impedance model includes a feature extraction module and a prediction module. The feature extraction module and the prediction module can refer to the structure shown in the above-mentioned training process.

[0103] In one possible implementation, the feature extraction module is an encoder, which includes multiple first model sub-units connected in series; the electronic device inputs the target charging data into the feature extraction module, extracts hidden features corresponding to the target charging data through the feature extraction module, and inputs the hidden features into the prediction module. The process can be as follows: the multi-dimensional charging parameters of the target charging data at each acquisition moment are input into each first model sub-unit of the encoder in sequence according to the time sequence of the target time series, and each first model sub-unit calculates the multi-dimensional charging parameters input by each first model sub-unit to obtain the target hidden state and target cell state of each first model sub-unit, and the target hidden state and target cell state output by one of the first model sub-units of each first model sub-unit are used as the hidden features corresponding to the target charging data, and the hidden features are input into the prediction module.

[0104] Optionally, the target hidden state and target cell state output by the previous first model subunit are used as the input of the next first model subunit, and each first model subunit calculates the multi-dimensional charging parameters of its own input, including: inputting a preset initial hidden state, a preset initial cell state, and the multi-dimensional charging parameters at the first acquisition moment into the first first model subunit for calculation to obtain the target hidden state and target cell state of the first first model subunit, and calculating each subsequent first model subunit in series order to obtain the target hidden state and target cell state of each first model subunit.

[0105] For example, the feature extraction module and the prediction module are as above Figure 4 The structural arrangement shown in the figure, when making predictions, the preset initial hidden state of the input is equivalent to the above H 0 The preset initial cell state is equivalent to the above C 0 , C 0 and H 0 The specific value of the preset impedance model can be randomly generated each time and is not limited here. The first LSTM model unit is calculated according to the preset initial hidden state, the preset initial cell state and the multi-dimensional charging parameters at the first acquisition moment to obtain the target hidden state and target cell state of the next LSTM model unit, and so on.

[0106] In one possible implementation, the target hidden state and target cell state are the hidden state and cell state output by a target first model subunit among multiple first model subunits; when the number of acquisition moments included in the initial charging data is greater than or equal to the number of multiple first model subunits, the target first model subunit is the last first model subunit among the multiple first model subunits; when the number of acquisition moments included in the initial charging data is less than the number of multiple first model subunits, the target first model subunit is the Tth first model subunit among the multiple first model subunits, and T is the number of acquisition moments included in the initial time series.

[0107] For example, the initial charging data is still U 初始 =[U 1 , U 2 ,…,U t ] For example, the target time series contains N acquisition moments, corresponding to N LSTM model units in the encoder. If t is greater than N, U is obtained based on the initial charging data by truncation. 目标 =[U 1 , U 2 , …, U N ], when the preset impedance model is executed here, the target first model subunit selected by the encoder is the last first model subunit among the multiple first model subunits; if t is less than N, the initial charging data is adjusted to U by padding. 目标 =[U 1 , U 2 , …, U N ], when the preset impedance model is executed here, the target first model subunit selected by the encoder is the input of the plurality of first model subunits is U t The corresponding Tth first model subunit.

[0108] Step 603, calculating the hidden features through the prediction module to obtain the predicted impedance spectrum of the battery.

[0109] Optionally, after the required hidden features are extracted by the feature extraction module, the hidden features are input into the prediction module, and the prediction module calculates the hidden features to obtain the predicted impedance spectrum of the battery.

[0110] Optionally, the prediction module is a decoder, and the decoder includes multiple second model sub-units connected in series; the electronic device calculates the hidden features through the prediction module to obtain the predicted impedance spectrum of the battery as follows: the target hidden state and the target cell state are input into the first model sub-unit of the decoder, and after the target hidden state and the target cell state are calculated by the first model sub-unit, the calculation result is input into the next second model sub-unit, and the calculation result is calculated by the next second model sub-unit until each second model sub-unit is traversed; the calculation result of each second model sub-unit includes the impedance output result of the corresponding second model sub-unit, the new hidden state and the new cell state; according to the impedance output result of each second model sub-unit, the predicted impedance spectrum of the battery is obtained.

[0111] Refer to the above Figure 4 In the structure, the electronics generates an initial impedance S for the first second model subunit in the decoder. 0 The first and second model subunits receive the target hidden state, target cell state, and initial impedance S from the encoder. 0 The calculation starts, and each subsequent second model subunit continues to calculate based on the calculation result of the previous second model subunit, and finally the impedance output result of each second model subunit in the decoder is obtained, and the predicted impedance spectrum of the battery is obtained. 1 , S 2 , ... S M ] represents the impedance output result of each second model subunit, and the predicted impedance spectrum is obtained by [S 1 , S 2 , ... S M ] composed of impedance spectrum.

[0112] Optionally, in the predicted impedance spectrum calculated by this scheme, each impedance output result includes multi-dimensional data of two dimensions, wherein the multi-dimensional data of one dimension includes the real part of the predicted impedance value under different charging states and different frequency bands, and the multi-dimensional data of the other dimension includes the imaginary part of the predicted impedance value under different charging states and different frequency bands.

[0113] It should be noted that the LSTM model unit used by the encoder and decoder in the above scheme is exemplary. In practical applications, other traditional machine learning model units may also be used, such as: Gaussian Process Regression (GPR) model unit, or other deep learning model units, such as: Recurrent Neural Network (RNN) model unit, which is not limited here. Moreover, the example diagram and specific description of the LSTM model unit in the above scheme are only multiple examples of a usable model. For various traditional machine learning models and deep learning models mentioned in the scheme expansion, the example diagrams and specific calculation methods of the model structure are different from the examples. (For example, it does not necessarily include the two parts of the encoder and decoder, or the model unit used is not necessarily a deep learning model unit), so the change of the model structure does not occur at the sub-unit level in the specification, such as the cancellation and replacement of the encoder and decoder, etc. It is also possible to use other model structures to implement the method of this scheme.

[0114] In summary, the target charging data of the battery in the current charging process is obtained, and the target charging data includes multi-dimensional charging parameters collected based on the target time series; the target charging data is input into the preset impedance model to obtain the predicted impedance spectrum of the battery, and the preset impedance model is obtained by training based on the training sample data set, and the training sample data set includes the historical multi-dimensional charging parameters of the battery collected during the historical charging process and the corresponding historical actual impedance spectrum. In an embodiment of the present application, the collected target charging data includes multi-dimensional charging parameters, and the target charging data including the multi-dimensional charging parameters is predicted by the preset impedance model to obtain the predicted impedance spectrum of the battery. This process does not require the design of complex integrated circuits, which reduces the time spent on the process, and the multi-dimensional charging parameters provide diversified features, which increases the accuracy of the preset impedance model and improves the efficiency of obtaining the electrochemical impedance spectrum of the battery.

[0115] In addition, based on the more interpretable traditional machine learning methods, this solution also introduces deep learning models that are better at mining complex data relationships into the prediction scenario of battery electrochemical impedance spectra, increasing the generalization and adaptability of the model. In addition, adding multidimensional features to the input sequence increases the diversity of information and data that the model can learn, and improves the learning ability and stability of the model. In addition, the present invention simplifies the data preprocessing steps, uses the complete charging curve as input, and simultaneously predicts the EIS of the full frequency range under the full SOC, saving time cost and computing consumption, which is conducive to improving the performance and robustness of the model.

[0116] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0117] Please refer to Figure 7 , which shows a structural block diagram of an impedance spectrum acquisition device provided by an exemplary embodiment of the present application. The impedance spectrum acquisition device 700 can be used in electronic equipment. Figure 2 or Figure 6 All or part of the steps in the method provided by the illustrated embodiment are performed by an electronic device. The impedance spectrum acquisition device 700 comprises:

[0118] A first acquisition module 701 is used to acquire target charging data of the battery in the current charging process, where the target charging data includes multi-dimensional charging parameters collected based on a target time series;

[0119] The second acquisition module 702 is used to input the target charging data into a preset impedance model to obtain the predicted impedance spectrum of the battery, wherein the preset impedance model is trained based on a training sample data set, and the training sample data set includes historical multi-dimensional charging parameters of the battery collected during historical charging processes and corresponding historical actual impedance spectra.

[0120] In summary, the target charging data of the battery in the current charging process is obtained, and the target charging data includes multi-dimensional charging parameters collected based on the target time series; the target charging data is input into the preset impedance model to obtain the predicted impedance spectrum of the battery, and the preset impedance model is obtained by training based on the training sample data set, and the training sample data set includes the historical multi-dimensional charging parameters of the battery collected during the historical charging process and the corresponding historical actual impedance spectrum. In an embodiment of the present application, the collected target charging data includes multi-dimensional charging parameters, and the target charging data including the multi-dimensional charging parameters is predicted by the preset impedance model to obtain the predicted impedance spectrum of the battery. This process does not require the design of complex integrated circuits, which reduces the time spent on the process, and the multi-dimensional charging parameters provide diversified features, which increases the accuracy of the preset impedance model and improves the efficiency of obtaining the electrochemical impedance spectrum of the battery.

[0121] Optionally, the preset impedance model includes a feature extraction module and a prediction module, and the second acquisition module 702 includes: a first acquisition unit, a first calculation unit;

[0122] The first acquisition unit is used to input the target charging data into the feature extraction module, extract hidden features corresponding to the target charging data through the feature extraction module, and input the hidden features into the prediction module;

[0123] The first calculation unit is used to calculate the hidden features through the prediction module to obtain the predicted impedance spectrum of the battery.

[0124] Optionally, the feature extraction module is an encoder, and the encoder includes a plurality of first model sub-units connected in series; and the first acquisition unit is further used for:

[0125] The multidimensional charging parameters of the target charging data at each acquisition moment are input into each first model sub-unit of the encoder in sequence according to the time sequence of the target time series, and each first model sub-unit calculates the multidimensional charging parameters inputted respectively to obtain the target hidden state and target cell state of each first model sub-unit, and uses the target hidden state and target cell state outputted by one of the first model sub-units as the hidden features corresponding to the target charging data, and inputs the hidden features into the prediction module.

[0126] Optionally, the target hidden state and the target cell state output by the previous first model subunit are used as the input of the next first model subunit, and the multi-dimensional charging parameters of the respective inputs are calculated by the respective first model subunits, including:

[0127] The preset initial hidden state, the preset initial cell state and the multi-dimensional charging parameters at the first acquisition moment are input into the first first model subunit for calculation to obtain the target hidden state and target cell state of the first first model subunit, and each subsequent first model subunit is calculated in turn according to the series order to obtain the target hidden state and target cell state of each first model subunit.

[0128] Optionally, the prediction module is a decoder, and the decoder includes a plurality of second model sub-units connected in series;

[0129] The first computing unit is further configured to:

[0130] Inputting the target hidden state and the target cell state into the first model subunit of the decoder, calculating the target hidden state and the target cell state through the first model subunit, inputting the calculation result into the next second model subunit, calculating the calculation result through the next second model subunit, until traversing each second model subunit;

[0131] The calculation result of each second model subunit includes the impedance output result of the corresponding second model subunit, the new hidden state and the new cell state;

[0132] The predicted impedance spectrum of the battery is obtained according to the impedance output result of each second model subunit.

[0133] Optionally, the number of acquisition moments included in the target time series is the same as the number of the plurality of first model sub-units, and the first acquisition module 701 includes: a second acquisition unit, a first processing unit, and a second processing unit;

[0134] The second acquisition unit is used to acquire initial charging data in the current charging process, where the initial charging data is a multi-dimensional charging parameter collected based on an initial time series;

[0135] The first processing unit is configured to, when the number of collection moments included in the initial charging data is greater than the number of the plurality of first model sub-units, perform data interception processing on the initial charging data to obtain the target charging data corresponding to the target time series;

[0136] The second processing unit is used to perform data supplement processing on the initial charging data when the number of collection moments included in the initial charging data is less than the number of the multiple first model sub-units, so as to obtain the target charging data corresponding to the target time series.

[0137] Optionally, the target hidden state and the target cell state are the hidden state and the cell state output by a target first model subunit among the plurality of first model subunits;

[0138] In a case where the number of collection moments included in the initial charging data is greater than or equal to the number of the plurality of first model subunits, the target first model subunit is the last first model subunit among the plurality of first model subunits;

[0139] When the number of collection moments included in the initial charging data is less than the number of the multiple first model subunits, the target first model subunit is the Tth first model subunit among the multiple first model subunits, and T is the number of collection moments included in the initial time series.

[0140] Optionally, each of the impedance output results includes multidimensional data of two dimensions, wherein the multidimensional data of one dimension includes the real part of the predicted impedance value under different charging states and different frequency bands, and the multidimensional data of another dimension includes the imaginary part of the predicted impedance value under different charging states and different frequency bands.

[0141] Optionally, the device further comprises:

[0142] A third acquisition module is used to acquire a training sample data set before acquiring charging data of the battery during the charging process, wherein the training sample data set includes historical multi-dimensional charging parameters of the battery collected during the historical charging process and the corresponding historical actual impedance spectrum;

[0143] The fourth acquisition module is used to train the preset impedance model according to the training sample data set, and obtain the trained preset impedance model when the loss function of the preset impedance model converges, and the loss function includes a root mean square error function.

[0144] Please refer to Figure 8 , which is a structural diagram of another example of the impedance spectrum acquisition device provided in the embodiment of the present application. Among them, the impedance spectrum acquisition device 800 can be an electronic device that can implement the functions in the method provided in the embodiment of the present application. Among them, the impedance spectrum acquisition device 800 can be a chip system. In the embodiment of the present application, the chip system can be composed of a chip, or it can include a chip and other discrete devices.

[0145] In terms of hardware implementation, the communication module may be a transceiver, which is integrated into the impedance spectrum acquisition device 800 to form a communication interface 803 .

[0146] The impedance spectrum acquisition device 800 includes at least one processor 801, which is used to implement or support the impedance spectrum acquisition device 800 to implement the function of the electronic device in the method provided in the embodiment of the present application. Exemplarily, the processor 801 can execute the steps of obtaining the target charging data of the battery during the current charging process; inputting the target charging data into the preset impedance model to obtain the predicted impedance spectrum of the battery. Please refer to the detailed description in the method example for details, which will not be repeated here.

[0147] The impedance spectrum acquisition device 800 may also include at least one memory 802 for storing program instructions and / or data. The memory 802 is coupled to the processor 801. The coupling in the embodiment of the present application is an indirect coupling or communication connection between devices, units or modules, which may be electrical, mechanical or other forms, for information exchange between devices, units or modules. The processor 801 may operate in conjunction with the memory 802. The processor 801 may execute program instructions stored in the memory 802. At least one of the at least one memory may be included in the processor.

[0148] The impedance spectrum acquisition device 800 may also include a communication interface 803, which is used to communicate with other devices through a transmission medium, so that the device in the impedance spectrum acquisition device 800 can communicate with other devices. Exemplarily, the other device may be a network side device. The processor 801 may use the communication interface 803 to send and receive data. The communication interface 803 may specifically be a transceiver.

[0149] The specific connection medium between the communication interface 803, the processor 801 and the memory 802 is not limited in the embodiment of the present application. Figure 8 In the embodiment, the memory 802, the processor 801 and the communication interface 803 are connected via a bus 804. Figure 8 The connections between the other components are shown in bold lines, which are only for illustration and are not intended to be limiting. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0150] In the embodiment of the present application, the processor 801 can be a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and can implement or execute the methods, steps and logic block diagrams disclosed in the embodiment of the present application. The general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware processor to be executed, or the hardware and software modules in the processor can be combined and executed.

[0151] In an embodiment of the present application, the memory 802 may be a non-volatile memory, such as a hard disk drive (HDD) or a solid-state drive (SSD), etc., or a volatile memory (volatile memory), such as a random access memory (RAM). The memory is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory in the embodiment of the present application may also be a circuit or any other device that can implement a storage function, for storing program instructions and / or data.

[0152] Optionally, an embodiment of the present application further provides an electronic device, the electronic device comprising a processor and a memory, the memory storing a computer program, the computer program being executed by the processor to implement all or part of the steps performed by the electronic device in the impedance spectrum acquisition method described in the above embodiments.

[0153] Optionally, an embodiment of the present application further provides a computer-readable medium storing a computer program, which is executed by a processor to implement all or part of the steps performed by the electronic device in the impedance spectrum acquisition method described in the above embodiments.

[0154] Optionally, an embodiment of the present application also provides a computer program product, which stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set are loaded and executed by the processor to implement the impedance spectrum acquisition method described in the above embodiments, and all or part of the steps performed by the electronic device.

[0155] It should be noted that: when the device provided in the above embodiment performs the control of the electronic device, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0156] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0157] A person skilled in the art will appreciate that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk, or an optical disk, etc.

[0158] The above description is only an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for obtaining impedance spectrum, It is characterized in that Applied to electronic equipment, the method comprises: Acquire target charging data of the battery during the current charging process, wherein the target charging data includes multi-dimensional charging parameters collected based on a target time series; The target charging data is input into a preset impedance model to obtain a predicted impedance spectrum of the battery, wherein the preset impedance model is trained based on a training sample data set, and the training sample data set includes historical multi-dimensional charging parameters of the battery collected during historical charging processes and corresponding historical actual impedance spectra.

2. The method according to claim 1, It is characterized in that The preset impedance model includes a feature extraction module and a prediction module, and the target charging data is input into the preset impedance model to obtain the predicted impedance spectrum of the battery, including: Inputting the target charging data into the feature extraction module, extracting hidden features corresponding to the target charging data through the feature extraction module, and inputting the hidden features into the prediction module; The hidden features are calculated by the prediction module to obtain a predicted impedance spectrum of the battery.

3. The method according to claim 2, It is characterized in that The feature extraction module is an encoder, and the encoder includes a plurality of first model subunits connected in series; The step of inputting the target charging data into the feature extraction module, extracting hidden features corresponding to the target charging data through the feature extraction module, and inputting the hidden features into the prediction module includes: The multidimensional charging parameters of the target charging data at each acquisition moment are input into each first model sub-unit of the encoder in sequence according to the time sequence of the target time series, and each first model sub-unit calculates the multidimensional charging parameters inputted respectively to obtain the target hidden state and target cell state of each first model sub-unit, and uses the target hidden state and target cell state outputted by one of the first model sub-units as the hidden features corresponding to the target charging data, and inputs the hidden features into the prediction module.

4. The method according to claim 3, It is characterized in that The target hidden state and the target cell state output by the previous first model subunit are used as the input of the next first model subunit, and the multi-dimensional charging parameters of the respective inputs are calculated by the respective first model subunits, including: The preset initial hidden state, the preset initial cell state and the multi-dimensional charging parameters at the first acquisition moment are input into the first first model subunit for calculation to obtain the target hidden state and target cell state of the first first model subunit, and each subsequent first model subunit is calculated in turn according to the series order to obtain the target hidden state and target cell state of each first model subunit.

5. The method according to claim 3 or 4, It is characterized in that The prediction module is a decoder, and the decoder includes a plurality of second model sub-units connected in series; The step of calculating the hidden features by the prediction module to obtain the predicted impedance spectrum of the battery includes: Inputting the target hidden state and the target cell state into the first model subunit of the decoder, calculating the target hidden state and the target cell state through the first model subunit, inputting the calculation result into the next second model subunit, calculating the calculation result through the next second model subunit, until traversing each second model subunit; The calculation result of each second model subunit includes the impedance output result of the corresponding second model subunit, the new hidden state and the new cell state; The predicted impedance spectrum of the battery is obtained according to the impedance output result of each second model subunit.

6. The method according to claim 3 or 4, It is characterized in that The number of acquisition moments included in the target time series is the same as the number of the plurality of first model subunits, and the acquiring target charging data of the battery in the current charging process includes: Acquire initial charging data during the current charging process, wherein the initial charging data is multi-dimensional charging parameters collected based on an initial time series; When the number of collection moments included in the initial charging data is greater than the number of the plurality of first model sub-units, intercepting data is processed on the initial charging data to obtain the target charging data corresponding to the target time series; When the number of collection moments included in the initial charging data is less than the number of the plurality of first model sub-units, data supplementation is performed on the initial charging data to obtain the target charging data corresponding to the target time series.

7. The method according to claim 6, It is characterized in that The target hidden state and the target cell state are the hidden state and the cell state output by the target first model subunit among the plurality of first model subunits; In a case where the number of collection moments included in the initial charging data is greater than or equal to the number of the plurality of first model subunits, the target first model subunit is the last first model subunit among the plurality of first model subunits; When the number of collection moments included in the initial charging data is less than the number of the multiple first model subunits, the target first model subunit is the Tth first model subunit among the multiple first model subunits, and T is the number of collection moments included in the initial time series.

8. The method according to any one of claims 1 to 4, It is characterized in that Each of the impedance output results includes multidimensional data of two dimensions, wherein the multidimensional data of one dimension includes the real part of the predicted impedance value under different charging states and different frequency bands, and the multidimensional data of another dimension includes the imaginary part of the predicted impedance value under different charging states and different frequency bands.

9. The method according to any one of claims 1 to 4, It is characterized in that Before acquiring charging data of the battery during the charging process, the method further includes: Acquire a training sample data set, wherein the training sample data set includes historical multi-dimensional charging parameters of the battery collected during historical charging processes and corresponding historical actual impedance spectra; The preset impedance model is trained according to the training sample data set, and the trained preset impedance model is obtained when the loss function of the preset impedance model converges, and the loss function includes a root mean square error function.

10. An impedance spectrum acquisition device, It is characterized in that Applied to electronic equipment, the device comprises: A first acquisition module is used to acquire target charging data of the battery in the current charging process, wherein the target charging data includes multi-dimensional charging parameters collected based on a target time series; A second acquisition module is used to input the target charging data into a preset impedance model to obtain a predicted impedance spectrum of the battery, wherein the preset impedance model is trained based on a training sample data set, and the training sample data set includes historical multi-dimensional charging parameters of the battery collected during historical charging processes and corresponding historical actual impedance spectra.

11. An electronic device, It is characterized in that The electronic device comprises a processor and a memory, wherein a computer program is stored in the memory, and the computer program is executed by the processor to implement the impedance spectrum acquisition method according to any one of claims 1 to 9.

12. A computer-readable storage medium, It is characterized in that The storage medium stores a computer program, and the computer program is executed by a processor to implement the impedance spectrum acquisition method according to any one of claims 1 to 9.

Images