A lithium battery life prediction method, device and electronic equipment

Abstract

The present invention provides a lithium battery life prediction method, device and electronic device, wherein the lithium battery life prediction method includes: collecting battery data of a target battery after the charging and discharging process; parsing the battery data of the target battery to obtain a charging and discharging data set of the target battery; calling a pre-trained prediction model to predict the battery life of the target battery according to the charging and discharging data set of the target battery to obtain a prediction result; the prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer and an output layer; the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module and an output gate. Through the present invention, the success rate and accuracy of battery life prediction can be improved, and it can be applied to different application scenarios and battery types, solving the problem of low accuracy of battery life prediction in existing related technologies.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery life prediction, and in particular to a lithium battery life prediction method, device and electronic equipment. Background Art

[0002] In the prior art, there are various methods for predicting the life of lithium batteries, each of which has its own characteristics and shortcomings. After searching, the Chinese invention patent with publication number CN105068009A discloses a method for predicting the cycle life of a battery, which includes the following steps: placing the battery to be evaluated in the cycle conditions to be evaluated for cycle testing, recording the cumulative number of cycles and cycle capacity retention rate of the battery, and at the same time, performing a small current charge and discharge test on the battery every certain number of cycles or capacity loss rate, recording the voltage and capacity data of the battery during this charge and discharge process, as well as the corresponding number of cycles and capacity retention rate; thereby, data fitting and calculation are performed based on the cumulative number of cycles and cycle capacity retention rate of the battery and the differential data of capacity to voltage, and the cycle life of the battery is predicted.

[0003] This empirical lithium battery life prediction method relies on the usage guidelines and empirical formulas provided by battery manufacturers to predict battery life through a simple mathematical model. Due to the simplistic assumptions, the prediction accuracy is low in actual application and it is difficult to adapt to different usage conditions and battery types.

[0004] There is currently no effective solution to the problem of low accuracy in battery life prediction in existing related technologies. Summary of the invention

[0005] The present invention provides a lithium battery life prediction method, device and electronic equipment, which are used to solve the defect of low accuracy of battery life prediction in the prior art.

[0006] In a first aspect, the present invention provides a lithium battery life prediction method, comprising:

[0007] Collect battery data of the target battery after the charging and discharging process;

[0008] Parsing the battery data of the target battery to obtain a charge and discharge data set of the target battery;

[0009] Calling the pre-trained prediction model to predict the battery life of the target battery according to the charge and discharge data set of the target battery to obtain a prediction result;

[0010] The prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer and an output layer; the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module and an output gate;

[0011] The forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor;

[0012] The input gate identifies key parameter items for life prediction of the target battery through key feature adjustment factors;

[0013] The cell status update module is used to cumulatively represent the health data of the target battery;

[0014] The output gate is used to determine a parameter item related to the life span of the target battery.

[0015] According to a lithium battery life prediction method provided by the present invention, training the prediction model includes:

[0016] Acquire the pre-built prediction model and sample battery data after the sample battery undergoes a preset charging and discharging process;

[0017] Parsing the sample battery data to obtain a sample data set of the sample battery;

[0018] The prediction model is trained using the sample data set with the goal of minimizing the model loss.

[0019] According to a lithium battery life prediction method provided by the present invention, the preset charging and discharging process includes:

[0020] Set different simulation conditions;

[0021] The sample battery is subjected to a simulated charge and discharge process according to the corresponding simulation conditions until the battery capacity of the sample battery is less than 50% of the initial capacity.

[0022] According to a lithium battery life prediction method provided by the present invention, the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module and an output gate;

[0023] The forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor;

[0024] The input gate identifies key parameter items for life prediction of the target battery through key feature adjustment factors;

[0025] The cell status update module is used to cumulatively represent the health data of the target battery;

[0026] The output gate is used to determine a parameter item related to the life span of the target battery.

[0027] According to a lithium battery life prediction method provided by the present invention, the determination of the accelerated elimination factor includes:

[0028] Obtaining battery health parameters of the sample battery after the same number of charge and discharge cycles under two different simulation conditions;

[0029] The accelerated rejection factor is determined based on the acquired battery health parameter.

[0030] According to a lithium battery life prediction method provided by the present invention, the accelerated elimination factor is determined based on the acquired battery health parameter, including:

[0031] Calculating the difference between the battery health parameters of the sample battery under two different simulation conditions;

[0032] Compare the difference with the first health change threshold and the second health change threshold:

[0033] If the difference is less than the first health change threshold, the accelerated elimination factor is 0.1;

[0034] If the difference reaches the first health change threshold but does not exceed the second health change threshold, the accelerated elimination factor is 0.2;

[0035] If the difference exceeds the second health change threshold, the accelerated elimination factor is 0.3.

[0036] According to a lithium battery life prediction method provided by the present invention, the determination of the key characteristic adjustment factor includes:

[0037] Obtaining battery health parameters of the sample battery after the same number of charge and discharge cycles under two different simulation conditions;

[0038] The key characteristic adjustment factor is determined based on the acquired battery health parameter.

[0039] According to a lithium battery life prediction method provided by the present invention, the key feature adjustment factor is determined based on the acquired battery health parameter, including:

[0040] Calculating the difference between the battery health parameters of the sample battery under two different simulation conditions;

[0041] Compare the difference with the third health change threshold and the fourth health change threshold:

[0042] If the difference is less than the third health change threshold, the key feature adjustment factor is 0.1;

[0043] If the difference reaches the third health change threshold but does not exceed the fourth health change threshold, the key feature adjustment factor is 0.2;

[0044] If the difference exceeds the fourth health change threshold, the key feature adjustment factor is 0.3.

[0045] According to a lithium battery life prediction method provided by the present invention, before predicting the battery life of the target battery based on the charge and discharge data set of the target battery, it includes: preprocessing the charge and discharge data set of the target battery; the preprocessing includes data cleaning and data standardization.

[0046] In a second aspect, the present invention further provides a lithium battery life prediction device, comprising:

[0047] The acquisition module is used to collect the battery data of the target battery after the charging and discharging process;

[0048] A parsing module, used to parse the battery data of the target battery to obtain a charge and discharge data set of the target battery;

[0049] A prediction module, used to call a pre-trained prediction model to predict the battery life of the target battery according to the charge and discharge data set of the target battery to obtain a prediction result;

[0050] The prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer and an output layer; the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module and an output gate;

[0051] The forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor;

[0052] The input gate identifies key parameter items for life prediction of the target battery through key feature adjustment factors;

[0053] The cell status update module is used to cumulatively represent the health data of the target battery;

[0054] The output gate is used to determine a parameter item related to the life span of the target battery.

[0055] In a third aspect, the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the lithium battery life prediction method as described in the first aspect above is implemented.

[0056] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the lithium battery life prediction method as described in the first aspect above.

[0057] In a fifth aspect, the present invention further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the lithium battery life prediction method as described in the first aspect above.

[0058] The lithium battery life prediction method provided by the present invention predicts the battery life of the target battery by calling the trained prediction model and taking the charge and discharge data set of the target battery as input to obtain the prediction result. The prediction model is a long short-term memory network model built based on the Keras library. The LSTM model adopts the architecture of input layer->LSTM layer->Dropout layer->fully connected layer->output layer. By introducing an accelerated elimination factor to adjust the determination of discarding parameter items, irrelevant parameter items can be discarded more resolutely during the training process to improve the efficiency of training. The key feature adjustment factor is used in the input gate to identify the key parameter items related to the target battery life prediction, improve the prediction success rate of the key parameter items for the target battery life prediction, reduce the amount of training required for the prediction model, improve the prediction accuracy of the prediction model, and can be applied to different application scenarios and battery types, solving the problem of low accuracy of battery life prediction in the existing related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 is a flow chart of the lithium battery life prediction method proposed by the present invention;

[0061] Figure 2 It is a structural block diagram of the lithium battery life prediction device provided by the present invention;

[0062] Figure 3 It is a schematic diagram of the connection and use of the battery life rapid simulation device provided by the present invention;

[0063] Figure 4 It is a structural schematic diagram of a battery life rapid simulation device provided by the present invention;

[0064] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] The present invention provides a lithium battery life prediction method. Figure 1 is a flow chart of the lithium battery life prediction method proposed by the present invention, such as Figure 1 As shown, the method comprises the following steps:

[0067] Step S101, collecting battery data of a target battery after the charging and discharging process.

[0068] Step S102, parsing the battery data of the target battery to obtain a charge and discharge data set of the target battery.

[0069] Step S103, calling the pre-trained prediction model, predicting the battery life of the target battery according to the charge and discharge data set of the target battery, and obtaining the prediction result. The prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer, and an output layer. The LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module, and an output gate; the forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor; the input gate identifies the key parameter items for the life prediction of the target battery through the key feature adjustment factor; the cell state update module is used to accumulate the health data of the target battery; the output gate is used to determine the parameter items related to the life of the target battery.

[0070] Exemplarily, when the battery life of the target battery needs to be predicted, first, the battery data of the target battery after the charging and discharging process is collected. Then, the battery data is parsed to obtain the charging and discharging data set of the target battery. The charging and discharging data set can be battery health data, voltage data, current data, temperature data, humidity data, external pressure data, etc. Then, the trained prediction model is called, and the charging and discharging data set of the target battery is used as input to predict the battery life of the target battery to obtain the prediction result. The prediction model is a long short-term memory network model (Long Short-Term Memory, LSTM) built based on the Keras library. The LSTM model adopts the architecture of input layer->LSTM layer->Dropout layer->fully connected layer->output layer. By introducing an accelerated elimination factor to adjust the determination of discarding parameter items, during the training process, irrelevant parameter items can be discarded more resolutely to improve the efficiency of training. Key feature adjustment factors are used in the input gate to identify key parameter items related to the target battery life prediction, thereby improving the prediction success rate of the key parameter items for the target battery life prediction, reducing the amount of training required for the prediction model, and improving the prediction accuracy of the prediction model. The model can also be applied to different application scenarios and battery types, solving the problem of low battery life prediction accuracy in existing related technologies.

[0071] In some of the embodiments, training the prediction model includes: obtaining a pre-built prediction model and sample battery data after the sample battery undergoes a preset charging and discharging process; parsing the sample battery data to obtain a sample data set of the sample battery; and training the prediction model using the sample data set with the goal of minimizing model loss.

[0072] In this embodiment, first, a pre-built prediction model and sample battery data of a sample battery after a preset charge and discharge process are obtained. Among them, the sample battery can select different types of lithium batteries, and the preset charge and discharge process can also be set according to different actual application scenarios. Through more different training scenarios, the scope of application and prediction accuracy of the prediction model can be improved. Then, the sample battery data is parsed to obtain a sample data set of the sample battery. Correspondingly, the sample data set also includes data sets such as battery health data, voltage data, current data, temperature data, humidity data, and external pressure data of the sample battery. Furthermore, the sample data set is divided to obtain a training set and a validation set. The training set is used to train the prediction model. After training, the prediction model is verified and fine-tuned through the validation set to evaluate the performance and generalization ability of the prediction model.

[0073] Specifically, the preset charge and discharge process includes: setting different simulation conditions; and performing a simulated charge and discharge process on the sample battery according to the corresponding simulation conditions until the battery capacity of the sample battery is less than 50% of the initial capacity.

[0074] For example, different simulation environments can be set during the simulation, such as temperature and humidity, air pressure, and external mechanical pressure, and the charge and discharge cycles are performed under different gradual environments until the battery capacity of the sample battery is less than 50% of the initial capacity. For example, when other environmental parameters are constant and different temperatures are set, such as 10°C-80°C, a simulation item is set every 10°C. This simulation form can obtain more detailed actual parameter items of the sample battery 100, providing more abundant parameter items for the subsequent training of the prediction model.

[0075] In the above embodiment, the calculation formula of the forget gate in the LSTM layer is as follows:

[0076]

[0077] in, Represents the output of the forget gate, indicating whether the parameter item should be used; Represents the weight matrix of the forget gate, which is used to calculate the input of the forget gate; Represents the hidden state of the previous time step, which is used to capture the information of the previous time step; Represents the input of the current time step, including the hidden state from the previous time step and the new input of the current time step; Represents the bias term of the forget gate, which is used to adjust the output of the forget gate; The sigmoid activation function compresses the weighted sum result to between 0 and 1 and outputs a probability value. c is the acceleration culling factor, c The value is 0.1, 0.2 or 0.3. And, ≈0 means that the information of this parameter item should be completely forgotten. ≈1 means that the information of this parameter item should be fully preserved.

[0078] For the accelerated elimination factor, it can accelerate the forgetting of parameter items that are not related to the sample battery life prediction. In this way, the forget gate can determine which parameter items are important and which are not important for the battery health status update at the current time step, thereby helping the prediction model learn long-term dependencies.

[0079] Determination of the accelerated rejection factor includes: obtaining battery health parameters of the sample battery after the same number of charge and discharge cycles under two different simulation conditions; and determining the accelerated rejection factor based on the obtained battery health parameters.

[0080] Specifically, the accelerated elimination factor is determined based on the acquired battery health parameters, including: calculating the difference in battery health parameters of the sample battery under two different simulation conditions; comparing the difference and the preset first health change threshold and second health change threshold: if the difference is less than the first health change threshold, the accelerated elimination factor is 0.1; if the difference reaches the first health change threshold but does not exceed the second health change threshold, the accelerated elimination factor is 0.2; if the difference exceeds the second health change threshold, the accelerated elimination factor is 0.3.

[0081] For example, let the battery health parameter after the nth charge and discharge cycle under the A simulation condition be SOHAn, let the battery health parameter after the nth charge and discharge cycle under the B simulation condition be SOHBn, first compare the size of SOHAn and SOHBn, the simulation condition corresponding to the larger battery health parameter is the parameter item in the discard operation in the forget gate, and then calculate the difference X=|SOHAn-SOHBn|, let the first health change threshold Xa and the second health change threshold Xb, if X<Xa, then choose c =0.1, if Xa≤X≤Xb, then c =0.2, if X>Xb then c =0.3.

[0082] Acceleration factor for a parameter term c The determination process requires comparing the battery health parameters after completing the same number of charge and discharge times under two simulation conditions. For the simulation condition corresponding to the larger battery health parameter, it means that this parameter item has less impact on the sample battery life than other parameter items, so it needs to be eliminated faster.

[0083] In the LSTM layer, the input gate is calculated as follows:

[0084]

[0085]

[0086] in, represents the output of the input gate, represents the weight matrix of the input gate, represents the hidden state of the previous time step, represents the input of the current time step, represents the bias term of the forget gate, is the sigmoid activation function, the output value is 0-1, d represents the key feature adjustment factor, d The value is 0.1, 0.2 or 0.3; represents the candidate health status, represents the weight matrix, represents the bias term of the candidate health state, yes function, The function compresses the result of the weighted sum to between -1 and 1 and outputs a temporary state.

[0087] For the key characteristic adjustment factor, the determination of the key characteristic adjustment factor includes: obtaining battery health parameters of the sample battery after the same number of charge and discharge cycles under two different simulation conditions; and determining the key characteristic adjustment factor based on the obtained battery health parameters.

[0088] Similarly, the key characteristic adjustment factor is determined based on the acquired battery health parameters, including: calculating the difference in battery health parameters of the sample battery under two different simulation conditions; comparing the difference with the preset third health change threshold and the fourth health change threshold: if the difference is less than the third health change threshold, the key characteristic adjustment factor is 0.1; if the difference reaches the third health change threshold but does not exceed the fourth health change threshold, the key characteristic adjustment factor is 0.2; if the difference exceeds the fourth health change threshold, the key characteristic adjustment factor is 0.3.

[0089] For example, let the battery health after the nth charge and discharge cycle under the E simulation condition be SOHEn, let the battery health after the nth charge and discharge cycle under the F simulation condition be SOHFn, first compare the size of SOHEn and SOHFn, the condition corresponding to the smaller battery health parameter is the parameter item that needs to be input in the input gate, and then calculate the difference =|SOHEn-SOHFn|, let the third health change threshold value Xc and the fourth health change threshold value Xd, if <Xc, then choose d =0.1, if Xc≤ ≤Xd, then d =0.2, if >Xd d =0.3.

[0090] In the prediction model, the calculation formula of the fully connected layer is as follows:

[0091]

[0092] in, represents the output of the fully connected layer, represents the weight of the fully connected layer, represents the hidden layer state of the current step, b represents the bias of the fully connected layer.

[0093] Since there may be some abnormal data in the charge and discharge data set obtained by direct analysis, in order to improve data quality, in some embodiments, step S103, before predicting the battery life of the target battery based on the charge and discharge data set of the target battery, includes: preprocessing the charge and discharge data set of the target battery; the preprocessing includes data cleaning and data standardization.

[0094] Exemplarily, after parsing the charge and discharge data set, the outliers and missing values ​​in the charge and discharge data set are removed, and the data therein is scaled to the range of 0-1, and then the corresponding parameter items are extracted, such as voltage, charge and discharge time, temperature change, health change, etc., to facilitate the division of data sets with different parameter items. Through the above data processing, the storage occupied and data volume of the charge and discharge data set can be effectively reduced, the data quality can be improved, and the accuracy of the prediction can be improved. Correspondingly, in the process of training the prediction model, the above data preprocessing operation can also be performed on the sample data set of the sample battery to improve the training effect.

[0095] The present invention also provides a lithium battery life prediction device. The lithium battery life prediction device provided by the present invention is described below. The lithium battery life prediction device described below and the lithium battery life prediction method described above can be referenced to each other. Figure 2 is a structural block diagram of the lithium battery life prediction device provided by the present invention, such as Figure 2 As shown, the device comprises:

[0096] The acquisition module 201 is used to acquire the battery data of the target battery after the charging and discharging process;

[0097] The parsing module 202 is used to parse the battery data of the target battery to obtain a charge and discharge data set of the target battery;

[0098] The prediction module 203 is used to call the pre-trained prediction model to predict the battery life of the target battery according to the charge and discharge data set of the target battery to obtain a prediction result;

[0099] The prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer, and an output layer; the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module, and an output gate;

[0100] The forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor;

[0101] The input gate identifies the key parameter items of the life prediction of the target battery through the key feature adjustment factors;

[0102] The cell status update module is used to cumulatively represent the health data of the target battery;

[0103] The output gate is used to determine a parameter term related to the life of the target battery.

[0104] When the battery life of the target battery needs to be predicted, first, the acquisition module 201 acquires the battery data of the target battery after the charging and discharging process. Then, the parsing module 202 parses the battery data to obtain the charging and discharging data set of the target battery. The charging and discharging data set can be battery health data, voltage data, current data, temperature data, humidity data, external pressure data, etc. Then, the prediction module 203 calls the trained prediction model, takes the charging and discharging data set of the target battery as input, predicts the battery life of the target battery, and obtains the prediction result. The prediction model is a long short-term memory network model (Long Short-Term Memory, LSTM) built based on the Keras library. The LSTM model adopts the architecture of input layer->LSTM layer->Dropout layer->fully connected layer->output layer, and is obtained by training a large number of samples in different application scenarios, which can improve the success rate and accuracy of battery life prediction, and can be applied to different application scenarios and battery types, solving the problem of low accuracy of battery life prediction in existing related technologies.

[0105] In order to realize the process of training the prediction model in the above method, the present invention also provides a battery life rapid simulation device, which is connected to the above lithium battery life prediction device for use. Figure 3 This is a schematic diagram of the connection and use of the battery life rapid simulation device provided by the present invention. Figure 4 Schematic diagram of the structure of the battery life rapid simulation device provided by the present invention. Figure 3 and Figure 4 As shown, the device includes multiple simulation stations 101, multiple sealed environment simulation chambers, multiple battery simulators 102 and a chip-based simulation data collector 103. When in use, multiple sample batteries 100 are connected to multiple simulation stations 101 respectively, and multiple simulation stations 101 are electrically connected to multiple battery simulators 102 respectively. After multiple sample batteries 100 are installed respectively, multiple sealed environment simulation chambers are buckled on multiple simulation stations 101 to provide sample batteries 100 with a simulation environment with adjustable temperature, humidity and pressure. The simulation data collector 103 is respectively connected to the data output end of the multiple battery simulators 102 to collect sample battery data of all rapid charge and discharge processes of the multiple sample batteries 100. The lithium battery life prediction device 104 is pre-installed with a prediction program designed using a lithium battery life prediction method based on deep learning. The lithium battery life prediction device 104 communicates with the simulation data collector 103 to exchange data sets and train and verify the prediction model.

[0106] The battery life rapid simulation device is implemented based on multiple simulation stations 101, multiple sealed environment simulation chambers 1 and multiple battery simulators 102. The multiple simulation stations 101 provide installation stations for multiple sample batteries 100. The sealed environment simulation chamber can provide external environment simulation for the sample battery 100, such as temperature and humidity, pressure and external mechanical pressure simulation. The battery simulator 102 can simulate rapid charge and discharge cycles. The simulation data collector 103 can collect simulation parameters and parameter items such as voltage value, current value and battery health value of the sample battery 100 during the charge and discharge cycle. Finally, these data sets are collected and sent to the life prediction computer 104. The life prediction computer 104 organizes the data to form a training set, a verification set and a test set. The training set and the verification set are input into the LSTM model. After the LSTM model is trained, the test set is used for testing. After the test is completed, it can be used to predict the life of the sample battery 100.

[0107] Specifically, the simulation data collector 103 includes a data collector 1031, a memory 1032 and a data analysis chip 1033. The multiple data input ports of the data collector 1031 are respectively communicated with the data output ends of the multiple battery simulators 102, the multiple data output ports of the data collector 1031 are respectively communicated with the memory 1032, the data analysis chip 1033 is communicated with the memory 1032, and the memory 1032 is communicated with the life prediction computer 104.

[0108] The sealed environment simulation chamber includes a chamber body 11, a temperature and humidity regulating device 12, an air pressure regulating device 13 and a mechanical pressure simulation device 14. The chamber body 11 can be opened and closed by a hinge and installed on the simulation workstation 101. When the chamber body 11 is buckled with the simulation workstation 101, the interior is a closed space. The temperature and humidity regulating device 12 and the air pressure regulating device 13 are respectively installed on the inner wall of the chamber body 11, and are used to adjust the temperature, humidity and air pressure values ​​in the closed space. The mechanical pressure simulation device 14 includes a servo electric cylinder 141 and a pressure block 142. The shell of the servo electric cylinder 141 is installed on the inner wall of the chamber body 11, and the pressure block 142 is installed at the end of the piston rod of the servo electric cylinder 141. When the piston rod of the servo electric cylinder 141 is extended, the pressure block 142 presses the upper surface of the sample battery 100.

[0109] Both the temperature and humidity regulating device 12 and the air pressure regulating device 13 can be implemented by using ready-made simulations, and the temperature, humidity and air pressure values ​​can be adjusted steplessly, while the mechanical pressure simulation device 14 can set the change of the external pressure value through the servo electric cylinder 141. Appropriate external pressure can maintain close contact between the various interfaces of the battery, prevent the stratification and deformation of the active material during the charging and discharging process, and thus have a favorable effect on the life of the lithium battery. However, if the external pressure is too large or applied improperly, it may also cause adverse effects. Excessive external pressure may cause internal short circuits, thermal runaway, life decay and other processes. This is because excessive pressure may interfere with the normal chemical and physical processes inside the battery. Therefore, providing a simulation of the external pressure value when simulating the environment is important for predicting the life of lithium batteries.

[0110] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5 As shown, the electronic device may include: a processor 501, a communication interface 502, a memory 503 and a communication bus 504, wherein the processor 501, the communication interface 502 and the memory 503 communicate with each other through the communication bus 504. The processor 501 may call the logic instructions in the memory 503 to execute the lithium battery life prediction method, which includes:

[0111] Collect battery data of the target battery after the charging and discharging process;

[0112] Parse the battery data of the target battery to obtain the charge and discharge data set of the target battery;

[0113] Call the pre-trained prediction model to predict the battery life of the target battery according to the charge and discharge data set of the target battery to obtain the prediction result;

[0114] The prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer, and an output layer; the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module, and an output gate;

[0115] The forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor;

[0116] The input gate identifies the key parameter items of the life prediction of the target battery through the key feature adjustment factors;

[0117] The cell status update module is used to cumulatively represent the health data of the target battery;

[0118] The output gate is used to determine a parameter term related to the life of the target battery.

[0119] In addition, the logic instructions in the above-mentioned memory 503 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. 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, etc. Various media that can store program codes.

[0120] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the lithium battery life prediction method provided by the above methods, the method includes:

[0121] Collect battery data of the target battery after the charging and discharging process;

[0122] Parse the battery data of the target battery to obtain the charge and discharge data set of the target battery;

[0123] Call the pre-trained prediction model to predict the battery life of the target battery according to the charge and discharge data set of the target battery to obtain the prediction result;

[0124] The prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer, and an output layer; the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module, and an output gate;

[0125] The forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor;

[0126] The input gate identifies the key parameter items of the life prediction of the target battery through the key feature adjustment factors;

[0127] The cell status update module is used to cumulatively represent the health data of the target battery;

[0128] The output gate is used to determine a parameter term related to the life of the target battery.

[0129] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the lithium battery life prediction method provided by the above methods, the method comprising:

[0130] Collect battery data of the target battery after the charging and discharging process;

[0131] Parse the battery data of the target battery to obtain the charge and discharge data set of the target battery;

[0132] Call the pre-trained prediction model to predict the battery life of the target battery according to the charge and discharge data set of the target battery to obtain the prediction result;

[0133] The prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer, and an output layer; the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module, and an output gate;

[0134] The forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor;

[0135] The input gate identifies the key parameter items of the life prediction of the target battery through the key feature adjustment factors;

[0136] The cell status update module is used to cumulatively represent the health data of the target battery;

[0137] The output gate is used to determine a parameter term related to the life of the target battery.

[0138] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0139] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A lithium battery life prediction method, characterized in that: include: Collect battery data of the target battery after the charging and discharging process; Parsing the battery data of the target battery to obtain a charge and discharge data set of the target battery; Calling the pre-trained prediction model to predict the battery life of the target battery according to the charge and discharge data set of the target battery to obtain a prediction result; The prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer and an output layer; the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module and an output gate; The forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor; The input gate identifies key parameter items for life prediction of the target battery through key feature adjustment factors; The cell status update module is used to cumulatively represent the health data of the target battery; The output gate is used to determine a parameter item related to the life of the target battery; The determination of the accelerated elimination factor includes: Obtain the battery health parameters of the sample battery after the same number of charge and discharge cycles under two different simulation conditions; Determining the accelerated rejection factor based on the acquired battery health parameter; The determination of the key characteristic adjustment factor includes: Obtaining battery health parameters of the sample battery after the same number of charge and discharge cycles under two different simulation conditions; The key characteristic adjustment factor is determined based on the acquired battery health parameter.

2. The lithium battery life prediction method according to claim 1, characterized in that: Training the prediction model includes: Acquire the pre-built prediction model and sample battery data after the sample battery undergoes a preset charging and discharging process; Parsing the sample battery data to obtain a sample data set of the sample battery; The prediction model is trained using the sample data set with the goal of minimizing the model loss.

3. The lithium battery life prediction method according to claim 2, characterized in that: The preset charging and discharging process includes: Set different simulation conditions; The sample battery is subjected to a simulated charge and discharge process according to the corresponding simulation conditions until the battery capacity of the sample battery is less than 50% of the initial capacity.

4. The lithium battery life prediction method according to claim 1, characterized in that: Determining the accelerated rejection factor based on the acquired battery health parameter includes: Calculating the difference between the battery health parameters of the sample battery under two different simulation conditions; Compare the difference with the preset first health change threshold and the second health change threshold: If the difference is less than the first health change threshold, the accelerated elimination factor is 0.1; If the difference reaches the first health change threshold but does not exceed the second health change threshold, the accelerated elimination factor is 0.2; If the difference exceeds the second health change threshold, the accelerated elimination factor is 0.

3.

5. The lithium battery life prediction method according to claim 4, characterized in that: Determining the key feature adjustment factor based on the acquired battery health parameter includes: Calculating the difference between the battery health parameters of the sample battery under two different simulation conditions; Compare the difference with the preset third health change threshold and the fourth health change threshold: If the difference is less than the third health change threshold, the key feature adjustment factor is 0.1; If the difference reaches the third health change threshold but does not exceed the fourth health change threshold, the key feature adjustment factor is 0.2; If the difference exceeds the fourth health change threshold, the key feature adjustment factor is 0.

3.

6. The lithium battery life prediction method according to claim 1, characterized in that: Before predicting the battery life of the target battery according to the charge and discharge data set of the target battery, it includes: preprocessing the charge and discharge data set of the target battery; the preprocessing includes data cleaning and data standardization.

7. A lithium battery life prediction device, characterized in that: include: The acquisition module is used to collect the battery data of the target battery after the charging and discharging process; A parsing module, used to parse the battery data of the target battery to obtain a charge and discharge data set of the target battery; A prediction module, used to call a pre-trained prediction model to predict the battery life of the target battery according to the charge and discharge data set of the target battery to obtain a prediction result; The prediction model includes an input layer, an LSTM layer, a Dropout layer, a fully connected layer and an output layer; the LSTM layer of the prediction model includes a forget gate, an input gate, a cell state update module and an output gate; The forget gate adjusts the determination of discarding parameter items by accelerating the elimination factor; The input gate identifies key parameter items for life prediction of the target battery through key feature adjustment factors; The cell status update module is used to cumulatively represent the health data of the target battery; The output gate is used to determine a parameter item related to the life of the target battery; The determination of the accelerated elimination factor includes: Obtain the battery health parameters of the sample battery after the same number of charge and discharge cycles under two different simulation conditions; Determining the accelerated rejection factor based on the acquired battery health parameter; The determination of the key characteristic adjustment factor includes: Obtaining battery health parameters of the sample battery after the same number of charge and discharge cycles under two different simulation conditions; The key characteristic adjustment factor is determined based on the acquired battery health parameter.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the lithium battery life prediction method as described in any one of claims 1 to 6 is implemented.

Images