Method and device for predicting residual life of lithium battery
By using multiple sets of training data and LSTM network combined with Monte Carlo sampling method, the problem of low prediction accuracy of lithium battery life is solved, and the reliability and prediction accuracy of battery management are improved.
Patent Information
- Application Number
- CN202411167235.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-11-29
AI Technical Summary
The existing lithium battery life prediction methods have low accuracy, especially the lack of reliability in management methods, resulting in inaccurate battery life prediction results.
The battery residual life prediction model obtained by machine learning training using multiple sets of training data is generated, combining long and short-term memory artificial neural network (LSTM) and Monte Carlo sampling method to generate multiple prediction samples, train and MC-LSTM network through training sets Perform parameter tuning to capture long-term dependencies and fast-changing patterns in time series data.
It improves the prediction accuracy and reliability of lithium battery life, achieves more accurate battery life prediction, and provides a reliable decision-making basis for battery management.
Smart Images

Figure CN119044806A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of battery management, and in particular to a method and device for predicting the remaining battery life of a lithium battery. Background Art
[0002] The life of lithium-ion batteries is affected by many factors, such as the number of charge and discharge cycles, charge and discharge rate, temperature, depth of charge and discharge, etc. Traditional life prediction methods based on statistical models are often limited by the amount of data and model accuracy, and it is difficult to accurately predict the remaining life of the battery. Therefore, more and more research is currently devoted to using machine learning, artificial intelligence and other methods, combined with a large amount of battery operation data, to achieve accurate prediction of the remaining life of lithium-ion batteries. These new methods can better capture the working status and health of the battery, and provide a more reliable decision-making basis for the battery management system.
[0003] At present, the main research methods for predicting the remaining service life of lithium batteries are generally divided into two categories: experience-based and performance-based methods. Among them, the experience-based RUL prediction method is a statistical method for predicting the remaining service life of the battery based on the experience knowledge during the use of the battery, which mainly includes the cycle number method, the ampere-hour method, the weighted ampere-hour method, and the event-oriented aging accumulation method. The ampere-hour method here is to calculate the remaining power of the battery by recording the discharge amount of the battery and subtracting the discharge amount from the actual capacity of the battery. Since the impact of battery discharge on the battery under different working conditions is different, corresponding weights are set for different working conditions. This is the weighted ampere-hour method. This method uses the discharge weight multiplied by the discharge ampere-hour as the total ampere-hour, and estimates the remaining power with the total ampere-hour. The cycle number method counts the cycle cycles of the battery. When the number of cycle cycles of the battery is equal to the failure threshold, the battery is considered to have reached its service life. This method needs to consider the impact of factors such as different cycle conditions and cycle states on the cycle life, and determine the battery life based on both experience and standard parameters. The event-oriented aging accumulation method sets the events that cause battery life damage, sets corresponding weights based on the damage to the battery life caused by these events, calculates the number of these events and multiplies them by the weights to get the damage value. When the damage value exceeds the failure threshold, the battery is considered to have reached the end of its service life. The experience-based RUL prediction method is simple and fast to calculate, but it has high requirements for measuring equipment such as sensors and is difficult to apply in practice outside the laboratory environment. At present, the mainstream direction of RUL prediction is still performance-based methods.
[0004] Performance-based RUL prediction methods can include: (1) Model-based methods: Model-based methods usually establish physical models based on mechanical failure mechanisms, or statistical models based on empirical knowledge and real-time measurements to describe the degradation process of LIBs. It is necessary to have a deep understanding of the composition of the model, in which mathematical expressions are used to describe complex electrochemical processes. It has little dependence on historical data, and can be used for prediction research even without much sample data. Commonly used models include electrochemical models, equivalent circuit models, and empirical models. The electrochemical model consists of three processes, namely mass transfer, conduction, and electrochemical reaction. The degradation model is mainly established by analyzing the electrochemical properties of lithium-ion batteries to achieve RUL prediction; the equivalent circuit model simulates the process of converting chemical energy into electrical energy inside the battery through circuit elements such as battery voltage, current, resistance, and capacitance, and represents the external characteristics of the battery through its change law; the empirical model is a mathematical model of system parameters and system variables, and is an empirical function that conforms to the statistical analysis law of system parameters and variables. Empirical models require less battery physical or electrochemical information. They are usually established based on regression analysis of battery degradation data. The empirical model is a model that is fitted by continuously updating parameters through measurement or estimation during the cycle, such as linear, exponential and polynomial models; the particle filter (PF) algorithm here mainly uses the unscented Kalman filter (UKF) to optimize the selection of the proposed density distribution and improve the prediction accuracy. In addition, when using the unscented particle filter method, the covariance matrix is first decomposed into singular values to obtain the equivalent eigenvalue expression, and then the UT transformation is performed again to make the prediction result more accurate. (2) Data-driven method: The data-driven method refers to the method that does not consider the complex internal electrochemical mechanism of the battery, and from the measured lithium-ion battery related data characteristics (capacity, internal resistance, voltage, current and temperature, etc.), through the change of parameters, finds the laws and related information of the detection data and behavior patterns, and uses machine learning methods to construct the state mapping relationship between degradation data and health status, so as to predict the RUL of lithium batteries. Data-driven RUL prediction methods have been widely studied and have many advantages, such as model-free, high accuracy and strong robustness. Unlike model-based methods that require specific model expressions, data-driven methods use machine learning or deep learning technology to predict the RUL of lithium-ion batteries. Data-driven RUL prediction usually includes data collection, data preprocessing, model training and prediction.At present, there are many machine learning methods for RUL prediction, including artificial neural networks (ANN), support vector machine (SVM), relevance vector machine (RVM), autoregressive integrated moving average (ARIMA), autoregressive model (AR), Gaussian process regression (GPR) and other methods. Among them, the battery degradation characteristics are used as the conditions for battery remaining life prediction, and the Bayesian theory is used to quantitatively analyze the battery remaining service life. A lithium-ion battery RUL prediction method based on dynamic grey relevance vector machine is proposed in the prior art. This method uses the trend prediction results obtained by the discrete grey model as the input of RVM regression prediction, and dynamically updates the RVM regression prediction model according to the regression prediction results, thereby improving the accuracy of long-term trend prediction. (3) Method based on model and data-driven fusion: The combined method overcomes the limitations of model-based and data-driven prediction, and can give full play to the advantages of both model-based and data-driven methods. More and more researchers prefer to use fusion methods to predict the remaining service life of batteries to improve the accuracy of prediction. In the existing technology, tent chaotic mapping, adaptive weight, reverse learning strategy and Cauchy mutation perturbation strategy are introduced to respectively initialize the population, update the position of the discoverer, improve the optimal solution of the sparrow population, optimize the relevant parameters of LSTM, complete the model construction and realize the prediction of RUL of lithium-ion batteries. The fusion method performs well in RUL prediction and can combine the advantages of various algorithms. It is an extremely practical estimation method if used properly, but how to combine multiple methods to make the best use of their strengths and avoid their weaknesses is still a difficult problem to solve.
[0005] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention
[0006] The embodiments of the present invention provide a method and device for predicting the remaining battery life of a lithium battery, so as to at least solve the technical problem that the method for managing lithium batteries in the related art has low reliability, resulting in low accuracy of the results of battery life prediction of the lithium battery.
[0007] According to one aspect of an embodiment of the present invention, a method for predicting the remaining battery life of a lithium battery is provided, comprising: obtaining battery characteristic data of a target lithium battery, wherein the target lithium battery is a battery for which remaining battery life prediction is required, and the battery characteristic data is used to represent characteristic information of the target lithium battery; filtering out decay characteristic data of the target lithium battery from the battery characteristic data; inputting the decay characteristic data into a battery remaining life prediction model, so as to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained by machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample decay characteristic data and sample battery remaining life corresponding to the sample decay characteristic data.
[0008] Optionally, before inputting the decay characteristic data into a battery remaining life prediction model to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, the battery remaining life prediction method for the lithium battery also includes: acquiring the multiple groups of training data including multiple sample decay characteristic data and multiple sample battery remaining lives; training the multiple groups of training data including multiple sample decay characteristic data and multiple sample battery remaining lives to obtain the battery remaining life prediction model.
[0009] Optionally, the multiple groups of training data including multiple sample decay feature data and multiple sample battery remaining life are trained to obtain the battery remaining life prediction model, including: determining that the basic model of the battery remaining life prediction model is a long short-term memory artificial neural network LSTM model; converting multiple sample decay feature data into sample decay feature sequences according to the passage of time, and determining multiple sample battery remaining life as prediction label data; inputting the sample decay feature sequence and the prediction label data into the LSTM model at the same time, so as to use the LSTM model to capture the sample decay characteristics of the sample decay feature sequence in the time series; comparing the sample decay feature with the prediction label data to obtain a comparison result; when the comparison result indicates that the error between the sample decay feature and the prediction label data is less than an error threshold, stopping the model training to obtain the battery remaining life prediction model.
[0010] Optionally, the multiple groups of training data including multiple sample decay characteristic data and multiple sample battery remaining lifespans are trained to obtain the battery remaining life prediction model, including: normalizing the multiple sample decay characteristic data and multiple sample battery remaining lifespans to obtain the multiple sample decay characteristic data and multiple sample battery remaining lifespans after normalization; dividing the multiple sample decay characteristic data and multiple sample battery remaining lifespans after normalization into data sets to obtain a training set, a validation set and a test set; and obtaining the battery remaining life prediction model through training with the training set, the validation set and the test set.
[0011] Optionally, after inputting the degradation characteristic data into a battery remaining life prediction model to process the degradation characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, the battery remaining life prediction method for the lithium battery further includes: evaluating the current remaining life by a predetermined evaluation method to obtain an evaluation result, wherein the predetermined evaluation method includes: root mean square error, mean absolute error and goodness of fit; when the evaluation result indicates that the current remaining life is within a predetermined range, determining the current remaining life to be the actual value of the current remaining life of the target lithium battery; when the evaluation result indicates that the current remaining life is not within the predetermined range, adjusting the current remaining life.
[0012] According to another aspect of an embodiment of the present invention, a device for predicting a remaining battery life of a lithium battery is provided, comprising: a first acquisition module, used to acquire battery characteristic data of a target lithium battery, wherein the target lithium battery is a battery for which a remaining battery life prediction is required, and the battery characteristic data is used to represent characteristic information of the target lithium battery; a screening module, used to screen out decay characteristic data of the target lithium battery from the battery characteristic data; and an input module, used to input the decay characteristic data into a battery remaining life prediction model, so as to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained through machine learning training using multiple sets of training data, and each of the multiple sets of training data comprises: sample decay characteristic data and sample battery remaining life corresponding to the sample decay characteristic data.
[0013] Optionally, the battery remaining life prediction device for a lithium battery further includes: a second acquisition module, used to acquire the multiple groups of training data including multiple sample decay characteristic data and multiple sample battery remaining lives before inputting the decay characteristic data into a battery remaining life prediction model to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery; a training module, used to train the multiple groups of training data including multiple sample decay characteristic data and multiple sample battery remaining lives to obtain the battery remaining life prediction model.
[0014] Optionally, the training module includes: a first determination unit, used to determine that the basic model of the battery remaining life prediction model is a long short-term memory artificial neural network (LSTM) model; a first processing unit, used to convert multiple sample decay feature data into a sample decay feature sequence over time, and determine multiple sample battery remaining lives as prediction label data; a second processing unit, used to input the sample decay feature sequence and the prediction label data into the LSTM model at the same time, so as to use the LSTM model to capture the sample decay characteristics of the sample decay feature sequence in the time series; a comparison unit, used to compare the sample decay feature with the prediction label data to obtain a comparison result; and a third processing unit, used to stop model training when the comparison result indicates that the error between the sample decay feature and the prediction label data is less than an error threshold, so as to obtain the battery remaining life prediction model.
[0015] Optionally, the training module includes: a normalization processing unit, used to normalize the multiple sample decay characteristic data and the multiple sample battery remaining lifespans to obtain the multiple sample decay characteristic data and the multiple sample battery remaining lifespans after normalization; a data set division unit, used to divide the multiple sample decay characteristic data and the multiple sample battery remaining lifespans after normalization to obtain a training set, a validation set and a test set; and an acquisition unit, used to obtain the battery remaining life prediction model obtained by training the training set, the validation set and the test set.
[0016] Optionally, the battery remaining life prediction device of the lithium battery also includes: an evaluation module, which is used to input the degradation characteristic data into a battery remaining life prediction model, so as to process the degradation characteristic data using the battery remaining life prediction model, and then evaluate the current remaining life by a predetermined evaluation method to obtain an evaluation result after obtaining the current remaining life of the target lithium battery, wherein the predetermined evaluation method includes: root mean square error, mean absolute error and goodness of fit; a determination module, which is used to determine that the current remaining life is the actual value of the current remaining life of the target lithium battery when the evaluation result indicates that the current remaining life is within a predetermined range; and an adjustment module, which is used to adjust the current remaining life when the evaluation result indicates that the current remaining life is not within a predetermined range.
[0017] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, the computer-readable storage medium comprising a stored program, wherein the program executes any one of the above-mentioned methods for predicting the remaining battery life of a lithium battery.
[0018] According to another aspect of an embodiment of the present invention, a processor is further provided, wherein the processor is used to run a program, wherein when the program is run, the method for predicting the remaining battery life of a lithium battery described in any one of the above is executed.
[0019] According to another aspect of an embodiment of the present invention, a computer program product is provided, comprising computer instructions, wherein the computer instructions are executed by a processor to implement any one of the above-mentioned methods for predicting the remaining battery life of a lithium battery.
[0020] In an embodiment of the present invention, battery characteristic data of a target lithium battery is obtained, wherein the target lithium battery is a battery for which remaining battery life prediction is required, and the battery characteristic data is used to represent characteristic information of the target lithium battery; decay characteristic data of the target lithium battery is screened out from the battery characteristic data; the decay characteristic data is input into a battery remaining life prediction model, so as to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained through machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample decay characteristic data and sample battery remaining life corresponding to the sample decay characteristic data. Through the technical solution provided by the present invention, a plurality of prediction samples are generated by the Monte Carlo sampling method, and the remaining life of the lithium battery is predicted by training the training set and tuning the parameters of the MC-LSTM network. At the same time, the performance and accuracy of the MC-LSTM network are evaluated, thereby achieving the technical effect of better capturing the long-term dependencies and rapidly changing patterns in time series data, improving the accuracy and reliability of the prediction of the life of the lithium battery, and further solving the technical problem that the method for managing lithium batteries in the related technology is of low reliability, resulting in low accuracy of the battery life prediction results of the lithium battery. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:
[0022] Figure 1 It is a hardware structure block diagram of a mobile terminal of a method for predicting the remaining battery life of a lithium battery according to an embodiment of the present invention;
[0023] Figure 2 is a flow chart of a method for predicting the remaining battery life of a lithium battery according to an embodiment of the present invention;
[0024] Figure 3 is a structural schematic diagram of a method for predicting the remaining battery life of a lithium battery according to an embodiment of the present invention;
[0025] Figure 4 is a schematic diagram of the internal structure of an LSTM network according to an embodiment of the present invention;
[0026] Figure 5 is a schematic diagram of an LSTM network structure according to an embodiment of the present invention;
[0027] Figure 6 is a flowchart of a model training process according to an embodiment of the present invention;
[0028] Figure 7is a schematic diagram of an energy-saving control device for a water heater according to an embodiment of the present invention. DETAILED DESCRIPTION
[0029] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 should fall within the scope of protection of the present invention.
[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0031] As introduced in the background technology, the method used to manage lithium batteries in the related art has low reliability, resulting in a technical problem of low accuracy in the result of battery life prediction of lithium batteries. In view of the above defects, a method and device for predicting the remaining battery life of a lithium battery, a computer-readable storage medium, a processor, and a computer program product are provided in an embodiment of the present invention.
[0032] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.
[0033] The method embodiments provided in the embodiments of the present invention can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 1 is a hardware structure block diagram of a mobile terminal of a method for predicting the remaining battery life of a lithium battery according to an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.
[0034] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the method for predicting the remaining battery life of a lithium battery in an embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.
[0035] According to an embodiment of the present invention, a method embodiment of a method for predicting the remaining battery life of a lithium battery is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.
[0036] Figure 2 is a flow chart of a method for predicting the remaining battery life of a lithium battery according to an embodiment of the present invention. Figure 2 As shown, the method comprises the following steps:
[0037] Step S202, obtaining battery characteristic data of a target lithium battery, wherein the target lithium battery is a battery for which remaining battery life prediction is required, and the battery characteristic data is used to represent characteristic information of the target lithium battery.
[0038] In this embodiment, characteristic information representing a target lithium battery of a battery for which remaining battery life prediction is required can be obtained, which facilitates subsequent analysis of battery characteristic data of the target lithium battery, thereby providing important information for remaining battery life prediction.
[0039] Step S204, filtering out the degradation characteristic data of the target lithium battery from the battery characteristic data.
[0040] It should be noted that the above-mentioned degradation characteristic data can be expressed as the residual of historical capacity, that is, the difference between the battery capacity at the current moment and the historical capacity. It can be selected according to the correlation coefficient. The correlation coefficient is calculated to evaluate the correlation between the historical capacity residual and the battery life degradation to determine the most predictive feature.
[0041] In this embodiment, the degradation characteristic data of the target lithium battery can be screened out from the battery characteristic data, and the remaining life prediction and battery management analysis can be performed based on this, which can help quantify the battery degradation condition and thus detect the battery degradation condition in a timely manner.
[0042] Step S206, inputting the decay characteristic data into the battery remaining life prediction model, so as to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained through machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample decay characteristic data and sample battery remaining life corresponding to the sample decay characteristic data.
[0043] In this embodiment, the decay characteristic data can be input into a battery remaining life prediction model obtained by machine learning training using multiple groups of sample decay characteristic data and the corresponding sample battery remaining life. By processing the decay characteristic data to obtain the current remaining life of the target lithium battery, the current remaining life of the target lithium battery can be predicted, providing important reference information for battery management and maintenance, and helping to extend the battery life.
[0044] Figure 3 is a schematic diagram of the structure of a method for predicting the remaining battery life of a lithium battery according to an embodiment of the present invention. Figure 3 As shown, the original measurement data of the lithium battery is preprocessed and then input into the battery remaining life prediction model, wherein the model is obtained by machine learning training and outputs the current remaining life of the target lithium battery.
[0045] As can be seen from the above, in an embodiment of the present invention, the battery characteristic data of the target lithium battery is obtained, wherein the target lithium battery is a battery for which the remaining battery life prediction is required, and the battery characteristic data is used to represent the characteristic information of the target lithium battery; the decay characteristic data of the target lithium battery is screened out from the battery characteristic data; the decay characteristic data is input into the battery remaining life prediction model, so as to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained by machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample decay characteristic data and sample battery remaining life corresponding to the sample decay characteristic data. The method achieves the purpose of generating multiple prediction samples using the Monte Carlo sampling method, training the training set and tuning the parameters of the MC-LSTM network to realize the prediction of the remaining life of the lithium battery, and at the same time, evaluating the performance and accuracy of the MC-LSTM network, thereby achieving the technical effect of better capturing the long-term dependencies and fast-changing patterns in the time series data, and improving the prediction accuracy and reliability of the lithium battery life.
[0046] The technical solution provided by the embodiment of the present invention solves the technical problem that the method for managing lithium batteries in the related art has low reliability, resulting in low accuracy of the results of battery life prediction of lithium batteries.
[0047] According to the above-mentioned embodiment of the present invention, before the degradation characteristic data is input into the battery remaining life prediction model to process the degradation characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, the battery remaining life prediction method of the lithium battery also includes: obtaining multiple groups of training data including multiple sample degradation characteristic data and multiple sample battery remaining lives; training the multiple groups of training data including multiple sample degradation characteristic data and multiple sample battery remaining lives to obtain the battery remaining life prediction model.
[0048] In this embodiment, multiple sets of training data including multiple sample decay characteristic data and multiple sample battery remaining life are obtained through multiple sampling to generate multiple prediction books, and the data are trained to obtain a battery remaining life prediction model. Multiple sampling can provide prediction results from different perspectives, thereby more comprehensively evaluating the relationship between the decay characteristics of the samples and the remaining battery life.
[0049] According to the above embodiment of the present invention, multiple groups of training data including multiple sample decay feature data and multiple sample battery remaining life are trained to obtain a battery remaining life prediction model, including: determining that the basic model of the battery remaining life prediction model is a long short-term memory artificial neural network LSTM model; converting multiple sample decay feature data into a sample decay feature sequence according to the passage of time, and determining multiple sample battery remaining life as prediction label data; inputting the sample decay feature sequence and the prediction label data into the LSTM model at the same time, so as to use the LSTM model to capture the sample decay characteristics of the sample decay feature sequence on the time series; comparing the sample decay feature with the prediction label data to obtain a comparison result; when the comparison result indicates that the error between the sample decay feature and the prediction label data is less than the error threshold, stopping the model training to obtain the battery remaining life prediction model.
[0050] In this embodiment, the basic model of the battery remaining life prediction model is determined to be a long short-term memory artificial neural network LSTM model. After the LSTM layer, the Monte Carlo sampling method is introduced, that is, a dropout layer is added to the network, and multiple prediction books are generated through multiple samplings, and the remaining battery life of multiple samples is determined as prediction label data, which can reduce the risk of overfitting and increase the generalization ability of the model; then the sample decay feature sequence and the prediction label data are simultaneously input into the LSTM model to capture the sample decay features of the sample decay feature sequence in the time series, and the sample decay features are compared with the prediction label data. When the comparison result indicates that the error between the sample decay features and the prediction label data is less than the error threshold, the model training is stopped to obtain the battery remaining life prediction model, and the error threshold is cited as the condition for stopping training. The training strategy of the model can be adjusted according to the error situation during the model training process.
[0051] It should be noted that the above-mentioned LSTM model is a recurrent neural network (RNN) suitable for processing time series data. It can capture long-term dependencies in sequence data and is suitable for predicting the remaining life of the battery. By introducing the LSTM layer, the model can effectively process time series data and capture the decay characteristics of the samples.
[0052] Figure 4 Schematic diagram of the internal structure of an LSTM network according to an embodiment of the present invention. Figure 4As shown in the figure, the LSTM neural network adds a hidden cell state and designs three "gate" structures. σ represents the sigmoid activation function, and the output upper and lower bounds of this function are (0,1), where 0 represents the discard state and 1 represents the successful pass state (i.e., the output state of the activation function). The output upper and lower bounds of tanh (i.e., the activation function) are (-1,1). Specifically: The forget gate is mainly used to decide whether some information needs to be eliminated. It can decide which information needs to be forgotten in the previous time step and which information needs to be retained in the cell state. t =σ(W f *[h t-1 ,x t ]+b f ), in which W f represents the weight matrix, h t-1 Represents the output state at the previous moment, x t Indicates the input status at the current moment, b f represents the bias value, f t Represents the output value; the input gate is mainly used to decide whether to retain some information. It can decide the new information added to the cell state based on the current input and the output of the previous time step. t =σ(W i *[h t-1 ,x t ]+b i ), C t =tanh(W C *[h t-1 ,x t ]+b C ), in which W i represents the weight matrix, b i Represents the bias value, W C The weight matrix representing the cell state, b C is the bias value of the cell state, C t Represents the candidate cell state; the output gate mainly calculates the cell state results of the input gate and the forget gate, and can determine the current output value based on the current input and cell state. t =σ(W o *[h t-1 ,x t ]+b o ), h t =o t *tanh(C t ), in which W o represents the weight matrix, b o represents the offset value, o tRepresents the output value, C t Represents the cell state at the current moment, h t Represents the output of the cell at time t; the cell state is used to store and transmit information, and is updated under the control of the forget gate and the input gate. The process of updating the cell state is: C t =f t *C t-1 +i t *C t , in which f t Represented as the output of the forget gate, i t represents the output of the input gate, C t-1 Indicates the cell state at the last moment; y (t) represents the output of the output layer at time t, and the output of the "output layer" at each moment in the middle of the sequence. The mathematical expression is as follows: (t) =σ(Vh (t) )+c, where V represents the fully connected matrix.
[0053] Figure 5 is a schematic diagram of an LSTM network structure according to an embodiment of the present invention, such as Figure 5 As shown in the figure, the LSTM network structure includes an input layer, an output layer and a hidden layer. The characteristics of the lithium battery are obtained through feature selection and feature processing. This data is input into the LSTM network model as a time series, and the capacity ratio is used as a prediction label. Both are input into the written program at the same time, and the data set is divided. The number of program neuron layers is continuously debugged by using the optimizer and activation function. The LSTM network learns and memorizes by capturing the features on the time series, completes the subsequent prediction and outputs the results.
[0054] It should be noted that the above-mentioned optimizer is used to update the weight coefficient and give the gradient of the weight update, that is, the direction and size. Using a reasonable optimizer can accelerate the convergence of the network and obtain the global optimal solution; the above-mentioned activation function is used to increase the nonlinearity of the entire network, that is, the expressive ability or abstract ability, so that the network can learn deeper content in the data. The activation functions used in this embodiment are Sigmoid, Relu, Tanh, etc.
[0055] According to the above embodiment of the present invention, multiple groups of training data including multiple sample decay characteristic data and multiple sample battery remaining life are trained to obtain a battery remaining life prediction model, including: normalizing the multiple sample decay characteristic data and the multiple sample battery remaining life to obtain the normalized multiple sample decay characteristic data and the multiple sample battery remaining life; dividing the normalized multiple sample decay characteristic data and the multiple sample battery remaining life into data sets to obtain a training set, a validation set and a test set; and obtaining a battery remaining life prediction model through training with the training set, the validation set and the test set.
[0056] In this embodiment, multiple sample decay feature data and multiple sample battery remaining life can be normalized, and then the data set can be divided to obtain a training set, a validation set, and a test set and train them to obtain a battery remaining life prediction model.
[0057] like Figure 5 As shown in the figure, through the normalization method, the features of different dimensions are converted into dimensionless data samples of equal time intervals, so that the data of different dimensions take smaller values in a similar range, avoiding large data input into the network, causing large gradient updates of the network, and thus making the network unable to converge and learning slower. The normalization calculation formula is as follows: In the formula, x is the original data, x min is the minimum value of the original data, x max is the maximum value of the original data.
[0058] Figure 6 is a flowchart of the model training process according to an embodiment of the present invention. The MC-LSTM network training process is as follows Figure 6 As shown in the figure, taking the maximum capacity percentage of lithium battery as the prediction target, the decay features are selected from the lithium battery data set, the decay features are preprocessed, the processed data set is input into the MC-LSTM model for training, and the optimal fusion features obtained by fusion of lithium battery features are divided into 3 parts: training set, validation set and test set. The first 70% of the data is defined as the training set, and the model is trained through repeated learning and memory. 20% of the data is the validation set, which is used to judge the quality of the previous training. The last 10% of the data set is the test set, which is mainly used to judge the quality of the model.
[0059] It should be noted that the above preprocessing is to preprocess the original data using a variational mode decomposition algorithm.
[0060] According to the above-mentioned embodiment of the present invention, after the degradation characteristic data is input into the battery remaining life prediction model, and the battery remaining life prediction model is used to process the degradation characteristic data to obtain the current remaining life of the target lithium battery, the battery remaining life prediction method of the lithium battery also includes: evaluating the current remaining life by a predetermined evaluation method to obtain an evaluation result, wherein the predetermined evaluation method includes: root mean square error, mean absolute error and goodness of fit; when the evaluation result indicates that the current remaining life is within a predetermined range, determining the current remaining life as the actual value of the current remaining life of the target lithium battery; when the evaluation result indicates that the current remaining life is not within the predetermined range, adjusting the current remaining life.
[0061] In this embodiment, after averaging or weighted averaging multiple prediction samples generated by the MC layer to obtain the final lithium battery remaining life prediction result, the predicted values of different model outputs (i.e., current remaining life) are evaluated by error analysis, i.e., root mean square error, mean absolute error, and goodness of fit. When the evaluation result indicates that the current remaining life is within a predetermined range, the current remaining life is determined to be the actual value of the current remaining life of the target lithium battery. Otherwise, the current remaining life is adjusted.
[0062] It should be noted that the above different models refer to the comparison between EMD-EMD-BiLSTM-AM and EMD-BiLSTM-AM networks.
[0063] The specific contents of the evaluation are as follows: The model uses the root mean square error (E RMS ), mean absolute error (E MA ) and goodness of fit (R2) are used to effectively evaluate the prediction results, namely Where: E RMS is the root mean square error, E MA is the mean absolute error, n is the number of samples, y i is the actual value, is the model predicted output value, is the average of the actual values.
[0064] It should be noted that the above samples are historical battery capacity values, and the above actual values are true battery capacity values.
[0065] Through the technical solution provided by the above embodiment of the present invention, firstly, the current status of relevant research on lithium battery life prediction is investigated, and the advantages and applications of MC-LSTM network in time series prediction are analyzed; historical data of lithium batteries are collected, preprocessed, and data sets are prepared for model training; the MC-LSTM network structure is designed, and the remaining service life of lithium batteries is predicted by training with training sets and parameter tuning; the model is simulated based on the data set, the simulation results are analyzed, and compared with other prediction methods, the performance and accuracy of MC-LSTM network in lithium battery life prediction are evaluated. Taking into account various factors such as model accuracy and data volume, the remaining life of lithium batteries can be accurately predicted. It avoids the shortcomings of high equipment requirements in traditional methods and the difficulty of combining multiple methods to maximize strengths and avoid weaknesses.
[0066] That is, the above technical solution provided by the embodiment of the present invention comprehensively considers various factors, adopts appropriate feature extraction methods and advanced prediction models, and reasonably selects and adjusts optimizers and activation functions, which can improve the accuracy and generalization ability of the model and more accurately predict the current remaining life of the lithium battery.
[0067] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.
[0068] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.
[0069] According to an embodiment of the present invention, there is also provided a device for predicting the remaining battery life of a lithium battery for implementing the above-mentioned method for predicting the remaining battery life of a lithium battery. Figure 7 Schematic diagram of an energy-saving control device for a water heater according to an embodiment of the present invention. Figure 7As shown, the device includes: a first acquisition module 701, a screening module 703 and an input module 705. The device for predicting the remaining battery life of a lithium battery is described below.
[0070] The first acquisition module 701 is used to acquire battery characteristic data of a target lithium battery, wherein the target lithium battery is a battery for which remaining battery life prediction is required, and the battery characteristic data is used to represent characteristic information of the target lithium battery.
[0071] The screening module 703 is used to screen out the degradation characteristic data of the target lithium battery from the battery characteristic data.
[0072] Input module 705 is used to input the decay characteristic data into the battery remaining life prediction model so as to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained through machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample decay characteristic data and sample battery remaining life corresponding to the sample decay characteristic data.
[0073] It should be noted here that the above-mentioned first acquisition module 701, screening module 703 and input module 705 correspond to steps S202 to S206 in the above-mentioned embodiment, and the three modules and the corresponding steps implement the same instances and application scenarios, but are not limited to the contents disclosed in the above-mentioned embodiment.
[0074] As can be seen from the above, in the scheme recorded in the above embodiment of the present invention, first, the battery characteristic data of the target lithium battery can be obtained by using the first acquisition module, wherein the target lithium battery is a battery for which the remaining battery life prediction is required, and the battery characteristic data is used to represent the characteristic information of the target lithium battery; secondly, the decline characteristic data of the target lithium battery can be screened out from the battery characteristic data by using the screening module; finally, the decline characteristic data can be input into the battery remaining life prediction model by using the input module, so as to process the decline characteristic data by using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained by machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample decline characteristic data and sample battery remaining life corresponding to the sample decline characteristic data, so as to achieve the purpose of generating multiple prediction samples by using the Monte Carlo sampling method, training the training set and tuning the parameters of the MC-LSTM network, realizing the prediction of the remaining life of the lithium battery, and at the same time, evaluating the performance and accuracy of the MC-LSTM network, thereby achieving the technical effect of better capturing the long-term dependencies and fast-changing patterns in the time series data, and improving the prediction accuracy and reliability of the lithium battery life.
[0075] The technical solution provided by the embodiment of the present invention solves the technical problem that the method for managing lithium batteries in the related art has low reliability, resulting in low accuracy of the results of battery life prediction of lithium batteries.
[0076] In an optional embodiment, the battery remaining life prediction device for a lithium battery further includes: a second acquisition module, used to acquire multiple sets of training data including multiple sample decay characteristic data and multiple sample battery remaining lives before inputting the decay characteristic data into a battery remaining life prediction model to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery; a training module, used to train the multiple sets of training data including multiple sample decay characteristic data and multiple sample battery remaining lives to obtain a battery remaining life prediction model.
[0077] In an optional embodiment, the training module includes: a first determination unit, used to determine that the basic model of the battery remaining life prediction model is a long short-term memory artificial neural network LSTM model; a first processing unit, used to convert multiple sample decay feature data into a sample decay feature sequence according to the passage of time, and determine the multiple sample battery remaining life as prediction label data; a second processing unit, used to input the sample decay feature sequence and the prediction label data into the LSTM model at the same time, so as to capture the sample decay characteristics of the sample decay feature sequence in the time series by using the LSTM model; a comparison unit, used to compare the sample decay feature with the prediction label data to obtain a comparison result; and a third processing unit, used to stop model training when the comparison result indicates that the error between the sample decay feature and the prediction label data is less than an error threshold, so as to obtain a battery remaining life prediction model.
[0078] In an optional embodiment, the training module includes: a normalization processing unit, which is used to normalize multiple sample decay characteristic data and multiple sample battery remaining life to obtain multiple sample decay characteristic data and multiple sample battery remaining life after normalization; a data set division unit, which is used to divide the multiple sample decay characteristic data and multiple sample battery remaining life after normalization into data sets to obtain a training set, a verification set and a test set; and an acquisition unit, which is used to obtain a battery remaining life prediction model trained by the training set, the verification set and the test set.
[0079] In an optional embodiment, the battery remaining life prediction device of the lithium battery also includes: an evaluation module, which is used to input the degradation characteristic data into the battery remaining life prediction model, so as to process the degradation characteristic data using the battery remaining life prediction model, and then evaluate the current remaining life by a predetermined evaluation method to obtain an evaluation result after obtaining the current remaining life of the target lithium battery, wherein the predetermined evaluation method includes: root mean square error, mean absolute error and goodness of fit; a determination module, which is used to determine that the current remaining life is the actual value of the current remaining life of the target lithium battery when the evaluation result indicates that the current remaining life is within a predetermined range; and an adjustment module, which is used to adjust the current remaining life when the evaluation result indicates that the current remaining life is not within the predetermined range.
[0080] According to another aspect of an embodiment of the present invention, a processor is further provided, and the processor is used to run a program, wherein when the program is run, any one of the above-mentioned methods for predicting the remaining battery life of a lithium battery is executed.
[0081] According to another aspect of an embodiment of the present invention, a computer program product is provided, including computer instructions, and the computer instructions are executed by a processor to implement any one of the above-mentioned methods for predicting the remaining battery life of a lithium battery.
[0082] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, the computer-readable storage medium comprising a stored program, wherein the program executes any one of the above-mentioned methods for predicting the remaining battery life of a lithium battery.
[0083] Optionally, in this embodiment, the computer-readable storage medium may be located in any one of the computer terminals in a computer terminal group in a computer network, or in any one of the communication devices in a communication device group.
[0084] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: obtaining battery characteristic data of a target lithium battery, wherein the target lithium battery is a battery for which remaining battery life prediction is required, and the battery characteristic data is used to represent characteristic information of the target lithium battery; filtering out decay characteristic data of the target lithium battery from the battery characteristic data; inputting the decay characteristic data into a battery remaining life prediction model, so as to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained through machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample decay characteristic data and sample battery remaining life corresponding to the sample decay characteristic data.
[0085] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: obtaining multiple sets of training data including multiple sample decay characteristic data and multiple sample battery remaining life; training the multiple sets of training data including multiple sample decay characteristic data and multiple sample battery remaining life to obtain a battery remaining life prediction model.
[0086] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: determining that the basic model of the battery remaining life prediction model is a long short-term memory artificial neural network LSTM model; converting multiple sample decay feature data into a sample decay feature sequence over time, and determining multiple sample battery remaining lives as predicted label data; inputting the sample decay feature sequence and the predicted label data into the LSTM model at the same time, so as to use the LSTM model to capture the sample decay features of the sample decay feature sequence in the time series; comparing the sample decay features with the predicted label data to obtain a comparison result; when the comparison result indicates that the error between the sample decay features and the predicted label data is less than an error threshold, stopping the model training to obtain a battery remaining life prediction model.
[0087] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: normalizing multiple sample decay characteristic data and multiple sample battery remaining lifespans to obtain multiple sample decay characteristic data and multiple sample battery remaining lifespans after normalization; dividing the multiple sample decay characteristic data and multiple sample battery remaining lifespans after normalization into data sets to obtain a training set, a validation set, and a test set; and obtaining a battery remaining life prediction model through training with the training set, the validation set, and the test set.
[0088] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for executing the following steps: evaluating the current remaining life by a predetermined evaluation method to obtain an evaluation result, wherein the predetermined evaluation method includes: root mean square error, mean absolute error and goodness of fit; when the evaluation result indicates that the current remaining life is within a predetermined range, determining the current remaining life as the actual value of the current remaining life of the target lithium battery; when the evaluation result indicates that the current remaining life is not within the predetermined range, adjusting the current remaining life.
[0089] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.
[0090] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0091] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.
[0092] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.
[0093] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0094] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or 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, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.
[0095] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for predicting the remaining battery life of a lithium battery, characterized in that: include: Acquire battery characteristic data of a target lithium battery, wherein the target lithium battery is a battery for which remaining battery life prediction is required, and the battery characteristic data is used to represent characteristic information of the target lithium battery; Filtering out the degradation characteristic data of the target lithium battery from the battery characteristic data; The decay characteristic data is input into a battery remaining life prediction model, and the decay characteristic data is processed by the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained by machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample decay characteristic data and sample battery remaining life corresponding to the sample decay characteristic data.
2. The method for predicting the remaining battery life of a lithium battery according to claim 1, characterized in that: Before inputting the decay characteristic data into the battery remaining life prediction model to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, the method further includes: Acquire the plurality of sets of training data including a plurality of the sample decay characteristic data and a plurality of the sample battery remaining lifespans; The plurality of groups of training data including the plurality of sample decay characteristic data and the plurality of sample battery remaining lifespans are trained to obtain the battery remaining lifespan prediction model.
3. The method for predicting the remaining battery life of a lithium battery according to claim 1, characterized in that: Training the plurality of sets of training data including the plurality of sample decay characteristic data and the plurality of sample battery remaining lifespans to obtain the battery remaining lifespan prediction model comprises: Determine that the basic model of the battery remaining life prediction model is a long short-term memory artificial neural network LSTM model; Converting the plurality of sample decay feature data into a sample decay feature sequence according to the passage of time, and determining the plurality of sample battery remaining lifespans as prediction label data; Inputting the sample decay feature sequence and the predicted label data into the LSTM model simultaneously, so as to capture the sample decay feature of the sample decay feature sequence in the time series by using the LSTM model; Comparing the sample decay feature with the predicted label data to obtain a comparison result; When the comparison result indicates that the error between the sample decay feature and the predicted label data is less than an error threshold, the model training is stopped to obtain the battery remaining life prediction model.
4. The method for predicting the remaining battery life of a lithium battery according to claim 2, characterized in that: Training the plurality of sets of training data including the plurality of sample decay characteristic data and the plurality of sample battery remaining lifespans to obtain the battery remaining lifespan prediction model comprises: Normalizing the plurality of sample decay characteristic data and the plurality of sample battery remaining lifespans to obtain the plurality of sample decay characteristic data and the plurality of sample battery remaining lifespans after normalization; Divide the normalized sample decay characteristic data and the sample battery remaining life into a data set to obtain a training set, a validation set, and a test set; The battery remaining life prediction model is obtained by training the training set, the validation set and the test set.
5. The method for predicting the remaining battery life of a lithium battery according to claim 1, characterized in that: After the degradation characteristic data is input into the battery remaining life prediction model, the degradation characteristic data is processed by using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, and the method further includes: Evaluate the current remaining life by a predetermined evaluation method to obtain an evaluation result, wherein the predetermined evaluation method includes: root mean square error, mean absolute error and fit; When the evaluation result indicates that the current remaining life is within a predetermined range, determining the current remaining life to be an actual value of the current remaining life of the target lithium battery; When the evaluation result indicates that the current remaining life is not within a predetermined range, the current remaining life is adjusted.
6. A device for predicting the remaining battery life of a lithium battery, characterized in that: include: A first acquisition module is used to acquire battery characteristic data of a target lithium battery, wherein the target lithium battery is a battery for which remaining battery life prediction is required, and the battery characteristic data is used to represent characteristic information of the target lithium battery; A screening module, used to screen out the degradation characteristic data of the target lithium battery from the battery characteristic data; An input module is used to input the decay characteristic data into a battery remaining life prediction model so as to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery, wherein the battery remaining life prediction model is obtained through machine learning training using multiple sets of training data, and each of the multiple sets of training data includes: sample decay characteristic data and sample battery remaining life corresponding to the sample decay characteristic data.
7. The device for predicting the remaining battery life of a lithium battery according to claim 6, characterized in that: The battery remaining life prediction device of the lithium battery also includes: A second acquisition module is used to acquire the multiple sets of training data including the multiple sample decay characteristic data and the multiple sample battery remaining life before inputting the decay characteristic data into the battery remaining life prediction model to process the decay characteristic data using the battery remaining life prediction model to obtain the current remaining life of the target lithium battery; The training module is used to train the multiple groups of training data including the multiple sample decay characteristic data and the multiple sample battery remaining life to obtain the battery remaining life prediction model.
8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method for predicting the remaining battery life of a lithium battery as described in any one of claims 1 to 5.
9. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the method for predicting the remaining battery life of a lithium battery as described in any one of claims 1 to 5.
10. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by the processor, the method for predicting the remaining battery life of a lithium battery as described in any one of claims 1 to 5 is performed.
Citation Information
Cited By
Method, device and equipment for predicting remaining service life of battery, medium and product
CN119575222A
Working performance detection method for power storage battery of new energy automobile
CN119959782A
