Battery life prediction method and device, electronic equipment and computer program product
By acquiring battery identification and feature parameters, combining cross feature parameters and dynamic weight training battery life prediction model, the problem of inaccurate battery life prediction of different suppliers is solved, and prediction accuracy and power supply reliability of data centers are improved.
Patent Information
- Application Number
- CN202510777390.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-11
- Publication Date
- 2025-08-15
AI Technical Summary
The prior art is difficult to accurately predict the health status of batteries produced by different suppliers, resulting in the possibility of insufficient power supply at critical moments, affecting the stable operation of the data center.
By obtaining battery identification, dynamic feature parameters and static feature parameters, cross feature parameters are determined, and prediction is performed using the battery life prediction model. The model is trained by dynamically assigning weight parameters to different battery identifications.
Improve the accuracy of battery life prediction, ensure stable supply of data center power, and reduce the risk of power outage.
Smart Images
Figure CN120490829A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of financial technology, and in particular to a battery life prediction method, device, electronic device and computer program product. Background Art
[0002] Data centers at banks and financial institutions have extremely high requirements for data security, accuracy, and real-time performance. They process vast amounts of customer transaction data and account information, and must ensure rapid and accurate responses to customer requests, such as online banking transfers and securities trading orders. A power failure can lead to serious issues such as transaction interruptions, data loss, and inconsistencies, impacting both the financial institution's operations and the interests of its customers. Therefore, to ensure stable power supply in data centers, an uninterruptible power supply (UPS) is required to safeguard the normal operation of communication systems and prevent data loss.
[0003] However, when the mains fails or maintenance is required, it is necessary to switch the mains power to the UPS to maintain temporary power supply via additional batteries. Because batteries come from different suppliers and have inconsistent standards, predicting the battery's state of health (SOH) is difficult. If the battery's SOH is inconsistent, the battery may not be able to provide sufficient power at a critical moment, leading to the risk of power outages. Therefore, how to accurately predict the battery lifespan is a key issue that needs to be addressed urgently. Summary of the Invention
[0004] The present application provides a battery life prediction method, device, electronic device and computer program product, which can improve the accuracy of battery life prediction.
[0005] In a first aspect, the present application provides a battery life prediction method, comprising:
[0006] Obtaining a battery identifier and battery parameters corresponding to the battery to be tested, wherein the battery parameters include dynamic characteristic parameters and static characteristic parameters;
[0007] Determining a cross-feature parameter according to the dynamic feature parameter and the static feature parameter;
[0008] The dynamic feature parameters, the static feature parameters, the cross feature parameters and the battery identification are input into a battery life prediction model to obtain the predicted life of the battery to be tested. The battery life prediction model is obtained by training by dynamically assigning weight parameters to different battery identifications.
[0009] In a second aspect, the present application provides a battery life prediction device, the device comprising:
[0010] A battery parameter acquisition module, used to obtain a battery identifier and battery parameters corresponding to the battery to be tested, wherein the battery parameters include dynamic characteristic parameters and static characteristic parameters;
[0011] a cross-parameter determination module, configured to determine a cross-feature parameter based on the dynamic feature parameter and the static feature parameter;
[0012] A battery life prediction module is used to input the dynamic feature parameters, static feature parameters, cross feature parameters and the battery identification into a battery life prediction model to obtain the predicted life of the battery to be tested. The battery life prediction model is obtained by training by dynamically assigning weight parameters to different battery identifications.
[0013] In a third aspect, the present application further provides an electronic device, comprising:
[0014] at least one processor; and
[0015] a memory communicatively connected to the at least one processor; wherein,
[0016] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the battery life prediction method described in any embodiment of the present application.
[0017] In a fourth aspect, the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the battery life prediction method described in any embodiment of the present application when executed.
[0018] In a fifth aspect, the present application also provides a computer program product, including a computer program, which implements the battery life prediction method described in any embodiment of the present application when executed by a processor. The battery life prediction scheme provided in the embodiment of the present application first obtains the battery identification and battery parameters corresponding to the battery to be tested, and the battery parameters include dynamic feature parameters and static feature parameters; then determines the cross-feature parameters based on the dynamic feature parameters and the static feature parameters; finally, the dynamic feature parameters, the static feature parameters, the cross-feature parameters and the battery identification are input into the battery life prediction model to obtain the predicted life of the battery to be tested. This scheme can mine the potential relationship between dynamic features and static features through cross-feature parameters, and more comprehensively reflect the actual situation and performance changes of the battery, thereby providing an effective data basis for battery life prediction; and the battery life prediction model provided in this embodiment is obtained by training by dynamically assigning weight parameters to different battery identifications. By dynamically assigning weight parameters to different battery identifications, the model can better adapt to batteries of different types and different production batches, and achieve the beneficial effect of improving the accuracy of predicting the life of batteries with different identifications.
[0019] It should be noted that the above-mentioned computer instructions may be stored in whole or in part on a computer-readable storage medium. The computer-readable storage medium may be packaged together with the processor of the battery life prediction device or separately from the processor of the battery life prediction device, and this application does not limit this.
[0020] The descriptions of the second, third, fourth and fifth aspects of this application can refer to the detailed description of the first aspect; and the beneficial effects of the descriptions of the second, third, fourth and fifth aspects can refer to the analysis of the beneficial effects of the first aspect, which will not be repeated here.
[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description.
[0022] It is understandable that before using the technical solutions disclosed in the embodiments of this application, the type, scope of use, and usage scenarios of the personal information involved in this application should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.
[0024] Figure 1 This is a flow chart of a battery life prediction method provided by an embodiment of the present application;
[0025] Figure 2 This is a schematic diagram of the structure of a battery life prediction device provided in an embodiment of the present application;
[0026] Figure 3 It is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0027] In order to enable those skilled in the art to better understand the present invention, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this embodiment. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.
[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence 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 comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0029] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the present application and are not intended to limit the present application. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions of the present application, not all of the structures.
[0030] Figure 1This is a flow chart of a battery life prediction method provided in an embodiment of the present application. This embodiment is applicable to predicting the battery life of individual batteries in a UPS system in a data center in the financial sector. The method can be performed by a battery life prediction device, which can be implemented in hardware and / or software and integrated into the electronic device that performs the method. Preferably, the electronic device in the embodiment of the present application can be a server, a computer, or the like.
[0031] refer to Figure 1 The battery life prediction method of this embodiment includes but is not limited to the following steps:
[0032] S110: Obtain a battery identifier and battery parameters corresponding to the battery to be tested.
[0033] The battery under test is the battery in a UPS system whose lifespan needs to be predicted. Because UPS systems sometimes contain multiple batteries from different suppliers, the parameters of the batteries produced by different suppliers are often inconsistent. Therefore, to accurately predict the lifespan of each battery, the solution provided in this embodiment requires obtaining the battery identifier and battery parameters corresponding to the battery under test.
[0034] The battery identifier is a code or label that uniquely identifies the battery, typically containing information such as the supplier code, production batch, and battery model. The battery identifier is used to distinguish between different suppliers of batteries in a certain period. Based on the current supplier information, the corresponding battery chemistry type or manufacturing process can be determined. This information can then be used to dynamically adapt the corresponding weight parameters when training the battery life prediction model in subsequent steps.
[0035] In this embodiment, battery parameters include dynamic characteristic parameters and static characteristic parameters. Dynamic characteristic parameters are parameters that change over time during battery use. In this embodiment, dynamic characteristic parameters include at least: charging current parameters, discharging current parameters, charging voltage parameters, discharging voltage parameters, battery temperature parameters, number of charge and discharge cycles, and current remaining capacity. These parameters reflect the battery's operating status and performance changes at different times. Static characteristic parameters are parameters obtained from the battery's production records or product manual. They are inherent properties of the battery and do not change over time. In this embodiment, static characteristic parameters include at least: supplier code (coded as 0 / 1), battery type parameters (such as lithium-ion battery or nickel-metal hydride battery), nominal capacity parameters, rated capacity parameters, electrode material composition parameters, initial internal resistance value, and initial manufacturing date. These parameters remain essentially unchanged throughout the battery's lifecycle.
[0036] In practical applications, dynamic characteristic parameters and static characteristic parameters may have different dimensions and orders of magnitude. For example, the units of current and voltage are different, and the numerical ranges may also vary greatly. If the obtained dynamic characteristic parameters and static characteristic parameters are directly applied, then in the subsequent model training, there is a problem of poor performance of the trained model, and the battery life prediction results ultimately output by the model have large differences. To improve the current problem, the solution provided in this embodiment can further perform the following solution after executing step S110:
[0037] According to the battery identification, the dynamic nominal values corresponding to the dynamic characteristic parameters and the static nominal values corresponding to the static characteristic parameters are obtained respectively; the dynamic characteristic parameters are first processed according to the dynamic nominal values to obtain preprocessed dynamic characteristic parameters; and the static characteristic parameters are second processed according to the static nominal values to obtain preprocessed static characteristic parameters.
[0038] Specifically, for each dynamic characteristic parameter, the corresponding dynamic nominal value is searched and obtained from the relevant database, specification sheet or historical data according to the battery identification. This dynamic nominal value is usually the typical value or reference value of the dynamic characteristic parameter under standard conditions or normal working conditions. The dynamic characteristic parameter is first processed according to the obtained dynamic nominal value to obtain the preprocessed dynamic characteristic parameter. A common method is to perform normalization processing, in which the dynamic characteristic parameter is mapped to the [0,1] interval to provide a unified parameter measurement standard.
[0039] Similarly, for each static characteristic parameter, its corresponding static nominal value is obtained based on the battery identification. The static nominal value is a fixed reference value of the static characteristic parameter under specific conditions, such as the rated capacity and nominal voltage of the battery. The static characteristic parameter is subjected to a second processing based on the static nominal value to obtain a preprocessed static characteristic parameter. The processing method is similar to the first processing, and may be normalization, standardization or other data transformation operations based on the static nominal value. For example, if the static characteristic parameter is the actual capacity of the battery, and the static nominal value is the rated capacity of the battery, the ratio of the actual capacity to the rated capacity can be calculated as the static characteristic parameter after the second processing to reflect the state of the battery capacity relative to the nominal value.
[0040] This embodiment preprocesses dynamic and static characteristic parameters separately, standardizing the dynamic and static characteristic parameters of different batteries based on their respective nominal values. This makes the same type of characteristics comparable across different batteries, facilitating subsequent data analysis and model processing. This enables the model to better learn the relationship between characteristics and battery life, reducing model bias caused by differences in data scale and distribution, thereby improving the accuracy and reliability of the battery life prediction model.
[0041] S120: Determine cross-feature parameters according to the dynamic feature parameters and the static feature parameters.
[0042] Cross-feature parameters are new feature parameters derived by combining, calculating, or correlating dynamic and static feature parameters. This embodiment analyzes cross-feature parameters to more comprehensively describe battery performance and status, providing richer input parameters for tasks such as battery life prediction. In this embodiment, cross-feature parameters include at least power, battery charge and discharge status, charge and discharge efficiency, thermal power, and voltage change rate.
[0043] Specifically, the rated capacity of the battery (static characteristic parameter) can be added or subtracted from the current remaining power (dynamic characteristic parameter). The new parameter obtained after addition can reflect the total capacity potential of the battery (rated capacity + remaining power); the parameter result obtained after subtraction reflects the power consumed by the battery (rated capacity - remaining power). The battery's charging current (dynamic characteristic parameter) and the proportion of a key component in the electrode material (static characteristic parameter) can also be multiplied to obtain a new cross-characteristic parameter, which may be related to the chemical reaction activity of the battery under a specific charging current; for example, the current remaining power is divided by the rated capacity to obtain the cross-characteristic parameter of the remaining power ratio, which can intuitively reflect the remaining usable degree of the battery. The specific calculation methods for obtaining each cross-characteristic parameter are not listed here one by one.
[0044] In a preferred embodiment, in this solution, the above step S120 can be implemented as follows:
[0045] At least two time windows are determined based on the parameter characteristics of the dynamic feature parameters, and the dynamic feature quantities of the dynamic feature parameters of the corresponding parameter characteristics within the corresponding time windows are determined; and cross feature parameters are obtained according to the static feature parameters and the dynamic feature quantities within each time window.
[0046] The parameter characteristics of dynamic characteristic parameters can be understood as the characteristics of the dynamic characteristic parameters in the time series, such as the change pattern, fluctuation amplitude, and change frequency. For example, some dynamic characteristic parameters (such as current and voltage) change frequently and fluctuate greatly, while some (such as the number of charge and discharge cycles) change relatively slowly and are cumulative. Therefore, the solution provided in this embodiment divides the time series into time windows of different lengths by determining at least two time windows based on the parameter characteristics of the dynamic characteristic parameters, so as to analyze and extract corresponding characteristic information within each window. For example, a short time window (such as a few minutes) can be set to capture rapidly changing parameter information, or a long time window (such as a few days or weeks) can be set to analyze the long-term trend of the parameters. Further, in each time window, the dynamic characteristic parameters are statistically calculated to obtain a dynamic feature vector. The dynamic feature quantities in this embodiment include at least the mean, standard deviation, maximum value, minimum value, and slope (reflecting the parameter change trend). For example, for the charging current data within a 10-minute time window, its dynamic feature quantities such as the mean, standard deviation, maximum value, and minimum value are calculated; the above steps are repeated to calculate the corresponding dynamic feature quantity for each time window and corresponding dynamic feature parameter.
[0047] Furthermore, new parameters are obtained by combining or calculating the static feature parameters with the dynamic feature quantities in each time window. These cross-feature parameters integrate the static properties and dynamic working status information of the battery, can more comprehensively describe the performance and characteristics of the battery, and provide richer input information for subsequent tasks such as battery life prediction. For example, if the static feature parameter is the rated capacity of the battery, and the dynamic feature quantity in a certain time window is the mean of the current remaining power, then the ratio of the mean remaining power to the rated capacity can be calculated as a cross-feature parameter; more complex calculations can also be performed, such as combining the battery type (static feature parameter) with the temperature change rate (dynamic feature quantity) in different time windows, and performing weighted calculations based on the sensitivity of the battery type to temperature changes to obtain new cross-feature parameters; further repeat the above operations to generate corresponding cross-feature parameters for each combination of static feature parameters and dynamic feature quantities in the time window.
[0048] The solution provided in this embodiment can comprehensively capture the dynamic operating state of the battery at multiple time scales. By combining static feature parameters to generate cross-feature parameters, the inherent properties of the battery are further integrated, making the extracted features more comprehensive and rich, and more accurately describing the actual battery conditions. The obtained cross-feature parameters provide higher-quality input information for the subsequent battery life prediction model, allowing the model to learn more about the potential relationship between battery performance and life, thereby improving the accuracy and reliability of the prediction and enhancing the overall performance of the model.
[0049] In another specific implementation, in this embodiment, the above-mentioned "obtaining cross-feature parameters based on static feature parameters and dynamic feature quantities in each time window" is implemented through the following steps A) to C):
[0050] A) The static feature parameters and the dynamic feature quantities in each time window are fused separately to obtain the cross feature vector corresponding to each time window.
[0051] In the current step, the methods of fusing the static feature parameters and the dynamic feature quantities in each time window may include direct splicing, weighted fusion, and logical operation fusion. The specific fusion method is not limited here.
[0052] As an example of fusion, using direct concatenation, the values of static feature parameters are directly concatenated with the dynamic feature quantities within each time window. Assuming there are m static feature parameters and n dynamic feature quantities within each time window, the dimension of the cross-feature vector corresponding to each time window after concatenation is m+n. For example, the static feature parameters are battery type (assuming they are encoded as a 1-dimensional vector) and rated capacity (a 1-dimensional vector), for a total of two; the dynamic feature quantities within a time window are the mean, standard deviation, and maximum battery temperature of the charging current, for a total of three. These are then concatenated sequentially to obtain a 5-dimensional cross-feature vector. Another fusion example, using weighted fusion, assigns weights to the static feature parameters and dynamic feature quantities based on their importance, and then performs a weighted sum. The weights can be determined using domain knowledge or machine learning algorithms (such as feature importance analysis). For example, if the battery rated capacity is considered more important for subsequent analysis, it is given a higher weight of 0.6, and the battery type is given a weight of 0.4. Among the dynamic features, the charging current mean is given a weight of 0.3, the standard deviation is given a weight of 0.3, and the maximum battery temperature is given a weight of 0.4. The corresponding eigenvalues are then weighted and summed according to the weights to obtain a cross-feature vector. Another fusion example uses logical operation fusion as an example. Logical operations such as AND, OR, and XOR are performed on static feature parameters and dynamic features to generate new eigenvalues, which are then combined to form a cross-feature vector. For example, when the battery type is a certain type (represented by a Boolean value) and the maximum battery temperature in the current time window exceeds a certain threshold (represented by a Boolean value), a new eigenvalue is obtained through a logical AND operation, which, together with other relevant eigenvalues, forms a cross-feature vector. The above fusion operation is performed on the dynamic features in each time window to obtain a cross-feature vector corresponding to each time window.
[0053] B) Arrange the cross feature vectors of each time window in chronological order to obtain the time series feature matrix.
[0054] Based on the chronological order of data collection, clearly identify the time identifier for each time window, such as a timestamp or time sequence number. Arrange the cross-feature vectors corresponding to each time window obtained in step A in chronological order. Assuming there are k time windows and the dimension of each cross-feature vector is d, the resulting time series feature matrix will have a shape of k × d. For example, the cross-feature vector of the first time window will be the first row of the matrix, the cross-feature vector of the second time window will be the second row, and so on, until the cross-feature vectors of the kth time window have been arranged.
[0055] C) Based on the time series model, the dependency relationship between the time series feature matrices is extracted to obtain the cross feature parameters.
[0056] The time series model in this embodiment can be implemented based on the long short-term memory network model (Long Short-Term Memory, referred to as LSTM). Usually, the time series feature matrix is input in sequence by row. The time series model captures the dependency relationship in the time series through the hidden state, and calculates the correlation between each position in the sequence through the self-attention mechanism to capture the dependency relationship; in the processing of the time series model, a feature representation that can reflect the dependency relationship between the time series feature matrices is extracted. This can be the hidden state of the model, the intermediate result of the output layer, or the feature vector obtained by specific calculation. For example, in the LSTM model, the hidden state of the last time step can be used as a feature representation reflecting the dependency relationship of the entire time series.
[0057] The method for determining cross-feature parameters provided in this embodiment can fully utilize the inherent properties of the battery and real-time working status information by fusing static feature parameters and dynamic feature quantities, so that the obtained cross-feature vector provides more powerful data support for subsequent analysis; the method for extracting the dependency relationship between time series feature matrices based on the time series model to obtain cross-feature parameters integrates various information and takes into account the dependency relationship of time series, and provides them as input to the subsequent battery life prediction model, which can improve the accuracy and generalization ability of the model.
[0058] S130 , inputting the dynamic characteristic parameters, the static characteristic parameters, the cross characteristic parameters and the battery identification into a battery life prediction model to obtain a predicted life of the battery to be tested.
[0059] In this embodiment, the trained battery life prediction model can be deployed in a battery management system or related monitoring equipment to obtain the battery identification and battery parameters of the battery under test in real time, and then use the obtained parameter data to predict the battery life. Optionally, if the predicted result indicates that the battery under test is nearing the end of its life, a warning signal can be issued in a timely manner to prevent losses caused by power outages.
[0060] In this embodiment, the battery life prediction model is obtained by training by dynamically assigning weight parameters to different battery identifiers. Specifically, the battery life prediction model can be trained in the following manner: first, a large amount of battery data with different battery identifiers is collected, including the battery identifiers, dynamic feature parameters, static feature parameters, cross-feature parameters and corresponding actual battery life data obtained above. These data are divided into training sets, validation sets and test sets; during the model training process, weight parameters are dynamically assigned to different battery identifiers. The weights can be determined based on factors such as the frequency of battery use and the reliability of historical data. For example, higher weights are assigned to battery identifiers with high frequency of use and rich historical data; lower weights are assigned to battery identifiers with low frequency of use or poor data quality; and then, based on the LSTM model architecture, one or more LSTM layers are constructed to capture long-term dependencies in time series data. Each LSTM layer contains multiple memory units, which control the flow and memory of information through input gates, forget gates and output gates; then design an independent sub-network whose input is the battery identification code and output is a weight vector corresponding to the output feature of the LSTM layer. This sub-network can be a simple multi-layer perceptron, which generates appropriate weights for different battery identifications through training and learning; finally, connect one or more fully connected layers after the LSTM layer to further transform and fuse the output features of the LSTM layer, and finally output the predicted value of the battery life; and use the training set to train the model, and adjust the model parameters by minimizing the error between the predicted life and the actual life (such as mean square error, mean absolute error, etc.). When the final output result reaches convergence, the battery life prediction model is obtained.
[0061] Another preferred embodiment, in this embodiment, the above-mentioned step S130 can be implemented in the following manner: the dynamic feature parameters, static feature parameters and cross-feature parameters are respectively standardized to obtain parameter feature vectors; the battery identification is encoded to obtain the battery identification code, and the weight vector of the battery to be tested is determined according to the battery identification code; the battery life prediction model calculates the parameter feature vector and the weight vector to obtain the predicted life of the battery to be tested.
[0062] Since dynamic feature parameters, static feature parameters, and cross-feature parameters have dimensional differences when actually used, the parameter feature vector is obtained by eliminating the dimensionality effect through standardization. Battery identification is usually categorical data. Each battery identification is converted into a vector with only one element being 1 and the rest being 0 by using one-hot encoding (for example, if there are 10 different battery identifications, each battery identification will be encoded as a 10-dimensional vector, where only the position corresponding to its category is 1 and the rest are 0). Then, based on the battery identification encoding, a specific mapping function is used to determine the corresponding weight vector. The determined weight vector will be used for subsequent weighting operations on the parameter feature vector. Different battery identifications will correspond to different weight vectors to reflect the impact of differences between different batteries on life prediction. Finally, the parameter feature vector and the weight vector are passed as input to the battery life prediction model. Within the model, the parameter feature vector is weighted, that is, each element of the parameter feature vector is multiplied by the corresponding element of the weight vector. This can highlight or weaken the importance of certain features in the prediction. The weighted feature vector is then used by the model to perform nonlinear transformation and feature extraction on the input features, and the final predicted life value is calculated and output. This embodiment fully accounts for the differences between different batteries by encoding the battery identifiers and determining weight vectors for each battery identifier. Batteries of different models and production batches may have different performance and lifespan characteristics. In this way, the model can better adapt to the characteristics of various batteries and improve the prediction accuracy for different batteries.
[0063] Another preferred embodiment, in order to reduce the computational complexity of the model and reduce the amount of model input data without affecting the accuracy of the model output results, the solution provided in this embodiment, before executing the above S130, preferably also performs the following operations: based on the correlation between the dynamic feature parameters, static feature parameters and cross-feature parameters and the battery life, obtain key feature parameters; then input the key feature parameters and battery identification into the battery life prediction model to obtain the predicted life of the battery to be tested. This embodiment can screen out parameters that truly have a significant impact on battery life by determining the key feature parameters, avoiding the interference of irrelevant information on the prediction results, thereby making the prediction more accurate; and by reducing the input data dimension, it reduces the input variables of the model, improves the training and prediction speed of the model, and makes it easier to deploy and apply.
[0064] Specifically, the above-mentioned “obtaining key characteristic parameters based on the correlations between the dynamic characteristic parameters, the static characteristic parameters, and the cross characteristic parameters and the battery life” may be implemented as follows:
[0065] Based on the correlation algorithm, the correlations between dynamic feature parameters, static feature parameters, cross-feature parameters and battery life are calculated respectively to obtain dynamic correlation scores, static correlation scores and cross-correlation scores respectively; the dynamic correlation values, static correlation scores and cross-correlation scores are assigned corresponding weights according to their contribution to life prediction to obtain the priority ranking corresponding to each feature parameter; the threshold score is determined, and based on the priority ranking, the parameters in each feature parameter that meet the threshold score are determined as key feature parameters.
[0066] In this embodiment, the degree of correlation can be expressed as a value between [-1, 1], with the closer the absolute value is to 1, the stronger the correlation. The correlation algorithm can be a Pearson correlation coefficient or a Spearman rank correlation coefficient, and the specific algorithm used is not limited here. Assuming that dynamic characteristic parameters include the charge / discharge current change rate and temperature change, a selected correlation algorithm is used to calculate the correlation coefficient between these parameters and the battery life data to obtain a dynamic correlation score. For static characteristic parameters such as battery capacity and internal resistance, the same correlation algorithm is used to calculate the correlation coefficient with battery life to obtain a static correlation score. For cross-characteristic parameters such as the interaction between charge / discharge depth and cycle number, the algorithm is used to calculate the correlation coefficient with battery life to obtain a cross-correlation score. Furthermore, the contribution of dynamic, static, and cross-characteristic parameters to battery life prediction can be determined using historical data verification or feature importance assessment methods in machine learning. For example, through multiple experiments, it was found that dynamic characteristic parameters contribute 40% to life prediction, static characteristic parameters contribute 30%, and cross-characteristic parameters contribute 30%. Weights are then assigned to the dynamic, static, and cross-correlation scores based on their contribution. For example, the dynamic correlation score weight is 0.4, the static correlation score weight is 0.3, and the cross-correlation score weight is 0.3; finally, the correlation score of each feature parameter is multiplied by the corresponding weight and added together to obtain a comprehensive score. Prioritize all feature parameters according to the comprehensive score. Set a threshold score based on actual needs and data characteristics. For example, taking into account factors such as model performance and data noise, the threshold is set to 0.6; traverse all prioritized feature parameters, and determine the parameters with a comprehensive score greater than or equal to the threshold score as key feature parameters. This embodiment quantifies the correlation between feature parameters and battery life, and assigns weights according to contribution, so as to accurately select parameters that play a key role in battery life prediction, eliminate interference from irrelevant or secondary parameters, and thus improve the accuracy of the prediction model.
[0067] The battery life prediction method provided in this embodiment first obtains the battery identification and battery parameters corresponding to the battery to be tested, and the battery parameters include dynamic feature parameters and static feature parameters; then determines the cross-feature parameters based on the dynamic feature parameters and the static feature parameters; finally, the dynamic feature parameters, the static feature parameters, the cross-feature parameters and the battery identification are input into the battery life prediction model to obtain the predicted life of the battery to be tested. This solution can explore the potential relationship between dynamic features and static features through cross-feature parameters, more comprehensively reflect the actual situation and performance changes of the battery, and thus provide an effective data basis for battery life prediction; and the battery life prediction model provided in this embodiment is obtained by training by dynamically assigning weight parameters to different battery identifications. By dynamically assigning weight parameters to different battery identifications, the model can better adapt to batteries of different types and different production batches, achieving the beneficial effect of improving the accuracy of predicting the life of batteries with different identifications.
[0068] Figure 2 This is a schematic diagram of the structure of the battery life prediction device provided by the embodiment of the present application, which is suitable for executing the battery life prediction method provided by the embodiment of the present application. Figure 2 As shown, the device may specifically include: a battery parameter acquisition module 210, a cross parameter determination module 220 and a battery life prediction module 230, wherein:
[0069] A battery parameter acquisition module 210 is configured to acquire a battery identifier and battery parameters corresponding to the battery to be tested, wherein the battery parameters include dynamic characteristic parameters and static characteristic parameters;
[0070] A cross-parameter determination module 220, configured to determine a cross-feature parameter based on the dynamic feature parameter and the static feature parameter;
[0071] The battery life prediction module 230 is used to input the dynamic feature parameters, static feature parameters, cross feature parameters and the battery identifier into a battery life prediction model to obtain the predicted life of the battery to be tested. The battery life prediction model is obtained by training by dynamically assigning weight parameters to different battery identifiers.
[0072] The battery life prediction device provided in the embodiment of the present application first obtains the battery identification and battery parameters corresponding to the battery to be tested, and the battery parameters include dynamic characteristic parameters and static characteristic parameters; then determines the cross-feature parameters based on the dynamic characteristic parameters and the static characteristic parameters; finally, the dynamic characteristic parameters, the static characteristic parameters, the cross-feature parameters and the battery identification are input into the battery life prediction model to obtain the predicted life of the battery to be tested. This solution can explore the potential relationship between dynamic characteristics and static characteristics through cross-feature parameters, and more comprehensively reflect the actual situation and performance changes of the battery, thereby providing an effective data basis for battery life prediction; and the battery life prediction model provided in this embodiment is obtained by training by dynamically assigning weight parameters to different battery identifications. By dynamically assigning weight parameters to different battery identifications, the model can better adapt to batteries of different types and different production batches, achieving the beneficial effect of improving the accuracy of predicting the life of batteries with different identifications.
[0073] In one embodiment, the cross parameter determination module 220 includes a time window determination unit and a feature parameter determination unit, wherein:
[0074] a time window determination unit, configured to determine at least two time windows based on the parameter characteristics of the dynamic feature parameters, and determine dynamic feature quantities of the dynamic feature parameters of the corresponding parameter characteristics within the corresponding time windows;
[0075] The feature parameter determination unit is used to obtain cross feature parameters according to the static feature parameters and the dynamic feature quantities in each of the time windows.
[0076] In one embodiment, the feature parameter determination unit is specifically used to fuse the static feature parameters and the dynamic feature quantities in each time window respectively to obtain the cross feature vectors corresponding to each time window; arrange the cross feature vectors of each time window in chronological order to obtain a time series feature matrix; and extract the dependency relationship between the time series feature matrices based on a time series model to obtain the cross feature parameters.
[0077] In one embodiment, the apparatus further includes a key parameter acquisition module, wherein:
[0078] A key parameter obtaining module, configured to obtain key characteristic parameters according to correlations between the dynamic characteristic parameters, the static characteristic parameters, and the cross characteristic parameters and the battery life;
[0079] The battery life prediction module 230 is further configured to input the key characteristic parameters and the battery identifier into a battery life prediction model to obtain a predicted life of the battery to be tested.
[0080] In one embodiment, the key parameter acquisition module is specifically used to calculate the correlation between the dynamic characteristic parameters, the static characteristic parameters and the cross-characteristic parameters and the battery life based on the correlation algorithm, and obtain the dynamic correlation score, the static correlation score and the cross-correlation score respectively; assign corresponding weights to the dynamic correlation values, the static correlation score and the cross-correlation score according to the life prediction contribution, and obtain the priority ranking corresponding to each characteristic parameter; determine the threshold score, and based on the priority ranking, determine the parameters among the characteristic parameters that meet the threshold score as the key characteristic parameters.
[0081] In one embodiment, the battery life prediction module 230 is specifically used to standardize the dynamic feature parameters, the static feature parameters and the cross feature parameters respectively to obtain parameter feature vectors; encode the battery identification to obtain a battery identification code, and determine the weight vector of the battery to be tested based on the battery identification code; the battery life prediction model calculates the parameter feature vector and the weight vector to obtain the predicted life of the battery to be tested.
[0082] In one embodiment, the device further includes a nominal value acquisition module and a characteristic parameter processing module, wherein:
[0083] a nominal value acquisition module, configured to acquire the dynamic nominal values corresponding to the dynamic characteristic parameters and the static nominal values corresponding to the static characteristic parameters according to the battery identification;
[0084] The characteristic parameter processing module is used to perform a first processing on the dynamic characteristic parameters according to the dynamic nominal values to obtain preprocessed dynamic characteristic parameters; and is also used to perform a second processing on the static characteristic parameters according to the static nominal values to obtain preprocessed static characteristic parameters.
[0085] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional modules is used as an example for illustration. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the functional modules described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.
[0086] It should be noted that the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data comply with relevant laws, regulations and standards in the relevant regions.
[0087] An embodiment of the present application also provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the battery life prediction method described in any embodiment of the present application.
[0088] An embodiment of the present application further provides a computer-readable medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the battery life prediction method described in any embodiment of the present application when executed.
[0089] Reference below Figure 3 , Figure 3 FIG1 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, which shows a schematic diagram of the structure of a computer system 500 suitable for implementing the electronic device in an embodiment of the present application. Figure 3 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.
[0090] like Figure 3 As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage unit 508 into a random access memory (RAM) 503. Various programs and data required for the operation of the system 500 are also stored in the RAM 503. The CPU 501, ROM 502, and RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.
[0091] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, and the like; an output section 507 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 508 including a hard disk; and a communication section 509 including a network interface card such as a LAN card or a modem. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 510 as needed, so that computer programs read therefrom can be installed into the storage section 508 as needed.
[0092] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 509, and / or installed from a removable medium 511. When the computer program is executed by the central processing unit (CPU) 501, the above-mentioned functions defined in the system of the present application are executed.
[0093] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this application, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, and optical cables, or any suitable combination thereof.
[0094] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of the boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0095] The modules and / or units involved in the embodiments described in this application may be implemented in software or hardware. The modules and / or units described may also be provided in a processor. For example, it may be described as follows: a processor includes a battery parameter acquisition module, a cross parameter determination module, and a battery life prediction module. The names of these modules do not, in some cases, constitute limitations on the modules themselves.
[0096] As another aspect, the present application also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently and not be assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device includes: obtaining a battery identification and battery parameters corresponding to a battery to be tested, wherein the battery parameters include dynamic characteristic parameters and static characteristic parameters; determining a cross-characteristic parameter based on the dynamic characteristic parameters and the static characteristic parameters; inputting the dynamic characteristic parameters, the static characteristic parameters, the cross-characteristic parameters and the battery identification into a battery life prediction model to obtain a predicted life of the battery to be tested, wherein the battery life prediction model is obtained by training by dynamically assigning weight parameters to different battery identifications.
[0097] According to the technical solution of this embodiment, the potential relationship between dynamic features and static features can be explored through cross-feature parameters, and the actual situation and performance changes of the battery can be more comprehensively reflected, thereby providing an effective data basis for battery life prediction; and the battery life prediction model provided by this embodiment is obtained by training by dynamically assigning weight parameters to different battery identifications. By dynamically assigning weight parameters to different battery identifications, the model can better adapt to batteries of different types and different production batches, achieving the beneficial effect of improving the accuracy of predicting the life of batteries with different identifications.
[0098] The above specific embodiments do not constitute a limitation on the scope of protection of this application. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the scope of protection of this application.
Claims
1. A battery life prediction method, characterized in that: include: Obtaining a battery identifier and battery parameters corresponding to the battery to be tested, wherein the battery parameters include dynamic characteristic parameters and static characteristic parameters; Determining a cross-feature parameter according to the dynamic feature parameter and the static feature parameter; The dynamic feature parameters, the static feature parameters, the cross feature parameters and the battery identification are input into a battery life prediction model to obtain the predicted life of the battery to be tested. The battery life prediction model is obtained by training by dynamically assigning weight parameters to different battery identifications.
2. The battery life prediction method according to claim 1, characterized in that: The determining of the cross-feature parameter according to the dynamic feature parameter and the static feature parameter includes: Determining at least two time windows based on the parameter characteristics of the dynamic feature parameters, and determining dynamic feature quantities of the dynamic feature parameters of the corresponding parameter characteristics within the corresponding time windows; A cross feature parameter is obtained according to the static feature parameter and the dynamic feature quantity in each of the time windows.
3. The battery life prediction method according to claim 2, characterized in that: The obtaining of cross-feature parameters according to the static feature parameters and the dynamic feature quantities in each of the time windows includes: Fusing the static feature parameters with the dynamic feature quantities in each time window to obtain a cross feature vector corresponding to each time window; Arrange the cross feature vectors of each time window in chronological order to obtain the time series feature matrix; The dependency relationship between the time series feature matrices is extracted based on a time series model to obtain the cross feature parameters.
4. The battery life prediction method according to claim 1, wherein: After determining the cross-feature parameter according to the dynamic feature parameter and the static feature parameter, the method further includes: Obtaining key characteristic parameters according to correlations between the dynamic characteristic parameters, the static characteristic parameters, the cross characteristic parameters, and the battery life respectively; Accordingly, the dynamic characteristic parameters, static characteristic parameters, cross characteristic parameters and the battery identifier are input into a battery life prediction model to obtain the predicted life of the battery to be tested, including: The key characteristic parameters and the battery identifier are input into a battery life prediction model to obtain a predicted life of the battery to be tested.
5. The battery life prediction method according to claim 4, characterized in that: The obtaining of key characteristic parameters according to the correlations between the dynamic characteristic parameters, the static characteristic parameters, the cross characteristic parameters and the battery life respectively includes: Calculate the correlations between the dynamic feature parameter, the static feature parameter, the cross feature parameter and the battery life based on a correlation algorithm, and obtain a dynamic correlation score, a static correlation score and a cross correlation score respectively; Assigning corresponding weights to the dynamic correlation values, static correlation scores, and cross-correlation scores according to their contribution to life prediction, to obtain a priority ranking corresponding to each characteristic parameter; A threshold score is determined, and based on the priority ranking, parameters among the feature parameters that meet the threshold score are determined as the key feature parameters.
6. The battery life prediction method according to claim 1, characterized in that: Inputting the dynamic characteristic parameter, the static characteristic parameter, the cross characteristic parameter, and the battery identifier into a battery life prediction model to obtain the predicted life of the battery to be tested includes: performing standardization processing on the dynamic feature parameters, the static feature parameters, and the cross feature parameters respectively to obtain parameter feature vectors; Encoding the battery identification to obtain a battery identification code, and determining a weight vector of the battery to be tested according to the battery identification code; The battery life prediction model calculates the parameter feature vector and the weight vector to obtain the predicted life of the battery to be tested.
7. The battery life prediction method according to claim 1, characterized in that: After obtaining the battery identification and battery parameters corresponding to the battery to be tested, the following steps are also included: Obtaining dynamic nominal values corresponding to the dynamic characteristic parameters and static nominal values corresponding to the static characteristic parameters according to the battery identification; Performing a first processing on the dynamic characteristic parameter according to the dynamic nominal value to obtain a pre-processed dynamic characteristic parameter; The static characteristic parameters are respectively subjected to second processing according to the static nominal values to obtain pre-processed static characteristic parameters.
8. A battery life prediction device, characterized in that: include: A battery parameter acquisition module, used to obtain a battery identifier and battery parameters corresponding to the battery to be tested, wherein the battery parameters include dynamic characteristic parameters and static characteristic parameters; a cross-parameter determination module, configured to determine a cross-feature parameter based on the dynamic feature parameter and the static feature parameter; A battery life prediction module is used to input the dynamic feature parameters, static feature parameters, cross feature parameters and the battery identification into a battery life prediction model to obtain the predicted life of the battery to be tested. The battery life prediction model is obtained by training by dynamically assigning weight parameters to different battery identifications.
9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor. The computer program is executed by the at least one processor to enable the at least one processor to perform the battery life prediction method according to any one of claims 1 to 7.
10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the battery life prediction method according to any one of claims 1 to 7.