Battery charge state prediction method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a battery charge state prediction method and device, electronic equipment and a storage medium, and relates to the technical field of computers. The method comprises the following steps: acquiring respective first sequences of at least two parameters of a battery, wherein each first sequence comprises first parameter values of the corresponding parameters at a plurality of first moments in a preset time period; decomposing the first sequence based on a preset decomposition coefficient to obtain a second sequence and a third sequence of the parameters; and inputting the second sequence and the third sequence of each parameter into the charge prediction hybrid model, and outputting a first prediction result about the charge state of the battery. A plurality of parameters are collected to generate a first sequence of each parameter, each first sequence is decomposed to obtain a second sequence and a third sequence, and the decomposed sequences are input into a charge prediction hybrid model for charge state prediction. Therefore, the prediction of the residual charge state of the target battery is realized by collecting the data of the plurality of parameters, so that the predicted residual charge state is more accurate.

Description

Technical Field

[0001] This application relates to the field of computer technology. Specifically, this application relates to a method, apparatus, electronic device, and storage medium for predicting the state of charge of a battery. Background Art

[0002] With the rapid development of electric vehicles, smart grids, and renewable energy storage systems, lithium battery packs, as energy storage media, have been widely used in various fields. The state estimation of battery packs, especially the accurate estimation of the state of charge (SOC), is crucial for ensuring the safety, reliability, and efficiency of energy storage systems. The accurate estimation of SOC can effectively prevent overcharging and over-discharging of batteries, extend the battery life, and improve the performance of the energy management system.

[0003] Traditional SOC estimation methods mainly include the coulomb counting method, open-circuit voltage method, etc. The coulomb counting method calculates the SOC by integrating the current; the open-circuit voltage method estimates based on the relationship between the battery voltage and SOC. However, in the above methods, when performing calculation and prediction, they only calculate and predict based on one of the parameters. In actual application scenarios, it is found that there is still a large difference between the SOC predicted by the above traditional methods and the true SOC. Summary of the Invention

[0004] The purpose of this application is to at least solve one of the above technical defects. The technical solutions provided by the embodiments of this application are as follows: In a first aspect, an embodiment of this application provides a method for predicting the state of charge of a battery, including: Obtaining a first sequence of at least two parameters of the battery, each first sequence including first parameter values of the corresponding parameter at multiple first moments within a preset time period; For the first sequence of each parameter, decomposing the first sequence based on a preset decomposition coefficient to obtain a second sequence of the parameter and at least one third sequence; the second sequence represents the overall relationship of the parameter values of the parameter changing with time, and the third sequence represents the local relationship of the parameter values of the parameter changing with time; Inputting the second sequence and the third sequence of each parameter into a charge prediction hybrid model, and outputting a first prediction result regarding the state of charge of the battery; wherein, the first prediction result includes first predicted values of the state of charge of the battery at multiple second moments outside the preset time period; the state of charge represents the relationship between the remaining battery power and the total battery power.

[0005] In a second aspect, an embodiment of this application provides a device for predicting the state of charge of a battery, including: A sequence acquisition module for acquiring first sequences of at least two parameters of a battery, each first sequence including first parameter values of the corresponding parameter at multiple first moments within a preset time period; A sequence decomposition module for decomposing, for the first sequence of each parameter, the first sequence based on a preset decomposition coefficient to obtain a second sequence of the parameter and at least one third sequence; the second sequence represents the overall relationship of the parameter values of the parameter changing with time, and the third sequence represents the local relationship of the parameter values of the parameter changing with time; A charge state prediction module for inputting the second sequence and the third sequence of each parameter into a charge prediction hybrid model to output a first prediction result regarding the charge state of the battery; wherein, the first prediction result includes first predicted values of the charge state of the battery at multiple second moments outside the preset time period; the charge state represents the relationship between the remaining power and the total power of the battery.

[0006] In a third aspect, an embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory; The processor executes the computer program to implement the method provided in the embodiment of the first aspect or any optional embodiment of the first aspect.

[0007] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the method provided in the embodiment of the first aspect or any optional embodiment of the first aspect.

[0008] The beneficial effects brought by the technical solution provided by the embodiment of the present application are: By collecting the first parameter values of multiple parameters, generating the first sequences of each parameter, decomposing each first sequence to obtain a second sequence representing the overall change relationship of the parameter with time and a third sequence representing the local change relationship of the parameter with time, and then inputting the decomposed sequences into the charge prediction hybrid model for charge state prediction. That is, it realizes the prediction of the remaining charge state of the target battery by collecting data of multiple parameters, making the predicted remaining charge state more accurate. Description of the Drawings

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description in the embodiments of the present application.

[0010] Figure 1 It is a schematic flowchart of a method for predicting the charge state of a battery provided by an embodiment of the present application; Figure 2 It is a schematic flowchart of a sequence decomposition method in an example of an embodiment of the present application; Figure 3Schematic diagram of the overall process of the battery charge state prediction method in an example of the embodiment of the present application; Figure 4 Block diagram of the structure of a battery charge state prediction device provided by the embodiment of the present application; Figure 5 Schematic diagram of the structure of an electronic device provided by the embodiment of the present application. Detailed implementation manners

[0011] The embodiments of the present application will be described below with reference to the accompanying drawings in the present application. It should be understood that the embodiments described below in conjunction with the accompanying drawings are exemplary descriptions for explaining the technical solutions of the embodiments of the present application, and do not constitute limitations on the technical solutions of the embodiments of the present application.

[0012] Those skilled in the art of the present technology can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the terms "including" and "comprising" used in the embodiments of the present application mean that the corresponding features can be implemented as the presented features, information, data, steps, operations, elements and / or components, but do not exclude being implemented as other features, information, data, steps, operations, elements, components and / or their combinations supported by the art of the present technology. It should be understood that when we say an element is "connected" or "coupled" to another element, the one element can be directly connected or coupled to the other element, or it can mean that the one element and the other element establish a connection relationship through an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein indicates at least one of the items defined by the term, for example, "A and / or B" can be implemented as "A", or implemented as "B", or implemented as "A and B".

[0013] To make the objectives, technical solutions and advantages of the present application clearer, the embodiments of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0014] The technical solutions of the embodiments of the present application and the technical effects produced by the technical solutions of the present application will be described below through the description of several exemplary embodiments. It should be noted that the following embodiments can refer to, draw on or combine with each other. For the same terms, similar features and similar implementation steps in different embodiments, they will not be described repeatedly.

[0015] Figure 1 Schematic diagram of the process of a method for predicting the battery charge state provided by the embodiment of the present application. The execution subject of this method can be a terminal (such as a computer, a mobile phone, etc.). As Figure 1 shown, this method may include: Step S101: Obtain the first sequence of at least two parameters of the battery, where each first sequence includes the first parameter values of the corresponding parameter at multiple first moments within a preset time period.

[0016] In the embodiments of the present application, the parameters of the battery may include voltage, current, temperature, internal resistance, etc. The preset time period is generally a certain historical time period, and the first moment is generally the moment when there are numerical records of the parameters within the preset time period. The first parameter value may be the specific value of the parameter. For example, when the parameter is voltage, the corresponding first parameter value may be 12V (volts), 6V, etc.; when the parameter is current, the corresponding first parameter value may be 3A (amperes), 2A, etc.

[0017] Specifically, in the embodiments of the present application, the future state of charge needs to be predicted based on the future parameter values of multiple parameters of multiple batteries. For the future parameter value of each parameter, it is necessary to analyze the change trend of the historical parameter values and then make a prediction based on the analyzed change trend. Therefore, in the solution of the embodiments of the present application, by obtaining the first parameter values of each parameter at multiple first moments within a certain historical preset time period, after obtaining the multiple first parameter values of each parameter, for the convenience of subsequent prediction processing, the first parameter values can be sorted in the order of the first moments corresponding to the first parameter values to generate the first sequence corresponding to each parameter. For example, for the temperature parameter, there are three first moments t1, t2, and t3 within the preset time period. The temperature corresponding to the t1 moment is 30 degrees, the temperature corresponding to the t2 moment is 35 degrees, and the temperature corresponding to the t3 moment is 40 degrees. The order of the above three moments in time is t1 < t2 < t3, so the first sequence of temperature obtained is [30, 35, 40].

[0018] Step S102: For the first sequence of each parameter, decompose the first sequence based on a preset decomposition coefficient to obtain the second sequence of the parameter and at least one third sequence; the second sequence represents the overall relationship of the parameter value of the parameter changing with time, and the third sequence represents the local relationship of the parameter value of the parameter changing with time.

[0019] In the embodiments of the present application, a filter bank can be used to decompose the sequence. Each filter bank contains a low-pass filter and a high-pass filter, and each filter has a corresponding decomposition coefficient. The selection of the filter can be freely selected according to the actual decomposition requirements, and the embodiments of the present application do not limit this here.

[0020] Specifically, in order to extract the overall relationship and local relationship representing the change of parameter values over time in the first sequence respectively, the embodiments of the present application adopt the method of passing the sequence through the same filter bank. Specifically, the first sequence can be input into the low-pass filter and high-pass filter in the same filter bank respectively. The low-pass filter is used to smooth the passed sequence, making the output sequence more stable and better reflecting the overall change trend of each parameter value in the sequence. The high-pass filter is used to specify the change situation between each parameter value, and better reflecting the details of the change between each parameter value in the sequence. In the actual application scenario, the filter bank with appropriate decomposition coefficients can be selected for decomposition according to the required degree of detail of the parameter value change reflected by each decomposed sequence obtained. The embodiments of the present application do not limit the selection process of the filter bank here.

[0021] In the embodiments of the present application, in order to make the extracted parameter value change details more abundant, the first sequence can also be decomposed multiple times, that is, a second sequence representing the overall relationship of the parameter values of the parameter changing over time and multiple third sequences representing the local relationships of the parameter values of the parameter changing over time are obtained.

[0022] Step S103, input the second sequence and the third sequence of each parameter into the charge prediction hybrid model, and output a first prediction result regarding the charge state of the battery; wherein, the first prediction result includes first predicted values of the charge state of the battery at multiple second moments outside the preset time period; the charge state represents the relationship between the remaining power and the total power of the battery.

[0023] In the embodiments of the present application, the charge prediction hybrid model can be pre-trained, and this model can predict the charge state of the battery at a future moment according to the sequences of each parameter of the battery provided. The second moment is a certain future moment.

[0024] Specifically, after obtaining the second sequence representing the overall relationship of the parameter values of the parameter changing over time and the third sequence representing the local relationships of the parameter values of the parameter changing over time through the decomposition operation in step S102, these sequences are input into the trained charge prediction hybrid model. This model can deduce multiple future second moments based on the current moment and predict the charge state of the tram at each second moment, obtaining the first predicted values of the charge state of the battery at each second moment, and taking these first predicted values and the corresponding second moments as the first prediction result and outputting them.

[0025] Optionally, when the charge state at an important historical moment is missing, the charge prediction hybrid model provided in the embodiments of the present application can also be used to predict the charge state of the battery at this historical moment. Specifically, parameter values of each parameter at multiple moments before this historical moment can be obtained, a first sequence of each parameter can be generated and decomposed, and the obtained second sequence and third sequence after decomposition are input into the charge prediction hybrid model for prediction to supplement the historically missing data.

[0026] The solution provided in the present application generates a first sequence of each parameter by collecting first parameter values of multiple parameters, decomposes each first sequence to obtain a second sequence representing the overall change relationship of the parameter over time and a third sequence representing the local change relationship of the parameter over time, and then inputs the decomposed sequences into the charge prediction hybrid model for charge state prediction. That is, it realizes the prediction of the remaining charge state of the target battery by collecting data of multiple parameters, making the predicted remaining charge state more accurate.

[0027] Based on the above various embodiments, as an optional embodiment, the preset decomposition coefficients include a first coefficient group and a second coefficient group. The first coefficient group is used to extract the overall relationship of each parameter in the sequence changing over time, and the second coefficient group is used to extract the local relationship of each parameter in the sequence changing over time; Decomposing the first sequence based on the preset decomposition coefficients to obtain a second sequence of the parameter and at least one third sequence specifically includes: Performing a preset number of wavelet decomposition operations on the first sequence based on the preset decomposition coefficients, and using the subsequence obtained from the last decomposition as the second sequence; Wherein, each wavelet decomposition operation includes: Obtaining the subsequence of this wavelet decomposition operation, wherein the subsequence of the first wavelet decomposition operation is the first sequence; Processing the parameter values included in the subsequence of this wavelet decomposition operation based on the first coefficient group, and generating the subsequence of the next wavelet decomposition operation based on the processed parameter values; Processing the parameter values included in the subsequence of this wavelet decomposition operation based on the second coefficient group, and generating the third sequence of this wavelet decomposition operation based on the processed parameter values.

[0028] In the embodiments of the present application, the first coefficient group can be the decomposition coefficients of the low-pass filter described above, and the second coefficient group can be the decomposition coefficients of the high-pass filter described above. The subsequence can be the sequence obtained after passing through the low-pass filter in each wavelet decomposition operation.

[0029] Specifically, such as Figure 2As shown, the step-by-step decomposition method adopted by the embodiments of the present application for the first sequence of each parameter is wavelet decomposition. For each parameter, when performing the first wavelet decomposition, the first sequence of the parameter is respectively input into a low-pass filter and a high-pass filter. The subsequence obtained by the output of the low-pass filter is the subsequence for the subsequent second wavelet decomposition operation (hereinafter denoted as A1), and the subsequence obtained by the output of the high-pass filter is the third sequence (hereinafter denoted as D1). Since A1 can better reflect the overall relationship of the parameter value changing with time compared with the first sequence, in the next wavelet decomposition, A1 can be directly input into the low-pass filter and the high-pass filter respectively. At this time, the subsequence obtained by the output of the low-pass filter is the subsequence for the subsequent third wavelet decomposition operation (hereinafter denoted as A2), and the subsequence obtained by the output of the high-pass filter is the new third sequence (hereinafter denoted as D2). This process is repeated until the subsequence obtained after the last wavelet decomposition operation (hereinafter denoted as Am) and the last third sequence (hereinafter denoted as Dm) are obtained. Finally, Am is used as the second sequence, and D1, D2... Dm are all used as the third sequences.

[0030] Based on the above various embodiments, as an alternative embodiment, each parameter value included in the subsequence is processed based on the first coefficient group, and the subsequence for the next wavelet decomposition operation is generated based on the processed parameter values. Specifically, it includes: Construct multiple parameter value groups based on the subsequence of the current wavelet decomposition operation; wherein, each parameter value group includes parameter values that are adjacent in the subsequence, and the number of parameter values included in each parameter value group is the same as the number of coefficients included in the first coefficient group; For each parameter value group, the weighted sum of the parameter values in the parameter value group is obtained by successively weighting each parameter value in the parameter value group with each coefficient in the first coefficient group; The weighted results of each parameter value group are arranged in the order of the parameter value at the head of each corresponding parameter value group in the subsequence to generate a weighted result sequence; A downsampling operation is performed on the weighted result sequence to obtain the subsequence for the next wavelet decomposition operation.

[0031] In the embodiments of the present application, the parameter value group is composed of several adjacent parameter values in the subsequence. The downsampling operation is to obtain a parameter value every other parameter value.

[0032] Specifically, when the subsequence passes through the low-pass filter, the low-pass filter divides the subsequence into multiple groups of parameter values. The specific grouping method is to divide each adjacent group of parameter values with the same number of coefficients as the filter into one group. For example, when the coefficients of the low-pass filter are [1 / 2, 1 / 2], the number of coefficients is 2, then every two adjacent parameter values in the subsequence are grouped into one group of parameter values. For example, if the subsequence is [20, 24, 28, 34], then the obtained groups of parameter values should be [20, 24], [24, 28], [28, 34].

[0033] After that, the coefficients of the low-pass filter are used to perform weighted summation on each group of parameter values, and the results of the weighted summation of each parameter value are integrated into a new sequence (i.e., the weighted result sequence). According to the example described above, the process of weighted summation can be (20 ×1 / 2 + 24 ×1 / 2) = 22, (24 ×1 / 2 + 28 ×1 / 2) = 26, (28 ×1 / 2 + 34 ×1 / 2) = 31. The weighted results are 22, 26, 31. Each weighted result is arranged in the order of the parameter value at the first position in the corresponding parameter group in the subsequence. The obtained weighted result sequence is [22, 26, 31]. After obtaining the weighted result sequence, downsampling operation is performed on this sequence, that is, the subsequence [22, 31] for the next wavelet decomposition operation is obtained.

[0034] Optionally, the calculation method when the subsequence passes through the high-pass filter is also similar to the above process. Here, a high-pass filter with coefficients [-1 / 2, 1 / 2] is taken as an example for a brief description. Based on the above subsequence [20, 24, 28, 34], the results obtained after weighted summation using the coefficients of the high-pass filter are (20 ×(-1 / 2) + 24 ×1 / 2) = 2, (24 ×(-1 / 2) + 28 ×1 / 2) = 2, (28 ×(-1 / 2) + 34 ×1 / 2) = 3. Then the obtained weighted result sequence is [2, 2, 3]. After the downsampling operation, the third sequence obtained is [2, 3].

[0035] Based on the above various embodiments, as an optional embodiment, the charge prediction hybrid model includes a first sub-model and a second sub-model; Input the second sequence and the third sequence of each parameter into the charge prediction hybrid model, and output a first prediction result regarding the charge state of the battery, specifically including: Input the second sequence and the third sequence of each parameter into the first sub-model, and output a second prediction result regarding the charge state of the battery; the second prediction result includes second predicted values of the charge state of the battery at multiple second moments; wherein, each second predicted value is different from the first predicted value corresponding to the second moment. Input the second predicted charge result, the second sequence and the third sequence of each parameter into the second sub-model, and output the first prediction result.

[0036] In the embodiments of the present application, the first sub-model may be a model based on N-BEATS (Neural Basis Expansion Analysis for Time Series), and the second sub-model may be a model based on LSTM (Long Short-Term Memory).

[0037] Specifically, the charge prediction hybrid model provided in the embodiments of the present application is jointly composed of two sub-models. The input of the first sub-model is the second sequence and the third sequence of each parameter obtained by the previous wavelet decomposition, and the output is the preliminary prediction result of the charge state at multiple second moments (i.e., the second predicted value); the input of the second sub-model is the prediction result of the charge state at multiple second moments output by the first sub-model and the second sequence and the third sequence of each parameter obtained by the previous wavelet decomposition, and the output is the final prediction result of the charge state at multiple second moments (i.e., the first predicted value).

[0038] Based on the above various embodiments, as an optional embodiment, inputting the second sequence and the third sequence of each parameter into the first sub-model and outputting a second prediction result regarding the charge state of the battery specifically includes: For each parameter, extract features from the second sequence of the parameter and at least one third sequence of the parameter to obtain a first feature representing the periodic change law of the parameter, and predict the second parameter value of each parameter at each second moment based on the first feature; For each second moment, determine the second predicted value of the charge state at the second moment based on the second parameter values of each parameter at the second moment.

[0039] Specifically, there are multiple stackable prediction blocks in the first sub-model provided by the embodiments of the present application. The prediction blocks can be divided into multiple types. Each type of prediction block can extract different features from each of the input sequences. For example, the first type of stacked block can extract features regarding the trend of parameter values changing over time from the second sequences of each parameter, and the second type of stacked block can extract detailed features regarding the changes between parameter values at adjacent moments from the third sequences of each parameter, etc. The embodiments of the present application do not limit this here. After the multiple stacked blocks respectively extract different features, these features are further integrated to obtain a first feature that can accurately reflect the periodic change law of the parameter. Then, based on this first feature, the parameter values of each parameter at the second moment are predicted to obtain multiple second predicted values of the parameter.

[0040] After predicting the second predicted values of each parameter at a certain second moment, the charge state prediction calculation for this moment can be performed according to the second predicted values at this second moment. When the second predicted values of the charge state at each second moment are calculated, the second predicted values and the corresponding second moments are output together in the form of a time series.

[0041] It should be noted that each stacked block in the first sub-model provided by the present application further includes a forward neural network and a backward neural network. The forward neural network is composed of multiple fully connected layers and processes each of the input sequences to capture the local features of each parameter changing over time. The backward neural network is also composed of multiple fully connected layers. Different from the forward neural network, the backward neural network mainly learns the residual part in each sequence, that is, the original sequence minus the part predicted by the forward neural network. The significance of this neural network is to complement the features that were not captured or missed during the prediction process of the forward neural network and further improve the local features captured by the forward neural network.

[0042] Based on the above various embodiments, as an alternative embodiment, the second predicted charge result, the second sequences of each parameter, and the third sequences are input into the second sub-model to output a first prediction result, which specifically includes: For each parameter, feature extraction is performed on the second sequence of the parameter and at least one third sequence of the parameter to obtain a second feature that simultaneously represents the overall relationship and the local relationship. Based on the second feature, the third parameter value of each parameter at each second moment is predicted; For each second moment, based on the third parameter values of each parameter at the second moment, the third predicted value of the charge state at the second moment is determined. The second predicted value and the third predicted value are respectively weighted and summed through a preset weight to obtain the first predicted value at the second moment; Based on the first predicted values at each second moment, a first prediction result is generated.

[0043] Specifically, in actual operation, it is found that if only the preliminary second prediction result output by the first sub-model is used as the final prediction result, there will still be a large prediction error. Therefore, in the embodiment of the present application, a second sub-model is added on the basis of the first sub-model to further correct the prediction result of the first sub-model, so as to improve the accuracy of the final prediction result. Specifically, the second sub-model inputs the second sequence and the third sequence of parameters into its own forgetting gate. For each sequence, the forgetting gate determines the unnecessary feature information in each sequence, removes the unnecessary feature information, and then stores the remaining information in its own memory unit through the update gate; when the information of all sequences is updated to the memory unit, the memory unit can obtain the second feature that simultaneously represents the overall relationship and the local relationship, and then predicts the parameter values of the parameter at each second moment according to the second feature, and the predicted result is the third parameter value.

[0044] When the second sub-model predicts the third parameter value of each parameter at each second moment, then predicts the third prediction value of the charge state at each second moment according to the predicted third parameter value of each parameter, and then performs weighted summation on the third prediction values of each second moment predicted by the second sub-model and the second prediction values of the corresponding moments predicted by the first sub-model respectively, that is, the process of correcting the preliminary prediction result of the first sub-model by the second sub-model is realized. The result obtained after the weighted summation is the first prediction value at each second moment, and then outputting the first prediction value at each second moment obtains the first prediction result.

[0045] It should be noted that the weights used for weighting the second prediction values output by the first sub-model and the third prediction values output by the second sub-model in the embodiment of the present application can be adjusted during the training of the model, that is, during the training of the model, by observing the prediction values output by the two sub-models respectively, and comparing the prediction values output by the model with the true values at the actual corresponding moments, the respective weights that can ensure the minimum error are determined.

[0046] Optionally, during the training of the second sub-model, the adam (Adaptive Moment Estimation) algorithm can be used, and the learning rate of the model can be set to 0.001 during training.

[0047] On the basis of the above various embodiments, as an optional embodiment, the method further specifically includes: Obtain the true value of the charge state at each second moment, and determine the difference between the true value and the first prediction value; Adjust the model parameters of the charge prediction hybrid model based on the difference.

[0048] Specifically, the solution of the embodiment of the present application also provides a method for real-time adjustment of the model. Specifically, after the charge prediction hybrid model predicts the first predicted values of the battery charge states at each second moment, the first predicted values will be recorded. When the current moment reaches the predicted second moment, the actual true value of the charge state at the second moment can be obtained, and the true value is compared with the corresponding recorded first predicted value and the difference between them is calculated. Then, the model parameters are adjusted according to the calculated difference. The calculation of the difference can be performed using the mean squared error loss function, and the calculation formula is as follows:

[0049] where n is the number of second moments, yi represents the true value of the charge state, represents the first predicted value of the charge state.

[0050] Through the above method, the charge state hybrid model can continuously optimize itself and further improve its prediction accuracy.

[0051] The battery charge state prediction method provided by the embodiment of the present application can be applied to multiple actual application scenarios. For example, in the field of electric vehicles, during the driving of an electric vehicle, battery parameters of the vehicle (such as voltage, current, temperature, etc.) can be obtained in real time through sensors. Then, a first sequence of each parameter can be generated based on the obtained battery parameters and decomposed to obtain a second sequence and a third sequence of each parameter. Next, each second sequence and third sequence are input into the charge prediction hybrid model to perform real-time prediction on the remaining charge of the vehicle battery, and the prediction result is presented to the driver, so that the driver can understand the future driving range of the vehicle in real time, and at the same time, it can also prevent over-discharge or over-charging of the vehicle battery and improve the service life of the vehicle battery.

[0052] Next, in conjunction with Figure 3 , the overall process of the battery charge state prediction method provided by the embodiment of the present application will be introduced. As Figure 3 shown, first, the parameter values of each parameter of the battery at multiple first moments (i.e., T1, T2, Tn in the figure) are collected, and a first sequence of each parameter is generated. Then, the first sequence of each parameter is decomposed to obtain a second sequence and at least one third sequence of each parameter (only two are drawn in the figure). Next, each sequence is input into the first sub-model, and the preliminary predicted values of the first sub-model for each second moment (i.e., Tx, Ty, Tz in the figure) are output. Then, the preliminary predicted values output by the first sub-model and the second sequence and third sequence of each parameter are input into the second sub-model, and the second sub-model outputs the final predicted values of each second moment.

[0053] Figure 4 is the structural block diagram of a battery charge state prediction device provided by the embodiment of the present application. AsFigure 4 As shown in Figure 4 , the prediction device 400 for the charge state of the battery may include: a sequence acquisition module 401, a sequence decomposition module 402, and a charge state prediction module 403. Among them, The sequence acquisition module 401 is configured to acquire a first sequence of at least two parameters of the battery, and each first sequence includes first parameter values of the corresponding parameter at a plurality of first moments within a preset time period; The sequence decomposition module 402 is configured to decompose the first sequence of each parameter based on a preset decomposition coefficient to obtain a second sequence of the parameter and at least one third sequence; the second sequence represents the overall relationship of the parameter value of the parameter changing with time, and the third sequence represents the local relationship of the parameter value of the parameter changing with time; The charge state prediction module 403 is configured to input the second sequence and the third sequence of each parameter into the charge prediction hybrid model, and output a first prediction result regarding the charge state of the battery; wherein, the first prediction result includes first prediction values of the charge state of the battery at a plurality of second moments outside the preset time period; the charge state represents the relationship between the remaining power and the total power of the battery.

[0054] The solution provided by this application generates the first sequence of each parameter by collecting the first parameter values of multiple parameters, decomposes each first sequence to obtain a second sequence representing the overall change relationship of the parameter with time and a third sequence representing the local change relationship of the parameter with time, and then inputs the decomposed sequences into the charge prediction hybrid model for charge state prediction. That is, it realizes the prediction of the remaining charge state of the target battery by collecting data of multiple parameters, making the predicted remaining charge state more accurate.

[0055] Based on the above various embodiments, as an optional embodiment, the preset decomposition coefficient includes a first coefficient group and a second coefficient group. The first coefficient group is used to extract the overall relationship of each parameter in the sequence changing with time, and the second coefficient group is used to extract the local relationship of each parameter in the sequence changing with time; The sequence decomposition module is specifically configured to: Perform a preset number of wavelet decomposition operations on the first sequence based on the preset decomposition coefficient, and use the subsequence obtained from the last decomposition as the second sequence; Among them, each wavelet decomposition operation includes: Obtain the subsequence of the current wavelet decomposition operation, wherein the subsequence of the first wavelet decomposition operation is the first sequence; Process the parameter values included in the subsequence of the current wavelet decomposition operation based on the first coefficient group, and generate the subsequence of the next wavelet decomposition operation based on the processed parameter values; Process each parameter value included in the subsequence of the current wavelet decomposition operation based on the second coefficient array, and generate a third sequence of the current wavelet decomposition operation based on the processed parameter values.

[0056] Based on the above various embodiments, as an alternative embodiment, the sequence decomposition module is further configured to: Construct multiple groups of parameter values based on the subsequence of the current wavelet decomposition operation; wherein, each parameter value included in each group of parameter values is adjacent in the subsequence, and the number of parameter values included in each group of parameter values is the same as the number of coefficients included in the first coefficient array; For each group of parameter values, perform weighted summation on each parameter value in the group of parameter values in turn through each coefficient in the first coefficient array to obtain a weighted result; Arrange the weighted results of each group of parameter values in the order of the parameter value at the head of the corresponding group of parameter values in the subsequence to generate a weighted result sequence; Perform a downsampling operation on the weighted result sequence to obtain a subsequence of the next wavelet decomposition operation.

[0057] Based on the above various embodiments, as an alternative embodiment, the charge prediction hybrid model includes a first sub-model and a second sub-model; The charge state prediction module is specifically configured to: Input the second sequence and the third sequence of each parameter into the first sub-model, and output a second prediction result regarding the charge state of the battery; the second prediction result includes second prediction values of the charge state of the battery at multiple second moments; wherein, each second prediction value is different from the first prediction value corresponding to the second moment; Input the second predicted charge result, the second sequence and the third sequence of each parameter into the second sub-model, and output a first prediction result.

[0058] Based on the above various embodiments, as an alternative embodiment, the charge state prediction module is further configured to: For each parameter, perform feature extraction on the second sequence of the parameter and at least one third sequence of the parameter to obtain a first feature representing the periodic change law of the parameter, and predict the second parameter value of each parameter at each second moment based on the first feature; For each second moment, determine the second prediction value of the charge state at the second moment based on the second parameter values of each parameter at the second moment.

[0059] Based on the above various embodiments, as an alternative embodiment, the charge state prediction module is further configured to: For each parameter, feature extraction is performed on the second sequence of the parameter and at least one third sequence of the parameter to obtain a second feature that simultaneously characterizes the overall relationship and the local relationship, and based on the second feature, a third parameter value of each parameter at the second moment is predicted; For each second moment, based on the third parameter values of the parameters at the second moment, a third predicted value of the charge state at the second moment is determined, and the second predicted value and the third predicted value are respectively weighted and summed through a preset weight to obtain a first predicted value at the second moment; A first prediction result is generated based on the first predicted values at each second moment.

[0060] Based on the above various embodiments, as an alternative embodiment, the device further includes a model parameter adjustment module, which is specifically configured to: Obtain the true value of the charge state at each second moment, and determine the difference between the true value and the first predicted value; Adjust the model parameters of the charge prediction hybrid model based on the difference.

[0061] The following refers to Figure 5 , which shows a schematic structural diagram of an electronic device (such as a terminal device or a server that executes the method shown in Figure 1 ) 500 suitable for implementing the embodiments of the present application. The electronic device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), vehicle terminals (such as vehicle navigation terminals), wearable devices, etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 5 The electronic device shown is only an example and should not bring any limitation to the functions and usage scopes of the embodiments of the present application.

[0062] The electronic device includes: a memory and a processor. The memory is used to store a program for executing the methods described in the above various method embodiments; the processor is configured to execute the program stored in the memory. Here, the processor may be referred to as the processing device 501 described below, and the memory may include at least one of the read-only memory (ROM) 502, random access memory (RAM) 503, and storage device 508 described below, as specifically shown below: As Figure 5As shown, the electronic device 500 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 501, which may perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage device 508 into the random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. The input / output (I / O) interface 505 is also connected to the bus 504.

[0063] Generally, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509. The communication device 509 may allow the electronic device 500 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 5 an electronic device with various devices is shown, it should be understood that it is not required to implement or have all the shown devices. Instead, more or fewer devices may be implemented or had.

[0064] Specifically, according to an embodiment of the present application, the process described above with reference to the flowchart may be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program codes for executing the method shown in the flowchart. In such an embodiment, the computer program may be downloaded and installed from the network through the communication device 509, or installed from the storage device 508, or installed from the ROM 502. When the computer program is executed by the processing device 501, the above functions defined in the method of the embodiment of the present application are executed.

[0065] It should be noted that the computer-readable storage medium described above in this application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of a computer-readable storage medium 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, which can be used by or in conjunction with an instruction execution system, apparatus, or device. And in this application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0066] In some embodiments, the client and the server can communicate using any currently known or future-developed network protocol such as HTTP (HyperText Transfer Protocol), and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks (“LAN”), wide area networks (“WAN”), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed network.

[0067] The above computer-readable medium can be included in the above electronic device; or it can exist separately and not be assembled into the electronic device.

[0068] The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by the electronic device, the electronic device is caused to: Obtain a first sequence for each of at least two parameters of the battery, where each first sequence includes first parameter values of the corresponding parameter at a plurality of first moments within a preset time period; for the first sequence of each parameter, decompose the first sequence based on a preset decomposition coefficient to obtain a second sequence of the parameter and at least one third sequence; the second sequence represents the overall relationship of the parameter values of the parameter changing with time, and the third sequence represents the local relationship of the parameter values of the parameter changing with time; input the second sequence and the third sequence of each parameter into a charge prediction hybrid model, and output a first prediction result regarding the state of charge of the battery; wherein, the first prediction result includes first predicted values of the state of charge of the battery at a plurality of second moments outside the preset time period; the state of charge represents the relationship between the remaining power and the total power of the battery.

[0069] Computer program code for performing the operations of this application can be written in one or more programming languages or combinations thereof. The above-mentioned programming languages include, but are not limited to, object-oriented programming languages - such as Java, Smalltalk, C++, and also include conventional procedural programming languages - such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network - including a local area network (LAN) or a wide area network (WAN) - or, it can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0070] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and this module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0071] The modules or units involved in the embodiments described in this application can be implemented in software or in hardware. Among them, the name of the module or unit does not necessarily limit the unit itself in some cases. For example, the first constraint acquisition module can also be described as "the module for acquiring the first constraint".

[0072] The functions described above in this article can be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that can be used include: Field Programmable Gate Array (FPGA), Application Specific Integrated Circuit (ASIC), Application Specific Standard Product (ASSP), System on Chip (SOC), Complex Programmable Logic Device (CPLD), and so on.

[0073] In the context of this application, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media would include electrical connections based on one or more wires, portable computer disks, hard disks, Random Access Memory (RAM), Read Only Memory (ROM), Erasable Programmable Read Only Memory (EPROM or Flash Memory), optical fibers, portable compact disc read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0074] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise clearly stated in this article, the execution of these steps is not strictly limited in order and can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings can include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same moment but can be executed at different moments, and their execution order is not necessarily sequential but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0075] The above description is only part of the embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A method for predicting a battery charge state, characterized in that: include: Acquire a first sequence of at least two parameters of the battery, each first sequence comprising first parameter values ​​of the corresponding parameter at a plurality of first moments within a preset time period; For each first sequence of parameters, the first sequence is decomposed based on a preset decomposition coefficient to obtain a second sequence of the parameters and at least one third sequence; the second sequence represents the overall relationship of the parameter values ​​of the parameters changing over time, and the third sequence represents the local relationship of the parameter values ​​of the parameters changing over time; The second sequence and the third sequence of each parameter are input into the charge prediction hybrid model, and a first prediction result about the charge state of the battery is output; wherein the first prediction result includes a first prediction value of the charge state of the battery at multiple second moments outside the preset time period; and the charge state characterizes the relationship between the remaining power of the battery and the total power.

2. The method according to claim 1, characterized in that The preset decomposition coefficients include a first coefficient group and a second coefficient group, the first coefficient group is used to extract the overall relationship of each parameter in the sequence changing with time, and the second coefficient group is used to extract the local relationship of each parameter in the sequence changing with time; Decomposing the first sequence based on a preset decomposition coefficient to obtain a second sequence and at least one third sequence of the parameters includes: Performing a preset number of wavelet decomposition operations on the first sequence based on the preset decomposition coefficients, and using a subsequence obtained by the last decomposition as the second sequence; Among them, each wavelet decomposition operation includes: Obtaining a subsequence of the current wavelet decomposition operation, wherein the subsequence of the first wavelet decomposition operation is the first sequence; Processing each parameter value included in the subsequence of the current wavelet decomposition operation based on the first coefficient group, and generating a subsequence of the next wavelet decomposition operation based on the processed parameter values; The parameter values ​​included in the subsequence of the current wavelet decomposition operation are processed based on the second coefficient group, and the third sequence of the current wavelet decomposition operation is generated based on the processed parameter values.

3. The method according to claim 2, characterized in that The step of processing each parameter value included in the subsequence based on the first coefficient group and generating a subsequence for the next wavelet decomposition operation based on each parameter value after the processing includes: Constructing a plurality of parameter value groups based on the subsequence of the current wavelet decomposition operation; wherein the parameter values ​​included in each parameter value group are adjacent in the subsequence, and the number of parameter values ​​included in each parameter value group is the same as the number of coefficients included in the first coefficient group; For each parameter value group, weighted summing each parameter value in the parameter value group is performed in turn by each coefficient in the first coefficient group to obtain a weighted result; Arrange the weighted results of each parameter value group according to the order of the first parameter value in each corresponding parameter value group in the subsequence to generate a weighted result sequence; A downsampling operation is performed on the weighted result sequence to obtain a subsequence for the next wavelet decomposition operation.

4. The method according to claim 1, characterized in that The charge prediction hybrid model includes a first sub-model and a second sub-model; The step of inputting the second sequence and the third sequence of each parameter into the charge prediction hybrid model and outputting a first prediction result about the charge state of the battery comprises: Inputting the second sequence and the third sequence of each parameter into the first sub-model, and outputting a second prediction result about the charge state of the battery; the second prediction result includes second prediction values ​​of the charge state of the battery at multiple second moments; wherein each second prediction value is different from the first prediction value corresponding to the second moment; The second predicted charge result, the second sequence of parameters and the third sequence are input into the second sub-model, and the first predicted result is output.

5. The method according to claim 4, characterized in that The step of inputting the second sequence and the third sequence of each parameter into the first sub-model and outputting a second prediction result about the charge state of the battery comprises: For each parameter, feature extraction is performed on the second sequence of the parameter and at least one third sequence of the parameter to obtain a first feature characterizing a periodic variation law of the parameter, and a second parameter value of the parameter at each second moment is predicted based on the first feature; For each second moment, a second predicted value of the state of charge at the second moment is determined based on the second parameter values ​​of each parameter at the second moment.

6. The method according to claim 5, characterized in that The step of inputting the second predicted charge result, the second sequence of parameters, and the third sequence into the second sub-model and outputting the first predicted result comprises: For each parameter, feature extraction is performed on the second sequence of the parameter and at least one third sequence of the parameter to obtain a second feature that simultaneously characterizes the overall relationship and the local relationship, and a third parameter value of the parameter at each second moment is predicted based on the second feature; For each second moment, based on the third parameter value of each parameter at the second moment, determine a third predicted value of the charge state at the second moment, and perform weighted summation of the second predicted value and the third predicted value by using preset weights to obtain a first predicted value at the second moment; The first prediction result is generated based on the first prediction value at each second moment.

7. The method according to claim 1, characterized in that The method further comprises: Obtaining a true value of the charge state at each second moment, and determining a difference between the true value and the first predicted value; Model parameters of the charge prediction hybrid model are adjusted based on the difference.

8. A battery charge state prediction device, characterized in that: include: A sequence acquisition module, used to acquire first sequences of at least two parameters of the battery, each first sequence including first parameter values ​​of the corresponding parameter at a plurality of first moments within a preset time period; A sequence decomposition module, for decomposing a first sequence of each parameter based on a preset decomposition coefficient to obtain a second sequence of the parameter and at least one third sequence; the second sequence represents the overall relationship of the parameter value of the parameter changing over time, and the third sequence represents the local relationship of the parameter value of the parameter changing over time; A charge state prediction module is used to input the second sequence and the third sequence of each parameter into a charge prediction hybrid model, and output a first prediction result about the charge state of the battery; wherein the first prediction result includes a first prediction value of the charge state of the battery at multiple second moments outside the preset time period; and the charge state characterizes the relationship between the remaining power of the battery and the total power.

9. An electronic device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images