Battery charge estimation method and system based on machine learning

Abstract

The present disclosure relates to a battery power estimation method and system based on machine learning, wherein the method comprises: obtaining battery operating parameters, the battery operating parameters comprising a plurality of operating parameters of different types; inputting the battery operating parameters into a power estimation model to obtain an estimated remaining battery power value; the power estimation model is obtained by pre-training a first machine learning model based on a plurality of sample operating parameter subsets, each sample operating parameter subset containing a plurality of sample operating parameters of the same type of battery obtained in advance, the sample operating parameters in different sample operating parameter subsets being of different types, the plurality of sample operating parameter subsets being obtained by dividing a first sample operating parameter set of the battery obtained in advance according to a parameter classification result, the parameter classification result being obtained by processing the first sample operating parameter set by a preset classification model, and the preset classification model being obtained by training a second machine learning model based on a second sample operating parameter set of the battery.

Description

Technical Field

[0001] This disclosure relates to the field of battery power estimation technology, and in particular to a battery power estimation method and system based on machine learning. Background Technology

[0002] Currently, battery remaining capacity estimation is one of the core technologies of battery management systems. Traditional methods for estimating battery remaining capacity mainly include the ampere-hour integration method. However, the ampere-hour integration method accumulates estimation errors over time, leading to a decrease in the accuracy of battery remaining capacity estimation. Related technologies have proposed capacity estimation models based on pre-trained machine learning using neural network models to estimate battery remaining capacity. However, current technologies perform poorly when using a large amount of diverse test data as training samples for model training, meaning that the accuracy of the trained model in predicting battery remaining capacity still needs improvement. Summary of the Invention

[0003] To address or at least partially address the aforementioned technical problems, this disclosure provides a battery power estimation method and system based on machine learning.

[0004] In a first aspect, embodiments of this disclosure provide a battery power estimation method based on machine learning, including:

[0005] Obtain battery operating parameters, which include multiple operating parameters of different types;

[0006] The battery operating parameters are input into a power estimation model to obtain an estimated remaining battery power value. The power estimation model is pre-trained on a first machine learning model based on multiple subsets of sample operating parameters. Each subset contains multiple sample operating parameters of the same type for the battery, and the types of sample operating parameters in different subsets are different. The multiple subsets of sample operating parameters are obtained by classifying the pre-acquired first set of sample operating parameters of the battery using parameter classification results. The parameter classification results are obtained by processing the first set of sample operating parameters using a preset classification model. The preset classification model is trained on a second machine learning model based on a pre-acquired second set of sample operating parameters of the battery.

[0007] In one embodiment, the battery operating parameters of different types include any two or more of the following: battery voltage, battery current, battery temperature, and battery internal resistance.

[0008] In one embodiment, the training process of the preset classification model includes: constructing a target loss function for the second machine learning model; inputting the second sample working parameters and their corresponding label values ​​from the second sample working parameter set into the second machine learning model to output corresponding predicted label values; iteratively updating the second machine learning model based on the predicted label values ​​until the loss value of the target loss function meets a preset condition, at which point training ends; wherein the target loss function includes a first loss function and a second loss function, the first loss function being used to characterize the difference between the label values ​​corresponding to the second sample working parameters and the predicted label values ​​output by the second machine learning model, and the second loss function being used to characterize the probability of a class prediction error corresponding to the second sample working parameters input into the second machine learning model.

[0009] In one embodiment, the loss value of the target loss function satisfies the following preset conditions: the loss value of the first loss function is less than a first threshold, and the loss value of the second loss function is less than a second threshold.

[0010] In one embodiment, the training process of the preset classification model further includes: determining the weight values ​​corresponding to the second sample working parameters in the second sample working parameter set, and iteratively updating the second machine learning model based on the predicted label values ​​and weight values ​​until the loss value of the target loss function meets the preset condition and the training ends.

[0011] In one embodiment, the weight value corresponding to the second sample working parameter is related to the type of the second sample working parameter, and different weight values ​​correspond to different types of second sample working parameters.

[0012] In one embodiment, the second sample operating parameters include battery voltage, battery current, battery temperature, and battery internal resistance.

[0013] Secondly, embodiments of this disclosure provide a battery power estimation system based on machine learning, comprising:

[0014] The parameter acquisition module is used to acquire battery operating parameters, which include multiple operating parameters of different types.

[0015] A power estimation module is used to input the battery operating parameters into a power estimation model to obtain an estimated remaining battery power value. The power estimation model is pre-trained on a first machine learning model based on multiple subsets of sample operating parameters. Each subset of sample operating parameters contains multiple sample operating parameters of the same type for the battery, and the types of sample operating parameters in different subsets are different. The multiple subsets of sample operating parameters are obtained by classifying the pre-acquired first set of sample operating parameters of the battery using parameter classification results. The parameter classification results are obtained by processing the first set of sample operating parameters using a preset classification model. The preset classification model is trained on a second machine learning model based on a pre-acquired second set of sample operating parameters of the battery.

[0016] Thirdly, embodiments of this disclosure provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the machine learning-based battery power estimation method described in any of the above embodiments.

[0017] Fourthly, embodiments of this disclosure provide an electronic device, including:

[0018] Processor; and

[0019] Memory, used to store computer programs;

[0020] The processor is configured to execute the machine learning-based battery power estimation method described in any of the above embodiments by executing the computer program.

[0021] The technical solution provided in this disclosure has the following advantages compared with the prior art:

[0022] The battery power estimation method and system based on machine learning provided in this disclosure acquire multiple operating parameters of different types of battery operating parameters, and input the battery operating parameters into a power estimation model to obtain an estimated remaining battery power value. The power estimation model is pre-trained on a first machine learning model based on multiple subsets of sample operating parameters. Each subset of sample operating parameters contains multiple sample operating parameters of the same type of the pre-acquired battery, and the types of sample operating parameters in different subsets are different. The multiple subsets of sample operating parameters are obtained by classifying the pre-acquired first set of sample operating parameters of the battery using parameter classification results. The parameter classification results are obtained by processing the first set of sample operating parameters using a preset classification model. The preset classification model is obtained by training a second machine learning model based on a second set of pre-acquired sample operating parameters of the battery. In this way, when faced with a large number of different types of battery operating parameters as training sample data for model training, the pre-trained preset classification model accurately classifies the large number of different types of battery operating parameters into multiple sample operating parameter subsets. Then, based on the sample operating parameters in the multiple sample operating parameter subsets, the first machine learning model is trained to obtain the power estimation model. That is, by accurately classifying the large number of different types of battery operating parameters and then training the power estimation model in a targeted manner based on the classified operating parameters, the training effect of the power estimation model is improved, and the accuracy of the trained power estimation model in predicting the remaining battery power is improved. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0024] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart of a machine learning-based battery power estimation method according to an embodiment of the present disclosure;

[0026] Figure 2 This is a schematic diagram illustrating the training process of the power estimation model according to an embodiment of the present disclosure;

[0027] Figure 3 This is a flowchart illustrating the training process of a preset classification model according to an embodiment of the present disclosure.

[0028] Figure 4This is a schematic diagram of a machine learning-based battery power estimation system according to an embodiment of the present disclosure;

[0029] Figure 5 This is a schematic diagram of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0030] To better understand the above-mentioned objectives, features, and advantages of this disclosure, the solutions disclosed herein will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0031] Numerous specific details are set forth in the following description in order to provide a full understanding of this disclosure, but this disclosure may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only some, and not all, of the embodiments of this disclosure.

[0032] It should be understood that in the following text, "at least one item" refers to one or more items, and "more than one item" refers to two or more items. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0033] Figure 1 This is a flowchart of a machine learning-based battery power estimation method according to an embodiment of the present disclosure. The method can be executed by an electronic device such as a smart terminal or a vehicle controller, and specifically includes the following steps:

[0034] Step S101: Obtain battery operating parameters, which include multiple operating parameters of different types.

[0035] For example, in one embodiment, the battery operating parameters of different types may include, but are not limited to, any two or more of the following: battery voltage V, battery current I, battery temperature T, and battery internal resistance R. Battery voltage V and battery current I may be, for example, discharge voltage and discharge current, or charging voltage and charging current, etc. The specific methods for obtaining these parameters can be understood by referring to existing technologies, and will not be elaborated here.

[0036] Step S102: Input the battery operating parameters into the power estimation model to obtain the estimated remaining battery power value; wherein, the power estimation model is pre-trained on a first machine learning model based on multiple subsets of sample operating parameters, each subset of sample operating parameters contains multiple sample operating parameters of the same type of pre-acquired battery, and the types of sample operating parameters in different subsets of sample operating parameters are different, the multiple subsets of sample operating parameters are obtained by dividing the first set of sample operating parameters of the pre-acquired battery by parameter classification results, and the parameter classification results are obtained by processing the first set of sample operating parameters by a preset classification model; wherein, the preset classification model is trained on a second machine learning model based on a second set of sample operating parameters of the pre-acquired battery.

[0037] For example, refer to Figure 2As shown, during the training process of the power estimation model 201, multiple second sample working parameters are pre-acquired to form a second sample working parameter set. These second sample working parameters may include, but are not limited to, battery voltage V, battery current I, battery temperature T, and battery internal resistance R. Then, based on these multiple second sample working parameters and their corresponding classification labels, a second machine learning model, such as a convolutional neural network, is trained to obtain a preset classification model 202. The classification labels can be numerical values ​​such as 1, 2, 3, and 4, representing different types of second sample working parameters, such as battery voltage V, battery current I, battery temperature T, and battery internal resistance R. Next, the preset classification model 202 is used to classify the sample working parameters in the first sample working parameter set to obtain parameter classification results. The first sample working parameter set may contain multiple sample working parameters of different types, such as battery voltage V, battery current I, battery temperature T, and battery internal resistance R, for example, multiple battery voltages V, multiple battery currents I, multiple battery temperatures T, and multiple battery internal resistances R. Based on this pre-defined classification model, the classification results of multiple sample working parameters of different types in the first sample working parameter set, such as battery voltage V, battery current I, battery temperature T, and battery internal resistance R, can be accurately obtained. The parameter classification results can include the classification results (labels) of each sample working parameter in the first sample working parameter set. Then, based on the parameter classification results, the first sample working parameter set is divided into multiple sample working parameter subsets. Each sample working parameter subset contains multiple sample working parameters of the same type, and the types of sample working parameters in different sample working parameter subsets are different. For example, it can be divided into four sample working parameter subsets: sample working parameter subset 1, sample working parameter subset 2, sample working parameter subset 3, and sample working parameter subset 4. Sample working parameter subset 1 contains multiple sample working parameters of the same type, such as multiple battery voltages V; sample working parameter subset 2 contains multiple sample working parameters of the same type, such as multiple battery currents; sample working parameter subset 3 contains multiple sample working parameters of the same type, such as multiple battery temperatures T; and sample working parameter subset 4 contains multiple sample working parameters of the same type, such as multiple battery internal resistances R. Finally, the power estimation model 201 is obtained by training the first machine learning model, such as a convolutional neural network, together with the sample operating parameters and corresponding remaining power values ​​from multiple sample operating parameter subsets. For example, the sample operating parameters and corresponding remaining power values ​​extracted from sample operating parameter subsets 1, 2, 3, and 4 are used to train the convolutional neural network.

[0038] In this embodiment, when faced with a large number of different types of battery operating parameters as training sample data for model training, a pre-trained classification model is used to accurately classify the large number of different types of battery operating parameters into multiple subsets of sample operating parameters. Then, based on the sample operating parameters in the multiple subsets of sample operating parameters, a first machine learning model, such as a convolutional neural network, is trained to obtain a power estimation model. That is, by accurately classifying the large number of different types of battery operating parameters and then training the power estimation model specifically based on the classified operating parameters, the training effect of the power estimation model is improved, thereby increasing the accuracy of the trained power estimation model in predicting the remaining battery power.

[0039] In one embodiment, for example, during each iteration, one working parameter can be randomly selected from each of multiple sample working parameter subsets to update and train the first machine learning model, such as a convolutional neural network. Specifically, one working parameter can be randomly selected from each of sample working parameter subsets 1, 2, 3, and 4, i.e., four different types of working parameters, such as battery voltage V, battery current I, battery temperature T, and battery internal resistance R, to update and train the first machine learning model, such as a convolutional neural network. This can further improve the training effect of the power estimation model, thereby increasing the accuracy of the trained power estimation model in predicting the remaining battery power.

[0040] Based on any of the above embodiments, combined with Figure 3 As shown, in one embodiment, the training process of the preset classification model 202 may include the following steps: i) constructing a target loss function F for the second machine learning model, such as a convolutional neural network; ii) inputting second sample working parameters from the second sample working parameter set, such as battery voltage V, battery current I, battery temperature T, or battery internal resistance R, and the corresponding label value d, into the second machine learning model to output the corresponding predicted label value D, and iteratively updating the second machine learning model, such as a convolutional neural network, based on the predicted label value D until the loss value of the target loss function F meets a preset condition, at which point the training ends. The preset classification model 202 is then obtained. The target loss function F includes a first loss function F1 and a second loss function F2. The first loss function F1 characterizes the difference between the label value d corresponding to the second sample working parameters and the predicted label value D output by the second machine learning model, and the second loss function F2 characterizes the probability of a class prediction error corresponding to the second sample working parameters input into the second machine learning model.

[0041] For example, the greater the difference between the label value d corresponding to the second sample working parameter and the predicted label value D output by the second machine learning model, the greater the loss value of the first loss function F1, and vice versa. The greater the probability of the category prediction error corresponding to the second sample working parameter input to the second machine learning model, the greater the loss value of the second loss function F2, and vice versa. In this embodiment, a specific target loss function F is constructed through the first loss function F1 and the second loss function F2, which can reflect the degree of influence of the training process of the preset classification model on the accuracy of the working parameter classification result from two different aspects. The second machine learning model, such as a convolutional neural network, is updated and trained in this way until the loss value of the target loss function F meets the preset condition, at which point the training ends. The preset classification model 202 obtained at this time can more accurately classify a large number of different types of working parameters of the battery. Then, based on the classified working parameters, the power estimation model is trained in a targeted manner, which can further improve the training effect of the power estimation model and further improve the accuracy of the power estimation model 201 in predicting the remaining power of the battery.

[0042] For example, in one embodiment, the loss value of the target loss function F may satisfy the preset conditions, including but not limited to the following conditions: the loss value Y1 of the first loss function F1 is less than a first threshold, i.e., the difference between the predicted label value and the actual label value is small, and the loss value Y2 of the second loss function F2 is less than a second threshold, i.e., the probability of category prediction error is small. The first and second thresholds can be set as needed and are not specifically limited. In this embodiment, training ends when the loss value Y1 of the first loss function F1 is less than the first threshold (i.e., the difference between the predicted label value and the actual label value is small), and the loss value Y2 of the second loss function F2 is less than the second threshold (i.e., the probability of category prediction error is small). The preset classification model 202 obtained at this time can more accurately classify a large number of different types of battery operating parameters. Then, based on the classified operating parameters, the power estimation model 201 is trained specifically, which can further improve the training effect of the power estimation model and further improve the accuracy of the trained power estimation model in predicting the remaining battery power.

[0043] To further improve the accuracy of the power estimation model 201 obtained from the training in predicting the remaining battery power, in one embodiment, based on the above embodiments, the training process of the preset classification model 202 may further include: determining the weight value corresponding to the second sample working parameter in the second sample working parameter set, and then iteratively updating the second machine learning model based on the predicted label value obtained in step ii) above and the weight value until the loss value of the target loss function meets the preset condition and the training ends. For example, in one embodiment, the weight value corresponding to the second sample working parameter is related to the type of the second sample working parameter. Different types of second sample working parameters have different weight values. For example, the battery internal resistance R has a different degree of influence on the estimated value of the remaining battery capacity compared with other working parameters such as battery voltage V, battery current I, and battery temperature T. Therefore, the weight values ​​corresponding to different types of second sample working parameters can be determined in advance through experiments. Then, in the iterative training process of the preset classification model 202, when a second sample working parameter is taken from the set of second sample working parameters each time, its corresponding weight value Q is obtained. Then, the taken second sample working parameter and its corresponding label value d are input into the second machine learning model to output the corresponding predicted label value D. Then, based on the predicted label value D and the weight value Q, the second machine learning model, such as a convolutional neural network, is iteratively updated until the loss value of the target loss function F meets the above preset conditions and the training ends. By adding the weight values ​​corresponding to the second sample working parameters during training, the output of the model can be adjusted, enabling the trained preset classification model 202 to more accurately classify a large number of different types of working parameters of the battery. Then, based on the classified working parameters, the power estimation model 201 is trained in a targeted manner, which can further improve the accuracy of the trained power estimation model in predicting the remaining power of the battery.

[0044] It should be noted that although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that these steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps. Furthermore, it is readily understood that these steps may be executed synchronously or asynchronously, for example, in multiple modules / processes / threads.

[0045] like Figure 4 As shown, this disclosure provides a battery power estimation system based on machine learning, including:

[0046] The parameter acquisition module 401 is used to acquire battery operating parameters, which include multiple operating parameters of different types.

[0047] The power estimation module 402 is used to input the battery operating parameters into a power estimation model to obtain an estimated remaining battery power value. The power estimation model is pre-trained on a first machine learning model based on multiple subsets of sample operating parameters. Each subset of sample operating parameters contains multiple sample operating parameters of the same type for the battery, and the types of sample operating parameters in different subsets are different. The multiple subsets of sample operating parameters are obtained by classifying the pre-acquired first set of sample operating parameters of the battery using parameter classification results. The parameter classification results are obtained by processing the first set of sample operating parameters using a preset classification model. The preset classification model is trained on a second machine learning model based on a pre-acquired second set of sample operating parameters of the battery.

[0048] In one embodiment, the battery operating parameters of different types include any two or more of the following: battery voltage, battery current, battery temperature, and battery internal resistance.

[0049] In one embodiment, the training process of the preset classification model includes: constructing a target loss function for the second machine learning model; inputting the second sample working parameters and their corresponding label values ​​from the second sample working parameter set into the second machine learning model to output corresponding predicted label values; iteratively updating the second machine learning model based on the predicted label values ​​until the loss value of the target loss function meets a preset condition, at which point training ends; wherein the target loss function includes a first loss function and a second loss function, the first loss function being used to characterize the difference between the label values ​​corresponding to the second sample working parameters and the predicted label values ​​output by the second machine learning model, and the second loss function being used to characterize the probability of a class prediction error corresponding to the second sample working parameters input into the second machine learning model.

[0050] In one embodiment, the loss value of the target loss function satisfies the following preset conditions: the loss value of the first loss function is less than a first threshold, and the loss value of the second loss function is less than a second threshold.

[0051] In one embodiment, the training process of the preset classification model further includes: determining the weight values ​​corresponding to the second sample working parameters in the second sample working parameter set, and iteratively updating the second machine learning model based on the predicted label values ​​and weight values ​​until the loss value of the target loss function meets the preset condition and the training ends.

[0052] In one embodiment, the weight value corresponding to the second sample working parameter is related to the type of the second sample working parameter, and different weight values ​​correspond to different types of second sample working parameters.

[0053] In one embodiment, the second sample operating parameters include battery voltage, battery current, battery temperature, and battery internal resistance.

[0054] Regarding the system in the above embodiments, the specific methods by which each module performs operations and the corresponding technical effects have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0055] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units. Components shown as modules or units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the disclosed solution according to actual needs. Those skilled in the art can understand and implement this without any inventive effort.

[0056] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the machine learning-based battery power estimation method described in any of the preceding embodiments.

[0057] For example, the readable storage medium may be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0058] The computer-readable storage medium may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, capable of transmitting, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0059] This disclosure also provides an electronic device including a processor and a memory for storing a computer program. The processor is configured to execute the machine learning-based battery power estimation method from any of the above embodiments by executing the computer program.

[0060] The following reference Figure 5 To describe an electronic device 600 according to this embodiment of the present invention. Figure 5 The electronic device 600 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0061] like Figure 5 As shown, the electronic device 600 is presented in the form of a general-purpose computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different system components (including storage unit 620 and processing unit 610), a display unit 640, etc.

[0062] The storage unit stores program code that can be executed by the processing unit 610, causing the processing unit 610 to perform the steps described in the above-described method embodiment section of this specification according to various exemplary embodiments of the present invention. For example, the processing unit 610 can perform, as follows: Figure 1 The steps of the method shown are as follows.

[0063] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache storage unit 6202, and may further include a read-only memory unit (ROM) 6203.

[0064] The storage unit 620 may also include a program / utility 6204 having a set (at least one) program module 6205, such program module 6205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0065] Bus 630 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0066] Electronic device 600 can also communicate with one or more external devices 700 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 600, and / or with any device that enables electronic device 600 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 650. Furthermore, electronic device 600 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 660. Network adapter 660 can communicate with other modules of electronic device 600 via bus 630. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0067] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, or network device, etc.) to execute the steps of the machine learning-based battery power estimation method according to the above embodiments of this disclosure.

[0068] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0069] The above description is merely a specific embodiment of this disclosure, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A battery power estimation method based on machine learning, characterized in that, include: Obtain battery operating parameters, which include multiple operating parameters of different types; The battery operating parameters are input into a power estimation model to obtain an estimated remaining battery power value. The power estimation model is pre-trained on a first machine learning model based on multiple subsets of sample operating parameters. Each subset contains multiple sample operating parameters of the same type for the battery, and the types of sample operating parameters in different subsets are different. The multiple subsets of sample operating parameters are obtained by classifying the pre-acquired first set of sample operating parameters of the battery using parameter classification results. The parameter classification results are obtained by processing the first set of sample operating parameters using a preset classification model. The preset classification model is trained on a second machine learning model based on a pre-acquired second set of sample operating parameters of the battery. The training process of the preset classification model includes: constructing a target loss function for the second machine learning model; inputting the second sample working parameters and their corresponding label values ​​from the second sample working parameter set into the second machine learning model to output corresponding predicted label values; iteratively updating the second machine learning model based on the predicted label values ​​until the loss value of the target loss function meets a preset condition, at which point training ends; wherein, the target loss function includes a first loss function and a second loss function, the first loss function being used to characterize the difference between the label values ​​corresponding to the second sample working parameters and the predicted label values ​​output by the second machine learning model, and the second loss function being used to characterize the probability of a class prediction error corresponding to the second sample working parameters input into the second machine learning model; The loss value of the target loss function satisfies the following preset conditions: the loss value of the first loss function is less than a first threshold, and the loss value of the second loss function is less than a second threshold; The training process of the preset classification model further includes: determining the weight values ​​corresponding to the second sample working parameters in the second sample working parameter set, and iteratively updating the second machine learning model based on the predicted label value and the weight value until the loss value of the target loss function meets the preset condition and the training ends. The weight values ​​corresponding to the second sample working parameters are related to the type of the second sample working parameters. Different types of second sample working parameters have different weight values. Specifically, the weight values ​​corresponding to different types of second sample working parameters are determined through pre-experimentation. Then, during the iterative training of the preset classification model, when a second sample working parameter is taken from the set of second sample working parameters each time, its corresponding weight value is obtained. The taken second sample working parameter and its corresponding label value are then input into the second machine learning model to output the corresponding predicted label value. The second machine learning model is then iteratively updated based on the predicted label value and the weight value until the loss value of the target loss function meets the above preset conditions, at which point the training ends.

2. The method according to claim 1, characterized in that, The battery operating parameters include multiple operating parameters of different types, including any two or more of the following: battery voltage, battery current, battery temperature, and battery internal resistance.

3. The method according to claim 1, characterized in that, The second sample's operating parameters include battery voltage, battery current, battery temperature, and battery internal resistance.

4. A battery power estimation system based on machine learning, characterized in that, include: The parameter acquisition module is used to acquire battery operating parameters, which include multiple operating parameters of different types. A power estimation module is used to input the battery operating parameters into a power estimation model to obtain an estimated remaining battery power value. The power estimation model is pre-trained on a first machine learning model based on multiple subsets of sample operating parameters. Each subset of sample operating parameters contains multiple sample operating parameters of the same type for the battery, and the types of sample operating parameters in different subsets are different. The multiple subsets of sample operating parameters are obtained by classifying the pre-acquired first set of sample operating parameters of the battery using parameter classification results. The parameter classification results are obtained by processing the first set of sample operating parameters using a preset classification model. The preset classification model is trained on a second machine learning model based on a pre-acquired second set of sample operating parameters of the battery. The training process of the preset classification model includes: constructing a target loss function for the second machine learning model; inputting the second sample working parameters and their corresponding label values ​​from the second sample working parameter set into the second machine learning model to output corresponding predicted label values; iteratively updating the second machine learning model based on the predicted label values ​​until the loss value of the target loss function meets a preset condition, at which point training ends; wherein, the target loss function includes a first loss function and a second loss function, the first loss function being used to characterize the difference between the label values ​​corresponding to the second sample working parameters and the predicted label values ​​output by the second machine learning model, and the second loss function being used to characterize the probability of a class prediction error corresponding to the second sample working parameters input into the second machine learning model; The loss value of the target loss function satisfies the following preset conditions: the loss value of the first loss function is less than a first threshold, and the loss value of the second loss function is less than a second threshold; The training process of the preset classification model further includes: determining the weight values ​​corresponding to the second sample working parameters in the second sample working parameter set, and iteratively updating the second machine learning model based on the predicted label value and the weight value until the loss value of the target loss function meets the preset condition and the training ends. The weight value corresponding to the second sample working parameter is related to the type of the second sample working parameter; different types of second sample working parameters correspond to different weight values.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the battery power estimation method based on machine learning as described in any one of claims 1 to 3.

6. An electronic device, characterized in that, include: processor; as well as Memory, used to store computer programs; The processor is configured to execute the machine learning-based battery power estimation method according to any one of claims 1 to 3 by executing the computer program.

Images