Data processing device and method of operating same

Abstract

A data processing device according to one embodiment disclosed in this document may include an acquisition unit that acquires experimental data of a battery and field data of a vehicle, a generation unit that generates synthetic data based on the experimental data and the field data, and a first learning unit that learns a de-noising model that removes noise from the synthetic data based on the experimental data and the synthetic data.

Description

[Technical Field]

[0001] This application claims the benefit of priority based on Korean Patent Application No. 10-2023-0051510, filed April 19, 2023, the entire contents of which are incorporated herein by reference.

[0002] FIELD OF THE INVENTION The embodiments disclosed herein relate to data processing apparatus and methods of operation. [Background technology]

[0003] Recently, research and development into secondary batteries has been actively pursued. Here, secondary batteries are batteries that can be charged and discharged, and include both conventional Ni / Cd batteries, Ni / MH batteries, and the latest lithium-ion batteries. Among secondary batteries, lithium-ion batteries have the advantage of having a much higher energy density than conventional Ni / Cd batteries, Ni / MH batteries, and other batteries. Furthermore, lithium-ion batteries can be manufactured to be compact and lightweight, making them popular as power sources for mobile devices. Recently, their use has expanded to include electric vehicles, drawing attention as a next-generation energy storage medium.

[0004] Generally, when diagnosing the condition of a secondary battery, experimental data obtained through battery charge / discharge experiments, etc. are used. In addition, a condition diagnosis model based on machine learning trained using such experimental data is used to diagnose the condition of a battery installed in an actual vehicle.

[0005] However, experimental data of a battery and field data of a battery installed in a vehicle have very different characteristics. For example, there are differences between the experimental data and the field data in terms of noise characteristics, cycle patterns, etc. Therefore, when a condition diagnosis model trained using experimental data is applied to the condition diagnosis of a battery installed in an actual vehicle, it is difficult to guarantee the reliability of the diagnosis results. To solve this problem, a new method for generating data that reflects the field data characteristics of a battery is needed. Summary of the Invention [Problem to be solved by the invention]

[0006] The embodiments disclosed herein may provide a data processing device and an operating method thereof that can generate synthetic data that reflects the field data characteristics of a battery actually placed in a vehicle in addition to experimental data of the battery.

[0007] The embodiments disclosed herein may provide a data processing apparatus and method of operation that may use a denoising model to remove (or reduce) noise from the synthetic data.

[0008] The embodiments disclosed herein may provide a data processing apparatus and a method for operating the same that may use output data of a denoising model and the composite data to evaluate the performance of the denoising model.

[0009] The technical problems of the embodiments disclosed in this document are not limited to the technical problems mentioned above, and other technical problems not mentioned can be clearly understood by those skilled in the art from the following description. [Means for solving the problem]

[0010] A data processing device according to one embodiment disclosed in this document may include an acquisition unit that acquires experimental data of a battery and field data of a vehicle, a generation unit that generates synthetic data based on the experimental data and the field data, and a first learning unit that learns a de-noising model that removes (or reduces) noise in the synthetic data based on the experimental data and the synthetic data.

[0011] In one embodiment of the data processing device disclosed in this document, the generation unit may include an extraction unit that extracts noise data from the field data, and a synthesis unit that synthesizes the experimental data and the noise data to generate the synthesized data.

[0012] A data processing device according to one embodiment disclosed in this document further includes a second learning unit that learns an extraction model that extracts the noise data reflecting the noise characteristics of the field data based on the experimental data and the field data, and the extraction unit can extract the noise data from the field data based on the extraction model.

[0013] A data processing device according to one embodiment disclosed in this document may further include a filtering unit that generates de-noised data by removing (or reducing) noise from the composite data using the denoising model.

[0014] A data processing device according to one embodiment disclosed in this document may further include a third learning unit that generates a first status diagnosis model by learning a status diagnosis model that diagnoses the status of the battery based on the noise-removed data.

[0015] In a data processing device according to one embodiment disclosed in this document, the third learning unit can generate a second condition diagnosis model by learning the condition diagnosis model based on the synthetic data.

[0016] A data processing device according to an embodiment disclosed herein may further include an evaluation unit that compares the first state diagnostic model with the second state diagnostic model to evaluate performance of the denoising model.

[0017] A data processing method according to one embodiment disclosed herein may include operations of acquiring battery experimental data and vehicle field data, generating synthetic data based on the experimental data and the field data, and training a de-noising model based on the experimental data and the synthetic data to remove (or reduce) noise in the synthetic data.

[0018] In one embodiment of the data processing method disclosed herein, the operation of generating the composite data may include an operation of extracting noise data from the field data, and an operation of combining the experimental data and the noise data to generate the composite data.

[0019] The data processing method according to one embodiment disclosed in this document further includes an operation of training an extraction model that extracts the noise data reflecting the noise characteristics of the field data based on the experimental data and the field data, and the operation of extracting the noise data from the field data may include an operation of extracting the noise data from the field data based on the extraction model.

[0020] The data processing method according to one embodiment disclosed herein may further include an operation of generating de-noised data from the synthetic data using the denoising model, in which noise has been removed (or reduced).

[0021] A data processing method according to one embodiment disclosed in this document may further include an operation of generating a first status diagnostic model by training a status diagnostic model that diagnoses the status of the battery based on the noise-removed data.

[0022] The data processing method according to one embodiment disclosed in this document may further include an operation of generating a second condition diagnostic model by training the condition diagnostic model based on the synthetic data.

[0023] The data processing method according to one embodiment disclosed herein may further include an operation of comparing the first state diagnostic model with the second state diagnostic model to evaluate performance of the denoising model. [Effects of the Invention]

[0024] According to the embodiments disclosed herein, by generating synthetic data that reflects the field data characteristics of the battery, the accuracy of the battery diagnostic results using the synthetic data can be improved.

[0025] According to the embodiments disclosed herein, removing (or reducing) noise in the synthetic data can improve the performance of a battery condition diagnostic model trained using the synthetic data.

[0026] In addition, various other effects may be provided that are directly or indirectly understood by this document. [Brief explanation of the drawings]

[0027] [Figure 1] 1 is a block diagram showing a configuration of a data processing device according to an embodiment; [Figure 2] 1 is a block diagram illustrating a configuration of a generating unit included in a data processing device according to an embodiment. [Figure 3] FIG. 2 is a diagram for explaining the operation of the configuration within the data processing device according to an embodiment. [Figure 4] 1 is a flowchart illustrating the operation of a data processing device according to an embodiment. [Figure 5] 1 is a flowchart illustrating the operation of a data processing device according to an embodiment. [Figure 6] 1 is a flowchart illustrating the operation of a data processing device according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0028] Various embodiments of the present invention will now be described with reference to the accompanying drawings, but it should be understood that this is not intended to limit the present invention to the particular embodiments, but rather to include various modifications, equivalents, and / or alternatives to the embodiments of the present invention.

[0029] The various embodiments and terms used herein in this document should not be understood to limit the technical features described in this document to a specific embodiment, but should be understood to include various modifications, equivalents, or alternatives of the embodiment. In connection with the description of the drawings, similar reference numerals may be used for similar or related components. The singular form of a noun corresponding to an item may include one or more of the item, unless the relevant context clearly dictates otherwise.

[0030] In this document, each of the phrases such as "A or B," "at least one of A and B," "at least one of A or B," "A, B or C," "at least one of A, B and C," and "at least one of A, B, or C" may include any one of the items listed together in the phrase, or all possible combinations thereof. Terms such as "first," "second," "primary," "second," "A," "B," "(a)," or "(b)" may be used simply to distinguish one element from another, and do not limit the element in other respects (e.g., importance or order) unless specifically stated to the contrary.

[0031] In this document, when a (e.g., first) component is referred to as being "coupled," "coupled," or "connected" to another (e.g., second) component, with or without the terms "functionally" or "communicatively," or when a reference is made to "coupled" or "connected," it means that the component may be coupled to the other component directly (e.g., by wire), wirelessly, or through a third component.

[0032] According to various embodiments, each of the above-described components (e.g., modules or programs) may include one or more entities, and some of the entities may be located separately in other components. According to various embodiments, one or more of the above-described components or operations may be omitted, or one or more other components or operations may be added. Generally, or in addition, multiple components (e.g., modules or programs) may be integrated into a single component. In such cases, the integrated component may perform one or more functions of each of the multiple components that are the same as or similar to those performed by the multiple components before the integration. According to various embodiments, operations performed by a module, program, or other component may be performed sequentially, in parallel, repeatedly, or heuristically, or one or more of the operations may be performed in a different order, omitted, or one or more other operations may be added.

[0033] The configuration of the data processing device will be described below with reference to FIGS.

[0034] Fig. 1 is a block diagram showing the configuration of a data processing device according to an embodiment, and Fig. 2 is a block diagram showing the configuration of a generating unit included in the data processing device according to an embodiment.

[0035] 1, data processing device 100 may include a communication circuit 110, a memory 120, and / or a processor 130. Depending on the embodiment, data processing device 100 illustrated in FIG. 1 may further include at least one component (e.g., a display, an input device, or an output device) other than the components illustrated in FIG.

[0036] According to one embodiment, the communication circuit 110 can establish a wired and / or wireless communication channel between the data processing device 100 and an external electronic device and / or an external server, and can transmit and receive data to and from the external electronic device and / or the external server via the established communication channel. According to one embodiment, the communication circuit 110 can receive battery experimental data and / or vehicle (e.g., an electric vehicle equipped with a secondary battery) field data from the external electronic device and / or the external server. Here, the experimental data can include data related to the status (voltage, current, temperature, internal resistance, SOC (State of Charge), and / or SOH (State of Health)) of a battery not installed in the vehicle. The field data can include data related to the status of a battery installed in the vehicle.

[0037] According to one embodiment, memory 120 may include volatile and / or non-volatile memory.

[0038] In one embodiment, memory 120 may store data used by at least one component (e.g., processor 130) of data-processing device 100. For example, the data may include software (or associated instructions), input data, or output data. In one embodiment, the instructions, when executed by processor 130, may cause data-processing device 100 to perform the operations defined by the instructions.

[0039] In one embodiment, memory 120 may include one or more pieces of software (e.g., an acquisition unit 121, a generation unit 123, a training unit 125, a filtering unit 127, and / or an evaluation unit 129).

[0040] Referring to FIG. 2, the generating unit 123 may include an extracting unit 210 and / or a synthesizing unit 220 .

[0041] Still referring to FIG. 1, processor 130 may include a central processing unit, an application processor, a graphics processing unit, a neural processing unit (NPU), an image signal processor, a sensor hub processor, or a communication processor.

[0042] In one embodiment, the processor 130 can execute software (e.g., the acquisition unit 121, the generation unit 123, the learning unit 125, the filtering unit 127, and / or the evaluation unit 129) to control at least one other component (e.g., a hardware or software component) of the data processing device 100 coupled to the processor 130, and can perform various data processing or calculations.

[0043] In the following, with reference to FIG. 3, a method for processing battery experimental data and vehicle field data by the data processing device 100 via the acquisition unit 121, the generation unit 123, the learning unit 125, the filtering unit 127, and / or the evaluation unit 129 will be described.

[0044] FIG. 3 is a diagram for explaining the operation of the configuration within the data processing device according to one embodiment.

[0045] 3, the acquisition unit 121 may acquire field data of the vehicle 310 and experimental data of the battery 320. Here, the vehicle 310 may be equipped with a battery of the same model as the battery 320, and the field data may include data related to the state of the battery equipped in the vehicle 310. According to one embodiment, the acquisition unit 121 may acquire the field data and experimental data from an external electronic device and / or an external server connected via a wired and / or wireless network using the communication circuit 110.

[0046] According to one embodiment, the generator 123 may generate synthetic data based on experimental data and field data.

[0047] According to an embodiment, the extraction unit 210 included in the generation unit 123 may extract noise data from the field data. According to an embodiment, the extraction unit 210 may extract noise data from the field data based on an extraction model learned by the second learning unit 125-2 or an extraction model stored in the memory 120. Here, the extraction model may be a learning model based on machine learning that extracts noise data that reflects noise characteristics of the field data.

[0048] According to one embodiment, the synthesis unit 220 included in the generation unit 123 may generate synthetic data by synthesizing the experimental data and the noise data. According to one embodiment, the generation unit 123 may generate synthetic data by further synthesizing random noise (e.g., Gaussian random noise) with the experimental data and the noise data.

[0049] In one embodiment, the trainer 125 can train various machine learning based learning models based on experimental, field, synthetic, and / or denoised data.

[0050] According to one embodiment, the training unit 125 may include a first training unit 125-1, a second training unit 125-2, and / or a third training unit 125-3.

[0051] According to one embodiment, the first learning unit 125-1 can learn a denoising model that removes (or reduces) noise in the synthetic data based on the experimental data and the synthetic data. The denoising learning model can be a machine learning-based learning model that is trained to compare the synthetic data with the experimental data and output denoised data from the synthetic data that is similar to the experimental data. According to one embodiment, the first learning unit 125-1 can store the learned denoising model in the memory 120.

[0052] According to one embodiment, the second learning unit 125-2 can train an extraction model that extracts noise data that reflects the noise characteristics of the field data based on the experimental data and the field data. The extraction model can be a machine learning-based learning model that is trained to compare the experimental data with the field data to understand the characteristics of the noise contained in the field data and extract noise data that reflects the characteristics from the field data. According to one embodiment, the second learning unit 125-2 can store the trained extraction model in the memory 120.

[0053] According to an embodiment, the filtering unit 127 may generate denoised data by removing (or reducing) noise from the composite data. According to an embodiment, the filtering unit 127 may generate denoised data by removing (or reducing) noise from the composite data using a denoising model trained by the first learning unit 125-1 or a denoising model stored in the memory 120.

[0054] According to an embodiment, the third learning unit 125-3 can learn a state diagnosis model for diagnosing the state of the battery based on the input data. The state diagnosis model can be a learning model based on machine learning that is trained to diagnose the state of the battery (e.g., internal resistance, charge capacity, SOC, SOH, etc.) based on the input data.

[0055] According to one embodiment, the third learning unit 125-3 can generate a first condition diagnosis model by training the condition diagnosis model based on the noise-removed data generated by the filtering unit 127. According to one embodiment, the third learning unit 125-3 can generate a second condition diagnosis model by training the condition diagnosis model based on the combined data generated by the combining unit 220. That is, the third learning unit 125-3 can distinguish between the condition diagnosis model trained based on the noise-removed data and the condition diagnosis model trained based on the combined data and generate different condition diagnosis models. The third learning unit 125-3 can store the trained first condition diagnosis model and / or second condition diagnosis model in the memory 120.

[0056] According to one embodiment, the evaluation unit 129 may evaluate the performance of the denoising model based on the first condition diagnostic model and the second condition diagnostic model trained by the third learning unit 125-3. The evaluation unit 129 may evaluate the performance of the denoising model by comparing the first condition diagnostic model trained based on data in which noise has been removed (or reduced) from synthetic data via the denoising model with the second condition diagnostic model trained based on synthetic data in which noise has not been removed (or reduced). Here, the performance of the denoising model may be related to how precisely noise has been removed (or reduced) from the synthetic data. For example, the evaluation unit 129 may evaluate the performance of the denoising model by considering the level of performance improvement of the first condition diagnostic model relative to the second condition diagnostic model. The evaluation unit 129 may use various methods related to learning model evaluation to compare the first condition diagnostic model and the second condition diagnostic model.

[0057] 4 is a flowchart of the operation of a data processing device according to an embodiment. Fig. 4 can be used to explain the operation of the data processing device 100 in Fig. 1, and can be explained using the configuration (e.g., data processing device 100) in Fig. 1.

[0058] The embodiment illustrated in FIG. 4 is one embodiment, and the order of steps according to various embodiments of the present invention may differ from that illustrated in FIG. 4, and some steps illustrated in FIG. 4 may be omitted, the order between steps may be changed, or steps may be merged.

[0059] 4, in operation 405, the data processing device 100 can acquire experimental data of a battery and field data of a vehicle (e.g., an electric vehicle equipped with a secondary battery). Here, the experimental data can include data related to the state of a battery not installed in a vehicle. The field data can include data related to the state of a battery installed in the vehicle.

[0060] In operation 410, the data processing device 100 may generate synthetic data based on the experimental data and the field data. Operation 410, in which the data processing device 100 generates synthetic data, may be described in more detail with reference to Figure 5 below.

[0061] In operation 415, a denoising model can be trained based on the experimental data and the synthetic data to remove (or reduce) noise in the synthetic data. The denoising trained model can be a machine learning based trained model that compares the synthetic data with the experimental data and is trained to output denoised data from the synthetic data that is similar to the experimental data.

[0062] Fig. 5 is a flowchart of the operation of a data processing device according to one embodiment. Fig. 5 can be used to explain the operation of the data processing device 100 in Fig. 1, and can be explained using the configuration (e.g., data processing device 100) in Fig. 1.

[0063] The embodiment illustrated in Fig. 5 is just one embodiment, and the order of steps according to various embodiments of the present invention may differ from that illustrated in Fig. 5, such that some steps illustrated in Fig. 5 may be omitted, the order between steps may be changed, or steps may be merged. For example, in Fig. 5, operation 505 may be omitted.

[0064] 5, in operation 505, the data processing device 100 can train an extraction model that extracts noise data that reflects the noise characteristics of the field data based on the experimental data and the field data. Here, the experimental data and the field data may be the data acquired in operation 405 of FIG. 4. The extraction model may be a learning model based on machine learning that is trained to compare the experimental data with the field data, grasp the characteristics of the noise included in the field data, and extract noise data that reflects the characteristics from the field data.

[0065] In operation 510, the data processing device 100 can extract noise data from the field data. According to one embodiment, the data processing device 100 can extract noise data from the field data based on the extraction model learned in operation 505 or the extraction model stored in the memory 120. Here, the extraction model can be a learning model based on machine learning that extracts noise data that reflects the noise characteristics of the field data.

[0066] In operation 515, the data processing device 100 may combine the experimental data and the noise data to generate the combined data. According to one embodiment, the data processing device 100 may further combine the experimental data and the noise data with random noise (e.g., Gaussian Random Noise) to generate the combined data.

[0067] Fig. 6 is a flowchart of the operation of a data processing device according to one embodiment. Fig. 6 can be used to explain the operation of the data processing device 100 in Fig. 1, and can be explained using the configuration (e.g., data processing device 100) in Fig. 1.

[0068] The embodiment illustrated in FIG. 6 is one embodiment, and the order of steps according to various embodiments of the present invention may differ from that illustrated in FIG. 6, and some steps illustrated in FIG. 6 may be omitted, the order between steps may be changed, or steps may be merged.

[0069] 6, in operation 605, the data processing apparatus 100 may generate denoised data in which noise has been removed (or reduced) from the synthetic data. Here, the synthetic data may be data generated in operation 410 of FIG. 4 or operation 515 of FIG. 5. According to one embodiment, the data processing apparatus 100 may generate denoised data in which noise has been removed (or reduced) from the synthetic data using the denoising model learned in operation 415 of FIG. 4 or the denoising model stored in the memory 120.

[0070] In operation 610, the data processing apparatus 100 can generate a first condition diagnostic model. According to one embodiment, the data processing apparatus 100 can generate the first condition diagnostic model by training a condition diagnostic model that diagnoses the condition of the battery based on the noise-removed data generated in operation 605. The condition diagnostic model can be a learning model based on machine learning that is trained to diagnose the condition of the battery (e.g., internal resistance, charge capacity, SOC, SOH, etc.) based on input data.

[0071] In operation 615, the data processing apparatus 100 may generate a second condition diagnostic model. According to one embodiment, the data processing apparatus 100 may generate the second condition diagnostic model by training a condition diagnostic model that diagnoses the condition of the battery based on the synthetic data generated in operation 410 of Figure 4 or operation 515 of Figure 5.

[0072] In operation 620, the data processing device 100 may evaluate the performance of the denoising model trained in operation 415 of FIG. 4 or the denoising model stored in the memory 120. According to one embodiment, the data processing device 100 may evaluate the performance of the denoising model based on the first condition diagnostic model trained and generated in operation 610 and the second condition diagnostic model trained and generated in operation 615. The data processing device 100 may evaluate the performance of the denoising model by comparing the first condition diagnostic model trained based on data from which noise has been removed (or reduced) through the denoising model from synthetic data with the second condition diagnostic model trained based on synthetic data from which noise has not been removed. Here, the performance of the denoising model may relate to how precisely noise has been removed from the synthetic data. For example, the data processing device 100 may evaluate the performance of the denoising model by considering the level of performance improvement of the first condition diagnostic model relative to the second condition diagnostic model. The data processing device 100 may use various methods related to learning model evaluation to compare the first condition diagnostic model and the second condition diagnostic model.

[0073] As used above, terms such as "comprise," "comprise," or "have," unless otherwise specified to the contrary, should be interpreted as meaning that the relevant element may be contained within the term, and as meaning that the term may further include other elements, rather than excluding other elements. All terms, including technical or scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the art to which the embodiments disclosed herein belong, unless otherwise defined. Commonly used terms, such as dictionary-defined terms, should be interpreted in a manner consistent with the context of the relevant art, and should not be interpreted in an idealized or overly formal sense unless expressly defined herein.

Claims

1. an acquisition unit that acquires battery experimental data and vehicle field data; a generation unit that generates synthetic data based on the experimental data and the field data; a first learning unit that learns a denoising model that removes noise from the synthetic data based on the experimental data and the synthetic data.

2. The generation unit an extraction unit that extracts noise data from the field data; 2. The data processing apparatus according to claim 1, further comprising a synthesis unit that synthesizes the experimental data and the noise data to generate the synthesized data.

3. a second learning unit configured to learn an extraction model that extracts the noise data reflecting noise characteristics of the field data based on the experimental data and the field data; The data processing device according to claim 2 , wherein the extraction unit extracts the noise data from the field data based on the extraction model.

4. The data processing device according to claim 1 , further comprising a filtering unit that generates denoised data by removing noise from the composite data using the denoising model.

5. The data processing device according to claim 4 , further comprising a third learning unit that generates a first state diagnostic model by learning a state diagnostic model that diagnoses the state of the battery based on the noise-removed data.

6. The data processing device according to claim 5 , wherein the third learning unit generates a second condition diagnosis model by learning the condition diagnosis model based on the composite data.

7. The data processing apparatus according to claim 6 , further comprising an evaluation unit that compares the first state diagnostic model with the second state diagnostic model to evaluate performance of the denoising model.

8. acquiring battery experimental data and vehicle field data; generating synthetic data based on the experimental data and the field data; and training a denoising model based on the experimental data and the synthetic data to remove noise from the synthetic data.

9. The operation of generating the composite data includes: extracting noise data from the field data; and combining the experimental data and the noise data to generate the combined data.

10. further comprising an operation of training an extraction model that extracts the noise data that reflects noise characteristics of the field data, based on the experimental data and the field data; The data processing method of claim 9 , wherein extracting noise data from the field data comprises extracting the noise data from the field data based on the extraction model.

11. The data processing method of claim 8 , further comprising the operation of generating denoised data by removing noise from the synthetic data using the denoising model.

12. The data processing method according to claim 11 , further comprising the operation of generating a first state diagnostic model by learning a state diagnostic model for diagnosing the state of the battery based on the noise-removed data.

13. The data processing method according to claim 12 , further comprising an operation of generating a second condition diagnostic model by training the condition diagnostic model based on the synthetic data.

14. The data processing method of claim 13 , further comprising the act of comparing the first condition diagnostic model and the second condition diagnostic model to evaluate performance of the denoising model.

Images