Method for estimating SOH with limited data when battery is discharged

Abstract

The present invention relates to a method for estimating an SOH with limited data when a battery is discharged, and to a method for estimating the SOH of a battery by estimating, with only the driving information of a vehicle, a current value supplied from the battery of the electric vehicle to the electric vehicle, even if direct data for estimating the SOH is not provided from the electric vehicle.

Description

How to estimate the state of health (SOH) of a battery during discharge with limited data

[0001] The present invention relates to a method for estimating the SOH of a battery when it is discharged with limited data, and more particularly, to a method for estimating the SOH of a battery by estimating the current value supplied from the battery of the electric vehicle to the electric vehicle using only the vehicle's driving information even when direct data for estimating the SOH from the electric vehicle is not provided.

[0002] An electric vehicle is a vehicle that uses electricity as a power source, unlike an internal combustion engine vehicle that uses chemical energy as a power source. It is equipped with a battery and supplies electricity from the battery to the motor to obtain the kinetic energy required for operation.

[0003] Batteries typically installed in electric vehicles are designed to operate over long periods of time through repeated charging and discharging. Therefore, ensuring safe and long-term battery operation is paramount. Furthermore, efficient battery management is crucial for long-term, safe battery operation.

[0004] Ultimately, in order to improve the safety of electric vehicles, it is very important to estimate and provide information (SOH) indicating how much the current performance of the battery has deteriorated compared to the initial performance of the battery during operation.

[0005] SOH (state of health), which is information indicating the lifespan or condition of a battery, is a performance index that compares the ideal condition of the battery (i.e., the initial condition) with the current condition of the battery, and is used as an indicator of how good the performance of the battery is.

[0006] Electric vehicles typically don't provide SOH, and even if they do, the method of calculating SOH varies from vehicle to vehicle. Furthermore, estimating SOH requires sensing data, such as current, voltage, charge, and temperature, directly measured from the electric vehicle's battery.

[0007] More specifically, in order to estimate the SOH during vehicle operation, the amount of charge provided from the battery to the electric vehicle during operation (passed charge) must be calculated, and to calculate the amount of charge, a current (discharge current) value is required.

[0008] However, in the case of general electric vehicles, rather than providing battery sensing data, they only provide limited data including driving speed, external and internal temperature, and whether the heating and cooling is set (on or off), and do not provide direct data that can be used to estimate SOH.

[0009] Here, the electric vehicle's operating speed is related to the electric motor's power, which in turn is correlated with the battery's current (discharge current). Additionally, the battery is also related to temperature (inside and outside the electric vehicle) and the operation of heating and cooling systems. Therefore, the interior and exterior temperatures, as well as heating and cooling settings, are also correlated with the battery's current.

[0010] Accordingly, the present invention proposes a method of estimating a current value using driving information of an electric vehicle including speed, external temperature, internal temperature, heating and cooling settings, or a combination thereof, and estimating the amount of charge transferred from a battery to an electric vehicle in a changed SOC section (i.e., a discharge cycle) according to the driving of the vehicle through the estimated current value, thereby estimating the SOH of the electric vehicle.

[0011] That is, the present invention aims to objectively and accurately estimate the SOH of a battery using limited data (information) provided by an electric vehicle.

[0012] Next, we will briefly explain the prior art existing in the technical field of the present invention, and then describe the technical details that the present invention seeks to achieve differently from the prior art.

[0013] First, Korean Patent Publication No. 2023-0101789 (July 6, 2023) relates to a battery SOH estimation system, a parameter extraction system and method therefor, and a battery SOH estimation system that receives and stores voltage and current values ​​of a battery measured at preset cycles, calculates a response function in the frequency domain for the stored voltage and current values, and a change rate of the voltage value and charge capacity of the battery to estimate the state of health (SOH) of the battery, and a parameter extraction system and method therefor.

[0014] That is, Korean Patent Publication No. 2023-0101789 receives voltage and current values, which are direct data for calculating SOH, when charging a battery.

[0015] On the other hand, the present invention estimates SOH when the battery is discharged, and estimates SOH only with driving information of the electric vehicle even if direct data for estimating SOH is not received, so the present invention and Korean Patent Publication No. 2023-0101789 have significant differences in their technical configuration, purpose, and effect.

[0016] In addition, Korean Patent No. 1172183 (August 1, 2012) relates to a device and method for estimating the SOH of a vehicle battery, which stores the battery temperature measured at regular time intervals as temperature distribution data over time, calculates the severity of temperature change if the temperature distribution data exceeds a reference period, and applies the calculated severity to a deterioration map for a driving distance stored in advance to calculate deterioration (SOH) according to the driving distance and temperature.

[0017] That is, Korean Patent No. 1172183 estimates the SOH using the temperature change of the battery and the total driving distance of the vehicle, and does not estimate the current value supplied from the battery to the electric vehicle according to the driving information of the electric vehicle proposed in the present invention, and does not describe at all a method for estimating the SOH of the battery by estimating the amount of charge supplied from the electric vehicle to the electric vehicle using the estimated current value. Therefore, the two inventions have significant differences in their technical configurations and effects.

[0018] The present invention was created to solve the above problems, and its purpose is to provide a method for estimating the SOH of a battery discharged during operation of an electric vehicle using limited data for estimating the SOH of the battery discharged during operation of an electric vehicle using the operation information of the electric vehicle even when direct data for estimating the SOH is not provided.

[0019] In addition, the present invention aims to provide a method for estimating the SOH of a battery by estimating a current value from driving information of an electric vehicle and estimating the amount of charge provided from the battery to the electric vehicle.

[0020] In addition, the present invention provides a method for generating a DNN model for current estimation by training a DNN (deep neural network) using each learning data generated by labeling each current value according to driving information of an electric vehicle, and estimating the current value through the current estimation DNN model when actual driving information of an electric vehicle is received.

[0021] In addition, the present invention aims to provide a method for estimating SOH by estimating the amount of charge provided to an electric vehicle using the estimated current value and calculating the ratio of the previously estimated amount of charge to the estimated amount of charge.

[0022] A method for estimating SOH of a battery when it is discharged with limited data according to one embodiment of the present invention includes a step of receiving driving information of an electric vehicle, a current value estimation step of estimating a current value of a battery using the received driving information, a charge amount estimation step of estimating an amount of charge provided from the battery to the electric vehicle using the estimated current value, and an SOH estimation step of estimating the SOH of the battery using the estimated charge amount, and is characterized in that the SOH is estimated through the received driving information without direct data for estimating the SOH.

[0023] In addition, the driving information is characterized in that it includes the speed of the electric vehicle, the internal temperature, the external temperature, the heating and cooling settings, or a combination thereof.

[0024] In addition, the current value estimation step is characterized by inputting the received driving information into a current estimation DNN model generated by learning learning data in which current values ​​are labeled according to driving information of the electric vehicle, thereby estimating the current value.

[0025] In addition, the charge estimation step is characterized by estimating the charge amount provided from the battery to the electric vehicle by using the current value estimated during the discharge cycle of the battery according to the operation of the electric vehicle.

[0026] In addition, the SOH estimation step estimates the SOH by calculating the ratio of the previously estimated charge amount to the estimated charge amount, and the previously estimated charge amount is characterized in that it is the initially estimated charge amount or the charge amount estimated in the previous discharge cycle.

[0027] In addition, the charge estimation step is characterized in that the charge amount is estimated for each section by dividing the speed into a speed increase section in which the speed increases according to the rate of change in the speed of the received driving information, a speed decrease section in which the speed decreases, and a constant speed section in which the absolute value of the rate of change in the speed is less than a predetermined threshold value.

[0028] In addition, the method for estimating the SOH further includes a constant-speed driving section detection step of detecting a constant-speed section in which the ratio of the charge amount estimated at the end of the constant-speed section to the charge amount estimated at the start of the constant-speed section among the classified constant-speed sections exceeds a predetermined charge amount change rate as a constant-speed driving section, and a driving distance calculation step of calculating a driving distance according to the speed of the electric vehicle in the detected constant-speed driving section, and the SOH estimation step is characterized in that the SOH is estimated by extracting a driving distance of the new vehicle corresponding to the speed of the electric vehicle by referring to a mapping table that maps the driving distance according to the speed when the electric vehicle is a new vehicle in the constant-speed driving section, and calculating a ratio of the calculated driving distance of the electric vehicle to the extracted driving distance of the new vehicle.

[0029] In addition, the mapping table is characterized in that it is configured according to the charge change rate, the internal temperature of the electric vehicle, the external temperature, or a combination thereof.

[0030] In addition, the SOH estimation step is characterized by further including extracting the driving distance of the new vehicle corresponding to the speed of the electric vehicle in the detected constant-speed driving section by referring to a mapping table corresponding to the change rate of the charge amount in the detected constant-speed driving section, the internal temperature, the external temperature, or a combination thereof, and calculating the ratio of the calculated driving distance of the electric vehicle to the extracted driving distance of the new vehicle, thereby estimating the SOH.

[0031] In addition, a device for estimating SOH of a battery during discharge with limited data according to one embodiment of the present invention is characterized in that it estimates SOH of a battery according to a method for estimating SOH of a battery with limited data.

[0032] As described above, the present invention has the effect of accurately estimating the SOH of a battery by estimating the SOH of the battery through the operating information of the electric vehicle when the electric vehicle is driven to discharge the battery mounted on the electric vehicle, even when direct data for estimating the SOH is not provided from the electric vehicle.

[0033] FIG. 1 is a diagram illustrating the difficulty in predicting SOH when discharging a battery of an electric vehicle according to one embodiment of the present invention.

[0034] FIG. 2 is a diagram illustrating a method for generating a DNN model for current estimation according to one embodiment of the present invention.

[0035] FIG. 3 is a diagram illustrating a method for estimating SOH using driving information according to one embodiment of the present invention.

[0036] FIG. 4 is a diagram illustrating a method for estimating SOH according to another embodiment of the present invention.

[0037] FIG. 5 is a diagram illustrating a mapping table of a new vehicle according to one embodiment of the present invention.

[0038] Fig. 6 is a block diagram showing the configuration of a DNN model generation device for current estimation according to one embodiment of the present invention.

[0039] FIG. 7 is a block diagram showing the configuration of a device for estimating SOH during battery discharge with limited data according to one embodiment of the present invention.

[0040] Figure 8 is a flowchart illustrating a procedure for generating a DNN model for current estimation according to one embodiment of the present invention.

[0041] FIG. 9 is a flowchart illustrating a procedure for estimating SOH during discharge of a battery with limited data according to one embodiment of the present invention.

[0042] FIG. 10 is a flowchart illustrating a procedure for estimating SOH during discharge of a battery with limited data according to another embodiment of the present invention.

[0043] [Description of symbols] 100: A device for estimating SOH when a battery is discharged with limited data; 110, 210: A driving information receiving unit; 120: A current value estimation unit; 130: A charge estimation unit; 140: A constant-speed driving section detection unit; 150: A driving distance calculation unit; 160: A SOH estimation unit; 200: A device for generating a DNN model for current estimation; 220: A learning data generation unit; 221: A preprocessing unit; 222: A labeling unit; 230: A learning unit; 300: A database.

[0044] Hereinafter, with reference to the attached drawings, a preferred embodiment of a method for estimating the SOH of a battery during discharge using limited data of the present invention will be described in detail. The same reference numerals in each drawing represent the same elements. In addition, specific structural and functional descriptions of embodiments of the present invention are merely illustrative for the purpose of explaining embodiments according to the present invention, and unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those of ordinary skill in the art to which the present invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning they have in the context of the related art, and are preferably not interpreted in an ideal or excessively formal sense unless explicitly defined herein.

[0045] FIG. 1 is a diagram illustrating the difficulty in predicting SOH when discharging a battery of an electric vehicle according to one embodiment of the present invention.

[0046] Typically, when estimating the SOH of a battery during discharge, it is done by calculating the ratio of charge.

[0047] However, as illustrated in Fig. 1, in order to calculate the amount of charge (passed charge) (i.e., changed SOC, delta SOC) provided from the battery to the electric vehicle during the discharge cycle (i.e., any SOC change section) of the battery according to the operation of the electric vehicle, the current value (I) during the SOC change section is required.

[0048] However, electric vehicles cannot calculate the charge amount because they provide only limited data, including driving information such as speed, inside and outside temperature, heating and cooling settings, or a combination of these.

[0049] Therefore, a method is needed to estimate the charge amount by estimating the current (i.e., discharge current) during discharge of the battery.

[0050] Here, the amount of charge (Q) supplied from the battery to the electric vehicle is calculated by integrating I(current)dt (current integration method). That is, the current (I) is defined as the amount of unit charge (Q) with respect to time.

[0051] Accordingly, the present invention provides a method for estimating the SOH of a battery by generating a DNN model for current estimation using an artificial intelligence method and estimating the current when the battery is discharged using the generated DNN model for current estimation.

[0052] FIG. 2 is a diagram illustrating a method for generating a DNN model for current estimation according to one embodiment of the present invention.

[0053] As illustrated in FIG. 2, a DNN model for current estimation according to one embodiment of the present invention is generated through a DNN model generation device (200) for current estimation (hereinafter referred to as a DNN model generation device).

[0054] The above DNN model generation device (100) labels current values ​​in driving information including the speed, external temperature, internal temperature, heating / cooling settings, or a combination thereof of an electric vehicle to generate learning data, and trains a pre-prepared DNN (deep neural network) using each of the generated learning data to generate a DNN model for current estimation.

[0055] Meanwhile, most electric vehicles are configured to provide driving information including speed, inside and outside temperature, heating and cooling settings, or a combination of these.

[0056] At this time, speed is related to the power of the electric vehicle's motor, which in turn is related to the battery current (discharge current). Furthermore, the battery current can vary not only with the motor power but also with external and internal temperatures, heating and cooling settings. Therefore, the DNN model generation device (200) of the present invention is configured to train the DNN using training data constructed according to the correlation between the current and driving information.

[0057] The above DNN is configured to include an input layer including a plurality of input nodes, a hidden layer including a plurality of hidden nodes, and an output layer including an output node.

[0058] Each input node of the input layer and each hidden node of the hidden layer, and each hidden node of the hidden layer and each output node of the output layer are each connected through a link having a predetermined weight.

[0059] The above input layer inputs the speed, outside temperature, outside temperature, whether heating or cooling is set, or a combination thereof, which constitute the operating information of the learning data.

[0060] The above DNN model generation device (200) generates learning data by arranging each driving information according to the input of the DNN and labeling each current value.

[0061] For example, if the speed is configured to be input to the first input node of the DNN, the speed in the training data is placed to be input to the first input node.

[0062] The output node of the above output layer is configured to output a learning result (current value) according to the input, and since the DNN model generation device (200) already knows the result (current value) according to the input input to the input layer during the learning process, it updates (adjusts) the weight of the link to reduce the error between the learning result output during the learning process and the actual current value.

[0063] The above weights refer to learning parameters, and the above learning is performed by updating the weights through the back propagation method, which reduces the error by back propagating the error to the DNN.

[0064] At this point, the DNN, which has completed training by learning all the training data, becomes the current estimation DNN model. The input of this DNN model is the actual driving information of the electric vehicle, and the output is the battery current value.

[0065] The creation of the DNN model for the above current estimation is described in detail with reference to Fig. 7.

[0066] FIG. 3 is a diagram illustrating a method for estimating SOH using driving information according to one embodiment of the present invention.

[0067] As illustrated in FIG. 3, the SOH estimation device (100) according to one embodiment of the present invention receives driving information of an electric vehicle from a driving information providing device.

[0068] The above driving information providing device is configured to be connected to the OBD (on board diagnostics) of an electric vehicle and receive driving information from the OBD and provide it to the SOH estimation device (100).

[0069] That is, the driving information providing device refers to a third party equipped with a communication function that provides driving information of an electric vehicle to an SOH estimation device (100). The driving information providing device may be installed in an electric vehicle.

[0070] Of course, the driving information providing device can be configured to receive driving information from the OBD and provide it to the SOH estimation device (100), and thus can be configured in various forms, such as a user terminal including a smartphone.

[0071] Ultimately, the SOH estimation device (100) receives driving information of the electric vehicle from the electric vehicle through the driving information providing device.

[0072] The above SOH estimation device (100) estimates a current (i.e., discharge current) value using the received operating information. The current is estimated for an SOC section that changes depending on the operation.

[0073] The above SOH estimation device (100) preprocesses the driving information into a format suitable for the DNN model for current estimation.

[0074] In addition, the SOH estimation device (100) loads a current estimation DNN model from a database (300) and estimates the current value by inputting the operation information into the loaded current estimation DNN model.

[0075] The above SOH estimation device (100) estimates the amount of charge provided from the battery to the electric vehicle using the estimated current value.

[0076] The above current value is estimated until the operation of the electric vehicle ends (e.g., parking, turning off the engine), and the charge amount is calculated by integrating the estimated current value over time.

[0077] That is, the above time refers to the period of time during the discharge cycle of the battery according to the operation of the electric vehicle (i.e., the period until the operation of the electric vehicle ends (e.g., parking, turning off the engine)).

[0078] At this time, the amount of charge provided to the electric vehicle during the discharge cycle can be represented by the changed SOC (delta SOC). For example, if the amount of charge provided to the electric vehicle during the discharge cycle is 2% of the actual SOC, the changed SOC is 2%.

[0079] Additionally, the SOH estimation device (100) estimates the SOH by calculating the ratio (i.e., a / b) of the previously estimated charge amount (a) to the estimated charge amount (b).

[0080] At this time, it is desirable to estimate the SOH as the charge amount per unit time (e.g., 10 minutes), because the time (operating time) of the discharge cycle may be different.

[0081] The previously estimated charge amount is accumulated and stored in the database (300).

[0082] Here, the previously estimated charge amount may be the charge amount estimated in the previous discharge cycle or the charge amount estimated in the first discharge cycle.

[0083] FIG. 4 is a diagram illustrating a method for estimating SOH according to another embodiment of the present invention.

[0084] As illustrated in FIG. 4, the SOH estimation device (100) according to one embodiment of the present invention can estimate SOH using the speed and driving distance of the electric vehicle according to the driving information of the electric vehicle.

[0085] At this time, the SOH estimation device (100) does not estimate the amount of charge in an arbitrary discharge cycle (operation) as described and illustrated in FIG. 3, but divides the driving section into sections according to the change in speed included in the driving information of the electric vehicle in an arbitrary discharge cycle (SOC change section) according to the operation of the electric vehicle, and estimates the amount of charge provided from the battery to the electric vehicle for each driving section.

[0086] The above driving section is composed of a speed increase section, a speed decrease section, and a constant speed driving section.

[0087] The above constant speed section refers to a section where the speed hardly changes, and corresponds to a case where the absolute value of the speed change rate (|dV / dt|) is less than a predetermined threshold value.

[0088] In addition, the speed increase section refers to a section where the rate of change in speed exceeds 0 ((2) and (4) in Fig. 4), and the speed decrease section refers to a section where the rate of change in speed is less than 0 ((3) in Fig. 4). In this case, the constant speed section ((1) and (5) in Fig. 4) takes precedence over the speed increase section and the speed decrease section. In other words, even if the rate of change in speed exceeds 0 or decreases, it becomes a constant speed section if the conditions of the constant speed section are satisfied.

[0089] In addition, the SOH estimation device (100) detects a constant-speed driving section among constant-speed sections. The constant-speed driving section refers to a case where the changed SOC exceeds a predetermined charge change rate among constant-speed sections in which the absolute value (|dV / dt|) for the speed change rate is less than a predetermined threshold value.

[0090] That is, it refers to a case where the estimated charge at the end of the constant speed section exceeds a predetermined charge change rate compared to the estimated charge at the start of the constant speed section. In this case, the estimated charge refers to the accumulated value of the charge estimated for each previous section. For example, in the constant speed driving section shown in (5) of Fig. 4, the estimated charge at the start of the constant speed section refers to the accumulated charge sum from (1) to (4) of Fig. 4, and the estimated charge at the end of the constant speed section refers to the accumulated charge sum from (1) to (5) of Fig. 4.

[0091] For example, assuming that a predetermined threshold value is 5 and a predetermined charge change rate is 2%, the SOH estimation device (100) first classifies it as a constant speed section if the absolute value of the speed change rate included in the driving information is 4, and if the ratio of the charge amount estimated at the end of the classified constant speed section to the charge amount estimated at the start of the section exceeds the charge change rate of 2%, the corresponding dependent section is detected as a constant speed driving section.

[0092] Additionally, the SOH estimation device (100) calculates the driving distance of the electric vehicle in a constant-speed driving section. Since the driving distance is calculated using speed and time, the driving distance of the electric vehicle in the constant-speed driving section is calculated using the speed of the driving information and the time for the constant-speed driving section.

[0093] In addition, the SOH estimation device (100) stores and manages a mapping table that maps the driving distance according to the speed in the constant-speed driving section according to the charge change rate when the electric vehicle is new. Here, the speed in the constant-speed driving section can use the average speed.

[0094] At this time, the SOH estimation device (100) extracts the driving distance of the new vehicle corresponding to the speed of the electric vehicle in the detected constant-speed driving section by referring to the mapping table corresponding to the rate of change in charge (i.e., changed SOC) in the detected constant-speed driving section.

[0095] In addition, the SOH estimation device (100) estimates the SOH by calculating the ratio of the driving distance of the electric vehicle calculated above to the driving distance of the new vehicle extracted above.

[0096] This is based on the idea that the longer an electric vehicle is driven, the more the battery depletion accelerates when traveling the same distance at the same speed in a constant-speed driving section due to battery deterioration.

[0097] FIG. 5 is a diagram illustrating a mapping table of a new vehicle according to one embodiment of the present invention.

[0098] As illustrated in FIG. 5, a mapping table of a new vehicle according to one embodiment of the present invention is configured by mapping the speed and driving distance of the new vehicle in a constant-speed driving section by charge change rate (delta SOC).

[0099] In addition, assuming that the speed in the detected constant-speed driving section of the electric vehicle (current vehicle) is 150 km / h, the driving distance is 10.5 km, and the charge change rate is 2%, the SOH estimation device (100) loads a mapping table corresponding to the charge change rate of 2% and extracts the driving distance of the new vehicle corresponding to the speed of the electric vehicle.

[0100] In addition, the SOH estimation device (100) estimates the SOH of the battery by calculating the ratio of the driving distance calculated for the electric vehicle to the driving distance of the new vehicle extracted above.

[0101] Meanwhile, the mapping table can be configured based on the charge change rate, the internal temperature of the electric vehicle, the external temperature, or a combination thereof.

[0102] Therefore, the SOH estimation device (100) can estimate the SOH by referring to a mapping table according to the internal temperature, external temperature, and charge change rate of the electric vehicle included in the driving information.

[0103] The above internal temperature and external temperature can be set as a temperature range to configure a mapping table.

[0104] Fig. 6 is a block diagram showing the configuration of a DNN model generation device for current estimation according to one embodiment of the present invention.

[0105] As illustrated in FIG. 6, a DNN model generation device (200) according to one embodiment of the present invention is configured to generate a DNN model for current estimation, and includes a driving information receiving unit (210), a learning data generation unit (220), and a learning unit (230).

[0106] The above driving information receiving unit (210) receives driving information of an electric vehicle. Here, the driving information of the electric vehicle is a feature value collected for learning, and can be received in real time or non-real time from a database (300) or a driving information providing device equipped in the electric vehicle.

[0107] The above learning data configuration unit (220) generates learning data for training a current estimation DNN for estimating current values ​​using the received operating information, and is configured to include a preprocessing unit (221) and a labeling unit (222).

[0108] The above preprocessing unit (221) arranges (arranges) each data (speed, internal temperature, external temperature, whether heating or cooling is described) that constitutes the driving information according to the structure of the DNN, and performs a process of preparing feature values ​​by normalizing, scaling, interpolating, or extracting statistical data for each data.

[0109] The above labeling unit (222) finally generates learning data by labeling the actual current value for each operation information.

[0110] The above actual current value is added based on driving information and labeled accordingly. The above actual current value may be added experimentally or empirically based on driving information obtained from electric vehicles operating in various environments. The present invention does not impose any restrictions on the method of labeling the actual current value with driving information.

[0111] The above DNN can be composed of various artificial intelligence learning networks such as a deep convolutional neural network (DCNN), a transformer, a temporal convolutional neural network (TCNN), a convolutional neural network (CNN), a recurrent neural network (RNN), and a general regression neural network (GRNN).

[0112] The above learning unit (230) trains the DNN with learning data generated through the learning data generation unit (220) to generate a DNN model for current estimation.

[0113] The above learning is performed in a way that minimizes the error between the output of the DNN (learning result) and the label of the input learning data when learning data is input during the learning process.

[0114] At this time, the DNN model generation device (200) performs the learning through a backpropagation method that updates the weights of the DNN to minimize the error by backpropagating the error to the DNN.

[0115] The DNN that has completed the above learning becomes the current estimation DNN model, and the input of the current estimation DNN model becomes the actual driving information of the electric vehicle, and the output becomes the current value according to the input. Finally, the current estimation DNN model that has completed learning and is created is stored in a database (300), and in the future, the device (100) that estimates SOH loads the current estimation DNN model and uses it to estimate SOH.

[0116] FIG. 7 is a block diagram showing the configuration of a device for estimating SOH during battery discharge with limited data according to one embodiment of the present invention.

[0117] As illustrated in FIG. 7, an SOH estimation device (100) according to one embodiment of the present invention is configured to include a driving information receiving unit (110), a current value estimation unit (120), a charge estimation unit (130), a constant speed driving section detection unit (140), a driving distance calculation unit (150), and an SOH estimation unit (160).

[0118] The above driving information receiving unit (110) receives driving information of an electric vehicle.

[0119] The above driving information is received through a driving information provision device and includes the electric vehicle's speed, outside temperature, inside temperature, whether heating or cooling is set, or a combination thereof.

[0120] The above driving information providing device can transmit driving information according to the operation of the electric vehicle in real time or periodically to the SOH estimation device (100).

[0121] The above current value estimation unit (120) estimates the current value of the battery using the received operating information.

[0122] The above current value estimation unit (120) inputs the operation information into the current estimation DNN model and estimates the current value according to the output result of the current estimation DNN model.

[0123] At this time, the current value estimation unit (120) can preprocess the operation information into a format suitable for the current estimation DNN model and input it into the current estimation DNN model.

[0124] The above charge estimation unit (130) estimates the amount of charge provided from the battery to the electric vehicle using the estimated current value.

[0125] The above charge estimation unit (130) estimates the charge amount by integrating the current value estimated during the discharge cycle section of the battery according to the operation of the electric vehicle with the time (discharge time) during the discharge cycle. In other words, the charge estimation unit (130) estimates the charge amount provided from the battery to the electric vehicle during the discharge cycle section of the battery.

[0126] Additionally, the SOH estimation unit (160) estimates the SOH of the battery by comparing the estimated charge amount with a previously estimated charge amount. Estimating the SOH is performed by calculating the ratio of the previously estimated charge amount to the estimated charge amount.

[0127] Here, the previously estimated charge amount may be estimated from the previous discharge cycle or may be estimated from the first discharge cycle.

[0128] The above constant speed driving section detection unit (140) and driving distance calculation unit (150) are components that operate when estimating SOH using the speed and driving distance of the electric vehicle in the constant speed driving section.

[0129] Meanwhile, when SOH is estimated using the speed and driving distance in a dependent driving section, the charge estimation unit (130) is configured to estimate the charge amount for each section by dividing the driving section according to the operation of the electric vehicle into a speed increase section in which the speed increases according to the rate of change in speed of the received driving information, a speed decrease section in which the speed decreases, and a constant speed section when the absolute value of the change in speed is less than a predetermined threshold value.

[0130] The above constant speed driving section detection unit (140) detects, among the above-described constant speed sections, a constant speed section in which the ratio (a / b) of the charge amount (b) estimated at the end of the constant speed section to the charge amount (a) estimated at the start of the constant speed section exceeds a predetermined charge amount change rate as a constant speed driving section.

[0131] The above driving distance calculation unit (150) calculates the driving distance of the electric vehicle in the detected constant speed driving section.

[0132] The above SOH estimation unit (160) extracts the driving distance of the new vehicle corresponding to the speed of the electric vehicle in the detected constant-speed driving section from the mapping table of the new vehicle corresponding to the charge change rate, the internal temperature of the electric vehicle, the external temperature, or a combination thereof.

[0133] At this time, the SOH estimation unit (160) estimates the SOH of the battery by calculating the rate of change in the driving distance of the electric vehicle calculated above with respect to the driving distance of the extracted new vehicle.

[0134] Figure 8 is a flowchart illustrating a procedure for generating a DNN model for current estimation according to one embodiment of the present invention.

[0135] As illustrated in FIG. 8, the procedure for generating a DNN model for current estimation according to one embodiment of the present invention first performs a driving information receiving step in which the DNN model generating device (200) receives driving information of an electric vehicle (S110).

[0136] Here, the driving information is received for learning purposes, and the driving information is received according to the operation of the electric vehicle in various environments.

[0137] In other words, the driving information reception step is to collect driving information of the electric vehicle during the discharge cycle according to the operation of the electric vehicle in the life cycle of the battery and prepare characteristic values.

[0138] Next, the DNN model generation device (200) performs a learning data configuration step of configuring learning data using driving information.

[0139] The above learning data configuration step performs preprocessing such as arranging each data that constitutes the driving information (i.e., speed, internal temperature, external temperature, heating / cooling setting, or a combination thereof) according to the DNN, normalizing the data, scaling, interpolating, or generating statistical information and adding it as a feature value (S120).

[0140] That is, the operating information data input to each input node of the DNN is determined, and the placement step is to place the operating information data according to the input of the input node.

[0141] Next, the learning data configuration step performs a labeling step to label actual current values ​​in the operation information placed through the placement step (S130).

[0142] That is, learning data is generated by arranging each data item of driving information to fit the input of the DNN and labeling the current value according to the driving information. The learning data may be generated in advance and stored in a database (300).

[0143] Next, the DNN model generation device (200) performs a learning step of inputting the configured learning data into the DNN and training it (S140).

[0144] The above learning step performs learning on the DNN using all of the configured learning data, and when learning is completed (S150), the DNN becomes a DNN model for current estimation.

[0145] As described above, the above learning is performed through the backpropagation method.

[0146] FIG. 9 is a flowchart illustrating a procedure for estimating SOH during discharge of a battery with limited data according to one embodiment of the present invention.

[0147] As illustrated in FIG. 9, the procedure for estimating SOH when discharging a battery with limited data according to one embodiment of the present invention first performs a driving information receiving step in which the SOH estimation device (100) receives driving information of an electric vehicle (S210).

[0148] Here, the driving information is received during actual driving of the electric vehicle to estimate the SOH of the electric vehicle, and can be received in real time or periodically.

[0149] Next, the SOH estimation device (100) inputs the received operating information into a current estimation DNN model to perform a current value estimation step of estimating the current value (S220).

[0150] The above current value is estimated during the battery discharge cycle during operation of the electric vehicle. The discharge cycle ends when the electric vehicle is no longer in operation.

[0151] Next, the SOH estimation device (100) performs a charge estimation step of estimating the amount of charge supplied from the battery to the electric vehicle using the estimated current value (S230).

[0152] The above charge amount is estimated by integrating the estimated current value over time according to the discharge cycle.

[0153] Next, the SOH estimation device (100) performs an SOH estimation step of estimating the SOH of the battery using the estimated charge amount and the previously estimated charge amount (S240).

[0154] The above SOH estimation step can estimate SOH using the amount of charge per unit time.

[0155] FIG. 10 is a flowchart illustrating a procedure for estimating SOH during discharge of a battery with limited data according to another embodiment of the present invention.

[0156] As illustrated in FIG. 10, a procedure for estimating SOH when discharging a battery with limited data according to another embodiment of the present invention first performs a driving information receiving step (S310) in which an SOH estimation device (100) receives driving information of an electric vehicle, and then performs a current value estimation step in which the driving information is input into a current estimation DNN model to estimate a current value (S320).

[0157] The above driving information receiving step and current value estimation step are identical to the driving information receiving step and current value estimation step of Fig. 9.

[0158] Next, the SOH estimation device (100) performs a charge estimation step of estimating the amount of charge provided from the battery to the electric vehicle according to the driving section of the electric vehicle using the estimated current value (S330).

[0159] That is, the above charge estimation step estimates the charge by dividing the driving section of the electric vehicle into a speed increase section, a speed decrease section, a constant speed section, or a combination thereof according to the rate of change in speed of the driving information.

[0160] The division of the above driving sections has been explained with reference to FIGS. 4 and 5, so it will be omitted here.

[0161] Estimating the above charge amount is performed by integrating the estimated current value for each section.

[0162] Next, the SOH estimation device (100) performs a constant speed driving section detection step for detecting the constant speed driving section of the electric vehicle (S340).

[0163] The above constant speed driving section detection step is performed by detecting a constant speed section in which the change rate of the charge amount estimated at the end of the constant speed section compared to the charge amount estimated at the start of the constant speed section exceeds a predetermined charge amount change rate.

[0164] Next, the SOH estimation device (100) performs a driving distance calculation step for calculating the driving distance of the electric vehicle in the detected constant speed section (S350).

[0165] The above driving distance is calculated by multiplying the speed of the electric vehicle in the constant-speed driving section by the time of the constant-speed driving section.

[0166] Next, the SOH estimation device (100) performs an SOH estimation step of estimating the SOH using the calculated driving distance.

[0167] The above SOH estimation step first performs a driving distance extraction step of extracting a driving distance corresponding to the speed of the electric vehicle in the constant-speed driving section by referring to a mapping table of a new vehicle corresponding to the rate of change in the amount of charge in the detected constant-speed driving section (S360).

[0168] The mapping table of the above new vehicle maps the driving distance according to the speed of the new vehicle in a constant-speed driving section by the change rate of charge (i.e., the changed SOC), and may be composed of the change rate of charge, internal temperature, external temperature, or a combination thereof.

[0169] Next, the SOH estimation step finally estimates the SOH of the battery using the extracted driving distance and the calculated driving distance (S370).

[0170] The SOH of the above battery is estimated by calculating the ratio of the calculated driving distance to the extracted driving distance.

[0171] As described above, the present invention has the effect of accurately estimating SOH by estimating the current during discharge (discharge current) using the driving information of the electric vehicle even if direct data for estimating SOH is not provided.

[0172] In addition, although the preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and various modifications can be implemented by a person having ordinary skill in the art to which the invention pertains without departing from the gist of the present invention claimed in the claims, and such modifications should not be understood individually from the technical idea or prospect of the present invention.

[0173] As described above, the present invention has industrial applicability because it can accurately estimate the SOH of a battery even when direct data for estimating the SOH is not provided from the electric vehicle by estimating the SOH of the battery through the operating information of the electric vehicle when the electric vehicle is operated to discharge the battery mounted on the electric vehicle.

Claims

1. Step of receiving driving information of an electric vehicle; A current value estimation step for estimating the current value of the battery using the above-mentioned received driving information; A charge estimation step for estimating the amount of charge provided from the battery to the electric vehicle using the estimated current value; and It includes an SOH estimation step for estimating the SOH of the battery using the estimated charge amount; A method for estimating SOH during battery discharge with limited data, characterized in that the SOH is estimated through the received driving information without direct data for estimating the SOH.

2. In claim 1, The above operation information is, A method for estimating SOH during battery discharge with limited data, characterized in that it includes speed, internal temperature, external temperature, heating / cooling setting of the electric vehicle, or a combination thereof.

3. In claim 1, The above current value estimation step is, A method for estimating SOH during battery discharge using limited data, characterized in that the current value is estimated by inputting the received driving information into a current estimation DNN model created by learning learning data labeled with current values according to the driving information of the electric vehicle.

4. In claim 1, The above charge estimation step is, A method for estimating SOH during discharge of a battery with limited data, characterized in that the amount of charge provided from the battery to the electric vehicle is estimated using the current value estimated during the discharge cycle of the battery according to the operation of the electric vehicle.

5. In claim 1, The above SOH estimation step is, The SOH is estimated by calculating the ratio of the previously estimated charge to the estimated charge. A method for estimating the SOH of a battery during discharge with limited data, wherein the previously estimated charge amount is a first estimated charge amount or a charge amount estimated in a previous discharge cycle.

6. In claim 1, The above charge estimation step is, A method for estimating SOH during battery discharge using limited data, characterized in that the charge amount is estimated for each section by dividing the speed into a speed increase section in which the speed increases according to the rate of change in the speed of the received driving information, a speed decrease section in which the speed decreases, and a constant speed section in which the absolute value of the rate of change in the speed is less than a predetermined threshold value.

7. In claim 6, The method for estimating the above SOH is as follows: A constant speed driving section detection step for detecting a constant speed section in which the ratio of the charge amount estimated at the end of the constant speed section to the charge amount estimated at the start of the constant speed section among the above-mentioned constant speed sections exceeds a predetermined charge amount change rate as a constant speed driving section; and It further includes a driving distance calculation step for calculating the driving distance according to the speed of the electric vehicle in the above-detected constant speed driving section; A method for estimating SOH when a battery is discharged with limited data, characterized in that the SOH estimation step extracts the driving distance of a new vehicle according to the speed of the electric vehicle by referring to a mapping table that maps the driving distance according to the speed when the electric vehicle is a new vehicle in the constant-speed driving section, and calculates the ratio of the calculated driving distance of the electric vehicle to the extracted driving distance of the new vehicle, thereby estimating the SOH as the ratio of the calculated driving distance.

8. In claim 7, The above mapping table is, A method for estimating SOH during discharge of a battery with limited data characterized by being configured according to the above charge change rate, the internal temperature of the electric vehicle, the external temperature, or a combination thereof.

9. In claim 8, The above SOH estimation step is, Referring to the mapping table corresponding to the change rate of charge in the above-detected constant-speed driving section, the internal temperature, the external temperature, or a combination thereof, Extract the driving distance of the new vehicle corresponding to the speed of the electric vehicle in the above-detected constant-speed driving section, A method for estimating SOH when discharging a battery with limited data, characterized in that it further includes estimating the SOH by calculating the ratio of the driving distance of the electric vehicle calculated to the driving distance of the new vehicle extracted above.

10. A device for estimating the SOH of a battery during discharge using limited data, characterized in that the SOH of the battery is estimated according to the method for estimating the SOH of the battery using limited data as set forth in any one of claims 1 to 9.

Images