KR20220082234A - Method and apparatus for diagnosing battery status through soc estimation - Google Patents

Abstract

According to an embodiment, receiving a state-of-charge (SOC) measurement value and at least one battery measurement value for a battery device from a battery management system; calculating an SOC estimation value by inputting the SOC measurement value and the at least one battery measurement value into a neural network; calculating an SOC prediction range by inputting the SOC estimation value to a fuzzy predictor; and comparing the SOC measurement value with the SOC prediction range, and determining that the battery device is abnormal when the SOC measurement value is out of the SOC prediction range.

Description

Method and device for diagnosing battery status through SOC estimation

This embodiment relates to a technology for diagnosing the state of a battery.

An energy storage system (ESS) is being used to supply stable power to electric devices or to store power generated from a renewable generator.

An energy storage device is a device that stores electrical energy as another type of energy and then converts another type of energy into electrical energy when necessary, and a battery that converts electrical energy into chemical energy and stores it is representative.

Lithium-ion batteries are widely used in large-capacity energy storage devices for electric vehicles and renewable generators because of their high energy storage density and fast energy conversion speed.

However, since such a battery is used in a sealed state, there is a problem in that it is difficult to check the internal state. Therefore, in the battery, it is difficult to accurately measure the remaining amount of energy, and even if a problem occurs inside the battery, it is difficult to accurately detect the problem.

In addition to the forward reaction that converts electrical energy into chemical energy or chemical energy into electrical energy, a battery frequently undergoes side reactions in which the surrounding materials react. Therefore, it is not easy to figure out the problem.

Against this background, an object of the present embodiment, in one aspect, is to provide a technique for diagnosing an internal state of a battery. In another aspect, an object of the present embodiment is to provide a technique for diagnosing the internal state of a battery in a non-destructive way without using a destructive method of releasing the sealed state of the battery. In another aspect, an object of the present embodiment is to provide a technology for diagnosing an internal state of a battery using basic information obtainable through an existing battery management system.

In order to achieve the above object, an embodiment provides the steps of: receiving a state-of-charge (SOC) measurement value and at least one battery measurement value for a battery device from a battery management system; calculating an SOC estimation value by inputting the SOC measurement value and the at least one battery measurement value into a neural network; calculating an SOC prediction range by inputting the SOC estimation value to a fuzzy predictor; and comparing the SOC measurement value with the SOC prediction range, and determining that the battery device is abnormal when the SOC measurement value is out of the SOC prediction range.

The at least one battery measurement value may include a terminal voltage measurement value of the battery device.

The at least one battery measurement value may include a temperature measurement value of the battery device.

The method for diagnosing the battery state further includes calculating a root-mean-square (RMS) error between the SOC measurement value and the SOC estimate value before the step of comparing the SOC measurement value with the SOC prediction range, In the step of calculating the RMS error, when the RMS error exceeds a reference value, the battery device may be determined to be abnormal.

The neural network includes a plurality of hidden layers, and each hidden layer may be composed of a linear function.

Another embodiment may include: a receiver configured to receive a state-of-charge (SOC) measurement value and at least one battery measurement value for a battery device from a battery management system; a neural network for calculating an SOC estimation value using the SOC measurement value and the at least one battery measurement value; a fuzzy predictor for calculating an SOC predictive range using the SOC estimate; and a control unit that compares the SOC measurement value with the SOC prediction range, and determines that the battery device is abnormal when the SOC measurement value is out of the SOC prediction range.

The battery management system may generate the SOC measurement value using a method different from the neural network method.

The battery management system may generate the SOC measurement value using a current integration method.

The battery device may be mounted in an electric vehicle.

The control unit may calculate a root-mean-square (RMS) error between the SOC measurement value and the SOC estimate value, and when the RMS error exceeds a reference value, determine that the battery device is abnormal.

As described above, according to the present embodiment, the internal state of the battery can be diagnosed. And, according to the present embodiment, it is possible to diagnose the internal state of the battery in a non-destructive way without using a destructive method of releasing the sealed state of the battery. And, according to the present embodiment, it is possible to diagnose the internal state of the battery using basic information obtainable through the existing battery management system.

1 is a configuration diagram of a battery system according to an embodiment.
2 is a configuration diagram of an apparatus for diagnosing a battery state according to an exemplary embodiment.
3 is a block diagram of a neural network according to an embodiment.
4 is a flowchart of a method for diagnosing a battery state according to an exemplary embodiment.
5 is a diagram illustrating an example of an aperiodic battery state diagnosis method.

Hereinafter, some embodiments of the present invention will be described in detail with reference to exemplary drawings. In adding reference numerals to the components of each drawing, it should be noted that the same components are given the same reference numerals as much as possible even though they are indicated on different drawings. In addition, in describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

In addition, in describing the components of the present invention, terms such as first, second, A, B, (a), (b), etc. may be used. These terms are only for distinguishing the elements from other elements, and the essence, order, or order of the elements are not limited by the terms. When it is described that a component is “connected”, “coupled” or “connected” to another component, the component may be directly connected or connected to the other component, but another component is formed between each component. It should be understood that elements may also be “connected,” “coupled,” or “connected.”

1 is a configuration diagram of a battery system according to an embodiment.

Referring to FIG. 1 , the battery system 100 may include an energy storage system 140 , a battery management system 120 , and a battery state diagnosis device 110 .

The energy storage system 140 may include a plurality of battery devices 130 .

The plurality of battery devices 130 may be connected to each other in series, or may be connected to each other in parallel. In the present specification, it is described that the battery devices 130 are connected in parallel to each other for convenience of description, but the present embodiment is not limited thereto.

The battery devices 130 may include a plurality of lithium-based battery cells, but may include a plurality of battery cells of different series or may include only one battery cell.

The battery management system 120 may monitor input/output of the battery devices 130 , measure a state-of-charge (SOC) of the battery devices 130 , and measure other parameters.

The battery management system 120 may measure the terminal voltage and charging/discharging current of the battery devices 130 . In addition, the battery management system 120 may measure the SOC by using the terminal voltage and the charging/discharging current of the battery devices 130 .

The SOC is the remaining energy of the battery device 130 and is known to be measured by an open circuit voltage (OCV). For example, if the remaining energy is high, the OCV is high, and if the remaining energy is low, the OCV is low. The OCV may be the same as the terminal voltage measured in a state where the battery device 130 is not charged and discharged and the battery device 130 is stable. The terminal voltage measured when the battery device 130 is not used for a long time is It can correspond to OCV.

The battery management system 120 monitors the battery device 130 and, when the terminal voltage of the battery device 130 corresponds to the OCV, measures the terminal voltage of the battery device 130 to obtain the OCV. In addition, the battery management system 120 may measure the SOC by using the OCV. When the OCV is confirmed while the battery management system 120 stores the OCV table, the battery management system 120 may find the SOC corresponding to the OCV from the OCV table and determine it as an SOC measurement value.

The SOC measurement method by OCV takes a lot of time, and in some cases, the battery device 130 may not reach such an OCV state. Therefore, the SOC measurement method by OCV is often used only auxiliary.

As a method that can replace the SOC measurement method by OCV, the battery management system 120 may use the current integration method. The current integration method, also called a coulomb counting method, is a method of counting the amount of electric charges input/output from the battery device 130 to measure the remaining energy of the battery device 130 . This is a method based on the principle that the sum of the amount of charges discharged from the battery device 130 and the amount of charged charges is the same as the remaining energy of the battery device 130 .

As described above, since not only the forward reaction but also the side reaction occur in the battery device 130 , the battery management system 120 may count the amount of charge by multiplying the amount of charge input/output from the battery device 130 by a certain gain.

Since the accumulating method can measure the SOC in real time, the battery management system 120 may use the accumulating method as the main method of measuring the SOC. Also, whenever an OCV condition occurs, the battery management system 120 may supplementally calibrate the SOC measurement value using the OCV method.

However, even by these methods, an error may occur in the SOC measurement value of the battery management system 120 . Such an error may be caused by a cumulative error of the accumulative integration method, a system error of the battery management system 120 , or an internal state problem of the battery device 130 .

The battery state diagnosis device 110 receives the SOC measurement value SOCm and at least one battery measurement value Vm and Tm from the battery management system 120 , and the SOC measurement value SOCm and the at least one battery measurement value An SOC estimation value of the battery device 130 may be calculated using (Vm, Tm).

The battery state diagnosis apparatus 110 may estimate the SOC in a method different from the method of measuring the SOC of the battery management system 120 . For example, the battery state diagnosis apparatus 110 may estimate the SOC of the battery apparatus 130 using a neural network technique.

The battery state diagnosis apparatus 110 or other learning apparatus may generate hidden layers of the neural network and learn internal parameters of the hidden layers by using the learning data. When the internal parameters are determined by such learning, the battery state diagnosis apparatus 110 may install the neural network in which the parameters are determined, and estimate the SOC of the battery apparatus 130 using the neural network.

The battery management system 120 mainly measures the SOC using the integrative integration method, and such an SOC measurement value may have a large error depending on the battery state. As described above, side reactions occur in the battery device 130 , and these side reactions may occur differently depending on the voltage and temperature of the battery device 130 . For example, when the battery state is a high temperature state, more side reactions may occur. However, since it is difficult for the battery management system 120 to measure the SOC by reflecting the battery state, an error may occur.

Since the battery state diagnosis apparatus 110 further uses the SOC measurement value and at least one battery measurement value Vm and Tm, it is possible to estimate the SOC more accurately than the battery management system 120 .

The battery measurement value may be a terminal voltage measurement value Vm of the battery device 130 , and may be a temperature measurement value Tm of the battery device 130 . The battery state diagnosis apparatus 110 may more accurately estimate the SOC by further using the battery measurement values Vm and Tm.

The battery condition diagnosis device 110 compares the SOC measured value (SOCm) received from the battery management system 120 with the SOC estimated value that it has estimated, and the difference between the SOC measured value and the SOC estimate is out of a certain range. In this case, it may be determined that the battery device 130 is abnormal.

The battery condition diagnosis apparatus 110 calculates a root-mean-square (RMS) error calculated over a certain time range in consideration of the possibility that a temporary difference may occur between the SOC measured value and the SOC estimated value, and the RMS error is the reference value. If it is out of , it may be determined that the battery device 130 is abnormal.

The battery state diagnosis apparatus 110 may input the SOC estimated value to the fuzzy predictor and calculate the SOC predicted value through the fuzzy predictor.

A fuzzy predictor is a technique that receives a series of data from the past and predicts the range of values at the next point in time. The battery state diagnosis apparatus 110 may input the SOC estimation value calculated at each predetermined calculation period to the fuzzy predictor. In addition, the battery state diagnosis apparatus 110 may calculate the SOC predicted value through the fuzzy predictor.

Here, the SOC predicted value may be composed of the SOC upper limit predicted value and the SOC lower limit predicted value. And, the SOC upper limit predicted value and the SOC lower limit predicted value may form an SOC predicted range in a range therebetween.

The battery condition diagnosis apparatus 110 may compare the thus formed SOC prediction range with the SOC measurement value, and when the SOC measurement value is out of the SOC prediction range, determine the battery apparatus 130 as abnormal.

If the comparison of the SOC measured value and the SOC estimated value is the primary normal judgment process for the battery device 130 , the comparison of the SOC measured value and the SOC predicted range is the secondary normal judgment process for the battery device 130 . can

The battery state diagnosis apparatus 110 may more accurately diagnose the battery apparatus 130 through the first and second normal determination processes.

2 is a block diagram of an apparatus for diagnosing a battery state according to an exemplary embodiment.

Referring to FIG. 2 , the battery state diagnosis apparatus 110 may include a receiver 210 , a neural network 220 , a fuzzy predictor 230 , and a controller 240 .

The receiver 210 may receive an SOC measurement value SOCm and at least one battery measurement value Vm and Tm from the battery management system while communicating with the battery management system.

The at least one battery measurement value Vm and Tm may include a terminal voltage measurement value Vm of the battery device and/or a temperature measurement value Tm of the battery device.

The neural network 220 may receive the SOC measurement value SOCm and at least one battery measurement value Vm and Tm from the receiver 210 and calculate the SOC estimation value SOCe using internal hidden layers. In addition, the neural network 220 may transmit the SOC estimation value SOCe to the fuzzy predictor 230 .

The fuzzy predictor 230 may calculate the SOC predicted value using the SOC estimated value (SOCe) input for every calculation period. The predicted SOC value may be composed of an SOC upper limit predicted value (SOCp+) and an SOC lower limit predicted value (SOCp-).

The controller 240 may receive the SOC measurement value SOCm from the receiver 210 and receive the SOC measurement value SOCm from the neural network 220 .

Then, the controller 240 may calculate an RMS error with respect to the SOC measured value SOCm and the SOC estimated value SOCe, and when the RMS error exceeds a reference value, the battery device may be determined to be abnormal.

When the RMS error of the SOC measured value SOCm and the SOC estimated value SOCe is smaller than the reference value, the controller 240 may receive the SOC predicted value from the fuzzy predictor 230 and analyze it.

When the SOC measurement value (SOCm) falls within the SOC prediction range - a range between SOCp- and SOCp+ -, the control unit 240 determines that the battery device is normal, and when the SOC measurement value (SOCm) is out of the SOC prediction range , it may be determined that the battery device is abnormal.

3 is a block diagram of a neural network according to an embodiment.

Referring to FIG. 3 , the neural network 220 may include an input layer 310 , a first hidden layer 320 , a second hidden layer 330 , and an output layer 340 .

In addition, each of the layers 310 to 340 may include at least one neural NR.

The input layer 310 may receive an SOC measurement value SOCm, a terminal voltage measurement value Vm, and a temperature measurement value Tm.

The calculation functions are mainly formed in the hidden radars 320 and 330, and these calculation functions may be linear functions.

4 is a flowchart of a method for diagnosing a battery state according to an exemplary embodiment.

Referring to FIG. 4 , the device may receive a state-of-charge (SOC) measurement value and at least one battery measurement value for the battery device from the battery management system (S402).

Here, the at least one battery measurement value may include a terminal voltage measurement value of the battery device and may include a temperature measurement value of the battery device.

Then, the device may input the SOC measurement value and at least one battery measurement value to the neural network to calculate the SOC estimation value (S404).

Here, the neural network includes a plurality of hidden layers, and each hidden layer may be composed of a linear function.

Then, the device can calculate the SOC prediction range by inputting the SOC estimation value to the fuzzy predictor (S406).

In addition, the device may calculate a root-mean-square (RMS) error between the SOC measurement value and the SOC estimate value.

Then, the device compares the RMS error with the reference value (S408), and when the RMS error exceeds the reference value (No in S408), it may determine that the battery device is abnormal (S412).

If the RMS error is less than or equal to the reference value (Yes in S408), the device compares the SOC measured value with the SOC predicted range (S410), and if the SOC measured value is out of the SOC predicted range (No in S410), the battery device It may be determined as abnormal (S412).

And, when the SOC measurement value does not deviate from the SOC prediction range (Yes in S410), the device may determine that the battery device is normal (S414).

The battery status diagnosis may be performed in real time, periodically or non-periodically.

5 is a diagram illustrating an example of an aperiodic battery state diagnosis method.

Referring to FIG. 5 , the battery device may be mounted on an electric vehicle such as an electric vehicle 500 . In addition, the battery condition diagnosis device 110 may be disposed outside the device on which the battery device is mounted.

In addition, the electric vehicle 500 in which the battery device is mounted may periodically or aperiodically transmit the battery management system data BMSDT to the battery state diagnosis device 110 .

The battery management system data BMSDT may include SOC measurement values and at least one battery measurement values stored at regular time intervals, for example, terminal voltage measurement values or temperature measurement values.

The battery state diagnosis apparatus 110 may diagnose the state of the battery device mounted in the electric vehicle 500 by using the SOC measurement values and at least one battery measurement value included in the battery management system data (BMSDT).

As described above, according to the present embodiment, the internal state of the battery can be diagnosed. And, according to the present embodiment, it is possible to diagnose the internal state of the battery in a non-destructive way without using a destructive method of releasing the sealed state of the battery. And, according to the present embodiment, it is possible to diagnose the internal state of the battery using basic information obtainable through the existing battery management system.

Terms such as "include", "comprise" or "have" described above mean that the corresponding component may be embedded unless otherwise stated, so it does not exclude other components. It should be construed as being able to further include other components. All terms, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs, unless otherwise defined. Terms commonly used, such as those defined in the dictionary, should be interpreted as being consistent with the meaning of the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present invention.

The above description is merely illustrative of the technical spirit of the present invention, and various modifications and variations will be possible without departing from the essential characteristics of the present invention by those skilled in the art to which the present invention pertains. Accordingly, the embodiments disclosed in the present invention are not intended to limit the technical spirit of the present invention, but to explain, and the scope of the technical spirit of the present invention is not limited by these embodiments. The protection scope of the present invention should be construed by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

Claims (10)

receiving a state-of-charge (SOC) measurement value and at least one battery measurement value for the battery device from the battery management system;
calculating an SOC estimation value by inputting the SOC measurement value and the at least one battery measurement value into a neural network;
calculating an SOC prediction range by inputting the SOC estimation value into a fuzzy predictor; and
comparing the SOC measurement value with the SOC prediction range, and determining that the battery device is abnormal when the SOC measurement value is out of the SOC prediction range
A battery condition diagnosis method comprising a. According to claim 1,
The at least one battery measurement value is
A method for diagnosing a battery state including a terminal voltage measurement value of the battery device. According to claim 1,
The at least one battery measurement value is
A method for diagnosing a battery condition including a temperature measurement value of the battery device. According to claim 1,
Before the step of comparing the SOC measurement value with the SOC prediction range,
The method further comprises calculating a root-mean-square (RMS) error between the SOC measurement value and the SOC estimate value,
In the step of calculating the RMS error,
When the RMS error exceeds a reference value, the battery state diagnosis method for determining that the battery device is abnormal. According to claim 1,
The neural network includes a plurality of hidden layers,
Each hidden layer is a battery condition diagnosis method composed of a linear function. a receiving unit for receiving a state-of-charge (SOC) measurement value and at least one battery measurement value for the battery device from the battery management system;
a neural network for calculating an SOC estimation value using the SOC measurement value and the at least one battery measurement value;
a fuzzy predictor for calculating an SOC predictive range using the SOC estimate; and
A control unit that compares the SOC measurement value with the SOC prediction range, and determines that the battery device is abnormal when the SOC measurement value is out of the SOC prediction range
A battery condition diagnosis device comprising a. 7. The method of claim 6,
The battery management system,
A battery condition diagnosis apparatus for generating the SOC measurement value using a method different from the neural network method. 7. The method of claim 6,
The battery management system,
A battery condition diagnosis device for generating the SOC measurement value using a current integration method. 7. The method of claim 6,
The battery device is a battery condition diagnosis device mounted on an electric vehicle. 7. The method of claim 6,
The control unit is
A battery condition diagnosis apparatus for calculating a root-mean-square (RMS) error between the SOC measurement value and the SOC estimation value, and determining the battery device as abnormal when the RMS error exceeds a reference value.

Images