CN111919128A

CN111919128A - Method for estimating state of charge of power storage device and system for estimating state of charge of power storage device

Info

Publication number: CN111919128A
Application number: CN201980021051.3A
Authority: CN
Inventors: 伊佐敏行; 千田章裕
Original assignee: Semiconductor Energy Laboratory Co Ltd
Current assignee: Semiconductor Energy Laboratory Co Ltd
Priority date: 2018-04-06
Filing date: 2019-03-29
Publication date: 2020-11-10
Also published as: WO2019193471A1; JP7446990B2; JPWO2019193471A1

Abstract

The present invention provides a method for estimating a state of charge of a secondary battery, which has high accuracy of estimating deterioration of the secondary battery. Further, the present invention provides a capacity measuring system for a secondary battery that estimates SOC in a short time and at low cost with high accuracy. A neural network is used in order to further improve the estimation accuracy of the SOC obtained by performing calculation processing using a regression model such as kalman filtering. The data obtained by the optimization algorithm is used as supervisory data in the neural network, and the initial parameters obtained by using the neural network can estimate the SOC with high accuracy.

Description

Method for estimating state of charge of power storage device and system for estimating state of charge of power storage device

Technical Field

One embodiment of the invention relates to an article, a method, or a method of manufacture. In addition, one embodiment of the present invention relates to a program (process), a machine (machine), a product (manufacture), or a composition (machine). One embodiment of the present invention relates to a semiconductor device, a display device, a light-emitting device, a power storage device, a lighting device, or an electronic apparatus. One embodiment of the present invention relates to a method for estimating a state of charge of a power storage device and a method for controlling the charge of the power storage device. More particularly, the present invention relates to a System for estimating a state of charge of an electric storage device, a System for charging an electric storage device, and a System for controlling an electric storage device (also referred to as a BMS (Battery Management System)).

Note that in this specification, the power storage device refers to all elements and devices having a power storage function. Examples of the secondary batteries include secondary batteries such as lithium ion secondary batteries (also referred to as secondary batteries), lithium ion capacitors, nickel hydride batteries, all-solid-state batteries, and electric double layer capacitors.

One embodiment of the present invention relates to a neural network and a control device for an electric storage device using the neural network. Further, one embodiment of the present invention relates to a vehicle having a BMS using a neural network. In addition, one embodiment of the present invention relates to an electronic device using a neural network. One embodiment of the present invention relates to a power storage device that is applicable not only to a vehicle but also to a power storage device for storing electric power obtained from a power generation device such as a solar power generation panel provided in a structure or the like.

Background

As a method of estimating the remaining capacity of the secondary battery, there are coulomb counting method and ocv (open Circuit voltage) method.

When the conventional method is used for a long period of time and charging or discharging is repeated, errors are accumulated, and the accuracy of estimation of the SOC (State of Charge), which is the charging rate, may be significantly reduced. Further, since the initial SOC (0) changes due to self-discharge even with the elapse of time when the battery is not in use, it is difficult to improve the SOC estimation accuracy. The coulomb counting method has a disadvantage that it cannot correct an error of the initial SOC (0) or an error of the accumulated current sensor. Patent document 1 discloses a technique of estimating the state of a secondary battery at a low temperature with high accuracy by a state estimating unit based on information including a parameter associated with a temperature.

[ Prior Art document ]

[ patent document ]

[ patent document 1] Japanese patent application laid-open No. 2016-80693

Disclosure of Invention

Technical problem to be solved by the invention

A vehicle mounted with a secondary battery may charge regenerative electric power generated at the time of braking or the like to the secondary battery, and there is a possibility that the secondary battery may not be properly used due to overcharge. In order to prevent the overcharge and overdischarge problems from occurring in advance, it is necessary to estimate the remaining amount of electricity of the secondary battery, that is, the SOC of the secondary battery with high accuracy. The invention provides a method for estimating a state of charge of a secondary battery or a method for controlling a power storage device, which have high estimation accuracy.

In the production of secondary batteries, even in the same production lot, slight individual differences may occur due to slight differences in the amount of active material, electrode size, and the like during assembly. In vehicles and the like, a plurality of secondary batteries are used, and when a plurality of batteries are combined, individual differences affect each other, and thus, a capacity difference between vehicles may be increased due to deterioration.

Further, as the deterioration of the secondary battery progresses, the accuracy of estimating the SOC sometimes significantly decreases. Further, the SOC is defined according to the ratio of the remaining capacity with respect to the maximum capacity of the secondary battery. When the maximum capacity of the secondary battery is calculated from the time integral of the current by discharging after full charge, it may take a long time to discharge.

The invention provides a method for estimating a state of charge of a secondary battery with high estimation accuracy even if the secondary battery is deteriorated. Further, the present invention provides a capacity measuring system for a secondary battery that estimates SOC in a short time and at low cost with high accuracy.

Means for solving the problems

The state-of-charge estimation method of a secondary battery disclosed in the present specification uses a neural network in order to further improve the estimation accuracy of SOC obtained by performing calculation processing using a regression model, such as kalman filtering. The charging rate (SOC) is estimated by using Artificial Intelligence (AI) such as a neural network.

The configuration of the invention disclosed in the present specification is a method for estimating a state of charge of an electric storage device, including the steps of: determining a circuit model of the power storage device; in a circuit model (a foster type circuit model) of an electric storage device, a current is taken as an input and a voltage is taken as an output; performing optimization to reduce an output error of a voltage of the power storage device, and calculating an initial parameter (first value) of a circuit model of the power storage device; storing initial parameter groups corresponding to input values of different currents; the initial parameter (second value) is determined by neural network processing using the initial parameter group obtained by optimization as supervisory data, and the state of charge (SOC) is estimated by kalman filter processing using the initial parameter (second value).

The initial parameter group corresponding to the input values of different currents does not have to use actually measured cycle characteristics, and data may be generated according to the type setting conditions of the power storage device used by the implementer. Another aspect of the present invention is a method for estimating a state of charge of a power storage device, including: determining a circuit model of the power storage device; in a circuit model of an electrical storage device, a current is taken as an input and a voltage is taken as an output; performing optimization to reduce an output error of a voltage of the power storage device, and calculating an initial parameter (first value) of a circuit model of the power storage device; generating an initial parameter group different from the initial parameter; initial parameters (second values) are determined by neural network processing using the initial parameter group as supervisory data, and the state of charge (SOC) is estimated by kalman filter processing using the initial parameters.

Kalman filtering is one of the infinite impulse response filters. Further, the multiple regression analysis is one of multivariate analysis in which an independent variable of the regression analysis is plural. As the multiple regression analysis, there is a least square method and the like. The regression analysis requires a time series of many observed values, and the kalman filter has an advantage that the most suitable correction coefficient can be obtained step by accumulating data to a certain extent. Furthermore, kalman filtering may also be applied to non-stationary time series.

As a method of estimating the internal resistance and the state of charge (SOC) of the secondary battery, a nonlinear kalman filter (specifically, a lossless kalman filter (also referred to as UKF)) can be used. In addition, extended kalman filtering (also known as EKF) may also be used.

The initial parameters obtained by the optimization algorithm are collected every n (n is an integer, for example, fifty) cycles, and these data groups are used as supervisory data to perform neural network processing, whereby the SOC can be estimated with high accuracy.

The learning system comprises a supervision generation device and a learning device. The supervision data generating means generates supervision data used when the learning means learns. The supervision data includes data of the processing object data identical to the identification object and evaluation of a label corresponding to the data. The supervisory data generation device includes an input data acquisition unit, an evaluation acquisition unit, and a supervisory data generation unit. The input data acquisition unit may acquire data from data stored in the storage device, or may acquire input data for learning, including a current value or a voltage value of the secondary battery, via the internet. The supervision data may not be actually measured data. For example, the charging rate (SOC) can be estimated by setting the initial parameters to have a variety according to the conditions, generating data close to actual measurement, and performing neural network processing using the predetermined characteristic database of these data as the supervision data. The SOC estimation of the same type of battery can be efficiently performed by generating data close to actual measurement based on the charge/discharge characteristics of any one of the batteries and performing neural network processing using the predetermined characteristic database for the supervision data.

When SOC estimation is performed using only the optimization algorithm, there are problems such as a large calculation amount of the optimization algorithm, convergence to a meaningless value, and divergence in which an optimal value cannot be determined. Of batteriesThe characteristics are nonlinear, and five initial parameters are obtained by a numerical optimization method of a nonlinear function. The five initial parameters are total capacity FCC (full Charge capacity), direct current resistance R_s(R₀) Resistance R generated by diffusion process_dDiffusion capacity C_dAnd an initial SOC (0). Note that FCC (also referred to as full charge capacity, total capacity) is a rated capacity at 25 ℃.

The optimization process for obtaining the five initial parameters may be performed using a means mounted on Python (registered trademark) or Matlab (Matrix Laboratory) (registered trademark).

When the deterioration of the secondary battery progresses, since an error in the SOC may occur when the FCC of the initial parameter changes greatly, the initial parameter used for estimating the operation of the SOC can be updated. The updated initial parameters are calculated by an optimization algorithm using data of charge-discharge characteristics measured in advance. By performing the calculation process using the updated regression model of the initial parameter, such as the kalman filter, the SOC can be estimated with high accuracy even after the degradation. In this specification, the calculation processing by using the kalman filter is also referred to as a kalman filter processing.

The timing of updating the initial parameters may be arbitrary, but in order to estimate the SOC with high accuracy, it is preferable that the updating frequency be as high as possible, and the updating be performed continuously at regular intervals.

The step of estimating the SOC is described more specifically below.

In the first stage, a voltage value or a current value of the secondary battery is measured by using a measuring unit (a voltage detecting circuit or a current detecting circuit). These data are acquired by a voltage measuring instrument or a current measuring instrument (also referred to as a current sensor) and stored in a storage device. The initial SOC (0) is calculated from the voltage value obtained by using the voltage measuring instrument, specifically, from the charge/discharge characteristic data. The initial SOC (0) is an initial value of SOC. The initial Rs is an initial value of the dc resistance Rs (also referred to as R)₀) And is the resistance created by the electrophoretic process of the ions. From the pre-measured charge-discharge characteristics by using an optimization algorithm, in particular by using the Nelder-Mead algorithmTo obtain five initial parameters, specifically initial SOC (0), FCC, R₀、R_dAnd C_d. Note that the Nelder-Mead algorithm is an algorithm that does not require a derivative function.

As another method for calculating the initial SOC (0), the initial SOC (0) may be determined by measuring the open circuit voltage of the battery before the start of use by using a voltage detection circuit and using a map or a correspondence table obtained in advance for the open end voltage OCV and the SOC. OCV is the voltage at which the battery is electrochemically in an equilibrium state, and corresponds to SOC.

And constructing a fully-connected neural network in the second stage. The charging voltage characteristic is used as an input to the neural network, five initial parameters can be calculated using the initial parameter group of the battery model calculated by the Nelder-Mead algorithm as supervision data, and the SOC can be estimated with high accuracy by performing kalman filter processing.

In addition, a configuration of a state of charge estimation device for a power storage device according to the present invention is a configuration of a state of charge estimation device for a power storage device, including: a measurement section; a storage unit; an estimation unit that measures the current or voltage of the power storage device by the measurement unit, stores data measured by the measurement unit in the storage unit, stores data obtained by the optimization algorithm as supervision data based on the data, and determines an initial parameter, and an operation unit that estimates the SOC by Kalman filtering using the initial parameter.

In the above configuration, the estimation unit includes a neural network. The neural network processing is performed using the data in the storage unit. In the above configuration, the optimization algorithm uses the Nelder-Mead algorithm.

In the present specification, a neural network (also referred to as an artificial neural network) refers to any model that has a problem solving ability by determining the strength of connection between neurons by learning, by simulating a neural network of a living body. The neural network includes an input layer, an intermediate layer (also referred to as a hidden layer), and an output layer.

In the present specification, when describing a neural network, an operation of determining the connection strength (also referred to as a weight coefficient) between neurons from existing information is sometimes referred to as "learning".

In the present specification, a work of constructing a neural network using the connection strengths obtained by learning and deriving a new conclusion therefrom is sometimes referred to as "inference".

The present invention is applicable to other batteries (for example, all-solid-state batteries). The present invention can estimate the SOC with high accuracy by appropriately converting a battery model according to the type of a battery.

In the present specification, the state of charge (SOC) is expressed as a percentage of the sum of the remaining amount of electricity and the charged amount of electricity of the capacity at the time of full charge of the secondary battery. The charging amount is required to be obtained in order to calculate the charging rate, and the charging amount may be calculated from the pulse number per short time, the current value of the charging current, and the on duty.

Effects of the invention

When the kalman filter process is performed to infer the SOC, a neural network that uses data obtained by the optimization algorithm as supervision data is used. By using this neural network, the SOC can be estimated with high accuracy. One embodiment of the present invention can estimate SOC with high accuracy using a relatively small calculation amount.

Drawings

FIG. 1 is an example of a block diagram showing an embodiment of the present invention.

FIG. 2 is an example of a flowchart showing an embodiment of the present invention.

FIG. 3 is an example of a flowchart showing an embodiment of the present invention.

Fig. 4 shows an example of a battery model according to an embodiment of the present invention.

FIG. 5 is a graph showing actually measured charge and discharge.

Fig. 6 is a graph showing the relationship between the total capacity FCC and the number of cycles.

[ FIG. 7 ]]Is to show a direct current resistance R_sGraph of the relationship to cycle number.

[ FIG. 8 ]]Is to show the diffusion capacity C_dGraph of the relationship to cycle number.

[ FIG. 9 ]]Is to show the resistance R produced by the diffusion process_dGraph of the relationship to cycle number.

FIG. 10 is a graph showing the relationship between the initial SOC (0) and the number of cycles.

Fig. 11 (a) and (B) are perspective views showing an example of the secondary battery, and (C) is a schematic cross-sectional view of the secondary battery.

Fig. 12 is a diagram showing an example of a mobile body.

Detailed Description

Embodiments of the present invention will be described in detail below with reference to the drawings. Note that the present invention is not limited to the following description, and a person of ordinary skill in the art can easily understand the fact that the modes and details thereof can be changed into various forms. The present invention should not be construed as being limited to the embodiments described below.

(embodiment mode 1)

In the present embodiment, an example in which the battery state estimating device is applied to an Electric Vehicle (EV) is described with reference to fig. 1A.

The electric vehicle is provided with a first battery 301 serving as a secondary battery for main driving and a second battery 311 that supplies electric power to an inverter 312 that starts the engine 304. In the present embodiment, state estimating section 300 driven by the power supply of second battery 311 monitors a plurality of secondary batteries constituting first battery 301 collectively. The state estimation unit 300 performs charge state estimation.

The state estimation unit 300 has a computer including a cpu (central Processing unit), a memory as a storage unit, and the like as main components. The CPU includes an arithmetic section that can correspond to the plurality of secondary batteries. The calculation unit determines a battery model of the secondary battery and estimates a numerical value using a neural network. The memory stores supervisory data, and estimates the SOC from the input current value or voltage value.

Fig. 2 shows an example of a flowchart for estimating SOC. The series of processes shown in fig. 2 includes the following steps: step 1(S1) of determining a circuit model of the secondary battery; storing the measured power of the secondary battery in a storage unitA step 2 of measuring a current value or a measured voltage value (S2); step 3(S3) of inputting the measured current value or the measured voltage value to the neural network using the five initial parameters as the supervisory data, and calculating the five initial parameters (FCC, R)_s(R₀)、R_d、C_dAnd initial SOC (0)) (S4). Here, if there is an abnormal value in the initial parameter in step 4, after step 5(S5) of changing the parameter indicating the abnormal value to a value in the abnormal range (for example, the previous calculation result or the like), step 6 of estimating SOC by the UKF is performed (S6). For example, an abnormal value may be detected due to a change in the ambient temperature of the secondary battery, and the abnormal detection may be performed based on whether or not an abnormal value is present in the initial parameter in step 4.

The supervision data is a data group obtained by obtaining the charge/discharge characteristics of several secondary batteries in advance through actual measurement and calculating five parameters using an optimization algorithm (in the present embodiment, the Nelder-Mead algorithm is used). When data actually measured in advance is used, the learning by the state estimation unit 300 is referred to as initial learning. Further, the case where the state estimating unit 300 learns data after repeating charge and discharge of the secondary battery and deteriorating to some extent a plurality of times, that is, the case where the supervision data is added or updated is also referred to as relearning.

Fig. 3A shows an example of the learning flow. The series of processes shown in fig. 3A includes the following steps: step 1(S1) of determining a circuit model of the secondary battery; step 2(S2) of actually measuring the charge-discharge characteristics of the secondary battery; a step 3 of calculating five initial parameters by optimizing the charge/discharge characteristics of the secondary battery by a Nelder-Mead algorithm (S3); and a step 4 of constructing a neural network for learning the five initial parameter groups as the supervision data (S4).

All or a part of the above-described respective processes may be set as a step of automatic operation. Further, a part of the steps may be performed manually at the timing of the user, or may be performed periodically. The order of steps shown in the drawings and information including various data or parameters may be arbitrarily changed. That is, the information shown in the drawings is not limited to the illustrated order of course.

The supervision data may be charge/discharge characteristics of the nth cycle (n is an integer of 2 or more) obtained in advance.

Fig. 3B shows an example of the learning flow. The series of processes shown in fig. 3B includes: step 1(S1) of determining a circuit model of the secondary battery; step 2(S2) of actually measuring the charge-discharge characteristics of the secondary battery; and a step 3 of calculating five initial parameters by optimizing the charge/discharge characteristics of the secondary battery using the Nelder-Mead algorithm (S3). If the actually measured cycle data can be obtained in advance, the method further includes a step 15(S15) of optimizing the actually measured data of the secondary battery for each n-cycle to calculate five initial parameter groups for each n-cycle, and a step 17(S17) of constructing a neural network for learning the five initial parameter groups as supervision data.

Further, a part of the supervision data may not be actually measured, and charge and discharge characteristics assumed by the implementer may be used. In this case, if the actually measured cycle data cannot be acquired in advance after step 3, step 16 of setting a virtual initial parameter group of each condition of five initial parameters may be performed before step 17 (S16). This step 16 may be referred to as an additional learning step. The virtual initial parameter group may also be referred to as virtual supervisory data. In addition, the change of the initial parameter group may be referred to as relearning because the supervisory data is added or updated.

The first battery 301 mainly supplies electric power to a 42V series (high voltage series) vehicle-mounted device, and the second battery 311 supplies electric power to a 14V series (low voltage series) vehicle-mounted device. In many cases, a lead storage battery is used as the second battery 311 because of its cost advantage. Lead storage batteries have a disadvantage that they have a large self-discharge as compared with lithium ion secondary batteries and are easily deteriorated by a phenomenon called sulfation. Although there is an advantage that maintenance is not required when the lithium-ion secondary battery is used as the second battery 311, an abnormality that cannot be identified at the time of manufacture may occur during a long period of use, for example, three years or more. Particularly, if the second battery 311, which activates the inverter, cannot operate, the generator cannot be activated even if the first battery 301 has a residual capacity. In order to prevent this, when the second battery 311 is a lead storage battery, electric power is supplied from the first battery to the second battery, and the second battery is charged so as to maintain a fully charged state.

This embodiment shows an example in which both the first battery 301 and the second battery 311 use lithium ion secondary batteries. The second battery 311 may also use a lead storage battery or an all-solid battery.

The regenerative energy generated by the rotation of the tire 316 is sent to the engine 304 through the transmission 305, and is supplied to the second battery 311 or the first battery 301 through the engine controller 303 and the battery controller 302.

The first battery 301 is mainly used to rotate the engine 304, and also supplies electric power to 42V-series vehicle-mounted components (an electric power steering system 307, a heater 308, a defogger 309, and the like) via a DCDC circuit 306. The first battery 301 is used to rotate the rear engine in the case where the rear wheel includes the rear engine.

The second battery 311 supplies power to 14V-series vehicle-mounted components (an audio device 313, a power window 314, lamps 315, and the like) via the DCDC circuit 310.

The first battery 301 is constituted by a plurality of secondary batteries. For example, a cylindrical secondary battery 600 is used. As shown in fig. 1B, a cylindrical secondary battery 600 may be sandwiched between a conductive plate 613 and a conductive plate 614 to form a module 615. Switches between the secondary batteries are not illustrated in fig. 1B. The plurality of secondary batteries 600 may be connected in parallel, may be connected in series, or may be connected in series after being connected in parallel. By constituting the module 615 including a plurality of secondary batteries 600, it is possible to extract a large electric power.

Although fig. 1A shows the battery controller 302 and the state estimation unit 300 as separate structures, they are not particularly limited and may be formed of one IC chip on the same substrate or may be combined into one unit. The state estimation unit 300 may be formed of an lsi (large Scale integration) integrally manufactured on one chip. The method of integration is not limited to LSI, and may be realized by a dedicated circuit or a general-purpose processor. In addition, an fpga (field Programmable Gate array) which is Programmable after LSI manufacturing or a reconfigurable processor which can reconfigure connection or setting of circuit cells inside LSI may be used. In addition, an IC (also referred to as an inference chip) incorporating an AI system may also be used. An IC incorporating an AI system is sometimes called a circuit (microcomputer) that performs a neural network operation. In addition, the battery controller 302 is sometimes also referred to as bmu (battery Management unit). Five initial parameters are stored in, for example, a memory of the state estimation unit 300 of the secondary battery, specifically, a rom (read Only memory) or a ram (random Access memory). The state estimating unit 300 can calculate the SOC of the secondary battery more accurately.

It is possible to realize a power storage control apparatus or a management apparatus including the state estimating unit 300 of the secondary battery. Further, it is possible to realize a power storage control method in which a plurality of processes including a neural network process are sequentially performed to constitute a plurality of steps. Further, each step included in the power storage control method may be realized as a computer program executed by a computer. Further, such a computer program may be stored and executed on a recording medium or on the cloud side of a communication network using the internet.

The program that executes the software of the computer program may be written using various programming languages such as Python, Go, Perl, Ruby, Prolog, Visual Basic, C, C + +, Swift, Java (registered trademark),. NET, and the like. Further, the application can be written using a framework such as Chainer (available in Python), Caffe (available in Python and C + +), TensorFlow (available in C, C + + and Python), and the like.

(embodiment mode 2)

Fig. 4B shows an example of a battery model.

In the present embodiment, the battery model shown in fig. 4B is a simplified model of the model shown in fig. 4A. The weber impedance portion is infinite, but fifty stages are shown simplified in fig. 4A. In the weibull impedance portion of fifty stages shown in fig. 4A, the fourth to fiftieth stages are defined as resistances, and the unit having a small time constant is summarized as the direct current resistance Rs in fig. 4B. Fig. 4B shows a series connection as a direct-current resistance model and a diffusion resistance model.

Resistance R generated by diffusion process_dRepresents a resistance component, diffusion capacity C_dRepresenting the capacity component terms. The diffusion resistance is a resistance component andthe parallel connection body of the capacity component is constituted by connecting a plurality of parallel connections (three stages in the drawing). An equivalent circuit formed by parallel connection of the resistance component and the capacity is called a foster type circuit model. The foster type circuit model is preferable because it is less computationally intensive than the coulter type circuit model.

In the simplified model of FIG. 4, three parameters (R) may be used_s、R_d、C_d) And (4) showing.

The OCV can be expressed by the following equation.

[ equation 1]

OCV＝f(SOC(t))

The soc (t) can be expressed by the following equation.

[ equation 2]

Further, the state variable x (t) can be expressed by the following equation.

[ equation 3]

x(t)＝[SOC(t) v₁(t) v₂(t) v₃(t)]

Further, the output equation may be expressed as follows.

[ equation 4]

V(t)＝f(SOC(t))+v₁(t)+v₂(t)+v₃(t)+R_sI(t)

Therefore, if these equations can find five initial parameters, the calculation can be performed from the state space.

[ equation 5]

[FCC Rs Rd Cd SOC(0)]

In the present embodiment, these five initial parameters are calculated from measured data of voltage and current by performing optimization. The least square method is used as an optimization algorithm, but since the secondary battery has nonlinear characteristics, the Nelder-Mead algorithm is used as the optimization algorithm. Here, as an example, five initial parameters were calculated by performing optimization using data of current shown in fig. 5A and data of voltage shown in fig. 5B.

And constructing a neural network which learns one of the five initial parameter values as supervision data.

The cycle test data was used for verification. The conditions of the cycle test data used were: the ambient temperature was 45 ℃, the charge cutoff voltage was 4.2V, the discharge cutoff current was 2.5V, the charge mode was CC-CV, the charge rate in CC was 0.5C (1.625A), and the discharge rate was 1C (3.25V).

Five initial parameters obtained by optimizing the data for each cycle are shown as comparative black lines. Fig. 6, 7, 8, 9, and 10 show the calculated numerical values, respectively.

The method adopts a fully-connected neural network, 700 input layers are adopted, 500 first hidden layers are adopted, 500 second hidden layers are adopted, and five output layers are adopted (FCC, R)_s、R_d、C_dAnd initial SOC (0)). When the hidden layers of the neural network overlap in multiple layers, it is also called deep learning. In the present embodiment, an example using a fully connected neural network is shown, but the structure of the neural network and the learning method are not particularly limited.

Five parameters were calculated by performing neural network processing also in the fifty-th cycle, 150-th cycle, 250-th cycle, 350-th cycle, 450-th cycle, 550-th cycle, and 650-th cycle, and the data are shown by circles in fig. 6, 7, 8, 9, and 10.

The initial parameters obtained by learning the neural network of the supervisory data obtained by performing the optimization algorithm can be used to obtain substantially the same values as those of the data actually subjected to the optimization. When the optimization algorithm is used, there is a possibility that unnecessary iterative processing, convergence to a physically meaningless value, or divergence problems occur, and it is difficult to use only the optimization algorithm for estimation, but the initial parameters can be calculated by neural network processing using the optimized supervisory data.

From this, it can be said that the initial parameter calculated by the neural network process using the optimized supervisory data is an appropriate value as the initial parameter for the kalman filter, so that the accuracy of the SOC obtained by the kalman filter process is improved.

(embodiment mode 3)

An example of a cylindrical secondary battery will be described with reference to fig. 11A and 11B. As shown in fig. 11A, the cylindrical secondary battery 600 includes a positive electrode cap (battery cap) 601 on the top surface thereof, and a battery can (outer packaging can) 602 on the side surface and the bottom surface thereof. These positive electrode cap 601 and battery can (outer packaging can) 602 are insulated by gasket (insulating gasket) 610.

Fig. 11B is a view schematically showing a cross section of a cylindrical secondary battery. Inside the hollow cylindrical battery can 602, a battery element in which a band-shaped positive electrode 604 and a band-shaped negative electrode 606 are wound with a separator 605 interposed therebetween is provided. Although not shown, the battery element is wound around a center pin. One end of the battery can 602 is closed and the other end is open. As the battery can 602, a metal such as nickel, aluminum, or titanium, an alloy thereof, or an alloy thereof with other metals (e.g., stainless steel) having corrosion resistance to an electrolyte can be used. In addition, in order to prevent corrosion by the electrolyte, the battery can 602 is preferably covered with nickel, aluminum, or the like. Inside the battery can 602, the battery element in which the positive electrode, the negative electrode, and the separator are wound is sandwiched between a pair of insulating

plates

608 and 609 that face each other. A nonaqueous electrolytic solution (not shown) is injected into the battery case 602 provided with the battery element. The secondary battery comprises lithium cobaltate (LiCoO)₂) And lithium iron phosphate (LiFePO)₄) And the like, a negative electrode made of a carbon material such as graphite capable of occluding and releasing lithium ions, and a positive electrode made of LiBF₄、LiPF₆And a nonaqueous electrolyte solution in which an electrolyte composed of a lithium salt is dissolved in an organic solvent such as ethylene carbonate or diethyl carbonate.

Since the positive electrode and the negative electrode for the cylindrical secondary battery are wound, the active material is preferably formed on both surfaces of the current collector. The positive electrode 604 is connected to a positive electrode terminal (positive electrode collector lead) 603, and the negative electrode 606 is connected to a negative electrode terminal (negative electrode collector lead) 607. A metal material such as aluminum can be used for both the positive electrode terminal 603 and the negative electrode terminal 607. The positive terminal 603 is resistance welded to the safety valve mechanism 612, and the negative terminal 607 is resistance welded to the bottom of the battery can 602. AnThe full valve mechanism 612 is electrically connected to the Positive electrode cap 601 through a PTC (Positive Temperature Coefficient) element 611. When the internal pressure of the battery rises to exceed a predetermined threshold value, the safety valve mechanism 612 cuts off the electrical connection between the positive electrode cover 601 and the positive electrode 604. In addition, the PTC element 611 is a heat sensitive resistor whose resistance increases at the time of temperature rise, and limits the amount of current by the increase of resistance to prevent abnormal heat generation. Barium titanate (BaTiO) can be used for the PTC element₃) Quasi-semiconductor ceramics, and the like.

A lithium ion secondary battery using an electrolytic solution includes a positive electrode, a negative electrode, a separator, the electrolytic solution, and an exterior body. Note that in a lithium ion secondary battery, since an anode and a cathode are exchanged for an oxidation reaction and a reduction reaction due to exchange of charge and discharge, an electrode having a high reaction potential is referred to as a positive electrode, and an electrode having a low reaction potential is referred to as a negative electrode. Therefore, in this specification, even when charging, discharging, a reverse pulse current flows, and a charging current flows, the positive electrode is referred to as "positive electrode" or "+ electrode", and the negative electrode is referred to as "negative electrode" or "— electrode". If the terms anode and cathode are used in connection with the oxidation reaction and the reduction reaction, the anode and cathode are opposite in charge and discharge, which may cause confusion. Therefore, in this specification, the terms anode and cathode are not used. When the terms of the anode and the cathode are used, it is clearly indicated whether charging or discharging is performed, and whether positive (+ pole) or negative (-pole) is indicated.

The two terminals shown in fig. 11C are connected to a charger to charge battery 1400. As the charging of the battery 1400 progresses, the potential difference between the electrodes increases. The positive direction in fig. 11C is: the direction from the terminal outside the battery 1400 to the positive electrode 1402, from the positive electrode 1402 to the negative electrode 1404 in the battery 1400, and from the negative electrode to the terminal outside the battery 1400. That is, the direction in which the charging current flows is the direction of the current.

In this embodiment, an example of a lithium ion secondary battery is shown, but the present invention is not limited to the lithium ion secondary battery. As a positive electrode material of the secondary battery, for example, a material containing an element a, an element X, and oxygen can be used. Element A is preferablyIs one or more elements selected from the group consisting of the first group elements and the second group elements. As the first group element, for example, an alkali metal such as lithium, sodium, or potassium can be used. Examples of the second group element include calcium, beryllium, and magnesium. As the element X, for example, one or more elements selected from a metal element, silicon, and phosphorus can be used. The element X is preferably one or more elements selected from cobalt, nickel, manganese, iron, and vanadium. Typically, lithium cobalt composite oxide (LiCoO) can be mentioned₂) And lithium iron phosphate (LiFePO)₄)。

The negative electrode includes a negative electrode active material layer and a negative electrode current collector. The negative electrode active material layer may contain a conductive assistant and a binder.

As the negative electrode active material, an element capable of undergoing charge-discharge reaction by alloying/dealloying reaction with lithium can be used. For example, a material containing at least one of silicon, tin, gallium, aluminum, germanium, lead, antimony, bismuth, silver, zinc, cadmium, indium, and the like can be used. The capacity of this element is greater than that of carbon, and in particular, the theoretical capacity of silicon is greater, being 4200 mAh/g.

In addition, the secondary battery preferably includes a separator. As the separator, for example, a separator formed of a fiber having cellulose such as paper, a nonwoven fabric, a glass fiber, a ceramic, or a synthetic fiber containing nylon (polyamide), vinylon (polyvinyl alcohol fiber), polyester, acrylic resin, polyolefin, polyurethane, or the like can be used.

Fig. 12 illustrates a vehicle using a state of charge estimation device for a secondary battery according to an embodiment of the present invention. A secondary battery 8024 of an automobile 8400 shown in fig. 12A can supply electric power to a light-emitting device such as a headlight 8401 or a room lamp (not shown), as well as driving an electric motor 8406. As the secondary battery 8024 of the automobile 8400, a module 615 in which the cylindrical secondary battery 600 shown in fig. 11B is sandwiched between a conductive plate 613 and a conductive plate 614 may be used.

In the automobile 8500 shown in fig. 12B, the secondary battery of the automobile 8500 can be charged by receiving electric power from an external charging device by a plug-in system, a non-contact power supply system, or the like. Fig. 12B shows a case where a secondary battery 8024 mounted in an automobile 8500 is charged from a charging device 8021 of the above-ground installation type through a cable 8022. In the case of charging, the charging method, the specification of the connector, and the like may be appropriately performed according to a predetermined method such as CHAdeMO (registered trademark) or a joint charging system. As the charging device 8021, a charging station installed in a commercial facility or a power supply of a home may be used. For example, the secondary battery 8024 installed in the automobile 8500 can be charged by supplying electric power from the outside using a plug-in technique. The charging may be performed by converting AC power into DC power by a conversion device such as an AC/DC converter.

Although not shown, the power receiving device may be mounted in a vehicle and charged by supplying electric power from a power transmitting device on the ground in a non-contact manner. When the non-contact power supply system is used, the power transmission device is incorporated in a road or an outer wall, and charging can be performed not only during parking but also during traveling. In addition, the transmission and reception of electric power between vehicles may be performed by the non-contact power feeding method. Further, a solar battery may be provided outside the vehicle, and the secondary battery may be charged when the vehicle is stopped or traveling. Such non-contact power supply may be realized by an electromagnetic induction method or a magnetic field resonance method.

Fig. 12C shows an example of a two-wheeled vehicle using the secondary battery according to one embodiment of the present invention. A scooter 8600 shown in fig. 12C includes a secondary battery 8602, a rearview mirror 8601, and a turn signal light 8603. The secondary battery 8602 may supply power to the direction lamp 8603.

In the scooter 8600 shown in fig. 12C, a secondary battery 8602 may be accommodated in the under-seat accommodation box 8604. Even if the under-seat storage box 8604 is small, the secondary battery 8602 may be stored in the under-seat storage box 8604.

This embodiment can be combined with the description of the other embodiments as appropriate.

[ description of symbols ]

300: state estimation unit, 301: a battery, 302: battery controller, 303: engine controller, 304: engine, 305: transmission, 306: DCDC circuit, 307: electric power steering system, 308: heater, 309: demister, 310: DCDC circuit, 311: battery, 312: inverter, 313: acoustic device, 314: power window, 315: lamps, 316: tire, 600: secondary battery, 601: positive electrode cover, 602: battery can, 603: positive electrode terminal, 604: positive electrode, 605: spacer, 606: negative electrode, 607: negative terminal, 608: insulating plate, 609: insulating plate, 611: PTC element, 612: safety valve mechanism, 613: conductive plate, 614: conductive plate, 615: module, 1400: storage battery, 1402: positive electrode, 1404: negative electrode, 8021: charging device, 8022: cable, 8024: secondary battery, 8400: car, 8401: headlights, 8406: electric motor, 8500: car, 8600: scooter, 8601: rearview mirror, 8602: secondary battery, 8603: turn signal, 8604: a storage box under the seat.

Claims

1. A method for estimating a state of charge of an electrical storage device, comprising:

determining a circuit model of the power storage device;

taking a current as an input and a voltage as an output in a circuit model of the power storage device;

performing optimization to reduce an output error of a voltage of the power storage device, and calculating initial parameters of a circuit model of the power storage device;

storing initial parameter groups corresponding to input values of different currents; and

the initial parameter group is used as supervision data, an initial parameter value is determined by neural network processing, and the initial parameter value is calculated by Kalman filtering processing to estimate the charging rate.

2. A method for estimating a state of charge of an electrical storage device, comprising:

determining a circuit model of the power storage device;

generating an initial parameter group different from the initial parameter; and

the initial parameter group is used as supervision data, initial parameter values are determined through neural network processing, and the initial parameter values are processed through Kalman filtering to estimate the charging rate.

3. The method of estimating a state of charge of a power storage device according to claim 1 or 2,

wherein the initial parameter value is FCC, R_s、R_d、C_dAnd an initial SOC (0).

4. A state-of-charge estimation device for an electrical storage device, comprising:

a measurement section;

a storage unit;

an estimation unit; and

a calculation unit for calculating the time of the operation,

wherein the measuring section measures a current or a voltage of the power storage device,

the storage unit stores the data measured by the measuring unit,

the presumption part stores data obtained by an optimization algorithm as supervision data based on the data, and determines an initial parameter value,

the calculation unit estimates a charging rate by using the initial parameter value through kalman filter processing.

5. The state-of-charge estimation device of a power storage device according to claim 4,

wherein the inference section includes a neural network.

6. The state-of-charge estimation device of the power storage device according to claim 4 or 5,

wherein the optimization algorithm uses the Nelder-Mead algorithm.