JP3520886B2 - Rechargeable battery status determination method - Google Patents

Description

Detailed Description of the Invention

[0001]

The present invention relates to an uninterruptible power supply (hereinafter,
It is related to a method of determining a battery state such as a capacity and a life of a secondary battery such as a lead storage battery used for UPS).

[0002]

2. Description of the Related Art A UPS is a device that supplies electric power to a load that normally uses a general electric power system as a power source by a secondary battery when an abnormality occurs in the electric power system. Therefore, if there is an abnormality in the secondary battery, the role of UPS cannot be fulfilled. From this, the reliability of the secondary battery is greatly related to the reliability of UPS itself. Therefore, it is necessary to judge the capacity and life of the secondary battery. Secondary battery capacity,
The service life is greatly affected by conditions such as operating temperature, operating period, charging voltage, and number of discharges. Therefore, the capacity and life of the secondary battery are not constant and it is difficult to judge them during use.

As a method of determining the capacity and life of a secondary battery, there is known a method of using an internal resistance which becomes high at the end of the life of the secondary battery or when a defect occurs. There is also a method of measuring the discharge voltage of the secondary battery for a long time to determine the capacity, and displaying according to the measurement result.

[0004]

When a secondary battery becomes old due to long-term use, the discharge voltage tends to drop more rapidly than that of a new secondary battery. Therefore, if the discharge voltage is simply measured to determine the capacity, an error will occur. Further, in order to always know the state of the secondary battery, it is necessary to discharge for a long time and measure the voltage each time. Therefore, there is a drawback that the life of the secondary battery is shortened. Further, when the secondary battery is used as the backup power source, the backup time becomes short when the life is judged.

Therefore, an object of the present invention is to provide a method for accurately determining the battery state such as capacity or life without discharging for a long time in order to constantly monitor the state of the secondary battery. It is to be able to properly maintain and manage the.

[0006]

In order to achieve the above object, the present invention according to claim 1 provides at least one of the remaining capacity, life and voltage of a secondary battery after a predetermined time has passed from the start of discharge. A method of determining the open circuit voltage (OCV) at a plurality of elapsed time points after the start of use of the secondary battery for learning having substantially the same configuration as the secondary battery to be determined for performing the determination, immediately after the start of discharge. Voltage (VO) and open circuit voltage (OCV) between the values at the start of use of the learning secondary battery and the values at the plurality of elapsed times (ΔO
CVO), the difference between the values of the open circuit voltage (OCV) at the plurality of elapsed time points and the value at the immediately preceding time point (ΔOC)
V), each difference (ΔVOO) between the value of the voltage (VO) immediately after the start of discharge of the secondary battery for learning start and the values of the plurality of elapsed time points, and the voltage (VO) immediately after the start of discharge ( (VO), the memory means for storing in advance information indicating the relationship between the battery state and the difference (ΔVO) between the values at the plurality of elapsed time points and the value at the immediately preceding time point, and the open circuit of the secondary battery to be judged. Voltage (OCV), voltage (VO) immediately after the start of discharge, each difference (ΔOCVO) between the value of the open circuit voltage (OCV) at the start of use of the secondary battery for learning and the values at the plurality of elapsed times, The open circuit voltage (OCV)
Difference (ΔOCV) between the values at the plurality of elapsed time points and the value at the immediately preceding time point, the voltage (V
O) each difference (ΔVOO) between the value at the start of use of the learning secondary battery and the values at the plurality of elapsed times, and the value of the voltage (VO) immediately after the start of discharge at the plurality of elapsed times. And the difference (ΔV
O) and means for predicting the battery condition using the information in the memory means, and
The open circuit voltage (OCV) of the judged secondary battery, the voltage (VO) immediately after the start of discharge, the value of the open circuit voltage (OCV) at the start time of use of the judged secondary battery, and the values at the plurality of elapsed time points. (ΔOCVO), the difference (ΔOCV) between the values of the open circuit voltage (OCV) at the plurality of elapsed time points and the value at the immediately preceding time point, and the voltage (VO) immediately after the start of the discharge. Each difference between the value at the start of use of the determination secondary battery and the values at the plurality of elapsed times (ΔV
OO), and a step of measuring a difference (ΔVO) between a value of the voltage (VO) immediately after the start of the discharge at the plurality of elapsed time points and a value at a time point immediately before this, and the judgment target obtained by the measurement. The open circuit voltage (OCV) of the secondary battery, the voltage (VO) immediately after the start of discharge, the value of the open circuit voltage (OCV) at the start of use of the judged secondary battery, and the plurality of elapsed time values. The difference (ΔOCVO), the value of the open circuit voltage (OCV) at the plurality of elapsed times, and
The difference (ΔOCV) from the value at the immediately preceding time point, the difference (Δ) between the value of the voltage (VO) immediately after the start of discharge of the secondary battery to be judged and the value at the plurality of elapsed times.
VOO), and the difference (ΔVO) between the values at the plurality of elapsed time points of the voltage (VO) immediately after the start of the discharge and the values at the previous time points, and the information in the memory means. Means for determining the battery state of the secondary battery to be determined. As described in claims 2 and 3 , the internal resistance R, or the difference ΔR between the internal resistance R and the internal resistance R from the start of use.
The determination can be made in consideration of the difference ΔR between O and the value at the immediately preceding time point.

The principle of the method for determining the state of a secondary battery according to the first and second aspects of the invention will be described with reference to FIG. FIG. 1 shows the change over time in the discharge curve of the secondary battery. 1A shows a discharge curve of the new battery, B shows a discharge curve of the secondary battery after aging, and C shows a discharge curve of the secondary battery after further aging. In FIG. 1, t0 indicates the discharge start point, Tb is the backup time required by UPS, and Th is half the backup time Tb. In the present application, the discharge is started at t0 and half the time T
It is defined that the life of the battery has ended when the battery voltage Vh at time t1 when h has elapsed becomes less than the discharge end voltage Ve. In this embodiment, since the backup time Tb is 10 minutes, the half time Th is 5 minutes.

As is apparent from FIG. 1, the open circuit voltage OCV1 before the discharge start time t0 according to the discharge curve A of the new battery shown by the solid line, that is, the battery voltage when the discharge circuit is opened is shown by the chain line and the dotted line. Open circuit voltage OC after aging
Lower than V2 and OCV3. On the other hand, the voltage VO immediately after starting the discharge at t0 is the highest value VO1 in the case of the new battery shown by the curve A, and the values VO2 and VO3 lower than VO1 in the case of the aged battery. In the invention of the present application, the open circuit voltage OCV of the battery and the voltage V immediately after the start of discharge
Paying attention to the secular change with O, at least one of the dischargeable time (capacity) Te of the secondary battery, the usable period (lifetime) L, and the voltage Vh after a predetermined time (for example, 5 minutes) from the start of discharge To decide. Further, in the invention of claims 1 and 2 of the present application, the dischargeable time Te, the usable period L, and the voltage Vh after a predetermined time from the discharge start are determined by paying attention to the change in the discharge curve after the discharge starts. . As is clear from the comparison of A, B and C in FIG. 1, the slope of the discharge curve becomes steeper as the secular change increases. Further, the internal resistance R used in claims 2 and 3 becomes high due to aging. Therefore,
In the present application, the change in internal resistance is used to determine the battery state.

By the way, the secular change varies depending on the environment in which the same secondary battery is used and the difference in performance of each secondary battery. For this reason, it is difficult to accurately determine the capacity and life of the secondary battery by the approximate expression. Therefore, in the invention of claim 4 , the characteristics of the secular change of the discharge curve shown in FIG. 1 are characterized by a neural network.
Accurate determination of the battery state can be performed by learning. The feature of this neural network is that it is a model of nerve cells in the brain of an animal, and it can be mapped by learning what cannot be represented by an algorithm.

When the state of the secondary battery is judged using the neural network, a learning secondary battery having substantially the same structure as the secondary battery to be judged is prepared, and data generally called teacher data (for example, capacity) is used. Te, life L, voltage Vh after a predetermined time from the start of discharge are given to the neural network, and open circuit voltage (hereinafter sometimes simply referred to as OCV) and voltage immediately after the start of discharge (hereinafter sometimes simply referred to as VO) Yes, the internal resistance of the secondary battery (hereinafter also referred to simply as R), the discharge voltage after 5 minutes of discharge (hereinafter also simply referred to as Vh), and the value at the time when the OCV first started to be used. And a plurality of elapsed time values (hereinafter, may be simply referred to as ΔOCVO), a difference from a value one time before the OCV (hereinafter, may be simply referred to as ΔOCV), and discharge start straight The difference between the later voltage VO and the value at the start of use (hereinafter, may be simply referred to as ΔVOO), and the difference with the value immediately before the VO (hereinafter simply referred to as ΔVO
It may be referred to as “(may be called)”, but a plurality of data selected from among these are input to the neural network. The neural network learns based on the given data, and determines the weight (weight) necessary for predicting the battery state (Te, L, Vh). This weight is based on the input data (OCV, VO, etc.) of the neural network and the battery state (T
e, L, Vh).
Therefore, when OCV, VO, etc. are measured from the secondary battery to be judged and input to the neural network, the battery state (eg Te, L, Vh) can be judged with relatively high accuracy.

[0011]

According to the inventions of the respective claims, the secondary battery to be judged is discharged for a long time, and the state of the battery (life, remaining capacity, etc.)
Is unnecessary, and the battery state can be determined by discharging in a relatively short time. Therefore, the determination result can be obtained in a short time. When the secondary battery is being used as a backup, the capacity of the secondary battery, that is, the backup time is not wasted. Also, ΔOCVO,
If ΔOCV, ΔVOO, and ΔVO are also taken into consideration, it is possible to make a judgment or prediction with a small error. Further, when R or R and ΔRO, ΔR are taken into consideration as in the second aspect of the present invention, the battery state can be predicted with higher accuracy. Further, as in the invention of claim 3 , the battery state can be predicted in a short time by R, ΔRO, and ΔR.

[0012]

[First Embodiment] Next, a secondary battery state determination method of the first embodiment will be described with reference to FIGS. In the first embodiment, as shown in FIG. 2, OCV and VO of a secondary battery composed of a lead storage battery are input to the neural network 1 and Vh, Te and L are obtained as outputs. Neural network 1 is a program ROM (read only memory) and RA
Mh (random access memory) and CPU (central processing unit) are included, and Vh, Te, and L are programs of CPU and ROM as predicting means based on the data accumulated in RAM by learning and OCV and VO. Predict. Figure 3
Shows a schematic circuit network of the neural network 1. The neural network 1 of this embodiment has one intermediate layer and five units, and follows the back propagation (error back propagation) method.

FIG. 4 shows in principle or functionally a state determining device for a secondary battery, wherein a charging circuit 3 is connected to a secondary battery 2 made of a lead storage battery via a switch 4. The load 5 is connected via the switch 6, and the voltage detecting means 7 is connected via the switch 8. An open circuit voltage extraction circuit 9 and a voltage extraction circuit 10 immediately after the start of discharge are connected to the output line of the voltage detection means 7. The open circuit voltage extraction circuit 9 turns off the switches 4 and 6, and the switch 8
When is turned on, OCV is extracted and output. Immediately after the start of discharge, the voltage extraction circuit 10 extracts VO and outputs it when the switch 4 is turned off and both the switches 6 and 8 are turned on. The neural network 1 equivalently includes a voltage determining means 11 after discharging for 5 minutes, a dischargeable time determining means 12, and a usable period determining means 13, and outputs Vh, Te and L. As is clear from the equivalent circuit of FIG. 4, the three decision means 11, 12, 13 are not only coupled to the input lines of OCV and VO, but also to each other. Note that Vh, Te, and L are used during learning for determining the weight of the neural network 1, but the voltage Vh, the capacity Te, and the life L after 5 minutes from the start of discharging the secondary battery to be determined are used.
When predicting, OC without using Vh, Te and L
Use only for V and VO. This also applies to other embodiments described later. In this way, when a plurality of pieces of information are taken into consideration for prediction, a highly accurate prediction result can be obtained.

When the neutral network 1 is used to predict Vh, Te, and L of a secondary battery to be judged which is a lead storage battery, first, data necessary for learning the neural network 1 is obtained. In order to obtain this learning data, the secondary battery 2 in FIG. 4 is a learning secondary battery composed of a lead storage battery. The learning secondary battery has substantially the same configuration as the to-be-determined secondary battery and has substantially the same discharge characteristics. Table 1 shows data for learning. In Table 1 and Tables 2 to 6 described later, NO. Is a sample number, and NO. 1 and NO. 2 are samples of two secondary batteries. M indicates the period of use of the battery for examining the secular change in months. Note that the secular change is examined with the switch 6 turned off and no discharge (non-loaded state) except for discharge for measurement. Further, although the switch 4 is turned off during the measurement period, the switch 4 is kept on during the period other than the measurement period. OCV is an open circuit voltage, VO is a voltage immediately after the start of discharge, R is an internal resistance, Vh is a voltage 5 minutes after the start of discharge, Te is a dischargeable period (capacity), and L is a usable period (life).

[0015]

[Table 1]

The learning data in Table 1 shows that the new learning secondary battery is connected in the same manner as in FIG. 4, and the OCV, after the start of use, that is, after the start of use and at a plurality of elapsed time points after that,
VO and R are obtained, and Vh and Te are measured at the start of use and at each elapsed time. That is, when the learning data is obtained, the secondary battery is actually discharged for 5 minutes, the battery voltage is measured, and this is taken as Vh. Further, when Te is obtained, the battery voltage is the discharge end voltage V from the discharge start time.
Measure up to the point of e. L was calculated by subtracting the elapsed time from the actually measured life. In addition, M
When the discharge is completed once at each value of, the state is returned to the full charge state again, and the switch 4 is kept on to keep the charge state.

When the data in Table 1 is obtained, the learning data is given to the CPU of the neural network 1 offline in the order of the elapsed time indicated by M. That is, at each time point, Vh, Te and L are given as teacher data (desired data) of the neural network 1, and OCV and VO are given as input data in the case of the first embodiment. Although FIG. 4 does not show a means for measuring the period of use M, this measuring means is composed of a counting circuit built in the neural network 1 and has a period of use M.
Data is also given to the CPU. Of course, the measuring means of M can be provided outside the neural network 1.
When the data for learning is given to the neural network 1, the weight (weight) necessary to obtain the teacher data can be obtained based on the input data OCV and VO. That is, when the principle of the neural network is extremely simplified and schematically shown, W1 * OCV + W2 * VO = Vh W3 * OCV + W4 * VO = Te W5 * OCV + W6 * VO = L, and in this relational expression, Vh, Te, L, OC
If V and VO are known, the weights W1 to W6 can be determined. In learning, the weights W1 to W6 are determined at each elapsed time, that is, at various values of M, and the ROM or RA is determined.
Accumulate in M.

When the weight is determined by learning, it is possible to determine Vh, Te and L of the secondary battery to be determined. Therefore, a secondary battery to be determined is connected to the secondary battery 2 in FIG. 4, and OCV and VO of the secondary battery to be determined are input to the neural network 1 in the same use period as M in Table 1. The OCV and VO are measured in the same manner as during learning. The neural network 1 predicts Vh, Te, and L of the secondary battery to be judged based on the weight accumulated by learning. Table 2 below shows the prediction or judgment results of Vh, Te, and L. It should be noted that the secondary battery to be judged is returned to the fully charged state at each completion of the judgment, the switch 4 is kept in the ON state, and the switch 6 is kept in the OFF state. As described above, when determining Vh, Te, and L of the secondary battery to be determined, the determination outputs of Vh, Te, and L are not correlated with each other in the three determining means 11, 12, and 13 of FIG.

[0019]

[Table 2]

As is clear from the comparison between the actually measured values shown in Table 1 and the judgment results shown in Table 2, it is possible to obtain a judgment result that is close to the actually measured values in both sample No. 1 and NO. Note that the error of the determination result with respect to the actual measurement value is the period of use M
Becomes smaller as it becomes larger.

[0021]

Second Embodiment Next, with reference to FIG. 5, a secondary battery state determination method of the second embodiment will be described. However, in the second embodiment and the third to fifth embodiments to be described later, description of matters common to the first embodiment and each embodiment will be omitted, and only different points will be described.

In the second embodiment, ΔOCVO, ΔOCV and ΔVO are used as input data to be given to the neural network 1.
The third embodiment is different from the first embodiment in that O and ΔVO are added, and is otherwise the same. Therefore, the determination device of FIG. 5 has the ΔOCVO forming circuits 9a and ΔO connected to the open circuit voltage extracting circuit 9.
Δ connected to the CV forming circuit 9b and the VO extracting circuit 10
It has a VOO forming circuit 10a and a ΔVO forming circuit 10b, which are also connected to the neural network 1 and differ from the determination device of FIG. 4, and are otherwise the same.
In addition, in the neural network 1 of FIG. 5, the intermediate layer is 1
The number of units in each layer is eight.

The ΔOCVO forming circuit 9a has an open circuit voltage OC
Difference ΔOCV between the value at the start of use of V and the value at an arbitrary time point
Form O. Referring to FIG. 1, ΔOCVO is OC
V1-OCV2 or OCV1-OCV3. Further, the ΔOCV forming circuit 9b forms a difference ΔOCV between the open circuit voltage OCV and the value at the immediately preceding time point. Referring to FIG. 1, ΔOCV is OCV1 −OCV2 or OCV2 −.
It is OCV3. In addition, the ΔVOO forming circuit 10a determines the difference ΔVO between the value at the start of use of the voltage VO immediately after the start of discharge.
Form O. Referring to FIG. 1, ΔVOO is VO1 −
VO2 or VO1 -VO3. Further, the ΔVO forming circuit 10b forms a difference ΔVO between the voltage VO immediately after the start of discharge and the value one time before. Referring to FIG. 1, ΔV
O is VO1 -VO2 or VO2 -VO3.

Vh, Te, L using the determination device of FIG.
When determining or predicting, the neural network 1 is made to perform the same learning as described in the first embodiment. In the case of the second embodiment, the input data OCV, V shown in Table 1
In addition to O, Vh, Te, and L, ΔOCVO, ΔOCV, ΔVOO, and ΔVO formed based on OCV and VO are also given to the neural network 1 offline. As a result, the weight in the neural network 1 is determined as in the first embodiment. Next, a secondary battery to be determined is connected to the secondary battery 2 in FIG. 5, and OCV, ΔOCVO, ΔOCV, VO, at the time points indicated by the use periods M shown in Table 1 and Table 3,
After obtaining ΔVOO and ΔVO, these are input to the learned neural network 1. As a result, Vh is calculated based on the weight accumulated in the neural network 1.
, Te, L are determined. Table 3 below shows the result of this determination.

[0025]

[Table 3]

As is clear from the comparison between the measured values of Table 3 and Table 1 and the judgment values of the first embodiment of Table 2, ΔOCVO, Δ
The determination error becomes smaller than that in the first embodiment by additionally determining OCV, ΔVOO, and ΔVO.

[0027]

[Third Embodiment] The secondary battery state determination method of the third embodiment is different from the first embodiment in that the neural network 1 inputs data indicating the internal resistance R in addition to OCV and VO. Are the same. The determination device of the third embodiment shown in FIG. 6 has an internal resistance measuring circuit 20 instead of the same circuit as in FIG. The internal resistance measuring circuit 20 measures the internal resistance R at each time point of the secondary battery 2 and sends it to the neural network 1. Note that this data can be used when the mutual relationship between the internal resistance R and the use period M is known in advance by another means.

The learning of the neural network 1 having the intermediate layer of one phase and five unit layers is based on the learning secondary battery.
CV, VO, R, Vh, Te and L are given and the same procedure as in the first embodiment is performed to determine the weight of the neural network 1 and store it in the memory. Vh, Te of secondary battery to be judged
, L, OCV, VO, and R are input to the learned neural network 1 for each use period M. As a result, OCV, Vth, Te, and L are determined.
R is used as well as VO. Since the weights for OCV, VO, and R are determined by learning, Vh,
It is possible to quickly determine Te and L. The internal resistance R increases as the usage period of the secondary battery increases, and has a correlation with the usage period and the life. Therefore, the accuracy of determination of Vh, Te, and L increases by considering R. Table 4 below
Shows the results of the third embodiment.

[0029]

[Table 4]

[0030]

[Fourth Embodiment] FIG. 7 shows a part of a battery state determination device according to a fourth embodiment. In the embodiment of FIG. 7, OCV, ΔOCVO, and ΔOC are added to the neural network 1 as in the embodiment of FIG.
In addition to inputting V, VO, ΔVOO, ΔVO, R, ΔR
Input O and ΔR and obtain Vh, Te and L as outputs. OCV, ΔOCVO, ΔOCV, VO, ΔV in FIG.
OO and ΔVO are formed similarly to the circuit of FIG. The internal resistance R is obtained by a circuit similar to the internal resistance measuring circuit 20 shown in FIG.
ΔRO is the difference between the value of the internal resistance R at the start of use and the value at the predetermined elapsed time, and ΔR is the difference between the value at the measurement time immediately before the internal resistance R. ΔRO and ΔR are calculated and determined based on the output of the internal resistance measuring circuit 20 as shown in FIG.

When the neural network 1 is trained, OCV and ΔO in each use period M similar to those in Table 1 are obtained.
CVO, ΔOCV, VO, ΔVOO, ΔVO, R, ΔR
Data of O and ΔR were obtained, and Vh and T shown in Table 1 and these were obtained.
The weights are determined by giving e and L to the CPU of the neural network 1 to learn the neural network 1. The neural network 1 shown in FIG. 7 has one intermediate layer and eight units. Vh of secondary battery to be judged
, Te, L, when measuring the OC of the secondary battery
V, ΔOCVO, VO, ΔVOO, ΔVO, R, ΔR
O and ΔR are obtained and given as input data to the learned neural network 1. Since the weights for determining Vh, Te and L have already been determined, Vh
, Te, L can be obtained quickly. In this embodiment, since a large amount of input data is given to the neural network 1, it is possible to make a more accurate determination.
Table 5 below shows the determination results of the fourth embodiment.

[0032]

[Table 5]

[0033]

[Fifth Embodiment] The battery state determining apparatus of FIG. 8 is formed to determine Vh, Te and L based on the internal resistance R of the secondary battery and its change. Therefore, FIG. 8 has an internal resistance detection circuit 20 similar to that of FIG. Further, in order to improve the determination accuracy of Vh, Te, and L based on the internal resistance R, Δ
An RO forming circuit 20a and a ΔR forming circuit 20b are provided. The ΔRO formation circuit 20a is the internal resistance detection circuit 20.
, And calculates ΔRO based on R. ΔRO represents the difference between the internal resistance value at each lapse of the use period M of the secondary battery shown in Tables 1 and 6 and the internal resistance value at the start of use (M = 0), and ΔR represents the internal resistance. The difference between the value at each elapsed time and the value at the immediately preceding time is shown. For example, if the internal resistance of the secondary battery having the characteristics of curves A, B, and C in FIG. 1 is R1, R2, and R3, ΔRO is R1−R2.
And R1 -R3, and ΔR is R1 -R2 and R2 -R
3 The neural network 1 shown in FIG. 8 has one intermediate layer and five units.

When determining Vh, Te, and L in the determination device of FIG. 8, first, the internal resistance R of the learning secondary battery during the use period M is obtained as shown in Table 1, and from this, ΔRO and ΔR are calculated. Ask. Next, the weights are determined by giving the neural network 1 the measured values of R, ΔRO, ΔR of the secondary battery for learning and Vh, Te, and L of Table 1 and learning the neural network 1. When determining Vh, Te, and L of the secondary battery for determination, the secondary battery 2 of FIG. 8 is used as the secondary battery for determination, the internal resistance R is measured in each of a plurality of use periods M, and ΔROO, Neural network 1 for ΔROO and ΔR
To enter. The neural network 1 determines Vh, Te, and L using the weight determined by learning.
Table 6 below shows the determination results of the fifth embodiment.

[0035]

[Table 6]

[0036]

MODIFICATION The present invention is not limited to the above-mentioned embodiments, and the following modifications are possible. (1) Many discharge curves similar to the discharge curves A, B and C as shown in FIG.
And a life L and a capacity Te are shown in a table, which is stored in a memory. When OCV and VO of the secondary battery to be judged are detected, L or Te corresponding to OCV and VO is detected based on the table of the memory. Can be read to determine L, Te. Also in this case, since OCV and VO are used for the determination, L and Te can be determined without discharging the secondary battery to be determined for a long time. (2) Alarm means can be connected to the output lines of Te and L of the neural network 1 to generate an alarm when Te and L reach predetermined values.

[Brief description of drawings]

FIG. 1 is a diagram showing discharge characteristics of a secondary battery.

FIG. 2 is a block diagram showing a neural network according to the first embodiment of the present invention.

FIG. 3 is a schematic diagram of the neural network of FIG.

FIG. 4 is a block diagram showing a battery state determination device of the first embodiment.

FIG. 5 is a block diagram showing a battery state determination device of a second embodiment.

FIG. 6 is a block diagram showing a battery state determination device of a third embodiment.

FIG. 7 is a block diagram showing a neural network according to a fourth embodiment.

FIG. 8 is a block diagram showing a battery state determination device of a fifth embodiment.

[Explanation of symbols]

1 Neural network 2 Secondary battery

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) G01R 31/36

Claims (4)

(57) [Claims] 1. A method for determining at least one of the remaining capacity, life, and voltage of a secondary battery after a lapse of a predetermined time from the start of discharge, which is a secondary battery to be determined. The open circuit voltage (OCV) at a plurality of time points after the start of use of the learning secondary battery having substantially the same configuration as
O), respective differences (ΔOCVO) between the value of the open circuit voltage (OCV) at the start time of use of the secondary battery for learning and the values of the plurality of elapsed time points, and the plurality of elapsed times of the open circuit voltage (OCV). The difference between the value at the time point and the value at the immediately preceding time point (ΔOCV), the voltage immediately after the start of the discharge (VO)
The respective differences (ΔVOO) between the value at the start of use of the learning secondary battery and the values at the plurality of elapsed times, and the values at the plurality of elapsed times of the voltage (VO) immediately after the start of discharge and Memory means for storing in advance information indicating the relationship between the difference (ΔVO) from the value at the immediately preceding time point and the battery state, the open circuit voltage (OCV) of the secondary battery to be judged, the voltage immediately after the start of discharge (VO ), The respective difference (ΔOCVO) between the value of the open circuit voltage (OCV) at the start of use of the learning secondary battery and the values of the plurality of elapsed time points, and the plurality of elapsed time points of the open circuit voltage (OCV). And the value in
The difference (ΔOCV) from the value at the immediately preceding time point, and the difference (Δ) between the value at the time when the learning secondary battery starts to be used and the value at the plurality of elapsed time points of the voltage (VO) immediately after the start of discharge.
VOO), and the difference (ΔVO) data between the values of the voltage (VO) immediately after the start of the discharge at the plurality of elapsed time points and the values at the immediately preceding time points, and the information of the memory means. A step of preparing a means for predicting the battery state, an open circuit voltage (OCV) of the secondary battery to be judged, a voltage (VO) immediately after the start of discharging, and a secondary battery of the open circuit voltage (OCV) to be judged. Differences (ΔOCVO) between the value at the start of use and the values at the plurality of elapsed times, and a difference (ΔOCV) between the values at the plurality of elapsed times of the open circuit voltage (OCV) and the values at the immediately preceding time points. ), The difference (ΔV) between the value of the voltage (VO) immediately after the start of discharge of the secondary battery to be judged at the start of use and the values at the plurality of elapsed times.
OO), and the step of measuring the difference (ΔVO) between the values of the voltage (VO) immediately after the start of the discharge at the plurality of elapsed time points and the values at the immediately preceding time points, and the judgment target obtained by the measurement. The open circuit voltage (OCV) of the secondary battery, the voltage (VO) immediately after the start of discharge, the value of the open circuit voltage (OCV) at the start of use of the judged secondary battery, and the plurality of elapsed time values. Difference (ΔOC
VO), the difference between the values of the open circuit voltage (OCV) at the plurality of elapsed time points and the value at the immediately preceding time point (ΔOC)
V), each difference (ΔVOO) between the value of the voltage (VO) immediately after the start of discharge of the secondary battery to be judged and the values of the plurality of elapsed time points, and the voltage (VO) immediately after the start of discharge ( VO) at the plurality of elapsed time points and the difference (ΔVO) between the value at the previous time point and the value at the previous time point, and the information in the memory means, the means for predicting the value of the secondary battery to be determined. A method of determining the state of a secondary battery, comprising the step of determining the state of the battery. 2. A method for determining a state of at least one of a remaining capacity, a life, and a voltage of a predetermined time after the start of discharge of a secondary battery, the secondary battery to be determined for performing the determination. The open circuit voltage (OCV) at a plurality of time points after the start of use of the learning secondary battery having substantially the same configuration as
O), respective differences (ΔOCVO) between the value of the open circuit voltage (OCV) at the start time of use of the secondary battery for learning and the values of the plurality of elapsed time points, and the plurality of elapsed times of the open circuit voltage (OCV). The difference between the value at the time point and the value at the immediately preceding time point (ΔOCV), the voltage immediately after the start of the discharge (VO)
The respective differences (ΔVOO) between the value at the start of use of the learning secondary battery and the values at the plurality of elapsed times, and the values at the plurality of elapsed times of the voltage (VO) immediately after the start of discharge and Difference from the value at the immediately preceding time point (ΔVO),
Internal resistance (R), a difference (ΔRO) between each value of the internal resistance (R) at the start of use of the secondary battery for learning and the values of the plurality of elapsed times, and the internal resistance (R) Memory means for storing in advance information indicating the relationship between the battery state and the difference (ΔR) between the values at the plurality of elapsed time points and the value at the immediately preceding time point, and the open circuit voltage of the secondary battery to be determined ( OCV), the voltage (VO) immediately after the start of discharge, and the difference (Δ) between the value of the open circuit voltage (OCV) at the start of use of the secondary battery to be determined and the values at the plurality of elapsed times.
OCVO), the difference between the values of the open circuit voltage (OCV) at the plurality of elapsed time points and the value at the immediately preceding time point (ΔO
CV), respective differences (ΔVOO) between the value of the voltage (VO) immediately after the start of discharge of the secondary battery to be judged and the values at the plurality of elapsed time points, and the voltage (VO) immediately after the start of discharge ( VO), the difference (ΔVO) between the values at the plurality of elapsed time points and the value at the immediately preceding time point, the internal resistance (R), and the internal resistance (R) at the start of use of the secondary battery to be determined. A difference (ΔRO) between each of the values and the values at the plurality of elapsed time points, and a difference (ΔR) between the values of the internal resistance (R) at the plurality of elapsed time points and the value at the immediately preceding time point. A step of preparing means for predicting the battery state using the data and the information of the memory means, a starting voltage (OCV) of the secondary battery to be judged, a voltage (VO) immediately after the start of discharging, Use of the secondary battery to be judged having an open circuit voltage (OCV) Differences (ΔOCVO) between the value at the start time and the values at the plurality of elapsed times, and a difference (ΔOCV) between the values at the plurality of elapsed times of the open circuit voltage (OCV) and the values at the immediately preceding time points. , The difference (ΔV) between the value of the voltage (VO) immediately after the start of discharge of the secondary battery to be measured and the value of the plurality of elapsed times.
OO), and the difference (ΔVO) between the values at the plurality of elapsed time points of the voltage (VO) immediately after the start of discharge and the values at the immediately preceding time points, the internal resistance (R), and the internal resistance (R). Differences (ΔRO) between the values at the start of use of the secondary battery to be determined and the values at the plurality of elapsed times, and the values at the plurality of elapsed times of the internal resistance (R) and
Measuring the difference (ΔR) from the value at the immediately preceding time point, the measured open circuit voltage (OCV) of the secondary battery to be determined,
Voltage (VO) immediately after the start of discharge, the open circuit voltage (OCV)
Difference (ΔOCVO) between the value at the start of use of the secondary battery to be determined and the values at the plurality of elapsed times, and the values at the plurality of elapsed times of the open circuit voltage (OCV) and A difference (ΔOCV) from the time value, a difference (ΔVOO) between the value of the voltage (VO) immediately after the start of discharge of the secondary battery to be judged and the values of the plurality of elapsed times, and Voltage (VO) immediately after the start of discharge
Of the difference (ΔVO) between the values at the plurality of elapsed time points and the value at the immediately preceding time point, the internal resistance (R), and the internal resistance (R) at the start of use of the secondary battery to be determined. The respective differences (ΔRO) from the values at the plurality of elapsed times, and the differences (ΔR) between the values at the plurality of elapsed times of the internal resistance (R) and the values at the immediately preceding times, and the memory means. And a step of determining the battery state of the secondary battery to be determined by the predicting means using the information and. 3. A method of determining at least one of the remaining battery capacity, life, and voltage after a lapse of a predetermined time from the start of discharge of the secondary battery, the secondary battery being a judgment target for making the judgment. The internal resistance (R) at a plurality of elapsed time points after the start of use of the secondary battery for learning having substantially the same configuration, and the value of the internal resistance (R) at the start of use of the secondary battery for learning and the plurality of values. Of the internal resistance (R) and the difference (ΔR) between the values of the internal resistances (R) at the plurality of elapsed times and the value at the immediately preceding time and the battery state. Memory means for storing information indicating a relationship in advance, an internal resistance (R) of the secondary battery to be determined, a value of the internal resistance (R) at a start time of use of the secondary battery to be determined, and the plurality of elapsed time points. Difference (ΔRO) with the value of Means for predicting the battery state using data indicating a difference (ΔR) between the values of (R) at the plurality of elapsed time points and the value at the immediately preceding time point and the information of the memory means. And a step of preparing the internal resistance (R) of the secondary battery to be determined, and a difference between the value of the internal resistance (R) at the start of use of the secondary battery to be determined and the values at the plurality of elapsed times. (ΔRO), and a step of measuring a difference (ΔR) between a value of the internal resistance (R) at the plurality of elapsed time points and a value at a time immediately before the plurality of elapsed times; and an internal resistance ( R), the respective differences (ΔRO) between the values of the internal resistance (R) at the start of use of the secondary battery to be determined and the values of the plurality of elapsed times, and the plurality of internal resistances (R). Difference (ΔR) between the value at the elapsed time and the value at the immediately preceding time And a step of determining the battery state of the to-be-determined secondary battery by the predicting means using the information of the memory means and the method of determining the state of the secondary battery. 4. The memory means and the predicting means are
Claim 1 or 2 or state judging method of the secondary battery 3 Symbol mounting, characterized in that a neural network.