JP5467010B2 - Learning-type storage battery management system - Google Patents

Description

  The present invention relates to a learning storage battery management system that manages charge / discharge of a storage battery.

  In recent years, the introduction of distributed power sources using renewable energy such as solar power generation and wind power generation has become active, and there is a concern about destabilization of the system. On the other hand, the spread of vehicles such as electric vehicles and plug-in hybrid vehicles (pHV) that use electric power from storage batteries as an energy source is expanding. As described above, not only distributed power sources that supply power but also devices that store power are increasing. Therefore, using storage batteries installed in electric vehicles, pHV, etc., to prevent future system instability and to prevent uneven supply and demand balances in the system to contribute to power stabilization measures on the demand side Strengthening cooperation between the energy management system and the power supply to the in-vehicle storage battery is required more than ever.

  Here, in strengthening the cooperation with the above regional energy management system, surplus power and nighttime power from solar power generation are utilized for each household by utilizing a solar power generation system that is gradually spreading to ordinary households and business entities. Or, in addition to i) self-consumption of the electric power stored in storage batteries installed in business units or regional units, ii) so-called “local production for local consumption model” Is required as a premise first, and in that case, for in-vehicle storage batteries, i) further improvement of its charge / discharge control technology and life diagnosis technology And ii) Further improvement and establishment of storage battery management technology utilizing information communication technology is a major issue.

  As described above, in order to realize smooth power supply to the in-vehicle storage battery while cooperating with the above regional energy management system, first, further improvement of the storage battery charge / discharge control technology and lifetime diagnosis technology, and There was an urgent need to further improve storage battery management technology using information and communication technology.

  By the way, conventionally, by combining an RFID as a communication means with a storage battery, i) identifying the storage battery, ii) managing information related to charging / discharging of the storage battery, and iii) managing the deterioration state of the storage battery is a patent document. 1 is disclosed.

FIG. 14 is a diagram illustrating a charging device and a charging system disclosed in Patent Document 1. Four mobile phone cradles 600, 621, 622, and 623 are connected to the host computer 631, and a storage unit 601 that stores management history information 602 is provided inside one cradle 600. In addition, the control unit 605 includes a charge processing unit 606, a charge management unit 607, and a deterioration determination unit 608, a charging unit 609, a panel display unit 610, a speaker 611, an RFID processing unit 603, and an antenna 604. Reference numeral 632 denotes a control unit of the host computer.
Further, there are an AC adapter 649 and an AC power source 650 as power sources, and the reader / writer 660 includes a battery 661 and an RFID tag 662 integrated with the battery.

  With the charging device and the charging system configured as described above, it is possible to determine the deterioration state of the battery of the reader / writer and perform management so that the deterioration of a specific battery among the plurality of batteries is not concentrated. .

  Patent Document 2 relates to an assembled battery 10 that is composed of a plurality of secondary battery modules 21 to 25 connected in series and is used as a power source for a vehicle such as an electric vehicle or a plug-in hybrid vehicle (see Patent Document 2). 1 and FIG. 2), a method is disclosed in which the assembled battery 10 is collected from the user side, reconfigured as an assembled battery 10 having sufficient performance, and then returned to the user side. .

More specifically,
i) A secondary battery (module) that can be reused in the reconfiguration of the assembled battery 10 is used as a reusable object of itself or another assembled battery 10, and the rechargeable secondary battery (module) is discharged to a fully discharged state. A discharging step,
ii) a storage step of storing the secondary battery (module) to be reused after the discharge step in a storage area so that a predetermined time or more elapses from the complete discharge state;
iii) The secondary battery (module) to be reused by itself or another assembled battery 10 that has undergone the above storage step, or the secondary battery (module) to be reused and a new secondary battery (module) And a regenerating step for reconfiguring the assembled battery 10;
A reconfiguration method of the assembled battery 10 including the above is disclosed.

  In Patent Document 3, as a method for diagnosing the deterioration state of a lithium ion storage battery or a method for suppressing deterioration of a lithium ion storage battery, an upper limit value of the charging current is set according to the degree of voltage increase during charging of the storage battery. A technique for changing is disclosed.

  Patent Document 4 discloses a technique using a Bayesian network as an example of configuring a learning database.

JP 2006-172848 A JP 2009-277627 A JP 2010-60408 A WO 2006/006533

Kazuo Shigeru, Masaomi Ueno, Yoichi Motomura (co-author), “Outline of Bayesian Network”, Baifukan, 2006. Bayesian network analysis system “Bayonet” manual, Mathematical Systems Co., Ltd., 2009.

However, in the method described in Patent Document 1, only the deterioration determination of the portable device battery is performed, and it is not assumed to suppress the deterioration of the storage battery or to handle the storage battery when it is determined to be deteriorated. In fact, Patent Document 1 does not disclose or suggest such contents.
Also regarding Patent Document 3, it is not assumed that the storage battery is handled when it is determined to be deteriorated. In fact, Patent Document 3 does not disclose or suggest such contents.
In addition, patent document 1 deals with a portable device battery, and it is applied to a vehicle storage battery, a household storage battery, a building storage battery, a collective housing storage battery, or a storage battery for a community energy management system including an electric vehicle larger than this. There has been no detailed study of the application.

  Moreover, in patent document 2, in order to recover charging / discharging characteristics (dissolving a memory effect) after complete discharge, the standing time of one week or more is required.

  Therefore, it is necessary to perform optimal charge / discharge control for the storage battery according to the deterioration state of the storage battery without leaving or storing the storage battery. It is also important to find an optimal secondary usage destination for deteriorated storage batteries. Furthermore, in the future, a large number of installation-type storage batteries for electric vehicles, plug-in hybrid vehicles, homes, buildings, etc. will provide a wide variety of usage data regarding storage batteries. It is very useful to meet the above requirements. However, a system that fully meets such demands has not been provided to the world so far.

Therefore, this invention makes it a main subject to provide the learning type storage battery management system which can charge / discharge a storage battery according to the deterioration state of a storage battery.
Moreover, this invention makes it a secondary subject to determine the residual value of a storage battery, and to provide the learning type storage battery management system which can recommend the optimal secondary usage place of the deteriorated storage battery.

  As a result of various studies to solve the above problems, the present inventor has found that the fixed information of the storage battery, the fluctuation information corresponding to the use status data of the storage battery, that is, the characteristic variables such as the number of times of charging / discharging, the discharge load, and the temperature In addition to formulating a charge / discharge plan for the storage battery by processing environmental variables such as history, vibration history, and travel history using a learning database, the storage battery is further diagnosed for deterioration or recommended for the optimal secondary usage. The present invention has been completed by finding out that the above-described problems can be solved by adopting the configuration to be performed.

The learning type storage battery management system according to claim 1 capable of solving the above problems is (1) a learning type storage battery management system,
A charging / discharging device for charging / discharging the storage battery;
Receiving means for receiving fixed information and fluctuation information of the storage battery;
By storing the fixed information and fluctuation information of the storage battery in a learning database, it is classified into a plurality of charge / discharge characteristic groups,
The charge / discharge characteristic group of the storage battery identified by the fixed information is extracted from the plurality of charge / discharge characteristic groups,
Define the storage battery charge / discharge control rule from the extracted storage battery charge / discharge characteristics group,
Processing means for formulating a charge / discharge plan of the storage battery using the obtained charge / discharge control rule,
A transmission means for transmitting a charge / discharge control signal instructing charge / discharge based on a charge / discharge plan established by the processing means to a charge / discharge device for charging / discharging the storage battery identified by the fixed information;
It is characterized by comprising.

Moreover, the learning type storage battery management system according to claim 2 of the present application is (2) the learning type storage battery management system according to claim 1,
The charge / discharge control rule has a charge / discharge control pattern for suppressing deterioration of the storage battery,
The processing means formulates a charge / discharge plan according to the charge / discharge control rule defined by the charge / discharge characteristic group of the storage battery identified by the fixed information extracted using the learning database. To do.

Moreover, the learning storage battery management system according to claim 3 of the present application is (3) a learning storage battery management system,
A charging / discharging device for charging / discharging the storage battery;
Receiving means for receiving fixed information and fluctuation information of the storage battery;
By storing the fixed information and fluctuation information of the storage battery in a learning database, it is classified into a plurality of charge / discharge characteristic groups,
Extracting the charge / discharge characteristic group of the storage battery identified by the fixed information from the plurality of charge / discharge characteristic groups,
Define the storage battery charge / discharge control rule from the extracted storage battery charge / discharge characteristics group,
Processing means for formulating a charge / discharge plan of the storage battery using at least the prescribed charge / discharge control rule;
The fixed information and the fluctuation information, and a plurality of charge / discharge characteristic groups of the storage battery obtained by accumulating and classifying the information, a charge / discharge control rule of the storage battery defined by the charge / discharge characteristic group, Storage means for storing a charge / discharge plan of the storage battery formulated using a charge / discharge control rule;
A charging / discharging device that charges / discharges the storage battery identified by the fixed information, and a transmission unit that transmits a charging / discharging control signal that instructs charging / discharging based on a charging / discharging plan formulated by the processing unit;
The said processing means prescribes | regulates the charging / discharging rule of the said storage battery with the charging / discharging control pattern which aims at the deterioration suppression of the said storage battery.

  As in the invention according to the above (2) or (3), by using at least the charge / discharge control rule of the storage battery defined by the charge / discharge characteristic group of the storage battery extracted using the learning type database, by the fixed information By formulating a charging / discharging plan for the identified storage battery, it is possible to charge / discharge the storage battery with an optimal pattern for suppressing deterioration of the storage battery.

Moreover, the learning type storage battery management system according to claim 4 of the present application is (4) the learning type storage battery management system according to any one of claims 1 to 3,
The processing means selects one charge / discharge control rule corresponding to the extracted charge / discharge characteristic group of the storage battery based on a learning result from a plurality of different charge / discharge control rules, thereby controlling the charge / discharge of the storage battery. It is characterized by defining rules.

Moreover, the learning type storage battery management system according to claim 5 of the present application is (5) the learning type storage battery management system according to any one of claims 1 to 4,
The processing means includes
Storage battery characteristic analysis determination means for analyzing and determining characteristics of each of the storage batteries from the charge / discharge characteristic group of the extracted storage batteries, and analysis determination result output means for outputting the determination results to the outside,
Have
The analysis determining means determines an estimated residual value or a recommended use when assuming secondary use of the storage battery.

  By providing such analysis determination means and analysis determination result output means, the results of the analysis and determination of the storage battery deterioration state, the estimated residual value when the secondary use of the storage battery is assumed, or other storage battery characteristics, It is possible to output to the outside through the analysis determination result output means.

Moreover, the learning type storage battery management system according to claim 6 of the present application is (6) the learning type storage battery management system according to claim 5,
The analysis determination means assumes the secondary use of the storage battery using the learning database that extracts the charge / discharge characteristic group of the storage battery identified by the fixed information and defines the charge / discharge control rule of the storage battery. In this case, the estimated residual value or recommended use is determined.

Moreover, the learning type storage battery management system according to claim 7 of the present application manages (7) charging / discharging of the storage battery in cooperation with a local energy management system that adjusts the supply and demand balance between the power supply side and the power demand side. A learning-type storage battery management system according to claim 1,
In addition to using the charge / discharge control rules when formulating the charge / discharge plan, the charge / discharge plan is formulated using one or both of supply and demand adjustment information from the local energy management system and the charge rate of the storage battery, thereby It is characterized by contributing to the supply-demand balance.

  Thereby, in addition to using the charge / discharge control rule when formulating the charge / discharge plan, further formulating the charge / discharge plan using either or both of supply and demand adjustment information from the regional energy management system and the charge rate of the storage battery. This makes it possible to optimize the balance of electric energy supply and demand.

Moreover, the learning type storage battery management system according to claim 8 of the present application is (8) the learning type storage battery management system according to claim 7,
The quasi-variation information and the fixed information of the storage battery are associated with the contract information of the user who is the user at the start of use or replacement of the storage battery.

Moreover, the learning type storage battery management system according to claim 9 of the present application is (9) the learning type storage battery management system according to any one of claims 1 to 8,
The storage battery is constituted by an assembled battery, and when a change occurs in part or all of the configuration, the charge / discharge plan that has been formulated for each storage battery is configured to be formulated again. It is a feature.

Moreover, the learning type storage battery management system according to claim 10 of the present application is (10) the learning type storage battery management system according to any one of claims 1 to 9,
Including the individual identification number of the storage battery as the fixed information,
Including environmental variables and feature variables as the variation information;
In the learning database,
An individual identification number node of the storage battery as a child of the environment variable node;
A characteristic variable node of the storage battery as a child of the individual identification number node of the storage battery;
The individual storage battery is identified by a storage battery individual identification rule expressed by a Bayesian network.

Moreover, the learning type storage battery management system according to claim 11 of the present application is (11) the learning type storage battery management system according to any one of claims 1 to 10,
The fixed information includes a part or all of a storage battery individual identification number composed of a type of storage battery classified by a storage principle, a manufacturing number, a manufacturing date, a weight at the start of service, and a charge amount. Is.

Moreover, the learning type storage battery management system according to claim 12 of the present application is (12) the learning type storage battery management system according to any one of claims 1 to 11,
The variation information includes an environment variable, a characteristic variable, a part of or all of a current charge amount and a chargeable capacity.

Moreover, the learning type storage battery management system according to claim 13 of the present application is (13) the learning type storage battery management system according to claim 12,
The environmental variable includes part or all of a temperature history, a humidity history, a vibration history, and a travel history of the storage battery.

Moreover, the learning type storage battery management system according to claim 14 of the present application is (14) the learning type storage battery management system according to claim 12 or 13,
The characteristic variable includes a part of or all of a service life, a charge / discharge count, a discharge load, a charge status history, a discharge status history, a current charge / discharge rule, and a weight of the storage battery.

Moreover, the learning type storage battery management system according to claim 15 of the present application is (15) the learning type storage battery management system according to claim 8,
The quasi-variation information is
In-vehicle storage batteries including electric vehicles, household storage batteries, building storage batteries, housing batteries for collective housing, storage batteries for local energy management systems and other uses of the storage batteries, and
Usage status of the storage battery due to household composition,
It is characterized by including at least one of these.

According to the present invention, a charge / discharge characteristic group of a storage battery identified by fixed information is extracted using a learning database, and a charge / discharge plan is formulated based on the extracted charge / discharge characteristic group. A learning-type storage battery management system capable of discharging can be provided.
Further, according to the present invention, the residual value of the storage battery is determined by analyzing and determining the characteristics of each storage battery from the extracted charge / discharge characteristics group of the storage battery, and the optimum secondary battery of the deteriorated storage battery is determined. It is possible to provide a learning type storage battery management system that can recommend a user.

1 is a diagram illustrating an entire system according to a first embodiment of the present invention. It is a figure which shows a local energy management system connection block part. It is another figure explaining about a regional energy management system connection block part. It is a figure which shows the prediction analysis block part of the determination / utilization scene of a charging / discharging control rule. It is another figure explaining about the prediction analysis block part of determination / utilization scene of a charging / discharging control rule. It is a figure which shows a storage battery control management block part. It is another figure explaining about a storage battery control management block part. It is a block diagram explaining 2nd Example of this invention. It is a figure which shows an example of the charge characteristic of a lithium ion battery, and a discharge characteristic. It is a flowchart explaining 2nd Example of this invention. It is another flowchart explaining the 2nd example of the present invention. It is a conceptual diagram explaining the Bayesian network used with the learning type database of this invention. It is another figure explaining the Bayesian network used with the learning type database of this invention. It is a figure explaining per prior art.

  Hereinafter, it attaches to the detail of the learning type storage battery management system of this invention, and demonstrates based on an Example.

[Constitution]
First, Embodiment 1 of the present invention will be described with reference to FIGS.
In addition, in each FIG. 1-7 used as the base of the following description, although the storage battery management system 400 and the charger 300 are each described independently, in the following description in this specification, it is based on this invention. The learning type storage battery management system is defined in a form including the storage battery management system 400 and the charger 300 in FIGS. The same applies to the second embodiment described below.
In this embodiment, the charger 300 is “charge / discharge device”, the communication unit 415 or 415 ′ is “reception unit” or “transmission unit”, the storage unit 409 is “storage unit”, and the processing unit 408 is “processing unit”. Equivalent to. The same applies to the second embodiment described below.

In addition, regarding the present embodiment, when the object of application is an electric vehicle, a plug-in hybrid vehicle, or other vehicle, i) a case where the storage battery is charged from outside the vehicle via an in-vehicle charger, and ii) the city There is a case where the storage battery is directly charged from the quick charging station outside the vehicle set up in the vehicle. At this time, the learning type storage battery management system of the present invention includes both the above-mentioned i) in-vehicle charger and ii) the quick charging station. It shall be mounted on.
In this case, in FIG. 1, in the case of i) the in-vehicle charger, the learning storage battery management system according to the present invention including the storage battery management system 400 and the in-vehicle charger 300 d is in a state of living together with the vehicle body 350. On the other hand, ii) In the case of the rapid charging station, the learning storage battery management system according to the present invention including the storage battery management system 400 and the rapid charger 300d is in the state of living together outside the vehicle body 350 as the rapid charging station.

  The outline of the learning type storage battery management system of the present embodiment described in FIGS. 1 to 7 will be described first. In the present embodiment, the charger 300 connected to the storage battery 200 communicates with the storage battery management system 400, and the learning database is stored. The storage battery management system 400 includes a storage unit 409 and a processing unit 408 to formulate a charge / discharge plan for the storage battery 200 and transmit a charge / discharge control signal to the charger 300 to which the storage battery 200 identified by the fixed information is connected. Is adopted. The same applies to the second embodiment described below.

FIG. 1 is a diagram illustrating the entire system according to the present embodiment. The learning-type storage battery management system according to the present invention is defined in a form including at least one of the storage battery management system 400 described in the lower center part of the figure and the chargers 300a to 300d shown on the right side of the figure. In addition, each of the storage batteries 200a to 200d shown on the right side of the drawing is a storage battery of a large-scale building / factory / university, a small / medium-sized building group, a house group, an electric vehicle, and other EV groups.
Between the storage battery management system 400 and the storage batteries 200a-d and the chargers 300a-d, fixed information and fluctuation information of the storage batteries 200a-d are transmitted from the chargers 300a-d to the storage battery management system 400, while the storage battery management system A charging / discharging instruction based on the charging / discharging plan of the storage battery 200 (ad) formulated from the 400 learning database is transmitted from the storage battery management system 400 to the charger 300 (ad). The communication unit 415 (415 ′) of the storage battery management system 400 transmits and receives information and commands to and from the chargers 300a to 300d, and receives information from the local energy management system 500. Details will be described with reference to FIG.

  In this embodiment, as shown in the center of FIG. 1, information is transmitted from the regional energy management system 500 to the storage battery management system 400. Thereby, in a present Example, a charging / discharging plan is drawn up from a local energy management system and storage battery information. Details of this point will be described with reference to FIG.

In addition, the regional energy management system 500 shown in the center of FIG. The electric energy supply source 550 including the solar power generation system 552, the wind power generation system 553, and the storage battery 554 is also managed and operated in close association.
Between the regional energy management system 500 and the storage batteries 200a to 200d and the chargers 300a to 300d, information on demand results of large and small buildings / factories / universities, small / medium buildings, houses, electric vehicles and other EVs A large-scale building / factory / university, such as a small / medium-sized building group, a house group, and an electric vehicle or other EV group transmits to the local energy management system 500, while the local energy management system 500 issues a command regarding demand control to the charger 300 ( a to d).
At this time, the quasi-variation information and the fixed information of the storage battery are associated with the contract information of the user who is the user at the start of the storage battery operation or the replacement time. This is based on grasping the contract level between each user and the electric power company or the regional energy management system management company, grasping the demand trend of the storage battery in each user, and other requests. Details of the quasi-variation information will be described later.
In addition, between the local energy management system 500 and the electric energy supply source 550, information on the power generation results of the electric energy supply sources 551 to 554 is transmitted from the electric energy supply sources 551 to 554 to the local energy management system 500, respectively. On the other hand, a command relating to power generation control is transmitted from the local energy management system 500 to each of the electric energy supply sources 551 to 554.

  In addition to the customer management unit 501 and the EV management unit 502, the local energy management system 500 itself includes a large-scale building / factory / university, a small / medium-sized building group, a house group, an electric vehicle and other EV groups, or an electric energy supply source 551- 554, a data acquisition unit 504 that obtains data from each of them, a supply and demand prediction unit 505 that performs supply and demand prediction from the acquired data, a large-scale building / factory / university, a small and medium-sized building group, a house group, and the like A supply and demand adjustment unit 506 that creates a supply and demand adjustment command for each of electric vehicles and other EV groups or electric energy supply sources 551 to 554, and a small and medium-sized building group such as a large-scale building / factory / university based on the created supply and demand adjustment command, Demand control unit 507 that performs real-time control of house groups, electric vehicles, and other EV groups by demand control Real-time control unit 508 and the above-mentioned data fetching unit 504, supply and demand prediction unit 505 and supply and demand adjustment unit 506 are managed in an integrated manner, and local energy management system 500 and large-scale buildings / factories / universities, small and medium-sized buildings, houses And a communication management unit 503 for managing communication with each of the group, the electric vehicle, and the other EV group.

FIG. 2 is a diagram showing the regional energy management system connection block portion 1 in the entire system according to the present embodiment shown in FIG.
As understood from the system configuration diagrams shown on the left side of FIGS. 1 and 2, in the learning type storage battery management system according to the present invention, the storage battery management system 400 includes storage batteries connected to them by chargers 300 a to 300 d. Each of the fixed information and the fluctuation information is received by the communication unit 415, and the received fixed information and the fluctuation information are processed by the learning database including the storage unit 409 and the processing unit 408, and the storage battery 200 (ad) is stored. A charge / discharge plan is formulated, and a charge / discharge control signal is transmitted to the charger 300 (ad) connected to the storage battery 200 (ad) identified by the fixed information. 2, each storage battery 200a to 200d includes a power storage unit 201 and a control unit 206, and each charger 300a to 300d includes a communication unit 315.
Here, in the storage unit 409 and the processing unit 408 of the storage battery management system 400 including the learning database,
i) processing for accumulating fixed information and variation information and classifying the information into a plurality of charge / discharge characteristic groups;
ii) A process of extracting a charge / discharge characteristic group of the storage battery 200 (ad) identified by the fixed information from the plurality of charge / discharge characteristic groups,
iii) a process for defining a charge / discharge control rule for the storage battery 200 (a to d) from the extracted charge / discharge characteristic group; and
iv) a process of formulating a charge / discharge plan for the storage battery 200 (ad) using the obtained charge / discharge control rule;
A specific process including is performed.

  In addition, in this embodiment, as described in the system configuration diagram shown on the left side of FIG. 2, in addition to using the charging / discharging rules obtained from the learning database when formulating the charging / discharging plan, the regional energy management It is configured to formulate a charge / discharge plan using supply and demand adjustment information from the system 500. As described above, in this embodiment, the charge / discharge rule of the storage battery 200 (a to d) is defined by the charge / discharge control pattern for suppressing the deterioration of the storage battery, and the power supply / demand balance is also achieved. The supply and demand adjustment information is input from the demand or supply / demand prediction unit 505 of the local energy management system 500 to the processing unit 408 of the storage battery management system 400 through the communication units 515 and 415 '.

  The flow of formulating a charge / discharge plan from the regional energy management system and storage battery information shown on the right side of FIG. 2 will be described in sequence. First, in steps 101-104, the storage battery management system 400 starts from the charger 300 (ad) to the storage battery. The communication unit 415 receives 200 (ad) fixed information and fluctuation information, and supply and demand adjustment information from the local energy management system. These are stored in the storage unit 409 (ST101-104).

Next, in step 105, the processing unit 408 processes some or all of these using the learning database,
i) a process of accumulating fixed information and fluctuation information and classifying the information into a plurality of charge / discharge characteristic groups; and ii) from the plurality of charge / discharge characteristic groups, the storage battery 200 (ad) identified by the fixed information. Process to extract charge / discharge characteristic group,
(ST105).

In steps 106 and 107, the processing unit 408 continues.
iii) a process for defining a charge / discharge control rule for the storage battery 200 (a to d) from the extracted charge / discharge characteristic group; and
iv) a process of formulating a charge / discharge plan for the storage battery 200 (ad) using the obtained charge / discharge control rule;
Is performed (ST106, 107).
In the present embodiment, each of the above processes is performed by analyzing the fixed information and the fluctuation information using a Bayesian network in the storage unit 409 and the processing unit 408 having a learning database. Details thereof will be described in the second embodiment.

In addition, in a present Example, also assuming the case where the storage battery 200 (ad) is comprised by the assembled battery which is illustrated in Example 2, in the next step 108, the storage battery 200 (ad) is concerned. It is confirmed whether there is a change in the configuration (ST108). Then, if some or all of the configuration of the storage battery 200 (a to d) has changed, the process returns to the previous step and the charge that has been formulated for each storage battery 200 (a to d). The discharge plan is formulated again.
Through the above processes, the process of formulating a charge / discharge plan for the storage battery 200 (a to d) ends.

FIG. 3 shows an image of a charge / discharge plan established through the above processing.
As you can see from the right half of Fig. 3,
i) a plurality of charge-discharge characteristics group _{G n,} the charge-discharge characteristics group _{G x} of the battery 200 is identified by the fixed information (to d) are extracted,
ii) the charge / discharge control rule _{Rx of the} storage battery 200 (a to d) is defined from the extracted charge / discharge characteristic group, and
iii) Using the obtained charge / discharge control rule, a charge / discharge plan of the storage battery 200 (ad) including the charge / discharge pattern and the charge / discharge time is formulated,
I can read.
That is, the processing unit 408 selects one charge / discharge control rule R _x corresponding to the charge / discharge characteristic group G _x of the storage battery extracted based on the learning result from a plurality of different charge / discharge control rules. The charging / discharging control rule of a storage battery is prescribed | regulated.

  In the present embodiment, the supply / demand adjustment information from the regional energy management system 500 illustrated in the left half of FIG. 3 is considered when the charge / discharge plan is formulated. Thus, as shown in the lower side of FIG. 3, the present embodiment defines the charge / discharge rule of the storage battery 200 (a to d) with the charge / discharge control pattern for suppressing the deterioration of the storage battery, as well as the power supply / demand balance. It is also configured to plan.

FIG. 4 is a diagram showing a prediction analysis block portion 2 of the determination / utilization scene of the charge / discharge control rule in the entire system according to the present embodiment shown in FIG.
As understood from the system configuration diagram shown above and on the left side of FIG. 4, in the learning type storage battery management system according to the present invention, the storage battery management system 400 is more than the communication unit 315 of the charger 300 (ad), The receiving unit 415 receives the fixed information and the fluctuation information of each of the storage batteries 200a to 200d connected thereto, and processes the received fixed information and the fluctuation information with the learning database including the storage unit 409 and the processing unit 408, and the storage battery 200. A charge / discharge plan of (a to d) is formulated, and a charge / discharge control signal is transmitted to the charger 300 (ad) to which the storage battery 200 (ad) identified by the fixed information is connected. Yes. Each storage battery 200a to 200d has a power storage unit 201 and a control unit 206.
Here, in the storage unit 409 and the processing unit 408 of the storage battery management system 400 including the learning database,
i) processing for accumulating fixed information and variation information and classifying the information into a plurality of charge / discharge characteristic groups;
ii) A process of extracting a charge / discharge characteristic group of the storage battery 200 (ad) identified by the fixed information from the plurality of charge / discharge characteristic groups,
iii) a process for defining a charge / discharge control rule for the storage battery 200 (a to d) from the extracted charge / discharge characteristic group; and
iv) a process of formulating a charge / discharge plan for the storage battery 200 (ad) using the obtained charge / discharge control rule;
A specific process including is performed.

Further, in the present embodiment, the processing unit 408 analyzes and determines the characteristics of each storage battery 200 (ad) from the charge / discharge characteristic group of the extracted storage battery 200 (ad). And an analysis determination result output means for outputting the determination result to the outside, and the analysis determination means is an estimated residual value in the case of assuming secondary use of the storage battery 200 (ad) or It is configured to determine the recommended use.
Moreover, the said analysis determination means uses the learning type database which extracts the charging / discharging characteristic group of the said storage battery 200 (ad), and prescribes | regulates the charging / discharging control rule of the said storage battery 200 (ad), The said storage battery It is configured to determine an estimated residual value or a recommended use when assuming secondary usage of 200 (a to d).

  FIG. 4 right side shows the charge / discharge control rule determination / predictive analysis of utilization scene block part 2, that is, the charge / discharge control rule is defined from the storage battery information, the charge / discharge plan is formulated, and the utilization scene predictive analysis In step 201, the charger 300 (ad) measures variation information of the storage battery 200 (ad) (ST201).

In step 202, the measured fluctuation information and the fixed information of the storage batteries 200 (ad) are transmitted from the communication unit 315 of the charger 300 (ad) to the storage battery management system 400 (ST202).
In step 203, the storage battery management system 400 receives the fixed information and the fluctuation information of the storage battery 200 (ad) from the charger 300 (ad) through the communication unit 415. These are stored in the storage unit 409 (ST203).
In this flow, the supply and demand adjustment information from the regional energy management system 500 is not described.

Next, in steps 204 and 205, the processing unit 408 processes some or all of these using the learning database, accumulates fixed information and fluctuation information, and classifies the information into a plurality of charge / discharge characteristic groups. Processing is performed (ST204, 205).
In the present embodiment, the above processing is performed by analyzing the fixed information and the variation information using a Bayesian network in the storage unit 409 and the processing unit 408 having a learning database (as described above, details thereof are described in the second embodiment. explain).

Thereafter, in steps 206 and 207,
i) A process of extracting the charge / discharge characteristic group of the storage battery 200 (ad) identified by the fixed information from the plurality of charge / discharge characteristic groups,
ii) a process for defining a charge / discharge control rule for the storage battery 200 (a to d) from the extracted charge / discharge characteristic group; and
iii) Based on the individual identification number of the classified storage battery 200 (ad), the deterioration state of the storage battery 200 (ad), the estimated residual value when secondary use is assumed, or the recommended use or other relevant storage battery 200 (a to d) characteristics are analyzed and determined (ST206, 207).
In this flow, the step of confirming whether or not there is a change in the configuration of the storage battery 200 (ad) is omitted.

  As a result of the determination, if reuse can be recommended apart from those that can be used in the current usage environment, the estimated state when the storage battery 200 (a to d) is deteriorated and secondary usage is assumed. In addition to evaluating the residual value, the optimal application is proposed as necessary, and the process ends. At this time, when the deterioration is severe, there is a case where recycling or disposal is recommended.

FIG. 5 stipulates the charge / discharge control rules based on the storage battery information shown on the right side of FIG. 4 and the determination / utilization scene prediction / analysis block portion 2 shown on the left side of FIG. It expresses the image of the flow for formulating a plan and performing predictive analysis of usage scenes.
In the upper right half of FIG. 5, specific examples of fixed information, fluctuation information, and other information stored in the storage unit 409 are shown. Here, the quasi-variation information is derived from in-vehicle storage batteries including electric vehicles, household storage batteries, building storage batteries, housing batteries for collective housing, storage batteries for community energy management systems and other storage batteries, and household composition It includes at least one of the usage status of the storage battery, and the content does not change frequently, but includes the possibility of changing.

In the lower right half of FIG. 5, i) processing in the processing unit 408 that accumulates fixed information and fluctuation information and classifies the information into a plurality of charge / discharge characteristic groups G _n , and
ii) from the plurality of charge-discharge characteristics group, the process of extracting the charge and discharge characteristics group G _x of the battery 200 is identified by the fixed information (to d),
As a result of determination by the analysis determination means of the processing unit 408, if reuse can be recommended apart from those that can be used in the current usage environment, the storage battery 200 (a to d) It represents an image in which an estimated residual value is evaluated when a deterioration state and secondary use are assumed, and an image in which an optimum application is proposed as necessary. The drawing also shows that recycling or disposal may be recommended when deterioration is severe.

FIG. 6 is a diagram showing the storage battery control management block portion 3 in the entire system according to the present embodiment shown in FIG. The system configuration diagram shown on the left side of FIG. 6 is the same as the system configuration diagram shown on the left side of FIG. 4 except that the signal flow is opposite.
As can be understood from the signal flow of FIG. 6, the storage battery 200 identified by the fixed information formulated from the learning database of the storage battery management system 400 in which the fixed information and fluctuation information of the storage batteries 200 a to 200 d are received and stored. The charging / discharging instruction based on the charging / discharging plan (a to d) is performed by transmitting a charging / discharging control signal from the storage battery management system 400 to the corresponding charger 300 (ad). The charger 300 (ad) receiving the charge / discharge instruction performs charge / discharge of the storage battery 200 (ad) based on the charge / discharge control signal.

The flow of the storage battery control management block part 3 shown on the left side of FIG. 6 will be described in sequence. First, in step 301, the communication unit 415 of the storage battery management system 400 sends a charge / discharge control signal to the corresponding charger 300 (ad). Transmit (ST301).
In step 302, communication unit 315 of charger 300 (ad) receives the charge / discharge control signal transmitted from storage battery management system 400 (ST302).
And in step 303 and 304, charger 300 (ad) charges / discharges the corresponding storage battery 200 (ad) based on the received charge / discharge control signal (ST303, 304).

FIG. 7 shows an image of charge / discharge control performed in the storage battery control management block portion 3 in which the left side of FIG. 6 is a block diagram and the right side of FIG.
In the example shown in FIG. 7, the charge / discharge plan P _x of the storage battery 200 (ad) formulated from the learning database of the storage battery management system 400 includes the individual identification number and the transmission destination in addition to the charge / discharge control signal. The fixed information of the storage battery 200 (ad) is included. Thereby, the charge / discharge instruction | indication based on the charging / discharging plan of the said storage battery 200 (ad) identified by fixed information from the storage battery management system 400 is sent to the charger 300 (ad) applicable charge / discharge control signal. This is done by sending The charger 300 (ad) receiving the charge / discharge instruction performs charge / discharge of the storage battery 200 (ad) based on the charge / discharge control signal.

[Constitution]
Next, Example 2 of the present invention will be described with reference to FIGS. The present embodiment also corresponds to an example in which more specific application is intended for the learning-type storage battery management system of the present invention described above as an example.
First, based on FIG. 8, the outline of the learning type storage battery management system of the present embodiment will be described. The learning type storage battery management system of the present embodiment includes lithium ion batteries 102 to 105 connected in series as an example and the assembled battery main body 101. RFID assembled battery main body 100 including a BMU (Battery Management Unit) 106, a charger 300 and a storage battery management system 400 not shown in FIG. 8 (see FIGS. 1 to 7 according to the first embodiment, the same applies hereinafter). . The RFID assembled battery main body 100 is used as a vehicle power source of an electric vehicle or a plug-in hybrid vehicle, and is mounted on these vehicles. As shown in FIG. 8, the assembled battery main body 101 is configured such that a plurality of cells are internally connected in series to form a module body, and lithium ion batteries 102 to 105 are connected again in series. The basic configuration of the assembled battery main body 101 is the same as that shown in FIGS.
Here, the assembled battery main body 100 with RFID corresponds to the storage battery 200 of the first embodiment, the assembled battery main body 101 corresponds to the power storage unit 201 of the first embodiment, and the BMU 106 corresponds to the control unit 206 of the first embodiment.
In the present embodiment, the charger 300 has a discharging function and a storage function in addition to a charging function and a communication function of fixed information and fluctuation information of the RFID assembled battery main body 100. The discharge function is used in the process of diagnosing the deterioration state of the storage battery as described later (see FIG. 10).

In the learning type storage battery management system of this embodiment, the storage battery management system 400 is combined with at least one assembled battery main body 100 with RFID and a charger 300, but a plurality of assembled battery main bodies 100 with RFID and chargers are combined. A case in which 300 is connected in combination is assumed.
In that case, it is performed between the charger 300 to which the assembled battery main body 100 with RFID is connected and the storage battery management system 400.
i) Reception of fixed information and variation information from the RFID-attached battery main body 100 via the charger 300, or
ii) transmission of a charge / discharge control signal formulated by the storage battery management system 400 using a learning database to the charger 300 to which the RFID assembled battery body 100 identified by the fixed information is connected;
As in the first embodiment, the communication is performed through a RFID system, the Internet, a cellular phone communication network, a wired or wireless network including a local area network (LAN) or a wide area network (WAN), and other communication means.

In addition, in a present Example, the process of fixed information and fluctuation information using the learning type database in the storage battery management system 400, specifically speaking,
i) processing for accumulating fixed information and variation information and classifying the information into a plurality of charge / discharge characteristic groups;
ii) A process of extracting the charge / discharge characteristic group of the assembled battery main body 100 with RFID identified by the fixed information from the plurality of charge / discharge characteristic groups,
iii) a process for defining a charging / discharging control rule for the RFID assembled battery body 100 from the extracted charging / discharging characteristic group; and
iv) a process for formulating a charge / discharge plan for the storage battery using the obtained charge / discharge control rule;
It is also assumed that the process including the process is performed by the storage unit 409 and the processing unit 408 that are physically separated from the communication unit 415 with the charger 300.
As described above, the storage battery management system 400 in the present embodiment includes the communication unit 415, the storage unit 409, and the processing unit 408 in addition to the communication unit 415, the storage unit 409, and the processing unit 408 provided in one place. It also includes those that are physically isolated bodies.
Hereinafter, an example in which processing of fixed information and variation information using a learning database is performed in the storage unit 409 and the processing unit 408 that are physically separated from the communication unit 415 with the charger 300 will be described.

  The BMU 106 includes a holding battery 107, a CPU 108, and a memory 109, and has a combination of a demodulator 110, a modulator 111, and an antenna 112, and has a signal transmission / reception function.

As described above, the RFID-assembled battery main body 100 including the BMU 106 and the charger 300 are signal-connected using the RFID in addition to the power connection using the power cable using the connection plug.
As described above, between the charger 300 and the storage battery management system 400, a wired or wireless network including an RFID system, the Internet, a mobile phone communication network, a local area network (LAN) or a wide area network (WAN), and the like. The signal is connected through the communication means.

  When it is assumed that the assembled battery main body 100 with RFID is mounted on an electric vehicle, a plug-in hybrid vehicle, or other vehicles, the BMU 106 usually has an antenna 112 and an in-vehicle antenna 113 or a CAN network stretched on the vehicle body. Communicating with a VCU (Vehicle Control Unit) 114 mounted on the vehicle as needed, features of the assembled battery body 101 including environmental variables of the assembled battery body 101 including the temperature history, charging status history, and discharging status history Variables are monitored and measured to manage the storage status and monitor abnormal conditions. Further, the BMU 106 performs SOC (State of Charge) estimation and abnormality determination based on the environmental variables and characteristic variables of the assembled battery main body 101 that are monitored and measured.

  Further, in this embodiment, the BMU 106 sends to the storage battery management system 400 via the charger 300 and the charger 300 the fixed information and fluctuation information of the storage battery, specifically, the fixed information and the assembled battery of the assembled battery body 100 with RFID. The fluctuation information of the main body 101 is transmitted.

  Therefore, in the learning type storage battery management system according to the second embodiment, the charger 300 connected to the RFID assembled battery body 100 and the storage battery management system 400 communicate with each other, and the storage unit 409 of the storage battery management system 400 including the learning database and The processing unit 408 formulates a charging / discharging plan for the RFID assembled battery body 100 and transmits a charging / discharging control signal to the charger 300 to which the RFID assembled battery body 100 identified by the fixed information is connected. ing.

  The charger 300 performs charge / discharge control of the assembled battery body 100 with RFID based on the received charge / discharge control signal. More specifically, the charger 300 controls charge / discharge of the assembled battery body 101 via the BMU 106 of the assembled battery body 100 with RFID based on the charge / discharge control signal.

  As described above, the fixed information includes the type of the RFID assembled battery main body 100 classified by the principle of power storage, the individual identification number of the RFID assembled battery main body 100 including the manufacturing number, the manufacturing date, the weight and the charge amount at the start of service. Part or all of Further, the storage battery capacity, storage battery structure, storage battery constituent material, and the like can be secondarily known from the individual identification number. The variation information includes some or all of environmental variables, characteristic variables, the current charge amount, and the chargeable capacity.

  The environmental variables include part or all of the temperature history, humidity history, vibration history, and travel history of the assembled battery main body 101. The characteristic variables include the service life of the battery pack body 101, the number of charge / discharge cycles, the discharge load, the charge status history, the discharge status history, the current charge / discharge rule, and part or all of the weight.

Next, in this embodiment, the characteristics of the lithium ion batteries 102 to 105 to which the learning type storage battery management system shown in FIG. 8 is applied, or the assembled battery body 101 in which these are connected in series will be described.
FIG. 9A shows an example of discharge characteristics, and FIG. 9B shows an example of charge characteristics.

With respect to the discharge characteristics of FIG. 9A, in the lithium ion battery (N) in the initial stage of service, although the discharge voltage first decreases, the region where the discharge voltage is stable continues.
On the other hand, in the used lithium ion battery (D), although the voltage drop immediately after the start is small, the drop in the discharge voltage is continuous.

As for the charging characteristics of FIG. 9B, in the lithium ion battery (N) in the initial stage of service, after the charging voltage is first increased, the region where the charging voltage is stable continues.
On the other hand, the used lithium ion battery (D) has a small increase in voltage immediately after the start and a large increase in the charging voltage.
The difference between the in-service initial stage (N) and the used (D) lithium ion battery is called a memory effect.

According to the present embodiment, i) optimization of the charging voltage so that such a memory effect is unlikely to occur, ii) efficiently selecting a reusable lithium ion assembled battery with uniform charge / discharge characteristics when reusing, Is possible.
In particular, in this embodiment, the assembled battery body 100 with RFID used as a power source for an electric vehicle or a plug-in hybrid vehicle is repeatedly charged and discharged on the basis of the SOC intermediate region. The significance of applying the present invention to a battery is great.
Note that the reuse lithium-ion battery has a variety of changes from the beginning of service according to the diversity of demand environment.

For example, even if a lithium-ion battery used for an electric vehicle with severe demand environmental conditions cannot be used for an electric vehicle, the use of the lithium ion battery is more mild (ex. There are various cases where the battery can be sufficiently used for the purpose of charging the storage battery to the storage battery and effectively using the stored power at night.
Thus, by selecting storage batteries according to the diversity of demand environments, it becomes possible to efficiently reuse lithium ion batteries whose consumption will dramatically increase in the future.

[Operation]
Hereinafter, the operation of the learning storage battery management system of the present embodiment shown in FIG. 8 will be described with reference to FIGS.

  FIG. 10 is a flowchart for explaining the operation of the learning type storage battery management system of the present embodiment. Mainly, storage and processing of storage battery fixed information and fluctuation information using a learning type database in the storage battery management system 400 are performed. A flow of a process performed between the BMU 106 and the charger 300 before the process is performed is illustrated.

First, in step 1, the RFID-attached battery body 100 and the charger 300 are connected (ST01).
In step 2, the charger 300 reads the fixed information and fluctuation information of the storage battery transmitted from the BMU 106, specifically, the fixed information of the assembled battery body 100 with RFID and the fluctuation information of the assembled battery body 101 (ST02). The storage battery fixed information and the fluctuation information are transmitted from the BMU 106 to the charger 300 through the RFID function provided in the RFID assembled battery body 100 or directly between the RFID assembled battery body 100 and the storage battery 300 via a connector. This is done by connecting cables.

In step 3, the charger 300 measures the terminal voltage of the RFID assembled battery main body 100 (ST03). In this embodiment, it is determined whether or not the terminal voltage measured in step 3 is within a predetermined range. When the terminal voltage is out of the predetermined range, the process proceeds to step 5 through step 4, and the charger 300 determines that the storage battery is abnormal. The process ends (ST05).
On the other hand, when the terminal voltage is within the predetermined range, charger 300 performs the discharge process (ST04) shown in step 4 with a predetermined load connected for a predetermined time (ST06). At this time, the discharge characteristics are measured as shown in step 7 (ST07), and in step 8, the storage battery fixed information and the fluctuation information including the discharge characteristics are transmitted from the charger 300 to the storage battery management system 400 (ST08).

In step 9, the charger 300 performs a charging process (ST09). In this charging process, the RFID assembled battery main body 100 is charged with a charging pattern for suppressing the deterioration of the storage battery according to the deterioration state of the RFID assembled battery main body 100 identified by the fixed information.
In addition, the said charge pattern was prescribed | regulated by the process using a learning type database being performed in the storage battery management system 400 based on the fixed information and fluctuation information of the said storage battery transmitted to the storage battery management system 400 at step 8. This is based on the charging / discharging control rule of the assembled battery main body 100 with RFID.

The charging characteristics are measured by the charger 300 as shown in step 7 (ST10).
In step 11, various information related to the fixed information and fluctuation information of the storage battery including the measured charging characteristics is transmitted from the charger 300 to the storage battery management system 400 as necessary (ST11).

Next, a flow of storing and processing storage battery fixed information and variation information using a learning database in the storage battery management system 400 will be described with reference to FIG.
As described above, here, an example in which the processing of the fixed information and the fluctuation information using the learning database is performed in the storage unit 409 and the processing unit 408 that are physically separated from the communication unit 415 with the charger 300 will be described. To do. In this case, the storage unit 409 and the processing unit 408 that are physically separated from the communication unit 415 have strong data center colors. Therefore, hereinafter, the storage unit 409 and the processing unit 408 of the storage battery management system 400 may be handled as a data center as necessary.

First, as step 12, in the storage battery management system 400, the storage battery fixed information and the fluctuation information and other data transmitted from the charger 300 and received by the communication unit 415 are stored in the data center (including the storage unit 409 and the processing unit 408). If the entire system is small, it may be a server) (ST12).
Here, as between the charger 300 and the storage battery management system 400, an RFID system, the Internet, a mobile phone communication line network, a local area network (LAN), or a wide area network is provided between the data center and the communication unit 415. Signal connection is made through a wired or wireless network including (WAN) or other communication means.

  In step 13, the data center receiving the data transfer receives characteristic variables (Ex. Charge status history, discharge status history) and environment variables (existing status information) corresponding to the fixed information including the individual identification number of the assembled battery body 100 with RFID. (Ex. Environmental temperature history, operation status history such as vibration) is input and arranged on the database.

In step 14, the learning type database in the data center
i) An individual identification number node of the RFID-attached battery body 100 as a child of the environment variable node;
ii) a characteristic variable node of the RFID assembled battery main body 100 as a child of the individual identification number node of the RFID assembled battery main body 100;
The individual identification of the assembled battery main body 100 with RFID, that is, identification by fixed information, is performed by the storage battery individual identification rule expressed by the Bayesian network having (ST14).
The outline of the Bayesian network is as follows, and is disclosed in Non-Patent Document 2 and others.

In step 16, the storage battery individual identification rule created when the analysis by the Bayesian network is performed in step 14 is updated. By performing this update, it is possible to update the number of feature variable nodes (number of feature variable classifications and classification ranges) of the RFID-attached battery main body 100 identified by individual identification, that is, fixed information.
That is, the number of classifications and the range of classification according to the characteristic variables of the RFID battery assembly 100 with individual identification are gradually optimized by the learning effect as the number of analyzes by the Bayesian network increases.

As described above, the processing using the learning database in the storage battery management system 400 allows the processing unit 408 to perform charge / discharge control that suppresses deterioration of the RFID-attached battery body 100 identified by individual identification, that is, fixed information. A charge / discharge control rule having a pattern is defined.
In this embodiment, based on the prescribed charge / discharge control rule, a charge / discharge instruction is given to the charger 300 that charges / discharges the RFID-attached battery body 100 identified by the fixed information via the communication unit 415. The

Furthermore, in the present embodiment, the processing means 108 analyzes and determines storage battery characteristics analysis and determination means for analyzing and determining the characteristics of each of the RFID assembled battery bodies 100 from the extracted charge / discharge characteristics group of the RFID assembled battery body 100, And an analysis determination result output means for outputting the determination result to the outside. The analysis determination means determines an estimated residual value or a recommended use when assuming secondary use of the RFID assembled battery body 100. It is configured to do.
Further, the analysis determination unit extracts the charge / discharge characteristic group of the RFID-attached battery body 100 and uses the learning database that defines the charge / discharge control rules of the RFID-attached battery body 100, and uses the RFID attached assembly. It is configured to determine an estimated residual value or a recommended use when assuming secondary use of the battery body 100.

Therefore, after the database in which the individual identification number of the RFID assembled battery main body 100 is classified in step 15 (ST15), the individual identification number of the classified RFID assembled battery main body 100 in step 17 is set. Based on this, reuse determination is performed (ST17).
In the reuse determination, the degradation state of the RFID assembled battery main body 100, the estimated residual value in the case of assuming secondary use, the recommended use, and other characteristics of the RFID assembled battery main body 100 are analyzed and determined.

  As a result of the determination, if reuse can be recommended apart from those that can be used in the current usage environment, the process proceeds to step 18 and the process ends (ST18). If the deterioration is severe, the process proceeds to step 19 in which recycling or disposal is recommended, and the process ends (ST19).

FIG. 12 shows a Bayesian network used in the learning database of this embodiment. Y ₁ , Y ₂ , Y ₃ ... Are environment variable nodes. For example, when an electric vehicle storage battery is assumed, environmental temperature history, vibration and other operation status histories are included. Z is an individual identification number node. X ₁ , X ₂ ... Are feature variable nodes, and include a charging status history, a discharging status history, and the like.

  In the present embodiment, in step 14 of FIG. 11, the learning type database composed of the storage unit 409 and the processing unit 408 estimates the candidate (the RFID-attached battery body 100 identified by the fixed information) to be charged / discharged. The storage battery individual identification rule to be used is configured by a Bayesian network. A Bayesian network is a form of expression of a simultaneous probability distribution of a plurality of variables, and using this, a probability distribution of unknown variable values can be calculated from known variable values.

As described above, in the example of FIG. 12, the individual identification number node Z is a child of the variable nodes Y ₁ , Y ₂ ,..., Y _m corresponding to the environmental temperature history, vibration, and other operation status histories that are environment variables. the have, as a child of the individual identification number node Z, the charging status history is characteristic variable, wherein the variable node X _1, X 2 corresponding to each such discharge status _history, ..., in Bayesian network structure with X _n is there.

In order to express the joint probability distribution by this Bayesian network, the probabilities P (Y ₁ ), P (Y ₂ ),..., P (Y _m ) and the conditional probability P (Z | Y ₁ , Y _{2, ···, Y m),} P (X 1 | Z), P (X 2 | is required value of Z) | Z), ···, P (X n.

These probabilities can be estimated from the appearance frequency. When the learning database estimates a candidate to be charged / discharged, the characteristic variables and environmental variables (X ₁ , X ₂ ,..., X _n , Y ₁ , Y ₂ ,. .., Y _m ) = (x ₁ , x ₂ ,..., X _n , y ₁ , y ₂ ,..., Y _m ) to Z posterior probability distribution P (Z | x ₁ , x ₂ , , X _n , y ₁ , y ₂ ,..., Y _m ) can be calculated by the following (Equation 1).

  If the set of variable Z values for arbitrarily narrowed candidates is S, the Z value z that maximizes the posterior probability in S can be obtained by the following (Equation 2). .

  In the learning type database composed of the storage unit 409 and the processing unit 408, the candidate corresponding to z thus obtained is recognized as a candidate to be instructed to charge / discharge.

  In step 16 of FIG. 11 for updating the storage battery individual identification rule, the obtained characteristic variable and environment variable are input, and an output when applied to the storage battery individual identification rule of step 14 of FIG. 11 is obtained by feedback. The storage battery individual identification rule is updated so as to be close to the storage battery individual identification result.

At that time, the updating method in step 16 differs depending on which storage battery individual identification rule is used. For example, in the case of FIG. 12, probabilities P (Y ₁ ), P (Y ₂ ,..., P (Y _m ) and conditional probabilities P (Z | Y ₁ , Y ₂ ,..., Y _m ), P (X ₁ | Z), P (X ₂ | Z),..., P (X _n | Z), which is updated by feedback. The inference result data can be estimated by adding to the data so far.

Here, the probabilities P (Y ₁ ), P (Y ₂ ),..., P (Y _m ) are not used in (Equation 1) and (Equation 2), but this value is Y ₁ , Necessary when Y ₂ ..., Y _m is unknown. For example, when the value of Y ₁ is unknown, calculation is performed using (Expression 3) and (Expression 4) instead of (Expression 1) and (Expression 2).

  Next, the specific example which estimates the candidate which should perform charge / discharge instruction | indication using the storage battery individual identification rule comprised by the Bayesian network is shown.

  For example, the environmental variables of the RFID-attached battery main body 100 that has been connected to the storage battery management system 400 including the learning database according to the present embodiment and has individual identification numbers 1 to 20 to which the fixed information and the variation information are transmitted. Suppose that the characteristic variables are as shown in Table 1. Tables 2 and 3 show the top groups obtained by grouping with discharge minimum voltages of 3.0 [V] and 3.1 [V]. Here, the discharge characteristics illustrated as characteristic variables in Tables 1 to 3 are useful data for obtaining the estimated residual value of the storage battery in the process of FIG.

  In addition, each numerical value shown in the following Tables 1-3 is an example used for the purpose of explaining the Bayesian network, and the terminal voltage of the assembled battery body 100 with RFID used as a power source of an actual electric vehicle or a plug-in hybrid vehicle. It should be noted in advance that this is not necessarily an example.

  Here, in the following example, for simplicity, only the minimum voltage during discharging and the maximum charging history current are used as characteristic variables, and only the minimum temperature history value, the average temperature history value, and the maximum temperature history value are used as environment variables. Shall be used. Also, for the sake of simplification, the candidate (the RFID-attached battery main body 100 identified by the fixed information) that the storage battery management system 400 should give a charge / discharge instruction through the communication unit 415 and the charger 300 at this time is 3 It is assumed that samples A and B (individual identification numbers 13 and 20) arbitrarily selected from those shown in the upper row are narrowed down (lower row in Table 3).

In this case, the characteristics of the samples A and B are similar with respect to the minimum voltage during discharging and the maximum charging history current, and it is difficult to narrow down the sample A or B from the minimum voltage during discharging and the maximum charging history current. .
However, it can be estimated that the candidate for which the storage battery management system 400 should give a charge / discharge instruction at the present time is likely to be the sample A, from the environmental history minimum value, the average temperature history value, and the maximum temperature history value.

  This estimation is automatically performed by the identification rule learned from the connection history so far including the data shown in Tables 1 to 3. Specifically, if various conditional probabilities are obtained from the connection history including the data in Tables 1 to 3 by Laplace transform, a Bayesian network as shown in FIG. 13 is created, and Table 3 is created using this Bayesian network. Looking at the probabilities of the parts that meet the lower conditions, sample A (3/5, 4/5), sample B (2/5, 1/5), and therefore, it is estimated that the probability that the sample is A is high. can do.

  Here, a specific example is shown in which only the temperature history minimum value, the temperature history average value, and the temperature history maximum value are used as environment variables, but the present invention is not limited to this.

[Modification]
As described above, the learning type storage battery management system of the present invention has been described based on each embodiment, but the present invention is not limited to the configuration described in each of the above embodiments, and various modifications can be made. .

In each of the above-described embodiments, first, in the learning-type storage battery management system according to Embodiment 1, the storage unit 409 of the storage battery management system 400 including the learning database and the charger 300 to which the storage battery 200 is connected communicate with the storage battery management system 400. The charge / discharge plan of the said storage battery was drawn up from the process part 408, and the structure which transmits the charge / discharge control signal to the charger 300 to which the said storage battery 200 identified by fixed information was connected was taken.
On the other hand, in the learning type storage battery management system according to the second embodiment, the charger 300 to which the assembled battery body 100 with RFID is connected communicates with the storage battery management system 400 (none of which is shown in FIG. 8), and includes a learning database. A charging / discharging plan of the RFID assembled battery body 100 is formulated from the storage unit 409 and the processing unit 408 of the management system 400, and charging / discharging is performed on the charger 300 to which the RFID assembled battery body 100 identified by the fixed information is connected. The structure which transmits a control signal was taken.
However, the present invention is not limited to the configurations of the above-described examples, and various modifications can be made as long as the same effects can be achieved.

  In each embodiment, an example using a Bayesian network is given as an example of a configuration of a learning database. However, the present invention is not limited to this. For example, as shown in Japanese Translation of PCT International Publication No. 2010-519691, any two or more parameters are artificial neural networks. A configuration for inputting a charge / discharge characteristic group of a storage battery that is input to a network and identified by fixed information through calculation by an artificial neural network, and that develops a charge / discharge plan for the storage battery based on this, and further, for the extracted storage battery A configuration that analyzes and determines the characteristics of each storage battery from the charge / discharge characteristics group, and analyzes and determines the deterioration state of the storage battery, the estimated residual value when the secondary use of the storage battery is assumed, or the recommended use and other storage battery characteristics It does not matter.

  Also, in each embodiment, temperature history is given as one of the environment variables, and for temperature history, the temperature history maximum value, the temperature history minimum value, the temperature history average value, and the temperature history variance value are input to the learning database as parameters. However, the present invention is not limited to this. For example, as shown in FIGS. 3 and 5 of Japanese Patent Application Laid-Open No. 2002-271929, a graph with time as a horizontal axis and temperature as a vertical axis is regarded as a waveform, and this is subjected to Fourier transform. The fundamental wave component and the subsequent harmonic component may be separated and the number of the harmonic components may be used as one of the parameters input to the learning database.

  In addition, regarding the point that the charge / discharge characteristic group of the storage battery identified by the fixed information is extracted from a plurality of charge / discharge characteristic groups, and the charge / discharge control rule of the storage battery is defined from the extracted charge / discharge characteristic group of the storage battery, A new charge / discharge control rule is constructed by combining two or more charge / discharge control rules selected from a plurality of charge / discharge control rules, and the charge / discharge control rule corresponding to the extracted charge / discharge characteristic group of the storage battery It may be added to one of the options.

  As described above, the present invention performs charging / discharging of a storage battery with a charge / discharge control pattern that suppresses deterioration, analyzes and determines the characteristics of each storage battery, and can perform analysis determination of the storage battery's deterioration state and other storage battery characteristics. It is clear that the new storage battery management system is new and useful.

G _n Multiple charge / discharge characteristic groups G _x Charge / discharge characteristic group of storage battery identified by fixed information P _x Charge / discharge plan R _x Charge / discharge control rule 1 Regional energy management system connection block part 2, 2 'Charge / discharge control rule Prediction / analysis block part of judgment / utilization scene 3, 3 ′ Storage battery control management block part 10 Storage battery management system linked to local energy management system 100 Assembled battery body with RFID 101 Assembled battery body 102 to 105 Lithium ion battery 106 BMU
107 Retention battery 108 CPU
109 Memory 110 Demodulator 111 Modulator 112 Antenna 113 Car-mounted Antenna 114 VCU
200, 200a-200d Storage battery 201 Power storage unit 206 Control unit 300, 300a-300d Charger 315 Communication unit 350 Car body 400 Storage battery management system 408, 408 'Processing unit 409 Storage unit 415, 415' Communication unit 500 Regional energy management system 550 Electricity Energy supply

Claims (15)

A learning-type storage battery management system,
A charging / discharging device for charging / discharging the storage battery;
Receiving means for receiving fixed information and fluctuation information of the storage battery;
By storing the fixed information and fluctuation information of the storage battery in a learning database, it is classified into a plurality of charge / discharge characteristic groups,
The charge / discharge characteristic group of the storage battery identified by the fixed information is extracted from the plurality of charge / discharge characteristic groups,
Define the storage battery charge / discharge control rule from the extracted storage battery charge / discharge characteristics group,
Processing means for formulating a charge / discharge plan of the storage battery using the obtained charge / discharge control rule,
A transmission means for transmitting a charge / discharge control signal instructing charge / discharge based on a charge / discharge plan established by the processing means to a charge / discharge device for charging / discharging the storage battery identified by the fixed information;
A learning-type storage battery management system comprising: The charge / discharge control rule has a charge / discharge control pattern for suppressing deterioration of the storage battery,
The processing means formulates a charge / discharge plan according to the charge / discharge control rule defined by the charge / discharge characteristic group of the storage battery identified by the fixed information extracted using the learning database. The learning storage battery management system according to claim 1. A learning-type storage battery management system,
A charging / discharging device for charging / discharging the storage battery;
Receiving means for receiving fixed information and fluctuation information of the storage battery;
By storing the fixed information and fluctuation information of the storage battery in a learning database, it is classified into a plurality of charge / discharge characteristic groups,
Extracting the charge / discharge characteristic group of the storage battery identified by the fixed information from the plurality of charge / discharge characteristic groups,
Define the storage battery charge / discharge control rule from the extracted storage battery charge / discharge characteristics group,
Processing means for formulating a charge / discharge plan of the storage battery using at least the prescribed charge / discharge control rule;
The fixed information and the fluctuation information, and a plurality of charge / discharge characteristic groups of the storage battery obtained by accumulating and classifying the information, a charge / discharge control rule of the storage battery defined by the charge / discharge characteristic group, Storage means for storing a charge / discharge plan of the storage battery formulated using a charge / discharge control rule;
A charging / discharging device that charges / discharges the storage battery identified by the fixed information, and a transmission unit that transmits a charging / discharging control signal that instructs charging / discharging based on a charging / discharging plan formulated by the processing unit;
The processing means defines a charge / discharge rule for the storage battery with a charge / discharge control pattern for suppressing deterioration of the storage battery.   The processing means selects one charge / discharge control rule corresponding to the extracted charge / discharge characteristic group of the storage battery based on a learning result from a plurality of different charge / discharge control rules, thereby controlling the charge / discharge of the storage battery. The learning type storage battery management system according to claim 1, wherein a rule is defined. The processing means includes
Storage battery characteristic analysis determination means for analyzing and determining characteristics of each of the storage batteries from the charge / discharge characteristic group of the extracted storage batteries, and analysis determination result output means for outputting the determination results to the outside,
Have
The learning type storage battery management system according to any one of claims 1 to 4, wherein the analysis determination unit determines an estimated residual value or a recommended use when the secondary use of the storage battery is assumed.   The analysis determination means assumes the secondary use of the storage battery using the learning database that extracts the charge / discharge characteristic group of the storage battery identified by the fixed information and defines the charge / discharge control rule of the storage battery. 6. The learning-type storage battery management system according to claim 5, wherein an estimated residual value or a recommended use in the case of being determined is determined. The learning type storage battery management system according to claim 1, wherein the charge / discharge of the storage battery is managed in cooperation with a local energy management system that adjusts the supply and demand balance between the power supply side and the power demand side.
In addition to using the charge / discharge control rule when formulating the charge / discharge plan, the charge / discharge plan is further formulated using one or both of supply and demand adjustment information from the regional energy management system and the charge rate of the storage battery. Learning type storage battery management system.   The learning type storage battery according to claim 7, wherein the quasi-variation information and the fixed information of the storage battery are associated with the contract information of a user who is a user at the start of use or replacement of the storage battery. Management system.   The storage battery is constituted by an assembled battery, and when a change occurs in part or all of the configuration, the charge / discharge plan that has been formulated for each storage battery is configured to be formulated again. The learning type storage battery management system according to any one of claims 1 to 8. Including the individual identification number of the storage battery as the fixed information,
Including environmental variables and feature variables as the variation information;
In the learning database,
An individual identification number node of the storage battery as a child of the environment variable node;
A characteristic variable node of the storage battery as a child of the individual identification number node of the storage battery;
The learning type storage battery management system according to any one of claims 1 to 9, wherein the individual identification of the storage battery is performed by a storage battery individual identification rule expressed by a Bayesian network.   The fixed information includes a part or all of a storage battery individual identification number composed of a type of storage battery classified by a storage principle, a manufacturing number, a manufacturing date, a weight at the start of service, and a charge amount. The learning storage battery management system according to claim 1.   The learning type storage battery management system according to any one of claims 1 to 11, wherein the fluctuation information includes part or all of an environmental variable, a characteristic variable, a current charge amount and a chargeable capacity.   The learning type storage battery management system according to claim 12, wherein the environmental variable includes a part or all of a temperature history, a humidity history, a vibration history, and a travel history of the storage battery.   The characteristic variable includes part or all of a service life, a charge / discharge count, a discharge load, a charge status history, a discharge status history, a current charge / discharge rule, and a weight of the storage battery. The described learning battery management system. The quasi-variation information is
In-vehicle storage batteries including electric vehicles, household storage batteries, building storage batteries, housing batteries for collective housing, storage batteries for local energy management systems and other uses of the storage batteries, and
Usage status of the storage battery due to household composition,
The learning type storage battery management system according to claim 8, comprising at least one of the following.

Images