JP7229955B2 - Program, device and method for grouping members with multiple attribute values - Google Patents

Description

The present invention relates to technology for grouping members having attribute values. As an application, it relates to a technique of grouping consumers having one or more storage batteries as members in order to control the charge/discharge power amount according to the supply and demand adjustment power amount.

FIG. 1 is a system configuration diagram in the prior art.

Conventionally, in order to balance the supply and demand of electric energy, there is a technology in which not only the power generation side unilaterally adjusts the supply and demand of electric energy, but also the storage battery on the consumer side is charged and discharged by remote control.
According to FIG. 1, it is assumed that the aggregator device (power intermediary business operator) 1 acquires in advance the power usage information of each storage battery 3 of the consumer.
The electric power company system 2 transmits a supply and demand adjustment request based on the supply and demand adjustment power amount to the aggregator device 1 . On the other hand, the aggregator device 1 calculates the charge/discharge power amount in the storage battery 3 of each consumer based on the supply and demand adjustment power amount, and transmits the charge/discharge power amount to each storage battery 3 . The storage battery 3 charges and discharges power according to the received charge/discharge power amount.

In addition, the management server (aggregator function) determines the charging/discharging ratio of the storage battery to be charged/discharged from the remaining amount of each storage battery of the consumer equipment (energy storage device), etc. There is also a technique for calculating (see Patent Document 1, for example). According to this technique, the consumer-side equipment receives the sharing ratio from the management server, and calculates the sharing power amount to be charged and discharged by the storage battery based on the sharing ratio. Then, the storage battery controls charge/discharge according to the shared power amount.

Here, since the aggregator device 1 calculates the charge/discharge power amount and the sharing ratio for each storage battery, the amount of calculation increases as the number of storage batteries to be controlled increases. As a result, the calculation time is lengthened, and the latest power usage information (for example, remaining amount) of each storage battery cannot be reflected in the calculation result. For example, a consumer's storage battery cannot be requested to discharge more than its remaining capacity. If it takes a long time (for example, two hours) to calculate the sharing ratios of all the storage batteries, the maximum is calculated from the power usage information of each storage battery before the calculation time. In that case, the remaining amount of the storage battery may already be low.
That is, if the calculation is not performed using the latest power usage information, an error will occur between the supply and demand adjusted power amount and the sum of the charge/discharge power amounts of the storage batteries.

On the other hand, as the aggregator device 1, in order to reduce the load of control processing for a large number of storage batteries 3, it is desired to group consumers having one or more storage batteries 3 as members for control. In this case, a hierarchical cluster can be used in which the most similar combinations are grouped (clustered) in order. A typical example is a dendrogram (see, for example, Non-Patent Document 1).

Japanese Patent No. 6183576

"Cluster analysis method 2 (hierarchical cluster analysis)", [online], [searched on January 8, 2020], Internet <URL: https://www.albert2005.co.jp/knowledge/data_mining/cluster /hierarchical_clustering>

For existing dendrograms in hierarchical clusters, there is basically one element value for retrieving the most similar combination. That is, combinations are constructed in order by comparing the similarities between the element values.

In the case of the charging/discharging system as shown in FIG. 1, since not only charging but also discharging is required, a consumer having one or more member storage batteries can handle multiple power consumption and generated power. should be grouped based on the element values of the .
However, as described above, existing dendrograms cannot be grouped by applying multiple element values.

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide a program, apparatus, and method for grouping members having multiple attribute values using hierarchical clusters.

According to the present invention, each member has a first attribute value and a second attribute value, a program to be executed by a computer mounted on a device for grouping a plurality of members,
Group definition means predefining constraints for forming m groups containing n members;
With one or more members as one group, a first group group based on the first attribute value, a second group group based on the second attribute value, the first attribute value and the second group initialization means for setting a third group of groups based on attribute values;
Combine the first groups whose first attribute values are closest to each other to reconstruct the group of first groups, and if any second group is included in any first group, the constraint is a first group reconstruction means for reconstructing the intersection as a third group as long as it satisfies
Combine the second groups whose second attribute values are closest to each other to reconstruct the group of second groups, and if any first group is included in any second group, the constraint is causing the computer to alternately repeat the first group reconstruction means and the second group reconstruction means as the second group reconstruction means for reconstructing the intersection as the third group as long as it satisfies characterized by

According to another embodiment of the program of the present invention,
The initial setting means, with one member as one group, a first group group based on the first attribute value, a second group group based on the second attribute value, the first attribute value and It is also preferred to have the computer function to preset a third group of groups based on the second attribute value.

According to another embodiment of the program of the present invention,
It is also preferable that the first group reconstruction means and the second group reconstruction means cause the computer to combine the groups step by step based on the dendrogram.

According to another embodiment of the program of the present invention,
The group definition means is defined as follows,
Number of group members: n, nmin ≤ n ≤ nmax
nmin≧1, nmax>1
Number of groups with the same number of members n: m
Level of groups with the same number of members n: k (ascending order from groups with fewer members)
Group member increment rate between stages: U (1 < U ≤ 2)
Number of members of group with k=1: scale(1)=nmin
Number of members of the kth group: scale(k)=floor{U×scale(k−1)}
floor{}: Rounded integer Number of members of the K-th group: scale (K) ≤ nmax
It is also preferred to have the computer function to iterate through K stages while incrementing n.

According to another embodiment of the program of the present invention,
Members are consumers with storage batteries,
The first attribute value is an attribute value based on power consumption,
It is also preferred to cause the computer to act like the second attribute value is an attribute value based on generated power.

According to another embodiment of the program of the present invention,
As an aggregator function, in order to control the charge/discharge power amount of each of multiple storage batteries according to the supply and demand adjustment power amount,
group selection means for selecting a combination of the third group such that the total chargeable/dischargeable power amount of each storage battery is close to the chargeable/dischargeable power amount of the entire request, which can be calculated from the demand adjustment power amount;
It is also preferable to have the computer function as storage battery control means for instructing the charge/discharge power amount for each storage battery in the combined group.

According to another embodiment of the program of the present invention,
Group selection means
select multiple combinations of groups for the same demand regulation energy,
It is also preferable to have the computer function to change the combination of groups each time it receives the same demand regulation power amount request.

According to another embodiment of the program of the present invention,
It is also preferable that the storage battery control means causes the computer to function so as to divide the charge/discharge power amount of the group equally among the storage batteries for each group.

According to the present invention, an apparatus for grouping a plurality of members, each member having a first attribute value and a second attribute value, comprising:
Group definition means predefining constraints for forming m groups containing n members;
With one or more members as one group, a first group group based on the first attribute value, a second group group based on the second attribute value, the first attribute value and the second group an initialization means for preconfiguring a group of third groups based on attribute values;
Combine the first groups whose first attribute values are closest to each other to reconstruct the group of first groups, and if any second group is included in any first group, the constraint is a first group reconstruction means for configuring the intersection as a third group as long as it satisfies
Combine the second groups whose second attribute values are closest to each other to reconstruct the group of second groups, and if any first group is included in any second group, the constraint is a second group reconstructing means for constructing the intersection as a third group as long as they satisfy characterized by

According to the present invention, a device grouping method for grouping a plurality of members, each member having a first attribute value and a second attribute value, comprising:
The device
Predefine constraints that constitute m groups containing n members,
With one or more members as one group, a first group group based on the first attribute value, a second group group based on the second attribute value, the first attribute value and the second group preconfiguring a group of third groups based on attribute values,
Combine the first groups whose first attribute values are closest to each other to reconstruct the group of first groups, and if any second group is included in any first group, the constraint is a first step of forming the intersection as a third group as long as it satisfies
Combine the second groups whose second attribute values are closest to each other to reconstruct the group of second groups, and if any first group is included in any second group, the constraint is A second step of forming the intersection as a third group is executed as long as the conditions are satisfied, and the first step and the second step are alternately executed.

According to the program, device, and method of the present invention, members having multiple attribute values can be grouped using hierarchical clusters.
As a result, regarding the determination of the charge/discharge power amount for each storage battery according to the supply and demand adjustment power amount, even if the number of storage batteries to be controlled increases, the calculation amount in selecting the storage battery and determining the charge/discharge power amount is suppressed. be able to.

1 is a system configuration diagram in a conventional technique; FIG. 1 is a system configuration diagram in the present invention; FIG. 1 is a functional configuration diagram of an aggregator device according to the present invention; FIG. It is a storage battery database representing power usage information for each storage battery. FIG. 10 is an explanatory diagram showing processing of a group definition unit in the present invention; FIG. 10 is an explanatory diagram of the 0th to 4th stages showing the process of group restructuring; FIG. 10 is an explanatory diagram of fifth to seventh stages showing the process of group restructuring; FIG. 12 is an explanatory diagram of the ninth and tenth stages showing the background of group reconfiguration; FIG. 4 is an explanatory diagram showing chargeable/dischargeable power amounts for each group; FIG. 4 is an explanatory diagram showing processing of a group selection unit in the present invention; It is a graph showing the processing time with respect to the number of storage batteries.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 2 is a system configuration diagram in the present invention.

<Power company system 2>
The electric power company system 2 transmits a “supply and demand adjustment request” to the aggregator device 1 .
The demand/supply adjustment request includes a “supply/demand adjustment power amount” and a “time”. For example:
Time 2018/4/1 12:00 Supply and demand adjustment request "2018/4/1 13:00-13:30, 140 kWh, discharge"
The time interval for updating the charge/discharge power amount is set to 30 minutes, but of course, any time interval (for example, 5 minutes, 10 minutes, 15 minutes, etc.) may be used. Also, the supply and demand adjustment power amount may be kW, which is electric power, instead of kWh, which is the amount of power.
In addition, even if the schedule is predetermined, for example, charging from 23:00 to 06:59 the next day and discharging from 07:00 to 22:59 for the time period outside the supply and demand adjustment good.

The supply and demand adjustment power amount means how much the power company needs to prepare the supply amount (=power generation amount). In other words, it means how much adjustment can be made compared to the normal supply amount.
“Supply and demand adjustment power amount” specifically refers to the difference (adjustment amount) based on the normal time when there is no supply and demand adjustment request.
The following description is based on the assumption that the storage battery is not charging or discharging during normal times when there is no demand-supply adjustment request. Therefore, since the charge/discharge power from the storage battery is zero during normal times when there is no demand/supply adjustment request, the supply/demand adjustment amount and the charge/discharge power amount from the storage battery match.
However, if the storage battery is charging/discharging normally, it is necessary to estimate how much adjustment can be made in consideration of the charge/discharge power amount, and perform charging/discharging control of the storage battery.

According to the embodiment of the present invention, the aggregator device 1 and the electric power company system 2 are described as separate systems, but of course the electric power company system 2 having the function of determining the supply and demand adjustment amount includes the aggregator device 1. It may be configured integrally as follows.

<Storage battery 3>
The storage battery 3 is configured for each group, and charges and discharges according to the received charge/discharge power amount. Also, the power usage information of the storage battery is transmitted to the aggregator device 1 as appropriate.
Storage battery 3 communicates with HEMS-GW (Home Energy Management System - GateWay), HOME-GW (HOME GateWay), smart meter (power meter equipped with communication function, or HEMS-GW installed on distribution board) possible wattmeter). The aggregator device 1 transmits the charge/discharge power amount of each storage battery to the HEMS-GW or the like. And HEMS-GW etc. control charging/discharging of each storage battery.
The HEMS-GW or HOME-GW may receive data in the OpenADR data format from the aggregator device 1 and transmit the data to the storage battery in the ECHONet-Lite data format.

<Aggregator device 1>
The aggregator device 1 receives a supply and demand adjustment request from the electric power company system 2 and learns the supply and demand adjustment power amount.
The aggregator device 1 controls the charge/discharge power amount of each of the plurality of storage batteries according to the supply and demand adjustment power amount. For this purpose, a plurality of storage batteries are grouped, and the charging/discharging power amount equally divided for each storage battery is determined according to the supply and demand adjustment power amount. Then, each storage battery 3 is notified of the charge/discharge power amount.
As a result, even if the number of storage batteries increases, it is possible to suppress an increase in the amount of calculation and calculate the charge/discharge power amount of each storage battery according to the supply and demand adjustment power amount.
In particular, the aggregator device 1 of the present invention groups a plurality of members with each consumer as a member based on the first attribute value and the second attribute value of each member.

FIG. 3 is a functional configuration diagram of an aggregator device according to the present invention.

According to FIG. 3, the aggregator device 1 includes a storage battery database 10, a group definition unit 11, an initial setting unit 12, a first group reconstruction unit 13, a second group reconstruction unit 14, a group selection It has a unit 15 and a storage battery control unit 16 . These functional configuration units are implemented by executing a program that causes a computer installed in the aggregator device to function. In addition, the processing flow of these functional components can also be understood as a supply and demand adjustment method for an aggregator device.

[Storage battery database 10]
The storage battery database 10 associates power usage information with each storage battery. The power usage information is received and collected periodically from each storage battery 3 .

FIG. 4 is a storage battery database representing power usage information for each storage battery.

The storage battery database 10 stores power usage information in association with each storage battery. It is assumed that the power usage information has at least two attribute values for each storage battery (member) that is an element of grouping.
First attribute value: consumption data (attribute value based on power consumption)
(For example, average power consumption, number of family members, presence or absence of all electrification)
Second attribute value: power generation data (attribute value based on power generation)
(For example, average power generation, address ID, generator type)
Since the consumed power and the generated power change depending on the time period (for example, the date of each day), they are expressed as average values in a predetermined time period.

[Group definition unit 11]
The group definition unit 11 predefines constraints for forming m groups containing n members. Specifically, the group size is set, and each group is defined by repeating k steps of forming m groups each containing n members while incrementing n.

FIG. 5 is an explanatory diagram showing the processing of the group definition section in the present invention.

The group definition unit 11 defines parameters as follows.
Number of group members: n, nmin ≤ n ≤ nmax
nmin≧1, nmax>1
Number of groups with the same number of members n: m
Level of groups with the same number of members n: k (ascending order from groups with fewer members)
Group member increment rate between stages: U (1 < U ≤ 2)
Number of members of group with k=1: scale(1)=nmin
Number of members of the kth group: scale(k)=floor{U×scale(k−1)}
floor{}: Rounded integer Number of members of the K-th group: scale (K) ≤ nmax

Then, the group definition unit 11 repeats k steps while incrementing n until the total number of members of all groups is greater than or equal to the total number of storage batteries. Specifically, it is configured step by step as follows.
scale(k) group ID (number of members)
k=1 A1(1), A2(1), A3(1)
A maximum of 3 groups consisting of one storage battery k=2 B1(2), B2(2), B3(2)
Maximum 3 groups consisting of 2 storage batteries k=3 C1(4), C2(4), C3(4)
A maximum of 3 groups consisting of 4 batteries K=k=4 D1(8), D2(8), D3(8)
Up to 3 groups consisting of 8 storage batteries By setting the upper limit k of storage batteries in each group in stages, even if the number of storage batteries to be controlled increases, the increase in the number of groups is suppressed. be able to.

According to the present invention, the amount of calculation is suppressed by calculating the charge/discharge power amount only for each group. In addition, a plurality of m groups with different number of members n are intentionally formed.
For example, when only groups with a large number of members are created, it is difficult to adjust the value close to the supply and demand adjustment power amount by combining those groups. For example, in the case of 140 kWh of power supply and demand adjustment, it is difficult to make fine adjustments such as 1 kWh or 2 kWh depending on the group combination. For this purpose, groups of various sizes are created, ranging from groups with a small number of members to groups with a large number of members.

Also, if there is only one group with a certain number of members, the group may be selected many times. Therefore, a plurality of m groups with the same number of members are created. This avoids selecting the same group multiple times.
For example, if the supply and demand adjustment power amount is not reached without requesting more than half of all the storage batteries, the group with the largest number of members will always be selected. For this reason, by preparing a plurality of groups having the maximum number of members, it is possible to prevent selection from being concentrated in one group.

[Initial setting part 12]
The initial setting unit 12 initially sets the following three groups with one or more members as one group.
a first group based on the first attribute value a second group based on the second attribute value a third group based on the first attribute value and the second attribute value wherein one The members may be set in advance as one group, or different numbers of members may be set in advance for each group. In other words, the group can be reconfigured to some extent from the stage when the group has been configured.

FIG. 6 is an explanatory diagram of stages 0 to 4 showing the process of group reconfiguration.

According to FIG. 6, groups are set as follows at the initial stage 0 (t0).
(t0) Six first groups with one member and six second groups with one member are set in advance.
First group (consumption): {a} {b} {c} {d} {e} {f}
Second group (power generation): {a} {b} {c} {d} {e} {f}
Third Group (Consumption/Generation): {a} {b} {c} {d} {e} {f}

[First group reconstruction unit 13 and second group reconstruction unit 14]
The first group reconstruction unit 13 classifies groups obtained by clustering members (storage batteries) having similar first attribute values (for example, power consumption) into hierarchical clusters (dendrogram ) are combined step by step.
Similarly, the second group reconstruction unit 14 classifies groups obtained by clustering members (storage batteries) having similar second attribute values (for example, generated power) into hierarchical clusters under the defined group constraints. Based on the (dendrogram), it is combined step by step.

The “similarity” between groups is based on the Euclidean distance represented by a vector for each item of consumption data or power generation data of members included in each group. Groups with the shortest Euclidean distance are combined.

As a result, the first group reconfiguring unit 13 reconfigures a group of first groups by combining first groups having the closest first attribute values. At this time, if an arbitrary second group is included in an arbitrary first group, its intersection (∩) is reconfigured as the third group as long as the constraint is satisfied.
The second group reconfiguring unit 14 reconfigures a group of second groups by combining second groups having the closest second attribute values. At this time, if an arbitrary first group is included in an arbitrary second group, the intersection is reconfigured as the third group as long as the constraint is satisfied.
Then, the first group reconstruction unit 13 and the second group reconstruction unit 14 are alternately repeated.
Ultimately, a plurality of storage batteries (members) belonging to the same group have similar power usage information (consumption data and power generation data), so the same charge/discharge power amount can be requested. That is, the aggregator device 1 only needs to select the group and calculate the charge/discharge power amount, and can suppress the amount of calculation.

According to the example of FIG. 6, the group definition unit 11 is finally under the following group group constraints.
3 groups with 1 or less members: {1} {1} {1}
3 groups with 2 or less members: {2} {2} {2}
3 groups with 4 or less members: {4} {4} {4}
At the initial setting stage, even if there are three or more groups with one member, it is allowed. Finally, it is made to fit within the restrictions of the group group of the group definition section 11 .
A set C of constraints on the number of members (where C is sorted in descending order by the number of members) (e.g. C = {4, 4, 4, 2, 2, 2, 1, 1, 1}), and
Compare the set S of the number of members of the group of the third group under reconstruction, where S is sorted in descending order by the number of members (e.g. S = {2, 1, 1, 1, 1}),
continuing the joint process of reconstruction if for every element si of S satisfies ci≧si for each element of C;
Otherwise, stop the merging operation, return to the state before merging, and restart reconstruction.

According to FIG. 6, the process of group restructuring is traced as follows.
(t1) Combine the first groups whose first attribute values are the closest to each other to reconstruct a group of first groups. At this time, the second groups {a} {b} are included in the first groups {a, b}, respectively, but since the number of members is 1 {a} {b}, the third group is exactly the same as t0.
First group (consumption): {a,b} {c} {d} {e} {f}
Second group (power generation): {a} {b} {c} {d} {e} {f}
Third Group (Consumption/Generation): {a} {b} {c} {d} {e} {f}
A set S of the number of members of the third group group = {1, 1, 1, 1, 1, 1}
A set of constraint member counts C={4, 4, 4, 2, 2, 2, 1, 1, 1}
Since ci ≥ si, the reconstruction process continues.

(t2) Next, second groups having the closest second attribute values are combined to reconstruct a group of second groups. At this time, since the first group {a, b} is included in the second group {a, b}, the common set is reconstructed as the third group.
First group (consumption): {a,b} {c} {d} {e} {f}
Second group (power generation): {a, b} {c} {d} {e} {f}
Third group (consumption/generation): {a,b} {c} {d} {e} {f}
A set S of the number of members of the third group group = {2, 1, 1, 1, 1}
A set of constraint member counts C={4, 4, 4, 2, 2, 2, 1, 1, 1}
Since ci ≥ si, the reconstruction process continues.

(t3) Next, the first groups having the closest first attribute values are combined to reconstruct a group of first groups. At this time, each of the second groups {c} {d} is included in the first group {c, d}, but since the number of members is 1 {c} {d}, the third group is exactly the same as t2.
First group (consumption): {a,b} {c,d} {e}{f}
Second group (power generation): {a, b} {c} {d} {e} {f}
Third group (consumption/generation): {a,b} {c} {d} {e} {f}
A set S of the number of members of the third group group = {2, 1, 1, 1, 1}
A set of constraint member counts C={4, 4, 4, 2, 2, 2, 1, 1, 1}
Since ci ≥si is satisfied, the reconstruction process continues.

(t4) Next, the second groups having the closest second attribute values are combined to reconstruct a group of second groups. At this time, the first group {e} is included in the second group {d, e}, but since the number of members {e} is 1, the third group is exactly the same as t3. .
First group (consumption): {a,b} {c,d} {e}{f}
Second group (power generation): {a, b} {c} {d, e} {f}
Third group (consumption/generation): {a,b} {c} {d} {e} {f}
A set S of the number of members of the third group group = {2, 1, 1, 1, 1}
A set of constraint member counts C={4, 4, 4, 2, 2, 2, 1, 1, 1}
Since ci ≥si is satisfied, the reconstruction process continues.

FIG. 7 is an explanatory diagram of fifth to seventh stages showing the background of group reconfiguration.

(t5) Next, the first groups having the closest first attribute values are combined to reconstruct the group of first groups. At this time, the second group {f} is included in the first group {e, f}, but since the number of members {f} is 1, the third group is exactly the same as t4. .
First group (consumption): {a,b} {c,d} {e,f}
Second group (power generation): {a, b} {c} {d, e} {f}
Third group (consumption/generation): {a,b} {c} {d} {e} {f}
A set S of the number of members of the third group group = {2, 1, 1, 1, 1}
A set of constraint member counts C={4, 4, 4, 2, 2, 2, 1, 1, 1}
Since ci ≥si is satisfied, the reconstruction process continues.

(t6) Next, the second groups having the closest second attribute values are combined to reconstruct a group of second groups. At this time, since the first group {a, b} is included in the second group {a, b, c}, the common set is reconstructed as the third group. However, since the third group has already constructed {a, b}, it is exactly the same as t5.
First group (consumption): {a,b} {c,d} {e,f}
Second group (power generation): {a, b, c} {d, e} {f}
Third group (consumption/generation): {a,b} {c} {d} {e} {f}
A set S of the number of members of the third group group = {2, 1, 1, 1, 1}
A set of constraint member counts C={4, 4, 4, 2, 2, 2, 1, 1, 1}
Since ci ≥si is satisfied, the reconstruction process continues.

(t7) Next, the first groups having the closest first attribute values are combined to reconstruct a group of first groups. At this time, since the second group {a, b, c} is included in the first group {a, b, c, d}, the common set is reconstructed as the third group.
First group (consumption): {a,b,c,d} {e,f}
Second group (power generation): {a, b, c} {d, e} {f}
Third group (consumption/generation): {a,b,c}{d}{e}{f}
A set S of the number of members of the third group group = {3, 1, 1, 1}
A set of constraint member counts C={4, 4, 4, 2, 2, 2, 1, 1, 1}
Since ci ≥si is satisfied, the reconstruction process continues.

(t8) Next, the second groups having the closest second attribute values are combined to reconstruct a group of second groups. At this time, since the first group {a, b, c, d} is included in the second group {a, b, c, d, e}, its intersection {a, b, c, d } as the third group.
First group (consumption): {a,b,c,d} {e,f}
Second group (power generation): {a, b, c, d, e} {f}
Third group (consumption/generation): {a, b, c, d} {e} {f}
A set S of the number of members of the third group group = {4, 1, 1}
A set of constraint member counts C={4, 4, 4, 2, 2, 2, 1, 1, 1}
Since ci ≥si is satisfied, the reconstruction process continues.

FIG. 8 is an explanatory diagram of the 9th and 10th stages showing the process of group reconfiguration.

(t9) Next, the first groups having the closest first attribute values are combined to reconstruct a group of first groups. At this time, since the second group {a, b, c, d, e} is included in the first group {a, b, c, d, e, f}, its intersection {a, b , c, d} as the third group.
First group (consumption): {a, b, c, d, e, f}
Second group (power generation): {a, b, c, d, e} {f}
Third group (consumption/generation): {a, b, c, d, e} {f}
A set S of the number of members of the third group group = {5, 1}
A set of constraint member counts C={4, 4, 4, 2, 2, 2, 1, 1, 1}
Since ci≧si is not satisfied, the state returns to (t8) and the reconstruction process is continued.

(t10) Finally, the second groups having the closest second attribute values are combined to reconstruct a group of second groups. At this time, since the first group {e, f} is included in the second group {a, b, c, d, e, f}, the intersection {e, f} is the third group Configure as
First group (consumption): {a,b,c,d} {e,f}
Second group (power generation): {a, b, c, d, e, f}
Third group (consumption/generation): {a,b,c,d} {e,f}
Although ci≧si is satisfied, since the connection processing by the dendrogram is finished, the reconstruction processing is also finished, and the first group, the second group, and the third group at (t10) are brought to the final state.

The first group reconstruction unit 13 and the second group reconstruction unit 14 described above may reconstruct groups at predetermined time intervals or when a predetermined condition is satisfied. For example, it is preferable to reconfigure the group for a certain period of time such as one week or one month, or at the timing when a storage battery to be controlled is added.

FIG. 9 is an explanatory diagram showing chargeable/dischargeable power amounts for each group.

According to FIG. 9A, groups defined by the group definition unit 11 are represented. The size of a rectangle representing a group and the numbers in parentheses represent the maximum number of members in each group.
FIG. 9B shows an example in which, for example, six groups are configured.
FIG. 9(c) shows the chargeable/dischargeable power amounts of all the storage batteries for each group. The size of the rectangle representing the group represents the chargeable/dischargeable power amount of each group.
According to FIG.9(d), each group is sorted according to chargeable/dischargeable electric energy.

Replacing the example of FIG. 9 with the group of the third group of t10 in FIG. 8 results in two groups as follows.
B1={e,f}
C1={a, b, c, d}

[Group selection unit 15]
If the charge/discharge power from the storage battery during normal operation is not zero, the group selection unit 15 determines that the sum of the chargeable/dischargeable power amounts of the storage batteries can be calculated from the demand adjustment power amount. Select combinations in the third group such that . The charge/discharge power amount for the entire request is calculated by the following formula.
Amount of charge/discharge power for the entire request =
Supply and demand adjustment power amount from electric power company system + total charge/discharge prediction amount in normal time
Charge/discharge power (positive is discharging, negative is charging)
Supply and demand adjustment power amount (plus is request for reduction, minus is request for increase)
Overall predicted charge/discharge amount during normal operation (plus indicates discharge, minus indicates charge)
The total charge/discharge predicted amount during normal operation is the amount of electric power predicted to be charged/discharged when there is no supply and demand adjustment.
Then, for example, the plurality of storage batteries are sorted as shown in FIG. Note that the order of priority for sorting the plurality of storage batteries may be a weighted average of remaining amounts.

FIG. 10 is an explanatory diagram showing the processing of the group selection section in the present invention.

According to FIG. 10, it is going to respond to the following supply and demand adjustment requests.
Time 2019/4/1 12:00 Supply and demand adjustment request "2019/4/1 13:00-13:30, 140 kWh, discharge"
Here, six groups in FIG. 9D will be explained.

(S1) First, select group C2. At this time, since the chargeable/dischargeable power amount of group C2 is lower than the demand adjustment power amount, it continues as it is.
(S2) Next, group C2+C1 is selected. At this time, since the chargeable/dischargeable power amount of group C2+C1 is greater than or equal to the demand adjustment power amount, group C2 selected before exceeding the amount is determined. Also, the group C2+C1 is stored as a combination candidate.
Combination candidate: C2 + C1
(S3) Next, group C2+A1 is selected. At this time, since the chargeable/dischargeable power amount of the group C2+A1 matches the demand adjustment power amount, the group C2+A1 is stored as a combination candidate.
Combination candidate: C2 + C1
C2 + A1
(S4) Next, group C2+B2 is selected. At this time, since the chargeable/dischargeable power amount of group C2+B2 is lower than the demand adjustment power amount, it continues as it is.
(S5) Next, group C2+B2+B3 is selected. At this time, since the chargeable/dischargeable power amount of the group C2+B2+B3 is equal to or greater than the demand adjustment power amount, the selected group C2+B2 is determined before exceeding the demand adjustment power amount. Also, the group C2+B2+B3 is stored as a combination candidate.
Combination candidate: C2 + C1
C2 + A1
C2+B2+B3
(S6) Next, group C2+B2+A2 is selected. At this time, since the chargeable/dischargeable power amount of the group C2+B2+A2 matches the demand adjustment power amount, the group C2+B2+A2 is stored as a combination candidate.
Combination candidate: C2 + C1
C2 + A1
C2+B2+B3
C2+B2+A2
Finally, the combination of the group with the lowest chargeable/dischargeable power amount (closest to the supply and demand adjustment power amount) is selected from all the combination candidates. According to FIG. 8, the following two combinations are selected.
Final selected combination: C2+A1
C2+B2+A2
In this way, the group selection unit 15 can select a plurality of combinations of groups for the same demand-adjusted power amount.

Moreover, the group selection unit 15 may select a plurality of combinations of groups for the same demand-adjusted power amount. Accordingly, the group combination is changed each time the same request for the demand adjustment power amount is received. It is possible to avoid situations in which requests are concentrated in a specific group.

[Storage battery control unit 16]
The storage battery control unit 16 instructs each storage battery in the combined group to charge/discharge electric energy. At this time, the storage battery control unit 16 equally divides the charge/discharge power amount of the group into each storage battery for each group. Since storage batteries with a high degree of similarity in power usage information are grouped together, the charge/discharge power amount can also be equally divided.

FIG. 11 is a graph showing the processing time with respect to the number of storage batteries.

According to FIG. 11, it is assumed that the following parameters are set for the present invention.
Number of group members: nmin = 1
nmax = 32,768
Number of groups with the same number of members n: m = 1
Group member increment rate between stages: U=2
Here, according to FIG. 11, when trying to control the number of storage batteries of 65,535, the calculation time of the charge/discharge power amount for each storage battery is 6,553.5 seconds (about 109 minutes). The power amount calculation time is 1.6 seconds. For example, if the power usage information of the storage battery can be acquired every 30 minutes, it is preferable that the calculation time of the charge/discharge power amount is also within 30 minutes. According to the above example, it takes 109 minutes to calculate for each storage battery, and it cannot be calculated within 30 minutes. Become.

As described in detail above, according to the program, apparatus and method of the present invention, hierarchical clusters can be used to group members having multiple attribute values.
As a result, regarding the determination of the charge/discharge power amount for each storage battery according to the supply and demand adjustment power amount, even if the number of storage batteries to be controlled increases, the calculation amount in selecting the storage battery and determining the charge/discharge power amount is suppressed. be able to.
In particular, even if the number of storage batteries increases, the amount of calculation is suppressed and the calculation time does not become long. Therefore, it is possible to calculate using the new power usage information in each storage battery, and it is possible to reduce the error between the supply and demand adjustment power amount and the sum total of the charge/discharge power amount of each storage battery.

For the various embodiments of the present invention described above, various changes, modifications and omissions within the spirit and scope of the present invention can be easily made by those skilled in the art. The foregoing description is exemplary only and is not intended to be limiting. The invention is to be limited only as limited by the claims and the equivalents thereof.

1 aggregator device 10 storage battery database 11 group definition unit 12 initial setting unit 13 first group reconstruction unit 14 second group reconstruction unit 15 group selection unit 15 storage battery control unit 2 electric power company system 3 storage battery

Claims (10)

Each member has a first attribute value and a second attribute value, a program to be executed by a computer mounted on a device for grouping a plurality of members,
Group definition means predefining constraints for forming m groups containing n members;
With one or more members as one group, a first group group based on the first attribute value, a second group group based on the second attribute value, the first attribute value and the second group initialization means for setting a third group of groups based on attribute values;
Combine the first groups whose first attribute values are closest to each other to reconstruct the group of first groups, and if any second group is included in any first group, the constraint is a first group reconstruction means for reconstructing the intersection as a third group as long as it satisfies
Combine the second groups whose second attribute values are closest to each other to reconstruct the group of second groups, and if any first group is included in any second group, the constraint is causing the computer to alternately repeat the first group reconstruction means and the second group reconstruction means as the second group reconstruction means for reconstructing the intersection as the third group as long as it satisfies A program characterized by The initial setting means, with one member as one group, a first group group based on the first attribute value, a second group group based on the second attribute value, the first attribute value and 2. The program of claim 1, operable to cause a computer to preset a set of third groups based on second attribute values. 3. The program according to claim 1, wherein the first group reconstruction means and the second group reconstruction means cause a computer to function so as to combine groups step by step based on a dendrogram. The group definition means is defined as follows,
Number of group members: n, nmin ≤ n ≤ nmax
nmin≧1, nmax>1
Number of groups with the same number of members n: m
Level of groups with the same number of members n: k (ascending order from groups with fewer members)
Group member increment rate between stages: U (1 < U ≤ 2)
Number of members of group with k=1: scale(1)=nmin
Number of members of the kth group: scale(k)=floor{U×scale(k−1)}
floor{}: Rounded integer Number of members of the K-th group: scale (K) ≤ nmax
4. A program according to any one of claims 1 to 3, causing the computer to repeat K steps while incrementing n. Members are consumers with storage batteries,
The first attribute value is an attribute value based on power consumption,
5. The program according to any one of claims 1 to 4, wherein the second attribute value causes the computer to function as an attribute value based on generated power. As an aggregator function, in order to control the charge/discharge power amount of each of multiple storage batteries according to the supply and demand adjustment power amount,
group selection means for selecting a combination of the third group such that the total chargeable/dischargeable power amount of each storage battery is close to the chargeable/dischargeable power amount of the entire request, which can be calculated from the demand adjustment power amount;
6. The program according to claim 5, causing the computer to function as storage battery control means for instructing charge/discharge power amounts for each storage battery in the combined group. Group selection means
select multiple combinations of groups for the same demand regulation energy,
7. The program according to claim 6, causing the computer to change the combination of groups each time it receives the same demand regulation power amount request. 8. The program according to claim 6, wherein the storage battery control means causes the computer to function so as to equally divide the charging/discharging power amount of the group to each storage battery for each group. An apparatus for grouping a plurality of members, each member having a first attribute value and a second attribute value, comprising:
group definition means predefining constraints for forming m groups containing n members;
With one or more members as one group, a first group group based on the first attribute value, a second group group based on the second attribute value, the first attribute value and the second group an initialization means for preconfiguring a group of third groups based on attribute values;
Combine the first groups whose first attribute values are closest to each other to reconstruct the group of first groups, and if any second group is included in any first group, the constraint is a first group reconstruction means for configuring the intersection as a third group as long as it satisfies
Combine the second groups whose second attribute values are closest to each other to reconstruct the group of second groups, and if any first group is included in any second group, the constraint is a second group reconstructing means for constructing the intersection as a third group as long as they satisfy A device characterized by: A device grouping method for grouping a plurality of members, each member having a first attribute value and a second attribute value, comprising:
The device
Predefine constraints that constitute m groups containing n members,
With one or more members as one group, a first group group based on the first attribute value, a second group group based on the second attribute value, the first attribute value and the second group preconfiguring a group of third groups based on attribute values,
Combine the first groups whose first attribute values are closest to each other to reconstruct the group of first groups, and if any second group is included in any first group, the constraint is a first step of forming the intersection as a third group as long as it satisfies
Combine the second groups whose second attribute values are closest to each other to reconstruct the group of second groups, and if any first group is included in any second group, the constraint is and a second step of forming the intersection as a third group as long as it satisfies the conditions, and alternately repeating the first step and the second step. .

Images