JP7615644B2 - Virtual power plant formation control system and formation control method - Google Patents

Description

The present invention relates to a virtual power plant formation control system and a formation control method thereof.

In recent years, in order to realize a more stable and less wasteful power system, development has been progressing on virtual power plants (VPPs) that integrate and control various energy resources owned by consumers, such as distributed power sources and power storage facilities, and function as if they were a single power plant (see Patent Document 1).
By forming a virtual power plant, it becomes possible to adjust the balance of supply and demand for electricity through demand response such as up-demand response (DR) and down-demand response (DR), instead of adjusting the output of large-scale power plants such as thermal power plants as has been done in the past.

JP 2020-108301 A

Comparing the formation of a virtual power plant in an urban area with the formation of a virtual power plant in a surrounding area surrounding the urban area, urban areas are often able to form a virtual power plant efficiently because they are rich in various types of energy resources such as power generation facilities and storage facilities, whereas surrounding areas are often biased toward certain types of energy resources (such as solar power generation facilities), making it difficult to form a virtual power plant.
As a result, the current situation is that energy resources available in the surrounding area are not being effectively utilized as targets for virtual power plants.

The present invention aims to provide a system and method for controlling the formation of a virtual power plant that efficiently utilizes various energy resources that are ubiquitous, regardless of whether the plant is in an urban area or a surrounding area, to form a virtual power plant and ensure a better balance between the supply and demand of electricity.

The main present invention for solving the above-mentioned problems is a virtual power plant formation control system that integrates and controls multiple energy resources owned by multiple consumers and distributed throughout a management area including urban areas and surrounding areas via a network, and functions as if they were a single power plant, and includes an attribute memory unit in which attribute information including information indicating the power usage status of the multiple consumers is stored, a learning model memory unit in which a learning model that learns the correlation between the attribute information, which is an explanatory variable, and the classification result of the attribute information, which is an objective variable, is stored, a learning unit that predicts the classification result for the attribute information using the learning model, and an integration control unit that, when a down DR or up DR occurs, selectively integrates from among the multiple energy resources, multiple energy resources that correspond to the classification result that is capable of responding to the down DR or up DR based on the prediction result by the learning unit, and controls the integrated energy resources.
Other features of the present invention will become apparent from the accompanying drawings and the description of this specification.

According to the present invention, it is possible to create a virtual power plant by efficiently utilizing various energy resources that are ubiquitous, regardless of whether they are in an urban area or a surrounding area, and to more reliably balance the supply and demand of electricity.

FIG. 1 is an overall view for explaining the formation of a virtual power plant according to an embodiment of the present invention. FIG. 2 is a block diagram showing the functions of an aggregator server. 11 is a diagram showing an example of table data of attribute information A stored in an attribute storage unit; FIG. 13 is a diagram showing an example of table data of attribute information B stored in an attribute storage unit; FIG. 11 is a diagram showing an example of table data of attribute information C stored in an attribute storage unit; FIG. FIG. 1 is a diagram illustrating an example of a power system. 10 is a flowchart showing an example of an operation of an integrated control unit when forming a virtual power plant. 11 is a flowchart showing an example of an automated demand response flow.

The present specification and accompanying drawings make clear at least the following:

FIG. 1 is an overall diagram for explaining the formation of a virtual power plant according to the present embodiment. The virtual power plant is operated by an aggregator. In this embodiment, the aggregator is a general term for a resource aggregator and an aggregation coordinator, and may refer to either the resource aggregator or the aggregation coordinator, or both. The resource aggregator controls the energy resources based on a contract with each consumer who owns the energy resources. The aggregation coordinator aggregates the amount of power controlled by each resource aggregator, and trades power with a general power transmission and distribution business operator and a retail electricity business operator. In this embodiment, the general power transmission and distribution business operator and the retail electricity business operator are collectively referred to as an electricity business operator. Therefore, the electricity business operator may refer to either the general power transmission and distribution business operator and the retail electricity business operator, or both.
In this embodiment, for example, it is assumed that a management area 10 in which a virtual power plant is to be formed is determined by an aggregator. A plurality of energy resources owned by a plurality of consumers, such as energy generation, energy storage, and energy saving resources, are distributed and installed in this management area 10. The scale of the management area 10 may be a relatively small one such as a city, ward, town, or village, or a relatively large one spanning prefectures.

In this embodiment, the management area 10 is roughly divided into, for example, an urban area 10A and a surrounding area 10B surrounding the urban area 10A. In the urban area 10A, commercial facilities and industrial facilities for economic activities are densely concentrated, and many energy resources 20 are installed in a concentrated manner. On the other hand, in the surrounding area 10B, there tends to be more residential buildings than facilities for economic activities, and the energy resources 20 are installed sparsely, although in smaller numbers than in the urban area 10A. Note that the multiple energy resources distributed in the urban area 10A are referred to as 20A, and the multiple energy resources distributed in the surrounding area 10B are referred to as 20B. These energy resources 20A and 20B are connected to a power grid (not shown), but are selectively integrated to be able to communicate via a network 30 such as IoT according to instructions from an aggregator server 50 described later.

The virtual power plant is formed by integrating, via the network 30, the energy resources 20A and 20B (for example, within the bold frame marked VPP in FIG. 1) that correspond to the information corresponding to the attribute information of a plurality of consumers and the energy resources 20A and 20B from among the energy resources 20A and 20B in the management area 10, regardless of the urban area 10A or the surrounding area 10B. The virtual power plant formed in this way intentionally increases or decreases the demand for electricity in response to demand responses such as up-DR and down-DR that occur according to the supply amount of electricity at that time, so as to contribute to balancing the demand amount of electricity by consumers with the supply amount of electricity. In other words, the virtual power plant functions as a virtual power plant. Here, in this embodiment, the attribute information is used as an explanatory variable, the classification result of the attribute information is used as an objective variable, and a learning model that learns the correlation between these attribute information and the classification result is prepared, and the energy resources 20A and 20B are selectively integrated according to the classification result of the attribute information predicted using this learning model (information corresponding to the attribute information) to form the virtual power plant.

The energy resources 20A and 20B are connected to a relay system 40 (also called a power management system) that relays communications between a GW (gateway), HEMS (Home Energy Management System), BEMS (Building Energy Management System), FEMS (Factory Energy Management System), etc., so that a virtual power plant can be formed through the network 30. Furthermore, the energy resources 20A and 20B are communicatively connected to an aggregator server 50 that integrates the energy resources 20A and 20B and thereby forms a virtual power plant, via the relay system 40 and the network 30.

Demand response is a demand response command for changing the power demand pattern by controlling the power demand of the multiple energy resources 20A and 20B forming the virtual power plant so that the power supply and the power demand by the consumer are balanced. There are two types of demand response: up DR and down DR. Up DR is a command for increasing the power demand (consumption) of the multiple energy resources 20A and 20B in the virtual power plant when the power demand is less than the power supply. For example, when there is a risk that the power supply of a solar power generation system connected to the power grid will be oversupplied due to the influence of fine weather, an up DR is generated. For example, when the energy resources 20A and 20B forming the virtual power plant include a storage battery (storage energy resource), the demand for the storage battery is increased to charge the storage battery with the amount of power that would be oversupplied. On the other hand, down DR is a command for decreasing the power demand of the multiple energy resources 20A and 20B in the virtual power plant when the power demand is greater than the power supply. For example, if there is a high demand for air conditioners and the like during a heatwave and there is a risk of a shortage in the supply of electricity, a down DR is generated to reduce the demand for one of the energy resources 20A, 20B in the virtual power plant. In this way, changing the demand pattern for multiple energy resources 20A, 20B in the virtual power plant can contribute to balancing the supply and demand of electricity.

The energy generating resource, which is one type of the energy resources 20A and 20B, is a resource that generates electricity, and examples thereof include a photovoltaic power generation system, a cogeneration system, a system that utilizes renewable energy heat (geothermal heat, solar heat, snow and ice heat, etc.), and a fuel cell vehicle. The energy storage resource, which is another type of the energy resources 20A and 20B, is a resource that stores electric energy, and examples thereof include a home storage battery, a system storage battery, and a heat pump water heater. The energy saving resource, which is another type of the energy resources 20A and 20B, is a resource that saves electricity, and examples thereof include a business/industrial EMS (Energy Management System), lighting equipment that has been converted to IoT, air conditioning equipment, and refrigeration equipment. The virtual power plant is formed by selectively integrating the energy generating resources, energy storage resources, and energy saving resources included in these multiple energy resources 20A and 20B based on the attributes of multiple consumers and the multiple energy resources 20A and 20B.

The aggregator server 50 is a server managed by an aggregator. As described above, aggregator is a general term for a resource aggregator and an aggregation coordinator, so there are servers managed by resource aggregators (also referred to as resource aggregator servers) and servers managed by aggregation coordinators (also referred to as aggregation coordinator servers). In this embodiment, the resource aggregator servers and aggregation coordinator servers are collectively referred to as aggregator servers 50. The aggregator server 50 is configured as hardware using a computer, and various functions as the aggregator server 50 are realized by software processing on the computer.

The electric utility server 60 is a server managed by an electric utility. As described above, electric utility refers collectively to general electricity transmission and distribution business operators and retail electricity business operators, and therefore includes a server managed by the general electricity transmission and distribution business operator (also referred to as the general electricity transmission and distribution business operator server) and a server managed by the retail electricity business operator (also referred to as the retail electricity business operator server). In this embodiment, the general electricity transmission and distribution business operator server and the retail electricity business operator server are collectively referred to as the electric utility server 60. The electric utility server 60 generates a demand response to the aggregator server 50, and receives the results of controlling the operation of the energy resources 20A and 20B according to the demand response from the aggregator server 50. The electric utility server 60 is configured as hardware by a computer, and various functions as the electric utility server 60 are realized by software processing of the computer.

The aggregator server 50 selectively integrates multiple energy resources 20A, 20B installed within the management area 10, regardless of whether it is an urban area 10A or a surrounding area 10B, via the network 30 based on prediction results classified by a learning model from multiple consumers who own the energy resources 20A, 20B and from predetermined attributes related to the energy resources 20A, 20B, thereby forming a virtual power plant.

FIG. 2 is a block diagram showing the functions of the aggregator server 50. As shown in FIG.
The aggregator server 50 includes, as means for forming a virtual power plant, an attribute storage unit 510, a learning model storage unit 520, a learning unit 530, and an integrated control unit 540. The aggregator server 50 is configured as hardware by a computer similar to the electric utility server 60, and realizes the functions of the attribute storage unit 510, the learning model storage unit 520, the learning unit 530, and the integrated control unit 540 through software processing of the computer.

The attribute storage unit 510 stores in advance attribute information required when the learning unit 530 obtains a predetermined classification result using the learning model stored in the learning model storage unit 520. For example, the attribute information includes, for a plurality of energy resources 20A, 20B distributed and installed in the management area 10, attribute information A indicating the power usage status of a plurality of consumers who own these energy resources 20A, 20B, attribute information B indicating the power system to which the plurality of energy resources 20A, 20B are connected, and attribute information C indicating the geographical positions of the plurality of energy resources 20A, 20B. The attribute storage unit 510 is configured using hardware such as a hard disk or a semiconductor non-volatile memory, and is capable of functions such as referring to and updating the attribute information by software processing.

Figure 3A shows an example of table data of attribute information A stored in the attribute storage unit 510. Figure 3B shows an example of table data of attribute information B stored in the attribute storage unit 510. Figure 3C shows an example of table data of attribute information C stored in the attribute storage unit 510.

The attribute information A is, for example, information that captures the power consumption (power consumption curve) for each predetermined time in one day of each of multiple consumers who own energy resources 20A and 20B as the power usage status. The predetermined time may be, for example, 15 minutes or one hour. In addition, this one day may be divided into, for example, weekdays and holidays and public holidays. In this case, the power consumption for each predetermined time on weekdays may be, for example, the average value of the power consumption from Monday to Friday for each predetermined time, while the power consumption for each predetermined time on public holidays and public holidays may be, for example, the average value of the power demand for Saturday, Sunday, and public holidays for each predetermined time. When the power consumption for each predetermined time in one day is divided into weekdays and holidays and public holidays, this information may be updated, for example, on a weekly basis and stored in the attribute storage unit 510. This makes it possible to effectively use information indicating the latest power usage status of multiple consumers for the integration of virtual power plants. For example, let multiple consumers be X1 to Xn, and let information showing the power usage status of each consumer X1 to Xn on weekdays be W11 to W1n, and information showing the power usage status of each consumer X1 to Xn on weekends and holidays be W21 to W2n.

Figure 4 shows an example of a power system, and all areas shown in Figure 4 correspond to the management area 10. Figure 4 shows, for example, two substations H1 and H2, buses K1 and K2 connected to the substations H1 and H2, distribution lines F1 and F2 branching off from the bus K1, and distribution lines F3 and F4 branching off from the bus K2. For example, the distribution lines F1 and F2 are formed so as to span an urban area 10A, and multiple energy resources 20A (white circles) are installed in a dispersed and dense manner within this urban area 10A. Furthermore, the distribution lines F3 and F4 are formed so as to span a surrounding area 10B, and multiple energy resources 20B (white triangles) are installed in a dispersed manner within this surrounding area 10B.

The attribute information B is, for example, information indicating the distribution lines F1 to F4 to which the multiple energy resources 20A and 20B are connected. For example, if the multiple energy resources distributed in the urban area 10A across the distribution lines F1 and F2 are 20A1 to 20An, the multiple energy resources 20A1 to 20An are associated with information on the distribution lines F1 and F2 to which they are connected and information indicating the consumers X1 to Xn who own the corresponding energy resources 20A1 to 20An, and are stored in the attribute storage unit 510 as attribute information B. Also, if the multiple energy resources distributed in the surrounding area 10B across the distribution lines F2 to F4 are 20B1 to 20Bn, the multiple energy resources 20B1 to 20Bn are associated with information on the distribution lines F2 to F4 to which they are connected and information indicating the consumers X1 to Xn who own the corresponding energy resources 20B1 to 20Bn, and are stored in the attribute storage unit 510 as attribute information B. Here, the information indicating consumers X1 to Xn may be the same information as consumers X1 to Xn in attribute information A, or may be information corresponding to consumers X1 to Xn in attribute information A. In other words, any information may be used as long as consumers X1 to Xn in attribute information A and B match.

Attribute information C is information indicating the geographical positions where the multiple energy resources 20A, 20B are installed in the urban area 10A and the surrounding area 10B, respectively. The multiple energy resources 20A1-20An distributed in the urban area 10A in FIG. 4 are associated with information indicating the geographical positions of the multiple energy resources 20A1-20An and information indicating the consumers X1-Xn who own the corresponding multiple energy resources 20A1-20An, and are stored in the attribute storage unit 510 as attribute information C. Similarly, the multiple energy resources 20B1-20Bn distributed in the surrounding area 10B in FIG. 4 are associated with information indicating the geographical positions of the multiple energy resources 20B1-20Bn and information indicating the consumers X1-Xn who own the corresponding multiple energy resources 20B1-20Bn, and are stored in the attribute storage unit 510 as attribute information C. Here, the information indicating the consumers X1 to Xn may be the same as the consumers X1 to Xn in the attribute information A, or may be information corresponding to the consumers X1 to Xn in the attribute information A. In other words, any information may be used as long as the consumers X1 to Xn in the attribute information A and C match. In addition, the information indicating the geographical positions of the multiple energy resources 20A and 20B may be, for example, address information or longitude and latitude information indicating the installation positions of the corresponding energy resources 20A and 20B, but in this embodiment, it is address information. For example, the address information of the energy resources 20A1 to 20An is AD11 to AD1n, and the address information of the energy resources 20B1 to 20Bn is AD21 to AD2n.

The learning model storage unit 520 pre-stores a learning model that uses attribute information A to C as explanatory variables and various classification results as objective variables, and that uses machine learning to extract features to determine correlations between attribute information A to C and various classification results. The learning model may be a model that obtains various classification results from each of attribute information A to C, or a model that obtains various classification results from a combination of two or more pieces of attribute information A to C.

The learning unit 530 uses the learning model stored in the learning model storage unit 520 to predict the classification to which the input attribute information A to C should fall. For example, when considering a classification for reducing power consumption from 12:00 to 15:00 on weekdays in the management area 10, the learning unit 530 uses the learning model to predict the energy resources 20A and 20B that fall into the above classification from, for example, the attribute information A. In addition, in order to improve the accuracy of the classification prediction result, the learning unit 530 may use the learning model to predict the energy resources 20A and 20B that fall into the above classification from, for example, the attribute information A and B. Furthermore, when taking into account the geographical conditions of the energy resources 20A and 20B, the learning unit 530 may use the learning model to predict the energy resources 20A and 20B that fall into the above classification by prioritizing the attribute information A and B over the attribute information C.

The integrated control unit 540 selectively integrates, via the network 30, from among the multiple energy resources 20A distributed within the urban area 10A and the multiple energy resources 20B distributed in the surrounding area 10B, the multiple energy resources 20A and 20B that are determined to be optimal when responding to a demand response according to the prediction results of the learning unit 530, to form a virtual power plant.

The integrated control unit 540 selectively integrates the multiple energy resources 20A and 20B according to the prediction results of the learning unit 530, which prioritizes attribute information A and B over attribute information C as described above.

Here, the attribute information A includes information indicating the power consumption of each of the consumers X1 to Xn for each predetermined time in a day. Therefore, the integrated control unit 540 selects the consumers X1 to Xn that consume similar power according to the prediction result of the learning unit 530 using the attribute information A, and integrates the energy resources 20A and 20B owned by the selected consumers X1 to Xn via the network 30. As an example, when it is necessary to reduce the power demand from 12:00 to 15:00 on weekdays by a lowering DR, the integrated control unit 540 selects only the information W11 to W1n included in the classification indicating the power usage situation in which the power consumption from 12:00 to 15:00 on weekdays becomes greater than a certain amount according to the prediction result of the learning unit 530 using the attribute information A, and integrates the energy resources 20A and 20B owned by the consumers X1 to Xn that correspond to the information W11 to W1n indicating the selected power usage situation via the network 30. The integrated control unit 540 can then efficiently reduce the amount of power consumed by the energy resources 20A and 20B in the integrated virtual power plant between 12:00 and 15:00 on weekdays in accordance with the down DR. In this way, the attribute information A is useful information for predicting how to control the demand of the energy resources 20A and 20B that constitute the virtual power plant in response to demand response. In addition, the consumers X1 to Xn selected from the prediction results of the learning unit 530 using the attribute information A have similar power usage conditions, and therefore are high-priority information used by the virtual power plant to operate efficiently in one demand response.

Attribute information B is information indicating which of the power distribution lines F1 to F4 the multiple energy resources 20A and 20B are connected to. The power distribution lines F1 and F2 are power lines branching off from a common bus K1, so they are strongly correlated, and the distance between the power distribution lines F1 and F2 is relatively short compared to other power systems (e.g., power distribution lines F3 and F4). Similarly, the power distribution lines F3 and F4 are power lines branching off from a common bus K2, so they are strongly correlated, and the distance between the power distribution lines F3 and F4 is relatively short compared to other power systems (e.g., power distribution lines F1 and F2). For this reason, the multiple energy resources 20A and 20B connected to the power distribution lines F1 and F2 (or power distribution lines F3 and F4) are likely to be affected by similar natural environments (weather, disasters, etc.). The same can be said about the energy resources 20A and 20B that are commonly connected to any one of the power distribution lines F1 to F4. Therefore, the energy resources 20A and 20B connected to at least one of the power distribution lines F1 and F2 and at least one of the power distribution lines F3 and F4 are likely to have similar power usage conditions, and therefore have a high priority as information used by the virtual power plant to operate efficiently in one demand response.

The attribute information C is information indicating the geographical positions of the multiple energy resources 20A, 20B, and depending on the demand response conditions, the energy resources 20A, 20B distributed in geographically close locations (e.g., the energy resources 20A, 20B connected to the distribution lines F1, F2) or the energy resources 20A, 20B distributed in geographically distant locations (e.g., the energy resources 20A, 20B connected to the distribution lines F1, F4) can be used as information for forming a virtual power plant. In order to balance the amount of power demand consumed by consumers against the amount of power supply, the virtual power plant is operated efficiently with a small number of demand responses, and in this embodiment, the attribute information A and B are used preferentially over the attribute information C.

In this embodiment, the attribute information A to C is used as the information input to the learning unit 530 as an explanatory variable, but the attribute information is not limited to this. In addition to the attribute information A to C, for example, weather forecast information, event information, hazard information, etc. for the area in which the energy resources 20A, 20B are divided may be used as appropriate as information for integrating the multiple energy resources 20A, 20B.

Figure 5 is a flowchart showing an example of the operation of the integrated control unit 540 when forming a virtual power plant. For ease of explanation, it is assumed that the attribute information A and B stored in the attribute memory unit 510 are input to the learning unit 530 when forming a virtual power plant. In addition, the demand response is, for example, a down DR, which is a command to reduce the amount of power consumption by a certain amount from 12:00 to 15:00 on weekdays as described above.

First, the integrated control unit 540 determines whether or not a demand response has been received from the electric utility server 60 (step S10). If the integrated control unit 540 has not received a demand response (step S10: NO), it repeats the determination operation of step S10 until a demand response is received. On the other hand, if the integrated control unit 540 has received a demand response (step S10: YES), it selects a plurality of energy resources 20A, 20B that correspond to the classification result that is capable of responding to the above-mentioned lowering DR from among the classification results obtained by the learning unit 530 using the learning model stored in the learning model storage unit 520 (step S20).

Next, the integrated control unit 540 integrates the selected energy resources 20A, 20B via the network 30 to form a virtual power plant. For example, as shown in FIG. 4, among the energy resources 20A, 20B connected to the power distribution line F2, the energy resources 20A, 20B in the area surrounded by a thick line are integrated by the integrated control unit 540 to form a virtual power plant. This virtual power plant is formed by selectively integrating the energy resource 20A included in the urban area 10A and the energy resource 20B included in the surrounding area 10B (step S50).

Figure 6 is a flowchart showing an example of an automated demand response process.

First, a demand response indicating the above-mentioned reduced DR is generated from the electric utility server 60 (step S100).

Next, when the aggregator server 50 receives this lowered DR on a weekday (step S110), it integrates the energy resources 20A and 20B that can respond to the lowered DR via the network 30 based on the attribute information A and B in the procedure shown in FIG. 5 to form a virtual power plant (step S120). Here, in FIG. 5, the procedure using the classification results of the learning unit 530 based on the attribute information A and B when forming the virtual power plant is shown, but depending on the conditions of the lowered DR, the classification results of the learning unit 530 based on the attribute information A to C may be used to more reliably respond to the lowered DR. For example, the attribute information C is useful information when it is desired to further narrow down the geographical area of the virtual power plant. In this way, the aggregator server 50 forms a virtual power plant using at least one of the attribute information A, B, and C. For example, depending on the conditions of the DR, the aggregator server 50 may form a virtual power plant using only the attribute information A without using the attribute information B, or may form a virtual power plant using the attribute information A and C.

When the formation of the virtual power plant is completed, the integrated control unit 540 instructs the multiple energy resources 20A, 20B that form the virtual power plant to reduce their power consumption by a certain amount between 12:00 and 15:00 on weekdays in accordance with the reduction DR. This instruction is given to the relay system 40 connected to the energy resources 20A, 20B, which controls the reduction in power consumption by the energy resources 20A, 20B to reach a certain amount (step S130). For example, if the energy resources 20A, 20B are lighting equipment or air conditioning equipment, control is automatically performed to suppress the output of these equipment.

When the period for reducing power consumption within the virtual power plant (12:00-15:00 on weekdays) has elapsed, the integrated control unit 540 acquires and compiles actual data indicating the amount of reduction in power consumption from the relay systems 40 connected to the energy resources 20A, 20B that have reduced their power consumption within the virtual power plant, and transmits this compiled data to the electric utility server 60 (step S140).

When the electric power company server 60 acquires the aggregated data from the integrated control unit 540, it performs a process to pay the aggregator server 50 a remuneration according to the amount of reduction in power consumption indicated by the aggregated data (step S150). After confirming through the aggregator server 50 that the process of receiving the remuneration according to the amount of reduction in power consumption has been completed, the aggregator pays the above-mentioned remuneration to the consumer who owns the energy resources 20A and 20B in the virtual power plant, thereby completing the series of negawatt transactions. Of course, whether or not a remuneration is paid to the consumer in response to a demand response depends on the contents of the consumer's contract, and this embodiment is merely one example.

As described above, the virtual power plant formation control system integrates and controls multiple energy resources 20A, 20B owned by multiple consumers X1 to Xn and distributed in various locations via a network 30, and functions as if they were a single power plant. The system includes an attribute storage unit 510 that stores attribute information including information A indicating the power usage status of the multiple consumers X1 to Xn, a learning model storage unit 520 that stores a learning model that learns the correlation between the attribute information, which is an explanatory variable, and the classification result of the attribute information, which is an objective variable, a learning unit 530 that predicts the classification result for the attribute information using the learning model, and an integration control unit 540 that selectively integrates the multiple energy resources 20A, 20B from the multiple energy resources 20A, 20B based on the prediction result by the learning unit 530, and controls the integrated multiple energy resources 20A, 20B in response to demand response.

The attribute information further includes information indicating the power system to which the multiple energy resources 20A and 20B are connected.

The attribute information further includes information indicating the geographical locations of the multiple energy resources 20A and 20B.

The integrated control unit 540 also prioritizes prediction results from the learning unit 530 regarding information indicating the power usage status of the multiple consumers X1 to Xn and information indicating the power grid to which the multiple energy resources 20A and 20B are connected, over prediction results regarding information indicating the geographical locations of the multiple energy resources 20A and 20B, and selectively integrates the multiple energy resources 20A and 20B.

The multiple energy resources 20A, 20B include energy generation resources, energy storage resources, and energy saving resources, and the integrated control unit 540 selectively integrates any one of the energy generation resources, energy storage resources, and energy saving resources from among the multiple energy resources 20A, 20B based on the prediction results of the learning unit 530.

In this embodiment, the learning unit 530 can use a machine learning function, such as artificial intelligence, to accurately classify attribute information A to C into multiple groups that can respond to demand response. This makes it possible to efficiently utilize various ubiquitous energy resources 20A, 20B to form a virtual power plant regardless of the urban area 10A or the surrounding area 10B, thereby making it possible to more reliably balance the supply and demand of electricity.

The above embodiment is intended to facilitate understanding of the present invention, and is not intended to limit the present invention. The present invention may be modified or improved without departing from the spirit thereof, and the present invention also includes equivalents. In this embodiment, the virtual power plant operates to reduce the amount of power consumption according to the lowering DR, but is not limited to this. The virtual power plant can also operate to increase the amount of power consumption according to the increasing DR. In other words, the integrated control unit 540 in the aggregator server 50 can form and control the virtual power plant so that it can efficiently respond to both the increasing DR and the decreasing DR.

REFERENCE SIGNS LIST 10 Management area 10A Urban area 10B Surrounding area 20A, 20B Energy resource 30 Network 40 Relay system 50 Aggregator server 60 Electricity supplier server 510 Attribute storage unit 520 Learning model storage unit 530 Learning unit 540 Integrated control unit

Claims (10)

A virtual power plant formation control system that integrates and controls a plurality of energy resources owned by a plurality of consumers and distributed throughout a management area including an urban area and surrounding areas via a network, and functions as if they were a single power plant,
an attribute storage unit that stores attribute information including information indicating power usage status of the plurality of consumers;
a learning model storage unit that stores a learning model that learns a correlation between the attribute information, which is an explanatory variable, and a classification result of the attribute information, which is a target variable;
a learning unit that predicts the classification result for the attribute information by using the learning model;
an integrated control unit that selectively integrates, from among the plurality of energy resources, a plurality of energy resources that correspond to a classification result that is capable of responding to the lowering DR or the raising DR based on a prediction result by the learning unit when a lowering DR or an raising DR occurs , and controls the integrated energy resources;
A formation control system for a virtual power plant equipped with The virtual power plant formation control system according to claim 1 , wherein the attribute information further includes information indicating a power grid to which the plurality of energy resources are connected. The virtual power plant formation control system according to claim 2 , wherein the attribute information further includes information indicating geographical locations of the plurality of energy resources. 4. The virtual power plant formation control system of claim 3, wherein the integrated control unit selectively integrates the energy resources by prioritizing prediction results by the learning unit regarding information indicating the power usage status of the multiple consumers and information indicating the power grid to which the multiple energy resources are connected over prediction results regarding information indicating the geographical locations of the multiple energy resources. The plurality of energy resources include an energy generating resource, an energy storing resource, and an energy saving resource,
The virtual power plant formation control system according to any one of claims 1 to 4, wherein the integrated control unit selectively integrates the energy generating resource, the energy storing resource, and the energy saving resource from among the plurality of energy resources based on the prediction result of the learning unit. A method for forming and controlling a virtual power plant in which a plurality of energy resources owned by a plurality of consumers and distributed throughout a management area including an urban area and surrounding areas are integrated and controlled via a network to function as if they were a single power plant, comprising:
predicting the classification result for the attribute information using a learning model that learns a correlation between attribute information including information indicating the power usage status of the plurality of consumers, which is an explanatory variable, and a classification result of the attribute information, which is a target variable;
A method for controlling the formation of a virtual power plant, comprising : when a down DR or an up DR occurs, selectively integrating from among the plurality of energy resources, a plurality of energy resources that correspond to a classification result that is capable of responding to the down DR or the up DR based on a prediction result by the learning model , and controlling the integrated energy resources. The method for controlling formation of a virtual power plant according to claim 6 , wherein the attribute information further includes information indicating a power grid to which the plurality of energy resources are connected. The method for controlling formation of a virtual power plant according to claim 7 , wherein the attribute information further includes information indicating geographical positions of the plurality of energy resources. 9. The method for controlling the formation of a virtual power plant according to claim 8, further comprising the step of selectively integrating the energy resources by prioritizing prediction results relating to information indicating the power usage status of the plurality of consumers and information indicating the power grid to which the plurality of energy resources are connected over prediction results relating to information indicating the geographical locations of the plurality of energy resources among the prediction results of the learning model. The plurality of energy resources include an energy generating resource, an energy storing resource, and an energy saving resource,
The method for controlling the formation of a virtual power plant according to any one of claims 6 to 9, further comprising selectively integrating the energy generating resource, the energy storing resource, and the energy saving resource from among the plurality of energy resources based on a prediction result of the learning model.

Images