Composition system, method, and program for power generation balancing groups

Abstract

To compose a balancing group in accordance with a prediction error in a prediction and an inclination of a removal direction in each region by a power generation or weather in a power which can be supplied in a power generation side to reduce the generation of an unnecessary imbalance.SOLUTION: A component system of a power generation balancing group, is constructed by: an actual data acquisition part 76a that measures a power amount generated by each power generation facility, and acquires an actual data related to position information on the power generation facility and power generation information related to a power generation system; an external information acquisition part 76b that acquires external information such as weather information or the like from an information service organization; a prediction part 74 that predicts of a power generation amount of the power generation facility on the basis of a correlation of the actual data and the external information; a prediction error calculation part 74c that compares an actual power generation amount of the power generation facility with the prediction of the power generation amount, and calculates a prediction error of the prediction of the power generation amount in each power generation facility; a deviation tendency analysis part 751 that analyzes a relation of the power generation information of each power generation facility and the prediction error, and the outer information as deviation tendency information; and a balancing group setting part 75 that divides the group of the power generation facilities into a plurality of balancing groups on the basis of the deviation tendency information of each power generation facility.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present invention relates to a composition system, method, and program for a power generation balancing group that can contribute to profit improvement by making the combination of power plants in the power generation balancing group the most effective combination.

Background Art

[0002] Conventionally, in the operation of a power system, if the balance between the supply amount and the demand amount of electric power deviates from the allowable range, the stable supply of electric power cannot be achieved, and due to the characteristic that electric energy cannot be stored, it is necessary to maintain the balance between electric power demand and supply within the allowable range, that is, the so-called "simultaneous same amount".

[0003] In addition, in a power system, due to the gap between the power supply plan and the prediction of the power demand amount (such as additional supply due to power source dropout or increase in demand), the balance between power demand and power supply (power supply-demand balance) may be disrupted. If the supply-demand balance collapses beyond a certain range, the frequency and voltage of the power system will fluctuate, the electrical equipment of power consumers may not operate normally, and in extreme cases, the power system may experience a major power outage, which will have a great adverse impact on maintaining the power quality of the power system.

[0004] To solve this problem, conventionally, as a mechanism for formulating a plan to procure power for power supply balance adjustment, for example, the system disclosed in Patent Document 1 has been proposed. In the system disclosed in this Patent Document 1, the demand and supply are made to match and the simultaneous same amount is realized by the supply-demand adjustment plan procured in the real-time market according to the difference between the actual demand and the supply adjustment plan.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

[0006] Incidentally, the amount of electricity that can be supplied by power generators varies depending on the power generation method, climate, and other factors in each region, resulting in differences in prediction errors and deviations. However, the system disclosed in Patent Document 1 creates a power supply plan based on predictions of electricity demand, but it does not take into account the prediction errors and deviations in each region, such as the power generation method and climate, which raised concerns about unnecessary imbalances.

[0007] Therefore, in view of the problems of the prior art, the object of the present invention is to provide a power generation balancing group composition system, method, and program that can reduce the occurrence of unnecessary imbalances by forming balancing groups that take into account the tendency of prediction errors and deviations for each region, such as the power generation method and climate, regarding the power that can be supplied on the power generation side. [Means for solving the problem]

[0008] To solve the above problems, the power generation balancing group composition system of the present invention is A performance data acquisition unit that, for a group of power plants consisting of multiple power plants, measures the amount of electricity generated by each power plant and acquires performance data that associates the measurement results with the location information and power generation information related to the power generation method of the measured power plants. An external information acquisition unit that acquires external information from an information service organization, including at least one of the following: weather information, calendar information, or time information. A prediction unit that predicts the amount of power generated by the power plant based on the correlation between the aforementioned performance data and the aforementioned external information, A prediction error calculation unit compares the actual power generation of the power plant with the power generation prediction made by the prediction unit and calculates the error in the power generation prediction for each power plant. A deviation trend analysis unit analyzes the correlation between the power generation information and prediction errors of each power plant and the external information as deviation trend information, Based on the deviation trend information of each of the aforementioned power plants, a balancing group setting unit divides the group of power plants into multiple balancing groups. It consists of.

[0009] Furthermore, the power generation balancing group composition method of the present invention is For a group of power plants consisting of multiple power plants, the amount of electricity generated by each power plant is measured, and the actual data acquisition unit acquires actual data by associating the measurement results with the location information and power generation information related to the power generation method of the measured power plants. In addition, the external information acquisition unit acquires external information from an information service organization, which includes at least one of the following: weather information, calendar information, or time information. The prediction unit performs a prediction step of predicting the amount of power generated by the power plant based on the correlation between the actual data and the external information, The prediction error calculation unit compares the actual power generation amount of the power plant with the power generation amount prediction made by the prediction unit and calculates the error of the power generation amount prediction for each power plant, and the deviation trend analysis unit analyzes the correlation between the power generation information and prediction error of each power plant and the external information as deviation trend information in a deviation trend analysis step, The balancing group setting unit performs a balancing group setting step in which it divides the group of power plants into multiple balancing groups based on the deviation trend information of each power plant. Includes.

[0010] In the above invention, it is preferable that the balancing group setting unit performs the classification such that the sum of the prediction errors of all balancing group candidates is less than a predetermined value when the sum of the absolute values ​​of the prediction errors of those power plants is divided by the sum of the absolute values ​​of the prediction errors of those power plants.

[0011] The above invention further comprises a learning unit that extracts the distribution of features relating to the actual power generation amount corresponding to a power generation amount prediction with a prediction error of more than a predetermined standard value, and external information related to that power generation amount prediction, and forms a neural network as the deviation trend information. The prediction unit has the neural network formed by the learning unit identify the actual data and external information related to the power generation forecast, and performs the power generation forecast according to the identification result. It is preferable.

[0012] The above invention includes a candidate area acquisition unit that acquires candidate areas for the construction of a power plant, A power generation method selection unit that selects a power generation method based on user input, Based on the power generation method selected by the power generation method selection unit and the location information of the proposed construction area, the deviation trend prediction unit refers to the deviation trend information and calculates the prediction error in the proposed construction area as deviation trend prediction information. It is preferable to have even more.

[0013] In the above invention, the deviation trend analysis unit has a learning unit that extracts the distribution of features related to the actual power generation amount corresponding to the power generation amount prediction with a prediction error of a predetermined standard value or more, and external information related to the power generation amount prediction, and forms a neural network as the deviation trend information. The deviation trend prediction unit, based on the power generation method selected by the power generation method selection unit and the location information of the proposed construction area, refers to the deviation trend information by having the neural network identify location information having similar features to the proposed construction area, and calculates the prediction error in the proposed construction area as deviation trend prediction information. It is preferable.

[0014] Furthermore, the systems and methods according to the present invention described above can be realized by executing a program of the present invention written in a predetermined language on a computer. That is, by installing the program of the present invention on the IC chip or memory device of a mobile terminal device, smartphone, wearable device, mobile PC or other information processing terminal, or a general-purpose computer such as a personal computer or server computer, and executing it on the CPU, a system having the above-described functions can be constructed, and the methods according to the present invention can be implemented.

[0015] In addition, the program of the present invention can be distributed, for example, through a communication line, and can also be transferred as a package application that operates on a stand-alone computer by being recorded on a computer-readable recording medium. Specifically, as this recording medium, various recording media such as magnetic recording media such as flexible disks and cassette tapes, or optical disks such as CD-ROMs and DVD-ROMs, as well as RAM cards can be used for recording. And according to the computer-readable recording medium on which this program is recorded, it becomes possible to easily implement the above-described system and method using a general-purpose computer or a dedicated computer, and it is also possible to easily perform the storage, transportation, and installation of the program.

Advantages of the Invention

[0016] According to the present invention, regarding the power that can be supplied on the power generation side, a balancing group is formed in consideration of the prediction error and the tendency of deviation for each region such as the power generation method and climate, and the occurrence of unnecessary imbalance can be reduced. As a result, by making the combination of power plants in the power generation balancing group the most effective combination, it is possible to contribute to the improvement of profits.

Brief Explanation of Drawings

[0017] [Figure 1] It is a conceptual diagram showing the overall configuration of the power system according to the embodiment. [Figure 2] It is a conceptual diagram regarding power control in the power system according to the embodiment. [Figure 3] It is a block diagram showing the device configuration of the user system according to the embodiment. [Figure 4] It is a block diagram showing the internal configuration of the control device according to the embodiment. [Figure 5] It is a block diagram showing the internal configuration of the power control server according to the embodiment. [Figure 6] It is a flowchart showing the operation of the power generation balancing group composition system according to the embodiment. [Figure 7] This is a flowchart illustrating in detail the operation of the power generation balancing group composition step according to the embodiment. [Figure 8] This is an explanatory diagram illustrating the correlation analysis of prediction errors for a power plant according to the embodiment. [Modes for carrying out the invention]

[0018] The embodiments of the power generation balancing group composition control system according to the present invention will be described in detail below with reference to the attached drawings. Note that the embodiments shown below are illustrative examples of devices, etc., for realizing the technical concept of this invention, and the technical concept of this invention does not limit the materials, shapes, structures, arrangements, etc., of each component to those described below. Various modifications can be made to the technical concept of this invention within the scope of the claims.

[0019] (Overview of the power control system) Figures 1 and 2 show the overall configuration of a power control system to which the power generation balancing group composition control system according to the present invention is applied. As shown in Figures 1 and 2, the power control system according to this embodiment is a system that supplies power from each power plant P1 to P3 to each demand unit through the power grid connected to the power receiving and transforming equipment 50 at the high-voltage receiving point 5. In this embodiment, the demand units include H1 and H2, which are general houses equipped with residential solar power generation and energy storage systems.

[0020] Furthermore, in this embodiment, the power receiving and transforming equipment 50 operated by the power generation company is located at the high-voltage receiving point 5, where the power company operating the power plants supplies electricity from each power plant P1 to P3 to each demand unit H1 and H2. Each power plant or power company is equipped with a power generation control terminal 8, and power generation by each power generation company is managed by this power generation control terminal 8.

[0021] In this embodiment, the power control server 2 provides energy management services that integrate the management of each demand unit H1 and H2. Through these energy management services provided by the power control server 2, control planning, demand forecasting, power generation forecasting, and management / settings are performed for each demand unit H1 and H2. The power control server 2 predicts future (e.g., the next day's) power consumption using a performance database while referring to external factors such as weather data, and generates a control schedule. In accordance with the generated schedule, the power control server 2 controls each storage battery to pre-discharge and secure enough power to absorb the next day's power generation.

[0022] The power control server 2 is connected to the HEMS (Home Energy Management Systems) 40 of each demand unit H1 and H2 via the communication network 3. Each demand unit H1 and H2 is then connected to the power grid from the high-voltage receiving point 5, and power is supplied to each demand unit H1 and H2 from each power plant P1 to P3. Control data is transmitted from the power control server 2 to the HEMS 40 of each demand unit H1 and H2, and performance data (solar power, battery storage, power data) from the user systems 4 on the demand unit H1 and H2 side is collected by the power control server 2.

[0023] More specifically, the power control system 1 is a system that manages and controls the power generation, discharge, or transmission and reception of power in multiple user systems 4,4… that control and manage power for each power demand unit H1 and H2. As shown in Figure 3, it comprises smart meters 41, which are performance data generation units installed in each user system 4,4…, etc., and the power control system 2 connected to the smart meters 41 via the internet, telephone lines, dedicated lines, etc.

[0024] In the power control system 1, each smart meter 41 measures the amount of electricity generated or consumed during each power usage period in each user system 4 for each customer and generates actual data D1. The power control server 2 manages power consumption within the user systems 4, 4... based on the actual data D1. In this embodiment, the power management actual and forecast results for each user system 4, 4... are made available to the power control server 2. In the power control system 1, the power control server 2 and the HEMS 40, which are power control devices for demand units H1 and H2, are interconnected via a communication network 3. In each customer H1 and H2, the smart meter 41 of each user system 4 is connected to the external power grid.

[0025] HEMS40, also known as a "Home Energy Management System," is a power control terminal that manages the energy used in homes (consumers). It is interconnected with home appliances and electrical equipment within the consumer's home via Wi-Fi or other communication methods, allowing for the "visualization" of electricity and gas usage on a monitor screen and the "automatic control" of home appliances. Specifically, HEMS40 consists of an information processing terminal equipped with a CPU and communication functions, and can comprehensively control the power equipment of various facilities, including individual consumers, power plants, PPS (Power Producers and Suppliers), power prosumers, and aggregators. It is also connected to smart meters 41 and distribution boards 45 within the user system for communication purposes. The equipment controlled by HEMS40 includes smart meters 41, storage batteries 42, and PV (Photovoltaics: solar power generation) 43, which are included in the user system 4 deployed within facilities such as consumers and power prosumers, and are devices that manage power generation, storage, and power consumption.

[0026] Furthermore, the various devices controlled by this HEMS 40 can be omitted as needed. For example, in user systems 4,4…, power consumption is measured by a smart meter 41. Some consumers have both power generation and energy storage facilities, some have either power generation or energy storage facilities, and some have neither power generation nor energy storage facilities, only a smart meter 41 is installed, and they only consume power. Also, while power prosumers consume power, they can also be positioned as power suppliers equipped with solar power generation or battery storage.

[0027] Communication network 3 is an IP network using the TCP / IP communication protocol, such as the Internet, and is a distributed communication network constructed by interconnecting various communication lines (public lines such as telephone lines, ISDN lines, ADSL lines, and optical lines; dedicated lines; third-generation (3G) communication methods such as WCDMA® and CDMA2000; fourth-generation (4G) communication methods such as LTE; and fifth-generation (5G) and later communication methods, as well as wireless communication networks such as Wi-Fi® and Bluetooth®). This IP network also includes LANs such as intranets (corporate networks) and home networks using 10BASE-T and 100BASE-TX.

[0028] (Configuration of each device) Next, we will explain the configuration of each device. In this explanation, the term "module" refers to a functional unit composed of hardware such as devices or equipment, software possessing that functionality, or a combination thereof, for achieving a predetermined operation.

[0029] (1) User System 4 The smart meter 41 is a performance data generation unit that comprehensively manages power consumption within demand unit H1 or H2, as well as power generation and storage as needed. In addition to measuring power consumption within demand unit H1 or H2, it also controls and manages other equipment within demand unit H1 or H2, such as storage batteries and solar power generation, and measures the amount of power generated, stored, or consumed by the customer during each power usage period to generate performance data D1, which is then sent to the power control server 2. This performance data D1 is transmitted to the power control server 2 via the communication network 3, telephone lines, dedicated lines, etc.

[0030] In this embodiment, a smart meter 41 is used as the performance data generation unit, but the present invention is not limited thereto. For example, any electronic device equipped with a control device such as a power control device (IoT device) that transmits its own status as performance data to a communication network can be used, such as HEMS, various home appliances, factory equipment, office equipment, etc., located within a consumer's premises.

[0031] Furthermore, as shown in Figure 3, the user system 4 according to this embodiment encompasses all power equipment owned by consumers and power prosumers whose power is already managed by the HEMS 40, and also serves as a unit for consuming power. Here, a consumer is a contract unit related to power equipment that receives and uses power, and includes high-voltage large-scale consumers with a contracted power of 500kW or more, high-voltage small-scale consumers with a contracted power of 50kW or more but less than 500kW, and low-voltage consumers of less than 50kW, such as general households. The user system 4 may also include power generation and energy storage equipment. Examples of power generation equipment include solar power generation and wind power generation. This user system 4 includes the HEMS 40 and a smart meter 41 as a performance data generation unit. Furthermore, power-consuming equipment includes not only various home appliances, factory equipment, and office equipment, but also all control devices such as power control devices (IoT devices).

[0032] The HEMS 40, installed at each customer's home, is connected to the distribution board 45 and is capable of acquiring current, voltage, power waveform, frequency, etc., of the power circulating within the user system 4. It is also a device that actually controls the power generation, charging, and discharging of each electrical appliance, PV 43, and storage battery 42 located within the user system 4 (customer). Specifically, the HEMS 40 is an information processing terminal equipped with communication functions and a CPU. Various functions can be implemented by installing an OS or firmware and various application software. In this embodiment, it functions as a power management unit by installing and running an application. This information processing terminal can be a personal computer, a smartphone, or a dedicated device with specialized functions, including tablet PCs, mobile computers, and mobile phones.

[0033] The smart meter 41 is a performance data generation unit that comprehensively manages power generation, storage, and power consumption within the user system, which is the demand unit. Within the user system 4 of a consumer, it measures the power consumption of each consumer, and also controls and manages other equipment within the user system, such as storage batteries and solar power generation. It measures the amount of power generated, stored, or consumed by the consumer during each power usage period, generates performance data D1, and periodically sends it to the power control server 2 via the HEMS 40 and gateway terminal 46. This performance data D1 is transmitted to the power control server 2 via the communication network 3, telephone lines, dedicated lines, etc.

[0034] (2) Power generation control terminal 8 The power generation control terminals 8 installed in each power plant P1 to P3 are devices that manage the amount of power generated within the power plant, which is the unit of power supply. Specifically, as shown in Figure 4, the power generation control terminal 8 includes a CPU 802, memory 803, input interface 804, storage 801, output interface 805, and communication interface 806. In this embodiment, these devices are connected via a CPU bus 800, enabling the exchange of data between them.

[0035] The memory 803 and storage 801 are storage devices that store data on a recording medium and read this stored data according to the requests of each device, and can be configured, for example, with a hard disk drive (HDD), a solid state drive (SSD), a memory card, etc. In particular, in this embodiment, the storage 801 functions as a data recording unit that records estimated history information D2, which records the power supply status of individual devices in chronological order based on the estimation results of the component identification unit 802d, and also functions as an actual power consumption storage unit that stores actual power consumption information D5 regarding the power actually consumed within the customer's premises.

[0036] The input interface 804 is a module that receives control signals from the power generation equipment of power plants P1 to P3. The received control signals are transmitted to the CPU 802 and processed by the OS and various applications. On the other hand, the output interface 805 is a module that outputs control signals to each piece of equipment within power plants P1 to P3.

[0037] The communication interface 806 is a module that transmits and receives data with other communication devices. Communication methods include, for example, public lines such as telephone lines, ISDN lines, ADSL lines, and fiber optic lines, dedicated lines, third-generation (3G) communication methods such as WCDMA® and CDMA2000, eighth-generation (8G) communication methods such as LTE, and fifth-generation (5G) and later communication methods, as well as wireless communication networks such as Wi-Fi® and Bluetooth®.

[0038] The CPU 802 is a device that performs various calculations necessary for controlling each part, and by executing various programs, it virtually constructs various modules on the CPU 11. The OS (Operating System) is started and executed on the CPU 802, and the basic functions of each power generation control terminal 8 are managed and controlled by this OS. In addition, various applications can be executed on this OS, and various functional modules are virtually constructed on the CPU when OS programs are executed on the CPU 802.

[0039] In this embodiment, by running browser software on the CPU 802, it is possible to view and input information on the system through this browser software. More specifically, this browser software is a module for viewing web pages, and it downloads HTML (HyperText Markup Language) files, image files, music files, etc. from the power control server 2 via the communication network 3, analyzes the layout, and displays and plays them. This browser software also allows users to send data to the web server using forms, and to run application software written in JavaScript, Flash, and Java (registered trademark), and through this browser software, each user can use the power management service provided by the power control server 2.

[0040] In this embodiment, a smart meter 41 is used as the performance data generation unit, but the present invention is not limited thereto. For example, it includes HEMS 40, various home appliances, factory equipment, office equipment, etc., located within a consumer's premises, and all electronic devices equipped with a control device such as a power control device (IoT device) that transmits its own status as performance data to a communication network.

[0041] (3) Power control server 7 The power control server 7 is a server device that manages and controls power generation at each power plant and power consumption at each consumer by coordinating with the power generation control terminals 8 of each power plant P1 to P3 and the HEMS 40 of each demand unit H1 and H2. As shown in Figure 5, it comprises a communication interface 73, an authentication unit 72, a balancing group setting unit 75, various databases 71a to d, a prediction unit 74, and a data management unit 76.

[0042] The communication interface 73 is a module that transmits and receives data with other communication devices via the communication network 3. In this embodiment, it is connected to the power generation control terminal 8, each HEMS 40 and smart meter 41, and external information sources on the Internet in order to provide this service.

[0043] The authentication unit 72 is a computer or software with that function that verifies the legitimacy of the accesser related to the power control service, and performs authentication processing based on the user ID that identifies the user. In this embodiment, the user ID and password are obtained from the accesser's terminal device via the communication network 3, and by comparing them with the user database 71b, it is confirmed whether the accesser has the authority to use the service and whether the accesser is the subscriber.

[0044] The balancing group setting unit 75 is a module that divides a group of power plants into multiple balancing groups based on the deviation trend information of each power plant. In this embodiment, the balancing group setting unit 75 divides the group into balancing groups by combining power plants whose prediction error is below a predetermined low threshold and power plants whose prediction error is above a predetermined high threshold, so that the prediction error of the entire balancing group exceeds a predetermined level value. For example, the balancing group setting unit 75 divides a group of power plants into multiple balancing groups such that the sum of the prediction errors of all balancing group candidates is less than a predetermined value when divided by the simple sum of the absolute values ​​of the prediction errors of those power plants. In this embodiment, the balancing group setting unit 75 also includes a deviation trend analysis unit 751 and a power generation method selection unit 752.

[0045] The deviation trend analysis unit 751 is a module that analyzes the correlation between power generation information and prediction errors for each power plant and external information as deviation trend information. The deviation trend analysis unit 751 also has a learning function that extracts the distribution of features related to the actual power generation amount corresponding to power generation amount predictions with prediction errors exceeding a predetermined threshold value, and the external information related to that power generation amount prediction, and forms a neural network as deviation trend information.

[0046] The power generation method selection unit 752 is a module that selects a power generation method based on user operation and also functions as a candidate area acquisition unit that acquires candidate construction areas for power plants selected by the user. The deviation trend prediction unit 74e refers to the deviation trend information based on the power generation method selected by the power generation method selection unit 752 and the location information of the candidate construction areas, and calculates the prediction error in the candidate construction areas as deviation trend prediction information.

[0047] The various databases 71a to d described above are organized collections of structured information or data electronically stored in an information processing terminal having a storage device such as a computer system. They may be a single database, or they may be divided into multiple databases and linked together to form a relational database. Specifically, the power control server 7 according to this embodiment includes a prediction result information database 71a, a user database 71b, a performance management database 71c, and a learning information database 71d.

[0048] The prediction result information database 71a is a storage device that analyzes collected historical data and external information to extract and classify necessary information, and stores prediction results related to that information. Each prediction result information is linked and stored with the underlying historical data, information from external information sources distributed on the network, and additional information such as type, time information, and text. The user database 71b is a storage device that stores information about each customer's user, simulation administrator, aggregator, and other vendors.

[0049] The performance management database 71c is a storage device that collects, stores, and manages performance data from power plants, consumers, aggregators, and other parties involved in the exchange of electricity. Each performance data received from each smart meter is stored in this performance management database and used for machine learning in the learning unit 74d. The learning information database 71d is a storage device that records the performance of machine learning at each consumer.

[0050] The aforementioned data management unit 76 is a module that collects, stores, and manages actual data and external information. By collecting and analyzing actual data D1 from each customer, it provides the information necessary for the forecasting process in the forecasting unit 74.

[0051] The analysis results from the data management unit 76, along with the correlation information extracted by the correlation extraction unit 74b, are input to the learning unit 74d of the prediction unit 74 and used for machine learning. In this embodiment, the data management unit 76 includes a historical data acquisition unit 76a and an external information acquisition unit 76b.

[0052] The prediction unit 74 is a module that analyzes the actual data collected by the actual data acquisition unit 76a to predict the predicted power generation amount and predicted power consumption for each unit measurement period in each time period. In this embodiment, it predicts the power generation amount of the power plant and the power consumption of each consumer based on the correlation between the actual data and external information. Specifically, in this embodiment, the prediction unit 74 comprises an external information acquisition unit 74a, a correlation extraction unit 74b, a prediction error calculation unit 74c, a learning unit 74d, and a deviation trend prediction unit 74e.

[0053] The external information acquisition unit 74a is a module that acquires external information provided by the Web or information service organizations, such as weather information, calendar information, and time information. The correlation extraction unit 74b is a module that is, for example, a nonlinear regression analyzer, in which the characteristics of multiple types of predicted values ​​are set as patterns, analyzes actual data, and detects specific feature points from a large amount of actual data. The prediction error calculation unit 74c is a module that compares the actual power generation amount of a power plant with the power generation amount prediction by the prediction unit 74 and calculates the error of the power generation amount prediction for each power plant.

[0054] The deviation trend prediction unit 74e is a module that, based on the power generation method selected by the power generation method selection unit 752 and the location information of the proposed construction area, refers to deviation trend information and calculates the prediction error in the proposed construction area as deviation trend prediction information.

[0055] In this embodiment, the prediction process for power generation and power consumption, as well as their standard deviation values, in the prediction unit 74 queries the learning unit 74d to obtain candidates calculated by machine learning. This learning unit 74d is a module that performs machine learning so that the deep learning recognition function, which is an artificial intelligence, can make appropriate judgments. In this embodiment, it generates training data from actual data D1 collected by the data management unit 76 and external information, and trains the deep learning recognition unit based on this training data.

[0056] Furthermore, the learning unit 74d plays the role of a comparison unit that compares the estimated history, which is the judgment result of deep learning recognition, with the actual data, based on the input actual data D1 and external information.

[0057] Furthermore, the prediction error calculation unit 74c compares the judgment result for the same event (target power plant, target consumer, target area, time of occurrence, etc.) input as training data to the deep learning recognition unit with the actual control data D1 at that time, determines whether the results of the comparison match, and calculates the degree of match as the prediction error. The prediction error calculated by this prediction error calculation unit 74c is sent to the deviation trend analysis unit 751 and the deviation trend prediction unit 74e. The deviation trend analysis unit 751 analyzes the correlation between the power generation information and prediction error of each power plant and external information as deviation trend information. It also analyzes the prediction results and checks which option was incorrect if it differs from reality, calculates the prediction error rate of the prediction result by the prediction unit 74, inductively verifies the validity of the various options selected by the prediction unit 74 when executed, and provides feedback to the deviation trend analysis unit 751.

[0058] The deviation trend analysis unit 751 is a module that performs judgments using so-called deep learning, and the learning data (training data) obtained by analyzing the deviation trend is used for functional verification. Specifically, the deep learning recognition unit analyzes the correlation between each prediction information and the actual data D1, as well as external information and correlation information, according to a predetermined deep learning algorithm, and stores the deep learning recognition result (price prediction AI model), which is the result of this analysis, in the prediction result information database 71a.

[0059] The algorithm implemented in the learning unit 74d in this embodiment is a learning and recognition system that mimics the mechanism of the human brain, featuring a multi-layered neural network, particularly one with three or more layers. When data such as an image is input to this recognition system, the data is propagated sequentially from the first layer, and learning is repeated sequentially in each subsequent layer. During this process, the features within the image are automatically calculated.

[0060] These features are essential variables necessary for solving the problem and are variables that characterize a particular concept. In the learning unit 74d, actual data, estimation history, external information, and correlation information are input, and multiple feature points are extracted hierarchically from this data. Patterns are recognized by the hierarchical combination patterns of the extracted feature points. The recognition function module of the learning unit 74d is a multi-class classifier, and multiple events are set, and it detects feature vectors (here, for example, "weather at a specific time period") which are objects containing specific feature points from among multiple objects. This recognition function module has an input unit (input layer), a first weight coefficient, a hidden unit (hidden layer), a second weight coefficient, and an output unit (output layer).

[0061] At this time, multiple feature vectors are input to the input unit. The first weight coefficient weights the output from the input unit. The hidden unit performs a nonlinear transformation on the linear combination of the output from the input unit and the first weight coefficient. The second weight coefficient weights the output from the hidden unit. The output unit calculates the identification probability for each class (e.g., equipment used, usage state, etc.). Three output units are shown here, but this is not limited to them. The number of output units is the same as the number of events that the pattern classifier can detect. Increasing the number of output units increases the number of events that the event classifier can detect, such as the recognizable equipment type.

[0062] The aforementioned deviation trend prediction unit 74e has the function of referring to deviation trend information by having a neural network identify location information having similar features to the construction candidate area based on the power generation method selected by the power generation method selection unit 752 and the location information of the construction candidate area, and calculating the prediction error in the construction candidate area as deviation trend prediction information.

[0063] (Operation of the power generation balancing group formation system) The power generation balancing group composition method of the present invention can be implemented by operating the power generation balancing group composition system described above. Figures 6 and 7 are flowcharts showing the operation of the power control system. Note that the processing procedure described below is merely an example, and each process may be modified as much as possible. Furthermore, depending on the embodiment, steps in the processing procedure described below can be omitted, replaced, and added as appropriate.

[0064] As shown in the figure, the power generation control terminal 8 on the power plant side and the HEMS 40 on the consumer side constantly measure the power being generated, stored, or consumed within the system. This measured performance data is sent to the power control server 2. The power control server 2 also collects external information (S101). In this power measurement, the temporal fluctuations of the power waveform are also measured and recorded as they occur. Meanwhile, the simulation management server constantly collects and classifies external information in conjunction with the power consumption measurement on the consumer's power generation control terminal 8 side.

[0065] Next, the amount of power generated at each power plant is predicted, and the trend of individual power consumption for each customer over the prediction period from the start of the prediction is also predicted. At this time, the power status of individual equipment at each customer may be recorded in a time series. That is, the temporal fluctuation of the total power consumption measured at the customer level is analyzed to estimate the individual equipment operating within the customer and its individual power consumption. Using this estimated historical information, the number of individual equipment units operating at the start of the prediction is estimated from the total power consumption measured in real time. At this time, the power waveform and its temporal changes are analyzed, and the characteristics of the frequency components and power fluctuation patterns are extracted to estimate the individual equipment that is in operation, its individual power consumption, and its duration.

[0066] These prediction results are collected by the power control server 2 as prediction history information along with actual data and stored in storage 801. The actual data and prediction history information collected here are compared with external information to extract their correlations and generate correlation information (S103).

[0067] Then, based on the collected performance data, prediction history information, correlation information, and external information, the prediction error is calculated and the deviation trend, i.e., the tendency to be off, is analyzed (S104). The performance data, prediction history information, correlation information, external information, and deviation trend are accumulated, and machine learning is performed on these as well (S105), and their history is updated. The results of this machine learning are reflected in the training data.

[0068] Furthermore, based on the deviation trend (outlier trend) analyzed in step S104, the balancing group for the power plant is set, and a balancing group on the consumer side that will be combined with the balancing group set for this power plant is also set (S106). Figure 7 shows the balancing group composition process in step S106.

[0069] As shown in the figure, first, it is determined whether or not there is power generation prediction error data (S201). This power generation prediction difference data is data calculated for each power plant by comparing the actual power generation amount of each power plant with the power generation amount prediction and determining the prediction error of the power generation amount prediction for each power plant. In this embodiment, it is generated by the prediction error calculation unit 74c and stored in the prediction result information database 71a. In step S201, if power generation prediction error data exists in the prediction result information database 71a ("Y" in step S201), that power generation prediction error data is acquired by the balancing group setting unit 75.

[0070] On the other hand, in step S201, if there is no power generation forecast error data in the forecast result information database 71a ("N" in step S201), external information such as weather forecasts and weather records is acquired (S203), and this is assumed to be a pseudo power generation forecast. This is then compared with the actual power generation to calculate a pseudo forecast error, which is converted into power generation forecast error data (S204). Then, matching is performed based on the deviation trend to select a combination of power plants, and the forecast error reduction effect of the power generation balancing group is analyzed for each combination (S205). The combination with the largest forecast error reduction effect is output as the optimal balancing group (S206).

[0071] In this analysis of the prediction error reduction effect, the correlation coefficient related to the power generation prediction error is calculated for all pairs of power plants. This allows us to determine the correlation coefficient ρ of the prediction error for each pair of power plants, for example, the i-th and j-th power plants. ijThis is required. In this embodiment, this correlation coefficient is defined as the outlier tendency of the prediction error. Here, when any M sites are taken from N power plant sites (N≧M), and a balancing group candidate is constructed by combining the M sites, the standard deviation σ of the prediction error of the entire candidate is defined using the formula for the standard deviation of a sum in statistics.

number

[0072] Here σ i ρ is the standard deviation of the prediction error for the i-th power plant. ij This is the correlation coefficient between the prediction errors of the i-th power plant and the j-th power plant. This allows us to calculate the standard deviation σ of the prediction error for the entire balancing group candidate for various combinations, and then simply sum this standard deviation σ with the standard deviation of the error at each location.

number

[0073] As a simplified method, for locations where historical power generation prediction error data is available, the prediction error correlation coefficient is calculated using this historical data. The prediction error correlation coefficient is correlated with distance, and an approximation curve is drawn using the least squares method. Based on this approximation curve, the prediction error correlation coefficient for new power plants or candidate power plants is predicted, and the optimal balancing group is selected based on all existing and new prediction error correlation coefficients.

[0074] For example, in one use case, when forming a new balancing group, the N power plant sites are existing power plants or planned power plant construction sites. By selecting any M sites from these, the standard deviation σ is calculated using the above formula, and the ratio of the simple sum of the error standard deviations of each site, as shown in the above formula 2, is calculated. The group with the smallest value of this ratio can be considered the optimal balancing group.

[0075] Another use case involves adding existing power plants or new power plants, including the selection of candidate sites, to an existing balancing group. For example, suppose a balancing group already exists consisting of N1 power plants. We consider adding N2 power plants or candidate sites to this balancing group, bringing the total to N1 + N2 sites. Since N1 sites already constitute the balancing group, they are always included in the calculation of the standard deviation σ of the prediction error for the entire balancing group candidate. Furthermore, from the N2 sites under consideration, we extract M sites (M ≤ N2) that could be added as candidates, and calculate σ for the total of N1 + M sites.

[0076] Furthermore, the ratio of σ to the simple sum of the error standard deviations of each site is calculated, and the difference from the value of that ratio before the addition of M sites is taken as the score. This score can be used, for example, to decide whether or not to add a site to the balancing group, or, in the case of a potential power plant site, to decide on construction. This score can also be used to determine the service fee when adding a site to the balancing group.

[0077] As shown in Figure 6, following step S105 described above, matching is performed on both the power supply and demand balancing groups, a power supply plan is formulated according to the matched power supply and demand relationship (S107), and power is supplied according to the formulated power supply plan (S108). Power supplied in this manner is also continuously measured for generation, storage, or consumption by the power generation control terminal 8 on the power plant side and the HEMS 40 on the consumer side ("N" in step S109), and this measured actual data is collected (S101) and recorded by the power control server 2.

[0078] Furthermore, the deviation trend information (outlier tendency) calculated for each power plant can be used when selecting candidate sites for power plant construction. Specifically, the power generation method selection unit 752 accepts user-initiated power generation method selection operations, and the user also inputs the regions that are candidates for power plant construction. Based on the power generation method selected by the user and the location information of the candidate construction regions, the deviation trend prediction unit 74e refers to the deviation trend information and calculates the prediction error in the candidate construction regions as deviation trend prediction information. Then, considering balancing with other power plants, it is possible to select a power generation method and candidate construction region that can offset the deviation trend, thereby supporting the power plant construction plan.

[0079] (Power Generation Balancing Group Formation Program) The power generation balancing group formation system and method according to the above-described embodiments and modifications can be realized by executing a power generation balancing group formation program written in a predetermined language on a computer. That is, by installing a power supply management program on a server device, a dedicated device such as a smartphone, tablet PC, or in-vehicle terminal, or on an IC chip, and executing it on their CPUs, a system having the above-described functions can be easily constructed. This program can be distributed, for example, via a communication line, and can also be transferred as a package application that runs on a standalone computer.

[0080] Such programs can then be recorded on recording media readable by personal computers. Specifically, they can be recorded on various recording media, including magnetic recording media such as flexible disks and cassette tapes, optical discs such as CD-ROMs and DVD-ROMs, as well as USB memory sticks and memory cards.

[0081] (Effects / Actions) According to the embodiment described above, the time fluctuations of power generation at the power plant and total power consumption at the consumer, as well as the power waveform (frequency), are measured and analyzed to predict the amount of power generated and total power consumption. Then, based on the collected actual data, prediction history information, correlation information, and external information, the prediction error is calculated, the deviation trend, i.e., the outlier trend is analyzed, and the balancing group of the power plant is set based on the analyzed deviation trend (outlier trend). Matching is performed on both the power supply side and the demand side balancing group, and power is supplied according to the matched power supply and demand relationship.

[0082] According to this embodiment, regarding the power that can be supplied on the power generation side, balancing groups can be formed by taking into account the prediction errors and deviations in forecasts for each region, such as the power generation method and climate, thereby reducing the occurrence of unnecessary imbalances. As a result, by making the combination of power plants in the power generation balancing group the most effective combination, it is possible to contribute to improving profits.

[0083] In particular, in this embodiment, since external information is converted into pseudo-power generation prediction error data, it is possible to formulate balancing groups with existing power plants even for new power plants for which there is no historical data, providing an optimal regional selection logic for power plant construction and enabling revenue forecasting when new power plants are built.

[0084] It should be noted that the present invention is not limited to the embodiments described above, and the components can be modified and implemented in practice without departing from the spirit of the invention. Furthermore, various inventions can be formed by appropriate combinations of the multiple components disclosed in the above embodiments. For example, some components may be deleted from all the components shown in the embodiments. [Explanation of symbols]

[0085] D1…Performance Data H1, H2... Units of demand P1~P3... Power Plant 1…Power control system 2…Power control server 3…Communication Networks 4…User System 5…High-voltage power receiving point 7…Power control server 8... Power generation control terminal 40…HEMS 41…Smart meter 42… Storage battery 43…PV 45... Distribution board 46… Gateway terminal 50... Substation equipment 71a…Prediction Result Information Database 71b...User Database 71c...Performance Management Database 71d...Learning Information Database 72…Authentication Department 73…Communication Interface 74…Prediction Department 74a…External information acquisition department 74b...Correlation extraction section 74c... Prediction error calculation unit 74d...Learning Department 74e…Divergence trend prediction section 75... Balancing Group Setting Section 76…Data Management Department 76a...Performance data acquisition unit 76b…External information acquisition department 751...Divergence trend analysis department 752...Power generation method selection section 800...CPU bus 801...Storage 802…CPU 803...Memory 804... Input Interface 805…Output Interface 806...Communication Interface

Claims

1. A performance data acquisition unit that, for a group of power plants consisting of multiple power plants, measures the amount of electricity generated by each power plant and acquires performance data that associates the measurement results with the location information and power generation information related to the power generation method of the measured power plants. An external information acquisition unit that acquires external information from an information service organization, including at least one of the following: weather information, calendar information, or time information. A prediction unit that predicts the amount of power generated by the power plant based on the correlation between the aforementioned performance data and the aforementioned external information, A prediction error calculation unit compares the actual power generation amount of the power plant with the power generation amount prediction made by the prediction unit, and calculates the prediction error of the power generation amount prediction for each power plant. A deviation trend analysis unit analyzes the correlation between the power generation information and prediction errors of each power plant and the external information as deviation trend information, Based on the deviation trend information of each of the aforementioned power plants, a balancing group setting unit divides the group of power plants into multiple balancing groups. A power generation balancing group composition system characterized by being composed of the following.

2. The power generation balancing group composition system according to claim 1, characterized in that the balancing group setting unit performs the classification such that the prediction error of the entire balancing group exceeds a predetermined average value by combining power plants whose prediction error is below a predetermined low threshold and power plants whose prediction error is above a predetermined high threshold.

3. The system further includes a learning unit that extracts the distribution of features related to the actual power generation amount corresponding to power generation amount predictions with prediction errors exceeding a predetermined threshold value, and external information related to those power generation amount predictions, and forms a neural network as the aforementioned deviation trend information. The prediction unit has the neural network formed by the learning unit identify the actual data and external information related to the power generation forecast, and performs the power generation forecast according to the identification result. The power generation balancing group composition system according to claim 1 or 2, characterized by the above.

4. The candidate site acquisition department acquires candidate sites for the construction of power plants, A power generation method selection unit that selects a power generation method based on user input, Based on the power generation method selected by the power generation method selection unit and the location information of the proposed construction area, the deviation trend prediction unit refers to the deviation trend information and calculates the prediction error in the proposed construction area as deviation trend prediction information. The power generation balancing group composition system according to claim 1, further comprising the above.

5. The deviation trend analysis unit includes a learning unit that extracts the distribution of features related to the actual power generation amount corresponding to the power generation amount prediction with a prediction error of a predetermined threshold value and external information related to that power generation amount prediction, and forms a neural network as the deviation trend information. The deviation trend prediction unit, based on the power generation method selected by the power generation method selection unit and the location information of the proposed construction area, refers to the deviation trend information by having the neural network identify location information having similar features to the proposed construction area, and calculates the prediction error in the proposed construction area as deviation trend prediction information. The power generation balancing group composition system according to feature 4.

6. For a group of power plants consisting of multiple power plants, the amount of electricity generated by each power plant is measured, and the actual data acquisition unit acquires actual data by associating the measurement results with the location information and power generation information related to the power generation method of the measured power plants. In addition, the external information acquisition unit acquires external information from an information service organization, which includes at least one of the following: weather information, calendar information, or time information. The prediction unit performs a prediction step of predicting the amount of power generated by the power plant based on the correlation between the actual data and the external information, The prediction error rate calculation unit compares the actual power generation amount of the power plant with the power generation amount prediction made by the prediction unit and calculates the prediction error of the power generation amount prediction for each power plant, and the deviation trend analysis unit analyzes the correlation between the power generation information and prediction error of each power plant and the external information as deviation trend information in a deviation trend analysis step, The balancing group setting unit performs a balancing group setting step in which it divides the group of power plants into multiple balancing groups based on the deviation trend information of each power plant. A method for forming a power generation balancing group, characterized by including the following.

7. The method for forming a power generation balancing group according to claim 6, characterized in that, in the balancing group setting step, the balancing group setting unit performs the classification such that the prediction error of the entire balancing group exceeds a predetermined average value by combining power plants whose prediction error is below a predetermined low threshold and power plants whose prediction error is above a predetermined high threshold.

8. The learning unit further includes a learning step of extracting the distribution of features related to the actual power generation amount corresponding to power generation amount predictions with prediction errors exceeding a predetermined threshold value, and external information related to the power generation amount prediction, and forming a neural network as the deviation trend information. In the prediction step, the neural network formed by the learning unit identifies the actual data and external information related to the power generation prediction, and the power generation prediction is performed according to the identification result. The method for forming a power generation balancing group according to claim 6 or 7, characterized by the features described above.

9. The method for forming a power generation balancing group according to claim 6, further comprising a deviation trend prediction step in which a candidate area acquisition unit acquires candidate areas for the construction of a power plant, a power generation method selection unit selects a power generation method based on user operation, and a deviation trend prediction unit calculates a prediction error in the candidate area as deviation trend prediction information based on the power generation method selected by the power generation method selection unit and location information of the candidate area for construction.

10. The learning unit further includes a learning step of extracting the distribution of features related to the actual power generation amount corresponding to power generation amount predictions with prediction errors exceeding a predetermined threshold value, and external information related to the power generation amount prediction, and forming a neural network as the deviation trend information. In the deviation trend prediction step, the deviation trend prediction unit refers to the deviation trend information by having the neural network identify location information having similar features to the construction candidate area based on the power generation method selected by the power generation method selection unit and the location information of the construction candidate area, and calculates the prediction error in the construction candidate area as deviation trend prediction information. The method for forming a power generation balancing group according to feature 9.

11. Computers A performance data acquisition unit that, for a group of power plants consisting of multiple power plants, measures the amount of electricity generated by each power plant and acquires performance data that associates the measurement results with the location information and power generation information related to the power generation method of the measured power plants. An external information acquisition unit that acquires external information from an information service organization, including at least one of the following: weather information, calendar information, or time information. A prediction unit that predicts the amount of power generated by the power plant based on the correlation between the aforementioned performance data and the aforementioned external information, A prediction error calculation unit compares the actual power generation amount of the power plant with the power generation amount prediction made by the prediction unit, and calculates the prediction error of the power generation amount prediction for each power plant. A deviation trend analysis unit analyzes the correlation between the power generation information and prediction errors of each power plant and the external information as deviation trend information, Based on the deviation trend information of each power plant, the balancing group setting unit divides the power plant group into multiple balancing groups. A power generation balancing group composition program characterized by functioning as such.

12. The power generation balancing group composition program according to claim 11, characterized in that the balancing group setting unit performs the classification such that the prediction error of the entire balancing group exceeds a predetermined average value by combining power plants whose prediction error is below a predetermined low threshold and power plants whose prediction error is above a predetermined high threshold.

13. The aforementioned computer, Further functioning as a learning unit that extracts the distribution of features related to the actual power generation amount corresponding to power generation amount predictions with prediction errors exceeding a predetermined threshold value, and external information related to those power generation amount predictions, and forms a neural network as the aforementioned deviation trend information. The prediction unit has the neural network formed by the learning unit identify the actual data and external information related to the power generation forecast, and performs the power generation forecast according to the identification result. The power generation balancing group composition program according to claim 11 or 12, characterized by the features described herein.

14. The aforementioned computer, The candidate site acquisition department acquires candidate sites for the construction of power plants, A power generation method selection unit that selects a power generation method based on user input, Based on the power generation method selected by the power generation method selection unit and the location information of the proposed construction area, the deviation trend prediction unit refers to the deviation trend information and calculates the prediction error in the proposed construction area as deviation trend prediction information. The power generation balancing group composition program according to claim 11, characterized in that it further functions as such.

15. The deviation trend analysis unit includes a learning unit that extracts the distribution of features related to the actual power generation amount corresponding to the power generation amount prediction with a prediction error of a predetermined threshold value and external information related to that power generation amount prediction, and forms a neural network as the deviation trend information. The deviation trend prediction unit, based on the power generation method selected by the power generation method selection unit and the location information of the proposed construction area, refers to the deviation trend information by having the neural network identify location information having similar features to the proposed construction area, and calculates the prediction error in the proposed construction area as deviation trend prediction information. The power generation balancing group composition program according to feature 14.

Images