SYSTEM FOR MANAGING ENERGY, METHOD OF MANAGING ENERGY, AND METHOD OF PREDICTING ENERGY DEMAND

• US 20180138742A1
• Filed: 11/16/2016
• Published: 05/17/2018
• Est. Priority Date: 11/16/2016
• Status: Active Grant

• Current Assignee: Korea Institute of Energy Research
• Original Assignee: Korea Institute of Energy Research
• Inventors: SONG, Yu-jin, SUNG, Yoon-dong
• Granted Patent: US 10,361,584 B2
• Time in Patent Office: Days
• CPC Class Codes: G05B 13/0265 the criterion being a learning criterion
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: H02J 2310/14 The load or loads being home appliances
• CPC Class Codes: H02J 3/003 Load forecast, e.g. methods or systems for forecasting future load demand
• CPC Class Codes: H02J 3/14 by switching loads on to, or off from, network, e.g. progressively balanced loading
• CPC Class Codes: H02J 3/322 the battery being on-board an electric or hybrid vehicle, e.g. vehicle to grid arrangements [V2G], power aggregation, use of the battery for network load balancing, coordinated or cooperative battery charging
• CPC Class Codes: Y02B 70/30 Systems integrating technologies related to power network operation and communication or information technologies for improving the carbon footprint of the management of residential or tertiary loads, i.e. smart grids as climate change mitigation technology in the buildings sector, including also the last stages of power distribution and the control, monitoring or operating management systems at local level
• CPC Class Codes: Y02B 70/3225 Demand response systems, e.g. load shedding, peak shaving
• CPC Class Codes: Y02B 90/20 Smart grids as enabling technology in buildings sector smart grids supporting the management or operation of end-user stationary applications in general, or like technologies with no associated climate change mitigation effect Y04S20/00
• CPC Class Codes: Y04S 20/00 Management or operation of end-user stationary applications or the last stages of power distribution; Controlling, monitoring or operating thereof
• CPC Class Codes: Y04S 20/12 Energy storage units, uninterruptible power supply [UPS] systems or standby or emergency generators, e.g. in the last power distribution stages
• CPC Class Codes: Y04S 20/222 Demand response systems, e.g. load shedding, peak shaving
• CPC Class Codes: Y04S 20/242 Home appliances

Abstract

Disclosed is a system for managing energy of a community in which at least one building exists. The system includes a community agent for managing community energy demand prediction data including building energy demand prediction data of each building and a community demand reaction load capacity including a building demand reaction load capacity of each building and managing a community demand reaction incentive policy applied to the community; and a machine learning device for generating a building demand reaction incentive policy to be applied to each building through a community-optimized machine learning model.

Claims

1. A system for managing energy of a community in which at least one building having at least one divided zone is located, the system comprising:

at least one zone agent for acquiring environment data of each zone from a sensor network installed in each zone, acquiring device energy usage data of an energy device and a personal device operating in each zone, and managing physical information of each zone;

a machine learning device for generating occupant estimation information of each zone through a first machine learning model including temperature data and CO₂ data of the environment data and the device energy usage data of the personal device as input data, generating zone energy demand prediction data of each zone through a second machine learning model including the occupant estimation information, the environment data, the device energy usage data of the energy device and the personal device, and the physical information of each zone as input data, and transmitting the generated occupant estimation information and zone energy demand prediction data to the zone agent,

wherein the zone agent manages energy of each zone based on the zone energy demand prediction data received for each zone.

2. The system of claim 1, wherein the first machine learning model further includes illumination data or wireless communication data with a user terminal as the input data.

3. The system of claim 1, wherein the physical information of each zone further includes location information of each zone, the machine learning device further acquires outdoor air data at a location of each zone from the zone agent or another device, and the second machine learning model generates the zone energy demand prediction data with the outdoor air data as additional input data.

4. The system of claim 1, further comprising a building agent for acquiring state information of at least one building device among a building load, a building distributed power supply a building Energy Storage System (ESS), and a building Electric Vehicle (EV) charging station, which do not belong to the zone in each building, and managing physical information of each building, wherein the machine device generates building energy demand prediction data through a third machine learning model including the zone energy demand prediction data of each zone in each building, the state information of the building device, and the physical information of each building as input data and transmits the generated building energy demand prediction data to the building agent.

5. The system of claim 4, further comprising a community agent for acquiring state information of at least one community device among a community load, a community distributed power supply, a community ESS, and a community EV charging station, which do not belong to the building in the community, wherein the machine device generates community energy demand prediction data through a fourth machine learning model including the building energy demand prediction data of each building, the state information of the community device, and the physical information of the community as input data and transmits the generated building energy demand prediction data to the community agent.

6. The system of claim 5, wherein the community agent manages the building energy demand prediction data of each building and building energy usage data and manages a community demand reaction incentive policy applied to the community, the machine learning device includes the building energy demand prediction data of each building, the building energy usage data, and the community demand reaction incentive policy as input data, generates a building demand reaction incentive policy to be applied to each building through a community-optimized machine learning model for calculating a best community demand management profit of the community, and transmits the generated building demand reaction incentive policy to the community agent, and the community agent transmits the building demand reaction incentive policy to a building agent corresponding to each building.

7. The system of claim 6, wherein the machine learning device generates a predicted value of the community demand management profit through the community-optimized machine learning model, receives an actually measured value of the community demand management profit from the community agent, and learns the community-optimized machine learning model based on error data according to a difference between the predicted value and the actually measured value.

8. The system of claim 6, wherein the community agent bids for a demand reaction system based on the community energy demand prediction data and a community demand reaction load capacity and receives the community demand reaction incentive policy as a result of the bid.

9. The system of claim 6, wherein the machine learning device learns the community-optimized machine learning model based on error data according to a difference between a load capacity of each building which actually participates in a demand reaction and a demand reaction load capacity calculated for each building as one element of the community demand reaction incentive policy.

10. The system of claim 6, wherein the community-optimized machine learning model further generates a predicted value of a demand reaction adaptation index of each building, and the machine learning device learns the community-optimized machine learning model based on error data according to a difference between the predicated value and an actually measured value of the demand reaction adaptation index.

11. The system of claim 6, wherein the building agent manages the zone energy demand prediction data and the zone energy usage data of each zone and manages the building demand reaction incentive policy applied to the building, the machine learning device generates a zone demand reaction incentive policy to be applied to each zone through a building-optimized machine learning model, which includes the zone energy demand prediction data and the zone energy usage data of each zone and the building demand reaction incentive policy as input data and calculates minimum energy costs for the building, and transmits the generated zone demand reaction incentive policy to the building agent, and the building agent transmits the zone demand reaction incentive policy to the zone agent corresponding to each zone.

12. The system of claim 11, wherein the zone agent manages user setting information, indoor/outdoor environment information, energy use pattern information of each device, and occupant information, and manages the zone demand reaction incentive policy, and the machine learning device generates an optimal control scenario for a device within the zone through a zone-optimized program that includes the user setting information, the indoor/outdoor environment information, the energy use pattern information of each device, and the occupant information as input data and calculates an optimal value of an objective function including user convenience and energy cost reduction as parameters, and transmits the optimal control scenario to the zone agent.

13. The system of claim 12, wherein the user convenience is measured with a factor corresponding to proximity between the number of settings of an energy device or an energy device setting value and an energy device measurement value.

14. A method of managing energy of a community in which at least one building having at least one divided zone is located, the method comprising:

acquiring environment data of each zone from a sensor network installed in each zone, acquiring device energy usage data of an energy device and a personal device operating in each zone, and acquiring physical information of each zone;

calculating a number of occupants in each zone through a first machine learning model including temperature data and CO₂ data of the environment data and the device energy usage data of the personal device as input data, and generating zone energy demand prediction data of each zone through a second machine learning model including the occupant estimation information, the environment data, the device energy usage data of the energy device and the personal device, and the physical information of each zone as input data; and

managing energy of each zone based on the zone energy demand prediction data.

15. The method of claim 14, further comprising:

acquiring state information of at least one building device among a building load, a building distributed power supply, a building Energy Storage System (ESS), and a building Electric Vehicle (EV) charging station, which do not belong to the zone in each building, and acquiring physical information of each building; and

generating building energy demand prediction data through a third machine learning model including the zone energy demand prediction data of each zone in each building, the state information of the building device, and physical information of each building as input data.

16. The method of claim 15, further comprising:

acquiring state information of at least one community device among a community load, a community distributed power supply, a community ESS, and a community EV charging station, which do not belong to the building in the community; and

generating community energy demand prediction data through a fourth machine learning model including the building energy demand prediction data of each building, the state information of the community device, and the physical information of the community as input data.

17. The method of claim 16, further comprising:

managing the community energy demand prediction data and a community demand reaction load capacity including a building demand reaction load capacity of each building and managing a community demand reaction incentive policy applied to the community; and

generating a building demand reaction incentive policy to be applied to each building through a community-optimized machine learning model that includes the building energy demand prediction data of each building, building energy usage data, and the community demand reaction incentive policy as input data and calculates a maximum community demand management profit for the community.

18. The method of claim 17, further comprising generating a predicted value of the community demand management profit through the community-optimized machine learning model, receiving an actually measured value of the community demand management profit from the community agent, and learning the community-optimized machine learning model based on error data according to a difference between the predicted value and the actually measured value.

19. A method of predicting energy for a community in which at least one building divided into at least one zone exists, the method comprising:

calculating a number of occupants in each zone through a first machine learning model including CO₂ data of each zone as input data;

generating zone energy demand prediction data of each zone through a second machine learning model including the number of occupants, environment data of each zone, and device energy usage data as input data;

generating building energy demand prediction data of each building through a third machine learning model including zone energy demand prediction data of each zone and state information of a building device, which does not belong to each zone, as input data; and

generating community energy demand prediction data of the community through a fourth machine learning model including building energy demand prediction data of each building and state information of a community device, which does not belong to each building, as input data.

20. The method of claim 19, wherein the second machine learning model further includes physical information of each zone as the input data, the third machine learning model further includes physical information of each building as the input data, and the fourth machine learning model further includes physical information of the community as the input data.