Real-Time Forecasting of Electricity Demand in a Streams-Based Architecture with Applications

• US 20170371308A1
• Filed: 06/22/2016
• Published: 12/28/2017
• Est. Priority Date: 06/22/2016
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Ghosh, Soumyadip, Hosking, Jonathan R., Natarajan, Ramesh, Subramanian, Shivaram, Zhang, Xiaoxuan
• Granted Patent: US 10,282,795 B2
• Time in Patent Office: Days
• CPC Class Codes: G05B 15/02 electric
• CPC Class Codes: G05B 2219/2639 Energy management, use maximum of cheap power, keep peak load low
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/0631 Resource planning, allocation, distributing or scheduling for enterprises or organisations
• CPC Class Codes: G06Q 50/06 Energy or water supply
• CPC Class Codes: Y04S 10/50 Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Abstract

A streams platform is used. Multiple streams of electricity usage data are received, each from an electrical meter providing periodic updates to electrical usage for devices connected to the electrical meter. Weather information is received corresponding to locations where the electrical meters are. Real-time predictive modeling of electricity demand is performed based on the received multiple streams of electricity usage data and the received weather information, at least by performing: updating a state space model for electrical load curves using the usage data from the streams and the weather, wherein the updating uses current load observations for the multiple streams for a current time period; and creating forecast(s) for the electricity demand. The forecast(s) of the electricity demand are output. Appliance-level predictions may be made and used, and substitution effects and load management functions may be performed.

Claims

1. A method, comprising:

receiving at a streams platform multiple streams of electricity usage data, each of the multiple streams from an electrical meter providing periodic updates to electrical usage for devices connected to the electrical meter;

receiving at the streams platform weather information corresponding to locations where the electrical meters are;

performing by the streams platform real-time predictive modeling of electricity demand based on the received multiple streams of electricity usage data and the received weather information, at least by performing the following;

updating a state space model for electrical load curves using the usage data from the streams and the weather, wherein the updating uses current load observations for the multiple streams for a current time period; and

creating one or more forecasts for the electricity demand;

outputting by the streams platform the one or more forecasts of the electricity demand.

2. The method of claim 1, wherein the multiple streams of electricity usage data are for customer premises in an electrical power distribution system, wherein the electrical meters are connected to customer premises, and wherein model training for the state space model, calibration of the state space model, and forecasting are performed entirely within the streams platform.

3. The method of claim 1, wherein the multiple streams of electricity usage data are for customer premises in an electrical power distribution system and wherein the method further comprises performing load management functions in the electrical power distribution system based on the one or more forecasts.

4. The method of claim 3, wherein:

updating a state space model for electrical load curves further comprises modeling demand substitution effects in the electrical power distribution system to enable shifting of demand from one time period to another in a daily load cycle; and

performing load management functions comprises incorporating the demand substitution effects across the daily load cycle.

5. The method of claim 3, wherein performing load management functions comprises incorporating load substitution effects across a daily load cycle.

6. The method of claim 1, wherein, updating the state space model and creating the one or more forecasts makes use of most recent observational data and forecasts of independent variables in at least the state space model for performing the forecasts.

7. The method of claim 1, wherein the multiple streams of electricity usage data are for customer premises in an electrical power distribution system, and wherein creating one or more forecasts for the electricity demand further comprises:

generating one or more primary forecasts at a primary forecasting level in a hierarchy comprising forecasting levels from individual customer premises to multiple additional forecasting levels, wherein the multiple additional forecasting levels comprise aggregated electricity demand from previous levels; and

generating forecasts at all levels below the primary forecasting level in the hierarchy, wherein the generating at the primary forecasting level and at all levels below the primary forecasting level uses non-hierarchical forecasting techniques.

8. The method of claim 7, wherein creating one or more forecasts for the electricity demand further comprises:

allocating the primary forecasts downward to the leaf level in the hierarchy in proportion to the forecasts at each lower level.

9. The method of claim 8, wherein:

creating one or more forecasts for the electricity demand further comprises aggregating forecasts for reporting.

10. The method of claim 1, wherein the multiple streams of electricity usage data are for customer premises in an electrical power distribution system and wherein the method further comprises for a customer premises:

receiving usage data streams from network-connected electrical devices in the customer premises;

monitoring usage data at least for time-level data;

building at least one model at customer-device level for time-of-day device-level predictions;

aggregating these time-of-day device-level predictions;

computing device-type based price sensitivity to create a personalized energy usage plan for this premises; and

sending back one or more real-time control signals to the network-connected electrical devices based on the personalized energy usage plan for this premises.

11. The method of claim 10, wherein the network-connected electrical devices are connected to a home area network.

12. The method of claim 10, wherein the time-level data comprise one or more of household identification, appliance type, on/off, usage rate, and duration.

13. An apparatus, comprising:

a memory including computer-readable program code; and

one or more processors, wherein the one or more processors are configured, in response to loading and executing the computer-readable program code, to cause the apparatus to perform operations comprising;

receiving multiple streams of electricity usage data, each of the multiple streams from an electrical meter providing periodic updates to electrical usage for devices connected to the electrical meter;

receiving weather information corresponding to locations where the electrical meters are;

performing real-time predictive modeling of electricity demand based on the received multiple streams of electricity usage data and the received weather information, at least by performing the following;

updating a state space model for electrical load curves using the usage data from the streams and the weather, wherein the updating uses current load observations for the multiple streams for a current time period; and

creating one or more forecasts for the electricity demand;

outputting the one or more forecasts of the electricity demand.

14. The apparatus of claim 13, wherein the multiple streams of electricity usage data are for customer premises in an electrical power distribution system, wherein the electrical meters are connected to customer premises, and wherein model training for the state space model, calibration of the state space model, and forecasting are performed entirely within the streams platform.

15. The apparatus of claim 13, wherein the multiple streams of electricity usage data are for customer premises in an electrical power distribution system and wherein the one or more processors are configured, in response to loading and executing the computer-readable program code, to cause the apparatus to perform the operation of performing load management functions in the electrical power distribution system based on the one or more forecasts.

16. The apparatus of claim 15, wherein:

updating a state space model for electrical load curves further comprises modeling demand substitution effects in the electrical power distribution system to enable shifting of demand from one time period to another in a daily load cycle; and

performing load management functions comprises incorporating the demand substitution effects across the daily load cycle.

17. The apparatus of claim 15, wherein performing load management functions comprises incorporating load substitution effects across a daily load cycle.

18. The apparatus of claim 13, wherein updating the state space model and creating the one or more forecasts makes use of most recent observational data and forecasts of independent variables in at least the state space model for performing the forecasts.

19. The apparatus of claim 13, wherein the multiple streams of electricity usage data are for customer premises in an electrical power distribution system, and wherein creating one or more forecasts for the electricity demand further comprises:

generating one or more primary forecasts at a primary forecasting level in a hierarchy comprising forecasting levels from individual customer premises to multiple additional forecasting levels, wherein the multiple additional forecasting levels comprise aggregated electricity demand from previous levels; and

generating forecasts at all levels below the primary forecasting level in the hierarchy, wherein the generating at the primary forecasting level and at all levels below the primary forecasting level uses non-hierarchical forecasting techniques.

20. The apparatus of claim 19, wherein creating one or more forecasts for the electricity demand further comprises:

allocating the primary forecasts downward to the leaf level in the hierarchy in proportion to the forecasts at each lower level.

21. The apparatus of claim 20, wherein:

creating one or more forecasts for the electricity demand further comprises aggregating forecasts for reporting.

22. The apparatus of claim 13, wherein the multiple streams of electricity usage data are for customer premises in an electrical power distribution system and wherein the one or more processors are configured, in response to loading and executing the computer-readable program code, to cause the apparatus to perform operations for a customer premises comprising:

receiving usage data streams from network-connected electrical devices in the customer premises;

monitoring usage data at least for time-level data;

building at least one model at customer-device level for time-of-day device-level predictions;

aggregating these time-of-day device-level predictions;

computing device-type based price sensitivity to create a personalized energy usage plan for this premises; and

sending back one or more real-time control signals to the network-connected electrical devices based on the personalized energy usage plan for this premises.

23. The apparatus of claim 22, wherein the network-connected electrical devices are connected to a home area network.

24. The apparatus of claim 22, wherein the time-level data comprise one or more of household identification, appliance type, on/off, usage rate, and duration.

25. A computer program product comprises a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer system to cause the computer system to perform the following:

receiving at a streams platform multiple streams of electricity usage data, each of the multiple streams from an electrical meter providing periodic updates to electrical usage for devices connected to the electrical meter;

receiving at the streams platform weather information corresponding to locations where the electrical meters are;

performing by the streams platform real-time predictive modeling of electricity demand based on the received multiple streams of electricity usage data and the received weather information, at least by performing the following;

updating a state space model for electrical load curves using the usage data from the streams and the weather, wherein the updating uses current load observations for the multiple streams for a current time period; and

creating one or more forecasts for the electricity demand;

outputting by the streams platform the one or more forecasts of the electricity demand.