Automated energy load forecaster

• US 10,197,984 B2
• Filed: 10/12/2015
• Issued: 02/05/2019
• Est. Priority Date: 10/12/2015
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bharti, Harish, Dutta, Binayak, Patra, Abhay K.
• Primary Examiner(s): Coyer, Ryan D.
• Application Number:
  US14/880,308
• Publication Number:
  US 20170102683A1
• Time in Patent Office: 1,212 Days
• Field of Search: 700291
• CPC Class Codes: G05B 19/042 using digital processors G05B19/05 takes precedence
• CPC Class Codes: G05B 19/0428 Safety, monitoring G05B19/0423 takes precedence
• 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 50/06 Energy or water supply
• CPC Class Codes: H02J 13/00017 using optical fiber
• CPC Class Codes: H02J 13/00028 involving the use of Internet protocols
• CPC Class Codes: H02J 13/00034 the elements or equipment being or involving an electric power substation
• CPC Class Codes: H02J 3/003 Load forecast, e.g. methods or systems for forecasting future load demand
• 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 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: Y02P 80/20 using renewable energy
• 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
• 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 20/20 End-user application control systems

Abstract

Energy load forecasts are generated via model(s) for the grid hierarchy elements for different forecast time scale periods as a function of different sets of prioritized contextual influencing factors for respective associated combinations of grid hierarchy elements and forecast time scale periods. Relative priority values of the sets of the contextual influencing factors are iteratively weighted until a revised energy load forecast generated as a function of the weighted values via the model(s) is within a threshold value of a historic energy load data value for the associated combination of the grid hierarchy element and forecast time scale period.

Claims

1. A computer-implemented method for prioritizing and weighting model contextual influencing factors for energy load forecasting, the method comprising executing on a computer processor:

identifying via an application executing on a client platform computing node a plurality of contextual influencing factors as each relevant to use for energy load forecasting for an energy grid infrastructure element as a function of at least one energy forecasting model, wherein the energy grid infrastructure element is one of a zonal substation, a sub-transmission feeder, a distribution substation, a distribution feeder and a sub-transmission substation;

rank prioritizing via the application executing on the client platform computing node the contextual influencing factors into each of a plurality of different prioritized sets that each have different relative priority values assigned to each of the contextual influencing factors within the different respective sets as a function of differences in relevance determined for each of different combinations of the grid hierarchy element with ones each of a plurality of different forecast time scale periods, wherein the rank prioritizing of the contextual influencing factors generates different assignments of the relative priority values to ones of the contextual influencing factors within at least two of the different prioritized sets;

generating, via the application executing on the client platform computing node applying the at least one model for energy load forecasting, energy load forecast values for the grid hierarchy element for each of the plurality of different forecast time scale periods as a function of respective ones of the different prioritized sets of the contextual influencing factor relative priority values that are each prioritized for respective associated ones of the different combinations of the grid hierarchy element with respective ones of the plurality of different forecast time scale periods; and

iteratively weighting the relative priority values of a selected set of the prioritized sets of contextual influencing factors, and generating via the at least one model for energy load forecasting, a revised energy load forecast value for the grid hierarchy element as a function of the weighted relative priority values via the application executing on the client platform computing node, until the revised energy load forecast is within a threshold value of a historic energy load data value for the combination of the grid hierarchy element and the forecast time scale period that is associated with the selected set of the prioritized sets of contextual influencing factors.

2. The method of claim 1, further comprising:

integrating computer-readable program code into a computer system comprising the processor that executing the application on the client platform computing node, a computer readable memory in circuit communication with the processor, and a computer readable storage medium in circuit communication with the processor; and

wherein the processor executes program code instructions stored on the computer-readable storage medium via the computer readable memory and thereby performs the steps of identifying the plurality of contextual influencing factors, rank prioritizing the contextual influencing factors into the different prioritized sets of relative priority values, generating the energy load forecast values for the grid hierarchy element for each of the plurality of different forecast time scale periods, and iteratively weighting the relative priority values of the selected set of the prioritized sets of contextual influencing factors and generating via the at least one model for energy load forecasting the revised energy load forecast value.

3. The method of claim 1, wherein the at least one model for energy load forecasting for the energy grid infrastructure element is a plurality of different models, the method further comprising:

identifying a plurality of different best weighted sets of the iteratively reweighted relative priority values of the contextual influencing factors of the selected set of the prioritized sets, one each for different respective ones of the plurality of different models, in response to generating most accurate forecast values relative to historic energy load data values for different respective ones of the plurality of different models.

4. The method of claim 3, further comprising:

choosing the different best weighted sets of the iteratively reweighted relative priority values of the selected set of the prioritized sets of contextual influencing factors as a function of generating lowest root mean squared error values relative to the historic energy load data values for different respective ones of the plurality of different models.

5. The method of claim 4, further comprising:

determining a final model as a weighted function of identified plurality of different best weighted sets of the iteratively reweighted relative priority values of the selected set of the prioritized sets of contextual influencing factors.

6. The method of claim 5, wherein the plurality of different forecast time scale periods comprises an immediate forecast horizon period, a short term forecast horizon period, a mid-term forecast horizon period and a long term forecast horizon period.

7. The method of claim 6, wherein the plurality of contextual influencing factors comprises:

a utilization history factor that quantifies an expected energy load or rate of usage for the grid hierarchy element;

a weather forecast factor that quantifies an effect of projected temperatures, wind amounts or precipitation amounts that differ from a seasonal norm over at least one of the time scale periods; and

an economic factor that quantifies a trend of change in an amount of economic activity projected for a region served by the grid hierarchy element that consumes energy and thereby imparts a corresponding change in an amount of energy load on the grid hierarchy element for at least one of the time scale periods.

8. A system, comprising:

a processor on a client platform computing node;

a computer readable memory in circuit communication with the processor; and

a computer readable storage medium in circuit communication with the processor;

wherein the processor executes program instructions stored on the computer-readable storage medium via the computer readable memory and thereby;

identifies via an application executing on the client platform computing node processor a plurality of contextual influencing factors as each relevant to use for energy load forecasting for an energy grid infrastructure element as a function of at least one energy forecasting model, wherein the energy grid infrastructure element is one of a zonal substation, a sub-transmission feeder, a distribution substation, a distribution feeder and a sub-transmission substation;

rank prioritizes via the application executing on the client platform computing node the contextual influencing factors into each of a plurality of different prioritized sets that each have different relative priority values assigned to each of the contextual influencing factors within the different respective sets as a function of differences in relevance determined for each of different combinations of the grid hierarchy element with ones each of a plurality of different forecast time scale periods, wherein the rank prioritizing of the contextual influencing factors generates different assignments of the relative priority values to ones of the contextual influencing factors within at least two of the different prioritized sets;

generates, via the application executing on the client platform computing node applying the at least one model for energy load forecasting, energy load forecast values for the grid hierarchy element for each of the plurality of different forecast time scale periods as a function of respective ones of the different prioritized sets of the contextual influencing factor relative priority values that are each prioritized for respective associated ones of the different combinations of the grid hierarchy element with respective ones of the plurality of different forecast time scale periods; and

iteratively weights the relative priority values of a selected set of the prioritized sets of contextual influencing factors, and generates via the at least one model for energy load forecasting a revised energy load forecast value for the grid hierarchy element as a function of the weighted relative priority values, via the application executing on the client platform computing node, until the revised energy load forecast is within a threshold value of a historic energy load data value for the combination of the grid hierarchy element and the forecast time scale period that is associated with the selected set of the prioritized sets of contextual influencing factors.

9. The system of claim 8, wherein the at least one model for energy load forecasting for the energy grid infrastructure element is a plurality of different models;

and

wherein the processor executes the program instructions stored on the computer-readable storage medium via the computer readable memory and thereby identifies a plurality of different best weighted sets of the iteratively reweighted relative priority values of the contextual influencing factors of the selected set of the prioritized sets, one each for different respective ones of the plurality of different models, in response to generating most accurate forecast values relative to historic energy load data values for different respective ones of the plurality of different models.

10. The system of claim 9, wherein the processor executes the program instructions stored on the computer-readable storage medium via the computer readable memory and thereby chooses the different best weighted sets of the iteratively reweighted relative priority values of the selected set of the prioritized sets of contextual influencing factors as a function of generating lowest root mean squared error values relative to the historic energy load data values for different respective ones of the plurality of different models.

11. The system of claim 10, wherein the processor executes the program instructions stored on the computer-readable storage medium via the computer readable memory and thereby determines a final model as a weighted function of identified plurality of different best weighted sets of the iteratively reweighted relative priority values of the selected set of the prioritized sets of contextual influencing factors.

12. The system of claim 11, wherein the plurality of different forecast time scale periods comprises an immediate forecast horizon period, a short term forecast horizon period, a mid-term forecast horizon period and a long term forecast horizon period.

13. The system of claim 12, wherein the plurality of contextual influencing factors comprises:

a utilization history factor that quantifies an expected energy load or rate of usage for the grid hierarchy element;

a weather forecast factor that quantifies an effect of projected temperatures, wind amounts or precipitation amounts that differ from a seasonal norm over at least one of the time scale periods; and

an economic factor that quantifies a trend of change in an amount of economic activity projected for a region served by the grid hierarchy element that consumes energy and thereby imparts a corresponding change in an amount of energy load on the grid hierarchy element for at least one of the time scale periods.

14. The system of claim 13, wherein the program instructions stored on the computer-readable storage medium are provided as a service in a cloud environment.

15. A computer program product for prioritizing and weighting model contextual influencing factors for energy load forecasting, the computer program product comprising:

a computer readable storage medium having computer readable program code embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the computer readable program code comprising instructions for execution by a processor on a client platform computing node that cause the processor to;

identify via an application executing on the client platform computing node processor a plurality of contextual influencing factors as each relevant to use for energy load forecasting for an energy grid infrastructure element as a function of at least one energy forecasting model, wherein the energy grid infrastructure element is one of a zonal substation, a sub-transmission feeder, a distribution substation, a distribution feeder and a sub-transmission substation;

rank prioritize via the application executing on the client platform computing node the contextual influencing factors into each of a plurality of different prioritized sets that each have different relative priority values assigned to each of the contextual influencing factors within the different respective sets as a function of differences in relevance determined for each of different combinations of the grid hierarchy element with ones each of a plurality of different forecast time scale periods, wherein the rank prioritizing of the contextual influencing factors generates different assignments of the relative priority values to ones of the contextual influencing factors within at least two of the different prioritized sets;

generate, via the application executing on the client platform computing node applying the at least one model for energy load forecasting, energy load forecast values for the grid hierarchy element for each of the plurality of different forecast time scale periods as a function of respective ones of the different prioritized sets of the contextual influencing factor relative priority values that are each prioritized for respective associated ones of the different combinations of the grid hierarchy element with respective ones of the plurality of different forecast time scale periods; and

iteratively weight the relative priority values of a selected set of the prioritized sets of contextual influencing factors, and generates via the at least one model for energy load forecasting a revised energy load forecast value for the grid hierarchy element as a function of the weighted relative priority values, via the application executing on the client platform computing node, until the revised energy load forecast is within a threshold value of a historic energy load data value for the combination of the grid hierarchy element and the forecast time scale period that is associated with the selected set of the prioritized sets of contextual influencing factors.

16. The computer program product of claim 15, wherein the at least one model for energy load forecasting for the energy grid infrastructure element is a plurality of different models;

and

wherein the computer readable program code instructions for execution by the processor further cause the processor to identify a plurality of different best weighted sets of the iteratively reweighted relative priority values of the contextual influencing factors of the selected set of the prioritized sets, one each for different respective ones of the plurality of different models, in response to generating most accurate forecast values relative to historic energy load data values for different respective ones of the plurality of different models.

17. The computer program product of claim 16, wherein the computer readable program code instructions for execution by the processor further cause the processor to choose the different best weighted sets of the iteratively reweighted relative priority values of the selected set of the prioritized sets of contextual influencing factors as a function of generating lowest root mean squared error values relative to the historic energy load data values for different respective ones of the plurality of different models.

18. The computer program product of claim 17, wherein the computer readable program code instructions for execution by the processor further cause the processor to determine a final model as a weighted function of identified plurality of different best weighted sets of the iteratively reweighted relative priority values of the selected set of the prioritized sets of contextual influencing factors.

19. The computer program product of claim 18, wherein the plurality of different forecast time scale periods comprises an immediate forecast horizon period, a short term forecast horizon period, a mid-term forecast horizon period and a long term forecast horizon period.

20. The computer program product of claim 19, wherein the plurality of contextual influencing factors comprises:

a utilization history factor that quantifies an expected energy load or rate of usage for the grid hierarchy element;

a weather forecast factor that quantifies an effect of projected temperatures, wind amounts or precipitation amounts that differ from a seasonal norm over at least one of the time scale periods; and

an economic factor that quantifies a trend of change in an amount of economic activity projected for a region served by the grid hierarchy element that consumes energy and thereby imparts a corresponding change in an amount of energy load on the grid hierarchy element for at least one of the time scale periods.