Power curve correlation system

• US 9,118,182 B2
• Filed: 01/04/2012
• Issued: 08/25/2015
• Est. Priority Date: 01/04/2012
• Status: Active Grant

• Current Assignee: GE Digital Holdings LLC (GE Vernova, Inc.)
• Original Assignee: General Electric Company
• Inventors: McDonald, Jason Louis, Alderson, William Michael, Grey, Audley
• Primary Examiner(s): Breene, John
• Assistant Examiner(s): Hagos, Eyob
• Application Number:
  US13/343,074
• Publication Number:
  US 20130173191A1
• Time in Patent Office: 1,329 Days
• Field of Search: 702/61
• US Class Current: 1/1
• CPC Class Codes: G01R 1/203 Resistors used for electric measuring, e.g. decade resistors standards, resistors for comparators, series resistors, shunts resistors in general H01C; microwave or radiowave terminations H01P1/26; coupling devices H01R
• CPC Class Codes: G01R 19/0092 measuring current only all subgroups of G01R19/00 take precedence
• CPC Class Codes: G01R 31/2642 Testing semiconductor operation lifetime or reliability, e.g. by accelerated life tests
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities
• CPC Class Codes: G06Q 50/06 Energy or water supply
• CPC Class Codes: H02J 3/00 Circuit arrangements for ac mains or ac distribution networks
• CPC Class Codes: H02J 3/003 Load forecast, e.g. methods or systems for forecasting future load demand
• 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 50/14 Marketing, i.e. market research and analysis, surveying, promotions, advertising, buyer profiling, customer management or rewards

Abstract

A power curve correlation system is disclosed. The power curve correlation system includes a system including: at least one computing device configured to verify a forecast of a power network load for a component by performing actions including: obtaining a set of load data for the component in the power network over a period; identifying similar load patterns in the set of load data using a pattern recognition technique, independent of at least one of: a scale for a component load, a minimum component load value or a maximum component load value; grouping the identified similar load patterns into distinct groups; categorizing at least one of the distinct groups according to a recurring event associated with a load pattern in the at least one distinct group; and providing the categorized group for verifying the forecast of the power network load for the component using the at least one categorized group.

Claims

1. A system comprising:

at least one computing device configured to verify a forecast of a power network load for a component of the power network by performing actions including;

prompting the power network to provide load data for the component of the power network;

obtaining, from the power network during operation, a set of load data for the component in the power network over a period;

identifying similar load patterns in the set of load data using a pattern recognition technique, independent of at least one of;

a scale for a component load, a predetermined minimum component load value or a predetermined maximum component load value, the identifying comprising;

identifying similar historic load patterns in a set of historic load data;

comparing the identified similar load patterns to the identified similar historic load patterns; and

determining, using a mean absolute percent error (MAPE) function and based on comparing the identified similar load patterns to the identified similar historic load pattern, that the identified similar load patterns are similar to the identified similar historic load patterns within a predetermined accuracy range;

grouping the identified similar load patterns into distinct groups based upon the identified similar load patterns in the set of load data, wherein the distinct groups are represented by at least one of;

similar load values for the component over a predefined period, similar sudden increases in the load patterns for the component, or similar sudden decreases in the load patterns for the component;

categorizing at least one of the distinct groups according to a recurring event associated with having a load pattern similar to a load pattern of the at least one distinct group;

providing the at least one categorized group for verifying the forecast of the power network load for the component using the at least one categorized group, the providing comprising;

providing the at least one categorized group and the set of load data to an interface;

transmitting the at least one categorized group and the set of load data to a library of the at least one computing device; and

storing the at least one categorized group and the set of load data in the library of the at least one computing device;

verifying the accuracy of the forecast of the power network load by comparing an initial forecast with a provided categorized group; and

reforecasting the power network load in response to inaccuracies in the forecast, wherein the reforecasting includes altering algorithms used in forecasting the power network load.

2. The system of claim 1, wherein the set of load data includes at least one of:

hourly load data, daily load data, weekly load data, monthly load data, alternative measurement device load data, or phase based load data.

3. The system of claim 1, wherein the at least one computing device further performing actions including:

storing the categorized groups within a library, the stored categorized groups accessible for verifying an additional forecast of a future power network load for the component.

4. The system of claim 1, wherein the recurring event includes at least one of:

a demand response management system (DRMS) event, a power supply system maintenance event, a power supply system outage event, or a power usage spiking event.

5. The system of claim 1, wherein the pattern recognition technique identifies the similar load patterns by identifying at least one of:

similarities in the component load values over a predefine period, similarities in a graphic representation of the load patterns, similarities in spikes in the component load, or similarities in drops in the component load.

16. The system of claim 1, wherein the grouping the identified similar load patterns into distinct groups includes grouping similar graphic representations for the load patterns on a single load graph, and wherein the categorizing includes comparing the single load graph to a load graph associated with the recurring event.

6. A program product stored on a non-transitory computer readable medium for verifying a forecast of a power network load for a component, the computer readable medium comprising program code for causing the computer system to:

prompt the power network to provide load data for the component of the power network;

obtain, from the power network during operation, a set of load data for the component in the power network over a period;

identify similar load patterns in the set of load data using a pattern recognition technique, independent of at least one of;

a scale for a component load, a predetermined minimum component load value or a predetermined maximum component load value, the identifying comprising;

identifying similar historic load patterns in a set of historic load data

comparing the identified similar load patterns to the identified similar historic load patterns; and

determining, using a mean absolute percent error (MAPE) function and based on comparing the identified similar load patterns to the identified similar historic load pattern, that the identified similar load patterns are similar to the identified similar historic load patterns within a predetermined accuracy range;

group the identified similar load patterns into distinct groups based upon the identified similar load patterns in the set of load data, wherein the distinct groups are represented by at least one of;

similar load values for the component over a predefined period, similar sudden increases in the load patterns for the component, or similar sudden decreases in the load patterns for the component;

categorize at least one of the distinct groups according to a recurring event having associated with a load pattern similar to a load pattern of the at least one distinct group;

provide the at least one categorized group for verifying the forecast of the power network load for the component using the at least one categorized group, the providing comprising;

providing the at least one categorized group and the set of load data to an interface;

transmitting the at least one categorized group and the set of load data to a library of the at least one computing device; and

storing the at least one categorized group and the set of load data in the library of the at least one computing device;

verifying the accuracy of the forecast of the power network load by comparing an initial forecast with a provided categorized group; and

reforecasting the power network load in response to inaccuracies in the forecast, wherein the reforecasting includes altering algorithms used in forecasting the power network load.

7. The program product of claim 6, wherein the set of load data includes at least one of:

hourly load data, daily load data, weekly load data, monthly load data, alternative measurement device load data, or phase based load data.

8. The program product of claim 6, further comprises program code for causing the computer system to:

store the categorized groups within a library, the stored categorized groups accessible for verifying an additional forecast of a future power network load for the component.

9. The program product of claim 6, wherein the recurring event includes at least one of:

a demand response management system (DRMS) event, a power supply system maintenance event, a power supply system outage event, or a power usage spiking event.

10. The program product of claim 6, wherein the pattern recognition technique identifies the similar load patterns by identifying at least one of:

similarities in the component load values over a predefined period, similarities in a graphic representation of the load patterns, similarities in spikes in the component load, or similarities in drops in the component load.

11. A system comprising:

a human machine interface (HMI) operably connected to a power network; and

at least one computing device operably connected to the HMI, the at least one computing device configured to verify a forecast of a power network load for a component of the power network by performing actions including;

prompting the power network to provide load data for the component of the power network;

obtaining, from the power network during operation, a set of load data for the component in the power network over a period;

identifying similar load patterns in the set of load data using a pattern recognition technique, independent of at least one of;

a scale for a component load, a predetermined minimum component load value or a predetermined maximum component load value, the identifying comprising;

identifying similar historic load patterns in a set of historic load data

comparing the identified similar load patterns to the identified similar historic load patterns; and

determining, using a mean absolute percent error (MAPE) function and based on comparing the identified similar load patterns to the identified similar historic load pattern, that the identified similar load patterns are similar to the identified similar historic load patterns within a predetermined accuracy range;

grouping the identified similar load patterns into distinct groups based upon the identified similar load patterns in the set of load data, wherein the distinct groups are represented by at least one of;

similar load values for the component over a predefined period, similar sudden increases in the load patterns for the component, or similar sudden decreases in the load patterns for the component;

categorizing at least one of the distinct groups according to a recurring event having associated with a load pattern similar to a load pattern of the at least one distinct group; and

providing the at least one categorized group for verifying the forecast of the power network load for the component using the at least one categorized group, the providing comprising;

providing the at least one categorized group and the set of load data to an interface;

transmitting the at least one categorized group and the set of load data to a library of the at least one computing device; and

storing the at least one categorized group and the set of load data in the library of the at least one computing device;

verifying the accuracy of the forecast of the power network load by comparing an initial forecast with a provided categorized group; and

reforecasting the power network load in response to inaccuracies in the forecast, wherein the reforecasting includes altering algorithms used in forecasting the power network load.

12. The system of claim 11, wherein the set of load data includes at least one of:

hourly load data, daily load data, weekly load data, monthly load data, alternative measurement device load data, or phase based load data.

13. The system of claim 11, wherein the at least one computing device further performing actions including:

storing the categorized groups within a library, the stored categorized groups accessible for verifying an additional forecast of a future power network load for the component.

14. The system of claim 11, wherein the recurring event includes at least one of:

a demand response management system (DRMS) event, a power supply system maintenance event, a power supply system outage event, or a power usage spiking event.

15. The system of claim 11, wherein the pattern recognition technique identifies the similar load patterns by identifying at least one of:

similarities in the component load values over a predefined period, similarities in a graphic representation of the load patterns, similarities in spikes in the component load, or similarities in drops in the component load.