Thermostat classification method and system

• US 9,547,316 B2
• Filed: 03/13/2013
• Issued: 01/17/2017
• Est. Priority Date: 09/07/2012
• Status: Active Grant

• Current Assignee: Opower, Inc. (Oracle Corporation)
• Original Assignee: Opower, Inc. (Oracle Corporation)
• Inventors: Shilts, Erik, Ben-Meshulam, Jonathan Z., Curtis, Richard Tyler
• Primary Examiner(s): Ali, Mohammad
• Assistant Examiner(s): Booker, Kelvin
• Application Number:
  US13/799,550
• Publication Number:
  US 20140074300A1
• Time in Patent Office: 1,406 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G05D 23/01 without auxiliary power
• CPC Class Codes: G06Q 50/06 Energy or water supply

Abstract

The disclosure provides a computer-implemented method and system of reducing commodity usage by providing tailored consumer information to the consumer. The method utilizes neural network and machine learning techniques to calculate and cluster statistical data to classify the premises for desired observable condition, including the presence of a programmed thermostat. A score is determined that corresponds to at least one of: (i) a present state of an observable condition, (ii) a non-present state of the observable condition, and (iii) a degree of a condition of the observable condition, to provide tailored consumer information associated to the consumer'"'"'s usage of the commodity.

Claims

1. A computer-implemented method, the computer-implemented method comprising:

receiving energy usage data of a consumer associated to a premises, the premises having a controller to regulate temperature in a controlled space within the premises, the energy usage data representing energy usage being associated, in part, to regulation of the controller;

analyzing, via a processor, the energy usage data to determine whether the controller is a programmed thermostat or an unprogrammed thermostat;

wherein the analyzing includes determining, via the processor, whether the energy usage data indicates that the energy usage adjusts in a pattern according to a time-of-day dependent setting;

generating, via the processor, a status that indicates the controller at the premises is the programmed thermostat in response to determining that the energy usage adjusts in the pattern; and

generating, via the processor, the status that indicates the controller is the unprogrammed thermostat in response to determining that the energy usage does not adjust in the pattern.

2. The computer-implemented method of claim 1, wherein determining whether the energy usage data indicates that the energy usage adjusts in the pattern according to the time-of-day dependent setting includes:

comparing a first statistical feature derived from the received energy usage data to a second statistical feature derived from at least one representative pattern of a premises with a controller that adjusts temperature according to a time-of-day dependent setting.

3. The computer-implemented method of claim 1, further comprising providing, based on the generated status, tailored consumer information to the customer associated with the premises including:

outputting a report to the consumer at the premises based on the generated status, the report including, at least one of;

a message indicating at least one of an estimated reduction in usage of the commodity and an estimated reduction in cost associated with the estimated reduction in usage if the programmed thermostat is employed at the premises; and

an indication of a presence of a malfunctioned HVAC at the premises.

4. The computer-implemented method of claim 1, wherein the energy usage data is retrieved from at least one of an advanced meter infrastructure (AMI) meter and a home display system located at the premises.

5. The computer-implemented method of claim 1, wherein analyzing the energy usage data includes:

calculating, from the energy usage data, statistical data related to at least one of seasonality, trend, or decomposed data;

assigning the calculated statistical data to a representative self-organized map, the self-organized map having at least one cluster corresponding to a second premises having a programmed thermostat or an unprogrammed thermostat;

wherein the representative self-organized map is determined from at least one of k-means, hierarchical clustering, or Kohonen self-organizing maps.

6. The computer-implemented method of claim 5, wherein the representative self-organizing map includes at least one of a Kohonen supervised self-organizing maps and a Kohonen unsupervised self-organizing map.

7. The computer-implemented method of claim 5, wherein the representative self-organizing map corresponds to a neighborhood associated to the premises.

8. The computer-implemented method of claim 1, wherein analyzing the energy usage data includes:

calculating, from the energy usage data, statistical data related to at least one of seasonality, trend, decomposed data, kurtosis, skewness, serial correlation, non-linear autocorrelation, long range dependence, or chaos;

applying the statistical data to a regression model, the regression model having a plurality of model coefficients corresponding to a predictor variable that represents if the energy usage data indicates that the energy usage adjusts in the pattern according to a time-of-day dependent setting, the applying being based on the received energy usage data;

scoring the premises using the model coefficients to determine whether the controller adjusts temperature according to the time-of-day dependent setting.

9. The computer-implemented method of claim 8, wherein the regression model includes minimizing a sum of squared error between a predicted and an actual value.

10. The computer-implemented method of claim 8 further comprising:

filtering, prior to applying the calculated statistical data to the regression model, a portion of the statistical data that is insufficiently predictive based upon radical change in the model coefficient.

11. The computer-implemented method of claim 10, wherein filtering is performed by at least one of a ridge estimator, a Lasso estimator, and a least angle regression estimator.

12. The computer-implemented method of claim 8, wherein the predictor variable includes at least one parameter corresponding to seasonality, intercept, trend, decomposed data, kurtosis, skewness, serial correlation, non-linear autocorrelation, long range dependence, chaos, and sensitivity to heating and cooling.

13. A system comprising:

one or more processors; and

memory including instructions that, upon being executed by the one or more processors, cause the system to;

receive energy usage data of a consumer associated to a premises, the premises having a controller to regulate temperature in a controlled space within the premises, the energy usage data being associated, in part, to regulation of the controller;

analyze the energy usage data to determine whether the controller is a programmed thermostat or an unprogrammed thermostat by determining, via the processor, whether the energy usage data indicates that the energy usage adjusts in a pattern according to a time-of-day dependent setting; and

generating, via the processor, a status that indicates the controller as the programmed thermostat or the unprogrammed thermostat based at least in part on the analysis.

14. The system of claim 13, wherein the memory further including instructions that, upon being executed by the one or more processors, cause the system to generate and provide tailored consumer information to a consumer associated with the premises based on the determined status of the controller, wherein the tailored information represents at least a usage of energy at the premises.

15. The system of claim 13, wherein the instructions upon being executed further cause the system to:

calculate a ratio of usage in an evening time period and usage in a morning time period; and

compare the calculated ratio to a user-specific threshold, wherein the comparison provides the status of whether the controller adjusts the temperature according to the time-of-day dependent setting.

16. The system of claim 13, wherein the instructions upon being executed further cause the system to:

compare a first statistical feature derived from the received energy usage data to a second statistical feature derived from at least one representative pattern of a premises with the controller.

17. The system of claim 13, wherein the energy usage data is retrieved from at least one of an advanced meter infrastructure (AMI) meter and a home display system located at the first premises.

18. The system of claim 13, wherein the memory further comprising instructions that when executed by the one or more processors cause the processors to analyze the energy usage data by:

calculating, from the energy usage data, statistical data related to at least one of seasonality, trend, decomposed data, kurtosis, skewness, serial correlation, non-linear autocorrelation, long range dependence, or chaos; and

assigning the calculated statistical data to a representative self-organized map, a portion of the representative self-organized map having a cluster having an associated regression model.

19. The system of claim 18, wherein the representative self-organizing map includes at least one of a Kohonen supervised self-organizing maps or a Kohonen unsupervised self-organizing map.

20. The system of claim 18, wherein the representative self-organizing map corresponds to a neighborhood associated to the first premises.

21. A computer-implemented method of remotely ascertaining presence of a programmed thermostat controlling an environmental conditioner at a premises, the method comprising:

fetching, by at least a processor using a network communication, energy usage data from a device located at the premises, wherein the energy usage data includes data representing usage of energy over a plurality of time periods;

performing, by at least the processor, a statistical analysis of the energy usage data including analyzing the energy usage data for an indication of a repeated temporal variation in the usage of the energy based on a time periodic pattern of the usage;

if the indication of the repeated temporal variation is found, generating information indicating that the premises includes a programmed thermostat for controlling the environmental conditioner of the premises; and

if the indication of the repeated temporal variation is not found, generating information indicating that the premises includes an unprogrammed thermostat for controlling the environmental conditioner of the premises.

22. The method according to claim 21, further comprising, if the stored information indicates the unprogrammed thermostat is present:

generating a message recommending usage of a programmed thermostat; and

sending the message to a consumer associated with the premises.

23. The method according to claim 21, further comprising,

if the stored information indicates the programmed thermostat is present:

further analyzing the energy usage data for an indication of excessive energy usage by the environmental conditioner;

if the indication of excessive energy usage is found, generating a message recommending a course of action for reducing energy usage; and

sending the message to a consumer associate with the premises.