Apparatus and method for determining days of the week with similar utility consumption profiles

• US 20030101009A1
• Filed: 10/30/2001
• Published: 05/29/2003
• Est. Priority Date: 10/30/2001
• Status: Abandoned Application

• Current Assignee: Johnson Controls Technology Company (Johnson Controls International plc)
• Original Assignee: Johnson Controls Technology Company (Johnson Controls International plc)
• Inventors: Seem, John E.
• Application Number:
  US10/021,382
• Publication Number:
  US 20030101009A1
• Time in Patent Office: Days
• US Class Current: 702/61
• CPC Class Codes: H02J 2203/20 Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
• 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: Y02E 60/00 Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
• 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 40/20 Information technology specific aspects, e.g. CAD, simulation, modelling, system security

Abstract

A method and apparatus for determining days of the week with similar consumption of energy or utility by a computerized system utilizes a pattern recognition algorithm. The algorithm utilizes a time series of energy or utility use data spanning a plurality of days to generate at least one feature of interest for each day. The features of interest may be any or all of average daily utility consumption, maximum utility use during a predefined time interval for the day, minimum utility use over the predefined time interval for the day, and the like. The algorithm transforms the at least one feature of interest for each day to remove the effects of any seasonal variation that may be present in the time series data. The features of interest are then grouped by day of the week to define seven clusters. The algorithm next performs an outlier analysis for each feature of interest in each of the seven clusters to identify and remove any abnormal data. The seven clusters are then analyzed using an modified agglomerative hierarchical clustering method to determine days of the week with similar utility consumption profiles.

Claims

1. A method for determining days of the week with similar consumption of a utility by a computerized system, comprising:

gathering data representative of utility consumption for a plurality of days; and

analyzing the data to determine days of the week having similar utility consumption profiles.

2. The method of claim 1, wherein the gathering step includes measuring the utility consumption using at least one sensor.

3. The method of claim 2, wherein the measuring step includes positioning the at least one sensor to monitor at least one of an electric meter, a gas meter, a water meter and a supply connection to a specific piece of equipment.

4. The method of claim 1, wherein the utility is at least one of electricity, gas and water.

5. The method of claim 1, wherein the gathering step includes reading the data from a database contained within a memory of the computerized system.

6. The method of claim 1, wherein the gathering step includes receiving the data via a communication link.

7. The method of claim 6, wherein the communication link is at least one of a direct cable connection, a local area network, a wide area network, and a telephone line.

8. The method of claim 1, wherein the data is a time series of building utility consumption data.

9. The method of claim 8, wherein the building utility consumption data is at least one of electricity use, natural gas consumption, district heating consumption, cooling requirements, and heating requirements.

10. The method claim 1, wherein the gathering step includes determining at least one predefined utility consumption feature for each day.

11. The method of claim 10, wherein the at least one predefined utility consumption feature for each day is selected from average daily utility consumption, peak utility use during a predefined time period for the day, and minimum utility use during a predefined time period for the day.

12. The method of claim 1, wherein the gathering step includes transforming the data to remove effects of seasonal change.

13. The method of claim 12, wherein the transforming step includes determining the difference between the data for a day and a one-week period of surrounding data.

14. The method of claim 13, wherein the determining step utilizes the following equation:

15. The method of claim 1, wherein the gathering step includes identifying and removing abnormal utility consumption data.

16. The method of claim 15, wherein the identifying and removing step includes performing an outlier analysis on the data.

17. The method of claim 16, wherein the data includes N utility consumption features for each day, and the outlier analysis is performed N times for each day of the week.

18. The method of claim 16, wherein the outlier analysis is conducted using a Generalized Extreme Studentized Deviate (GESD) statistical procedure.

19. The method of claim 18, wherein the GESD statistical procedure utilizes two user selected parameters comprising:

a probability (α

) of incorrectly declaring one or more outliers when no outliers exist; and

an upper bound (n_u) on the number of potential outliers.

20. The method of claim 19, wherein the outliers are determined with α

=0.1 and with n_u set to the largest integer that satisfies the following inequality;

n_u≦

0.5(n−

1).

21. The method of claim 1, wherein the analyzing step utilizes a clustering algorithm.

22. The method of claim 21, wherein the clustering algorithm comprises a form of an agglomerative hierarchical clustering method.

23. The method of claim 21, wherein the clustering algorithm commences by defining each day of the week as a separate cluster.

24. The method of claim 23, wherein the clustering algorithm continues by determining a measure of dissimilarity between each pair of clusters.

25. The method of claim 24, wherein the measure of dissimilarity between any two clusters is a dissimilarity coefficient.

26. The method of claim 25, further including defining the dissimilarity coefficient between the two clusters as the average distance between every pair of observations in the clusters, where one observation of the pair belongs to one cluster and the other observation belongs to the other cluster.

27. The method of claim 26, wherein the dissimilarity coefficient between each pair of clusters C_i and C_j is determined from:

28. The method of claim 27, wherein the dissimilarity measure between observations x and y is the Euclidean distance:

d(x,y)={square root}{square root over ((x₁−

y₁)²+(x₂−

y₂)²+ . . . +(x_p−

y_p)²)}

where x_i is the value of the i^th variable of observation x.

29. The method of claim 23, wherein the clustering algorithm continues by finding a nearest pair of clusters among all possible pairs of clusters.

30. The method of claim 29, wherein the finding step comprises identifying the nearest pair of clusters using a measurement of dissimilarity between all possible pairs of clusters.

31. The method of claim 29, wherein the clustering algorithm utilizes a stopping rule for determining a final number of clusters.

32. The method of claim 29, wherein the clustering algorithm utilizes a stopping rule for determining whether the nearest pair of clusters should be joined into a combined cluster.

33. The method of claim 32, wherein the stopping rule determines that the nearest pair of clusters C_i and C_j should be joined if the following inequality is satisfied:

34. The method of claim 32, further including, if the stopping rule determines that the nearest clusters should be joined, combining the nearest pair of clusters into a combined cluster.

35. The method of claim 34, further including, if the stopping rule determines that the nearest clusters should be joined, updating a dissimilarity coefficient between the combined cluster and each remaining cluster using the following updating equation:

36. The method of claim 32, further including, if the stopping rule determines that the nearest clusters should be joined and there are two clusters remaining, setting a number of day types to one.

37. The method of claim 32, further including, if the stopping rule determines that the nearest clusters should not be joined, setting a number of day types to the number of remaining clusters.

38. A method for determining days of the week with similar consumption of a utility by a computerized system, comprising:

receiving a time series of utility use data spanning a plurality of days;

generating at least one feature of interest for each day in the time series;

transforming the at least one feature of interest for each day to remove any seasonal variation present therein;

grouping the features of interest by day of the week to define seven clusters;

identifying and removing outliers from the seven clusters for each feature of interest; and

analyzing the seven clusters to determine days of the week with similar utility consumption profiles.

39. An apparatus for determining days of the week with similar consumption of a utility, comprising:

a processor running a program to perform the steps of;

gathering time series data representative of utility consumption for a plurality of days; and

analyzing the time series data to determine days of the week having similar utility consumption profiles.

40. The apparatus of claim 39, wherein the apparatus includes a memory containing the time series data, and to perform the gathering step, the program causes the processor to read the time series data from the memory.

41. The apparatus of claim 39, wherein the apparatus is connected by a communication link to a source containing the time series data, and to perform the gathering step, the program causes the processor to receive the time series data from the source.

42. The method claim 39, wherein to perform the gathering step, the program causes the processor to generate at least one feature of interest for each day in the time series data.

43. The apparatus of claim 39, wherein to perform the gathering step, the program causes the processor to transform the time series data to remove effects of seasonal change.

44. The apparatus of claim 43, wherein to remove the effects of seasonal change from the time series data, the program causes the processor to determine the difference between the data for a day and a one-week period of surrounding data.

45. The apparatus of claim 39, wherein to perform the gathering step, the program causes the processor to identify and remove abnormal utility consumption data.

46. The apparatus of claim 45, wherein to identify and remove abnormal utility consumption data, the program causes the processor to perform an outlier analysis.

47. The apparatus of claim 39, wherein to perform the analyzing step, the program causes the processor to perform a clustering algorithm.

48. The apparatus of claim 47, wherein to perform the clustering algorithm, the program causes the processor to perform the steps of:

defining the data for each day of the week as a separate cluster;

determining a measure of dissimilarity between each pair of clusters;

combining a nearest pair of clusters and updating the dissimilarity measures when a stopping rule indicates the nearest pair of clusters should be combined; and

terminating when the stopping rule indicates the nearest clusters should not be combined or the number of clusters equals one.

49. An apparatus for determining days of the week with similar consumption of a utility, comprising:

means for receiving a time series of utility use data spanning a plurality of days;

means for generating at least one feature of interest for each day in the time series;

means for transforming the at least one feature of interest for each day to remove any seasonal variation present therein;

means for grouping the features of interest by day of the week to define seven clusters;

means for identifying and removing outliers from the seven clusters for each feature of interest; and

means for analyzing the seven clusters to determine days of the week with similar utility consumption profiles.

50. The apparatus of claim 49, wherein the means for analyzing includes a processor running a clustering algorithm in the form of an agglomerative hierarchical clustering method.