Automated data alignment based upon indirect device relationships

• US 8,024,390 B2
• Filed: 10/31/2007
• Issued: 09/20/2011
• Est. Priority Date: 07/01/2005
• Status: Active Grant

• Current Assignee: Schneider Electric USA, Inc. (Schneider Electric SE)
• Original Assignee: Schneider Electric USA, Inc. (Schneider Electric SE)
• Inventors: Bickel, Jon A.
• Primary Examiner(s): Mai; Tan V
• Application Number:
  US11/981,428
• Publication Number:
  US 20080065712A1
• Time in Patent Office: 1,420 Days
• Field of Search: 708/422
• US Class Current: 708/422
• CPC Class Codes: G01D 4/002 Remote reading of utility meters
• CPC Class Codes: H02J 13/00002 characterised by monitoring
• 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: Y02E 60/00 Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
• CPC Class Codes: Y04S 10/30 State monitoring, e.g. fault, temperature monitoring, insulator monitoring, corona discharge
• CPC Class Codes: Y04S 20/30 Smart metering, e.g. specially adapted for remote reading

Abstract

A noisy data alignment algorithm for determining cycle count offsets for noisy pairs of n monitoring devices. A direct cycle count offset matrix is determined based upon the highest correlation coefficients produced by correlating frequency variation data from each device pair D_ij. For each direct cycle count offset M_ij, indirect cycle count offsets are calculated as a function of at least M_k, where k≠i≠j, to produce indirect cycle count offsets. The statistical mode of these indirect offsets is compared with the corresponding M_ij in the matrix. When they differ, M_ij in the direct matrix is adjusted to be equal to the statistical mode. All indirect cycle count offsets for all other unique device pairs, M_ij, are calculated to iterate to a single solution in which all indirect cycle count offsets are equal to the corresponding direct cycle count offset. An optional verification algorithm is also provided.

Claims

1. A method of automatically aligning data measured by a number, n, of monitoring devices in a power monitoring system, comprising:

receiving from each of said monitoring devices respective signal data representing at least frequency variations measured by respective ones of said monitoring devices, said monitoring devices including a reference monitoring device and a second monitoring device;

correlating said signal data from said reference monitoring device with said signal data from said second monitoring device to determine a reference cycle count, M_i, associated with said reference monitoring device corresponding to a maximum correlation coefficient and a second cycle count, M_j, associated with said second monitoring device corresponding to said maximum correlation coefficient;

automatically calculating a direct cycle count offset, M_ij, as a function of a difference between said reference cycle count, M_i, and said second cycle count, M_j, and storing said direct cycle count offset in a direct cycle count offset matrix; and

automatically calculating an indirect cycle count offset as a function of at least M_k, where k≠

i≠

j and 1≦

k≦

n, said M_k being a cycle count associated with a monitoring device of said n monitoring devices other than said reference monitoring device and other than said second monitoring device.

2. The method of claim 1, wherein said automatically calculating said indirect cycle count offset is carried out with respect to at least two other of said n monitoring devices except said reference monitoring device and except said second monitoring device to produce at least two indirect cycle count offsets including said indirect cycle count offset.

3. The method of claim 2, further comprising determining which of said at least 2 indirect cycle count offsets differs from said direct cycle count offset.

4. The method of claim 2, further comprising:

determining a statistical mode from a set comprised of said at least 2 indirect cycle count offsets to produce a mode value; and

determining whether the mode value differs from said direct cycle count offset.

5. The method of claim 1, further comprising, responsive to said direct cycle count offset differing from said indirect cycle count offset, producing a modified direct cycle count offset equal to said indirect cycle count offset.

6. The method of claim 5, further comprising storing an indication that said signal data associated with said reference monitoring device and said signal data associated with said second monitoring device are aligned.

7. The method of claim 5, further comprising communicating said modified direct cycle count offset to said reference monitoring device or to said second monitoring device to cause said reference monitoring device or said second monitoring device to adjust a cycle counter in said reference monitoring device or in said second monitoring device by a value corresponding to said modified direct cycle count offset.

8. The method of claim 1, wherein said indirect cycle count offset is a function of at least M_ik−

M_jk.

9. The method of claim 1, wherein said indirect cycle count offset is a function of at least M_m, where m≠

i≠

j and 1≦

m≦

n, said M_m being a cycle count associated with a monitoring device of said n monitoring devices other than said reference monitoring device and other than said second monitoring device and other than said monitoring device associated with M_k.

10. The method of claim 1, further comprising

determining a first verification cycle count, M_i′

, associated with said reference monitoring device corresponding to a first correlation coefficient that is less than said maximum correlation coefficient;

determining a second verification cycle count, M_j′

, associated with said second monitoring device corresponding to said first correlation coefficient;

automatically calculating a verification cycle count offset, M_ij′

, based upon the difference between said first verification cycle count, M_i′

, and said second verification cycle count, M_j′

, and storing said verification cycle count offset; and

responsive to said verification cycle count offset equaling said indirect cycle count offset, storing an indication of a level of confidence in said indirect cycle count offset based upon said first correlation coefficient.

11. The method of claim 10, wherein said first correlation coefficient is above a predetermined threshold below said maximum correlation coefficient.

12. The method of claim 1, wherein said direct cycle count offset matrix is an (n×

n) skew-symmetric matrix.

13. The method of claim 1, wherein said frequency variations represented by said first signal data are variations in fundamental frequency or variations in harmonic frequency, wherein said variations are associated with a voltage or a current.

14. The method of claim 1, wherein each of said n monitoring devices includes a cycle counter, the method further comprising communicating simultaneously a signal to said n monitoring devices to reset their respective cycle counters.

15. A method of automatically aligning data measured by a number, n, of monitoring devices in a power, monitoring system, comprising:

receiving from each of said monitoring devices respective signal data, S_n, representing at least frequency variations measured by respective ones of said monitoring devices;

for each of a plurality of device pairs within a set comprising said n number of monitoring devices, wherein each device pair is termed D_ij, where i≠

j, where 1≦

i≦

n, and where 1≦

j≦

n, correlating said signal data S_i with said signal data S_j for each of said device pairs, D_ij, to determine respective cycle counts, M_i and M_j, associated with D_i and D_j, respectively, said cycle counts corresponding to a maximum correlation coefficient produced by said correlating;

for each of said device pairs, D_ij, automatically calculating a direct cycle count offset, M_ij=M_i−

M_j, and storing said direct cycle count offset in a direct cycle count offset matrix;

for each of said direct cycle count offsets, M_ij, automatically calculating a plurality of indirect cycle count offsets as a function of at least M_k, said M_k being a cycle count associated with a monitoring device of said n monitoring devices other than said reference monitoring device and other than said second monitoring device, where 1≦

k≦

n, and k≠

i≠

j;

storing said plurality of indirect cycle count offsets for each M_ij in an indirect cycle count offset matrix;

for each of said direct cycle count offsets, M_ij, in said direct cycle count offset matrix, determining whether M_ij is not equal to at least one of said plurality of indirect cycle count offsets, and, responsive thereto, storing an indication that M_ij in said direct cycle count offset matrix is equal to a value corresponding to said at least one of said indirect cycle count offsets.

16. The method of claim 15, further comprising communicating an indication of said value to D_i or D_j to cause D_i or D_j to adjust a cycle counter by an amount corresponding to said value.

17. The method of claim 15, further comprising communicating a signal to each of said n monitoring devices to reset respective cycle counters in each of said n monitoring devices at substantially the same time.

18. The method of claim 15, further comprising:

determining a statistical mode from a set comprised of said indirect cycle count offsets for M_ij to produce a mode value; and

determining whether said mode value equals M_ij in said direct cycle count offset matrix.

19. The method of claim 15, wherein said correlating produces a plurality of correlation coefficients including said maximum correlation coefficient, the method further comprising:

for said devices D_i and D_j corresponding to said modified M_ij, determining a plurality of probable cycle count offsets associated with a predetermined number of said plurality of correlation coefficients that are less than said maximum correlation coefficient; and

responsive to one of said plurality of said probable cycle count offsets equaling said modified M_ij, storing an indication of a level of confidence in said modified M_ij commensurate with how proximal a correlation coefficient of said plurality of said correlation coefficients corresponding to said one of said plurality of said probable cycle count offsets is to said maximum correlation coefficient.

20. The method of claim 15, further comprising:

receiving from at least one of said monitoring devices second signal data representing at least frequency variations measured by said at least one of said monitoring devices; and

adjusting a cycle count associated with said at least one of said monitoring devices by said value corresponding to said at least one of said indirect cycle count offsets.

21. A method of automatically aligning data monitored by a number, n, of monitoring devices, D, associated with a power monitoring system, comprising:

automatically correlating respective signal data from a pair of said monitoring devices, D_i and D_j, to determine a maximum correlation coefficient associated with respective cycle counts for said pair of monitoring devices, D_ij, said respective signal data representing frequency variations monitored by said D_i and said D_j, respectively;

determining a first cycle count offset, M_ij, by calculating a difference between said respective cycle counts;

determining a second cycle count offset from a function that includes at least a cycle count associated with a monitoring device of said monitoring devices other than D_i or D_j; and

storing an indication that said first cycle count offset is equal to a value corresponding to said second cycle count offset.

22. The method of claim 21, further comprising communicating an instruction to said D_i or to said D_j to adjust a cycle counter, which stores said respective cycle counts, by an amount corresponding to said value.

23. The method of claim 21, wherein said automatically correlating includes determining whether said second cycle count offset equals a cycle count offset associated with a correlation coefficient that is less than said maximum correlation coefficient.