METHOD AND SYSTEM FOR MONITORING APPARATUS

• US 20080161946A1
• Filed: 09/18/2007
• Published: 07/03/2008
• Est. Priority Date: 12/28/2006
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Hatano, Hisaaki, Kubota, Kazuto
• Granted Patent: US 7,653,456 B2
• Time in Patent Office: Days
• US Class Current: 700/90
• CPC Class Codes: H04Q 9/00 Arrangements in telecontrol or telemetry systems for selectively calling a substation from a main station, in which substation desired apparatus is selected for applying a control signal thereto or for obtaining measured values therefrom

Abstract

There is provided with a monitoring method of monitoring plural apparatuses each having a sensor, including: selecting two or more sample apparatuses out of the plural apparatuses;

• • calculating averages and standard deviations of sensor values detected by the sample apparatuses for each sample apparatus; calculating a confidence interval of the averages and an average of the standard deviations; and calculating an average of sensor values detected by an arbitrary apparatus among the plural apparatuses as an object average and calculating a normal range of the sensor of the arbitrary apparatus from the object average, the confidence interval, and the average of the standard deviations.

Claims

1. A monitoring method of monitoring plural apparatuses each having a sensor, comprising:

selecting two or more sample apparatuses out of the plural apparatuses;

calculating averages and standard deviations of sensor values detected by the sample apparatuses for each sample apparatus;

calculating a confidence interval of the averages and an average of the standard deviations; and

calculating an average of sensor values detected by an arbitrary apparatus among the plural apparatuses as an object average and calculating a normal range of the sensor of the arbitrary apparatus from the object average, the confidence interval, and the average of the standard deviations.

2. The method according to claim 1, wherein

when the object average is included in the confidence interval, a normal range of the sensor is calculated from the object average and the average of the standard deviations and

when the object average is not included in the confidence interval, a normal range of the sensor is calculated from a value within the confidence interval closest to the object average and the average of the standard deviations.

3. The method according to claim 2 wherein as the normal range, a range of ±

(a predetermined coefficient×

the average of the standard deviations) with respect to the object average or the value within the confidence interval closest to the average calculated.

4. The method according to claim 1, further comprising,

calculating, when sensor values in a number designated in advance are collected in the arbitrary apparatus or when a predetermined period designated in advance has elapsed, an average and a standard deviation of the sensor values detected by the arbitrary apparatus, and

updating, by using the average and the standard deviation calculated for the arbitrary apparatus, the confidence interval and the average of the standard deviations.

5. The method according to claim 1, further comprising:

calculating desired statistics from the sensor values of each sample apparatus for each of the sample apparatuses;

clustering the sample apparatuses and divide the sample apparatuses into plural groups according to the statistics calculated for the sample apparatuses;

generating, based on plural attribute values of each of the sample apparatuses and the groups to which the sample apparatuses belong, an inference rule for inferring a group from one or more attribute values; and

inferring, from the inference rule, a group to which the arbitrary apparatus belongs, wherein

for each of the groups, the confidence interval and the average of the standard deviations are calculated, and

the normal range of the sensor of the arbitrary apparatus is calculated from the sensor values detected from the arbitrary apparatus and the confidence interval and the average of the standard deviations obtained from the group to which the arbitrary apparatus belongs.

6. A monitoring method that monitors plural apparatuses each having a sensor, comprising:

selecting two or more sample apparatuses out of the plural apparatuses;

calculating, by sensor values detected from the sample apparatuses,

a sample average obtained by averaging averages of sensor values of each sample apparatus,

a sample standard deviation that is a standard deviation of the averages of the sensor values of each sample apparatus, and

a standard deviation average obtained by averaging standard deviations of the sensor values of each sample apparatus; and

calculating a mean of a normal distribution of averages of sensor values of an arbitrary apparatus among the plural apparatuses or a value within a threshold range with respect to the mean by using a number and an average of the sensor values detected from the arbitrary apparatus, the sample average, the sample standard deviation, and the standard deviation average, and calculating a normal range of the sensor of the arbitrary apparatus from the mean or the value within the threshold range and the standard deviation average.

7. The method according to claim 6, wherein averages of the sensor values in each sample apparatus and standard deviations of the sensor values in each sample apparatus are calculated;

and

the sample average, the sample standard deviation, and the standard deviation average are calculated from the averages and the standard deviations calculated for each sample apparatus.

8. The method according to claim 6, wherein

a range of ±

(a predetermined coefficient×

the standard deviation average) with respect to the mean or the value within the threshold range is calculated as the normal range.

9. The method according to claim 6, wherein the normal range is calculated every time a predetermined number of sensor values are detected from the arbitrary apparatus,

sensor values not included in the normal range are detected from the sensor values used for calculation of the normal range, and

the sensor values detected are not used for calculation of a normal range in next and subsequent times.

10. The method according to claim 7, further comprising;

calculating, when sensor values in a number designated in advance are collected in the arbitrary apparatus or when a predetermined period designated in advance has elapsed, an averages and a standard deviation of the sensor values detected by the arbitrary apparatus, and

updating the sample average, the sample standard deviation, and the standard deviation average using the average and the standard deviation calculated for the arbitrary apparatus.

11. The method according to claim 6, further comprising:

calculating desired statistics from the sensor values of each sample apparatus for each of the sample apparatuses;

clustering the sample apparatuses and divide the sample apparatuses into plural groups according to the statistics calculated for the sample apparatuses;

generating, based on plural attribute values of each of the sample apparatuses and the groups to which the sample apparatuses belong, an inference rule for inferring a group from one or more attribute values; and

inferring, from the inference rule, a group to which the arbitrary apparatus belongs, wherein

for each of the groups, the sample average, the sample standard deviation, and the standard deviation average are calculated, and

the mean or the value within the threshold range with respect to the mean is calculated from the number and the average of the sensor values detected from the arbitrary apparatus, the sample average, the sample standard deviation and the standard deviation average obtained from the group to which the arbitrary apparatus belongs.

12. A monitoring system that monitors plural apparatuses each having a sensor, comprising:

a selecting unit configured to select two or more sample apparatuses out of the plural apparatuses;

a first calculating unit configured to calculate averages and standard deviations of sensor values detected by the sample apparatuses for each sample apparatus;

a second calculating unit configured to calculate a confidence interval of the averages and an average of the standard deviations; and

a normal-range calculating unit configured to calculate an average of sensor values detected by an arbitrary apparatus among the plural apparatuses as an object average and calculate a normal range of the sensor of the arbitrary apparatus from the object average, the confidence interval, and the average of the standard deviations.

13. The system according to claim 12, wherein the normal-range calculating unit calculates, when the object average is included in the confidence interval, a normal range of the sensor from the object average and the average of the standard deviations and calculates, when the object average is not included in the confidence interval, a normal range of the sensor from a value within the confidence interval closest to the object average and the average of the standard deviations.

14. The system according to claim 13, wherein the normal-range calculating unit calculates, as the normal range, a range of ±

(a predetermined coefficient×

the average of the standard deviations) with respect to the object average or the value within the confidence interval closest to the object average.

15. The system according to claim 12, wherein

the first calculating unit calculates, when sensor values in a number designated in advance are collected in the arbitrary apparatus or when a predetermined period designated in advance has elapsed, an average and a standard deviation of the sensor values detected by the arbitrary apparatus, and

the second calculating unit updates, using the average and the standard deviation calculated for the arbitrary apparatus, the confidence interval and the average of the standard deviations.

16. The system according to claim 12, further comprising:

a statistics calculating unit configured to calculate desired statistics from the sensor values of each sample apparatus for each of the sample apparatuses;

a clustering unit configured to cluster the sample apparatuses and divide the sample apparatuses into plural groups according to the statistics calculated for the sample apparatuses;

a rule generating unit configured to generate, based on plural attribute values of each of the sample apparatuses and the groups to which the sample apparatuses belong, an inference rule for inferring a group from one or more attribute values; and

an inferring unit configured to infer, from the inference rule, a group to which the arbitrary apparatus belongs, wherein

the second calculating unit calculates, for each of the groups, the confidence interval and the average of the standard deviations, and

the normal-range calculating unit calculates the normal range of the sensor of the arbitrary apparatus from the sensor values detected from the arbitrary apparatus and the confidence interval and the average of the standard deviations obtained from the group to which the arbitrary apparatus belongs.

17. A monitoring system that monitors plural apparatuses each having a sensor, comprising:

a selecting unit configured to select two or more sample apparatuses out of the plural apparatuses;

a third calculating unit configured to calculate, using sensor values detected from the sample apparatuses,

a sample average obtained by averaging averages of sensor values of each sample apparatus,

a sample standard deviation that is a standard deviation of the averages of the sensor values of each sample apparatus, and

a standard deviation average obtained by averaging standard deviations of the sensor values of each sample apparatus; and

a normal-range calculating unit configured to calculate a mean of a normal distribution of averages of sensor values of an arbitrary apparatus among the plural apparatuses or a value within a threshold range with respect to the mean by using a number and an average of the sensor values detected from the arbitrary apparatus and the sample average, the sample standard deviation, and the standard deviation average calculated in the third calculating unit, and calculate a normal range of the sensor of the arbitrary apparatus from the mean or the value within the threshold range and the standard deviation average.

18. The system according to claim 17, wherein the third calculating unit includes:

a first unit configured to calculate averages of the sensor values in each sample apparatus and standard deviations of the sensor values in each sample apparatus; and

a second unit configured to calculate the sample average, the sample standard deviation, and the standard deviation average from the averages and the standard deviations calculated for each sample apparatus.

19. The system according to claim 17, wherein

the normal-range calculating unit calculates, as the normal range, a range of ±

(a predetermined coefficient×

the standard deviation average) with respect to the mean or the value within the threshold range.

20. The system according to claim 17, wherein the normal-range calculating unit

calculates the normal range every time a predetermined number of sensor values are detected from the arbitrary apparatus,

detects sensor values not included in the normal range from among the sensor values used for calculation of the normal range, and

does not use the sensor values detected for calculation of a normal range in next and subsequent times.

21. The system according to claim 18, wherein

the first unit in the third calculating unit calculates, when sensor values in a number designated in advance are collected in the arbitrary apparatus or when a predetermined period designated in advance has elapsed, an averages and a standard deviation of the sensor values detected by the arbitrary apparatus, and

the second unit in the third calculating unit updates the sample average, the sample standard deviation, and the standard deviation average using the average and the standard deviation calculated for the arbitrary apparatus.

22. The system according to claim 17, further comprising:

a statistics calculating unit configured to calculate desired statistics from the sensor values of each sample apparatus for each of the sample apparatuses;

a clustering unit configured to cluster the sample apparatuses and divide the sample apparatuses into plural groups according to the statistics calculated for the sample apparatuses;

a rule generating unit configured to generate, based on plural attribute values of each of the sample apparatuses and the groups to which the sample apparatus belong, an inference rule for inferring a group from one or more attribute values; and

an inferring unit configured to infer, from the inference rule, a group to which the arbitrary apparatus belongs, wherein

the third calculating unit calculates, for each of the groups, the sample average, the sample standard deviation, and the standard deviation average, and

the normal-range calculating unit calculates the mean or the value within the threshold range with respect to the mean from the number and the average of the sensor values detected from the arbitrary apparatus, the sample average, the sample standard deviation and the standard deviation average obtained from the group to which the arbitrary apparatus belongs.