Data processing: measuring – calibrating – or testing – Measurement system – Measured signal processing
Reexamination Certificate
2000-04-24
2003-06-24
Barlow, John (Department: 2863)
Data processing: measuring, calibrating, or testing
Measurement system
Measured signal processing
Reexamination Certificate
active
06584434
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to data filtering and anomaly detection, and more particularly to improved change-detect data compressing using a rolling average of the data as a low pass filter and mode based statistical process control for anomaly detection.
Optimal operational characteristics for modern gas turbine systems include high operational efficiency, low exhaust and long operational life. To obtain these operational characteristics, monitoring the operational parameters of the gas turbine system becomes desirable. When monitoring the operational parameters of the gas turbine system, data relating to the physical and operational conditions of the gas turbine system are collected and analyzed. The data are collected from a large number of locations on, in or near the gas turbine system to accurately assess the operational characteristics of the gas turbine system. The data relating to the operational parameters are particularly meaningful when the data are collected at high frequencies (i.e., one data point every one or two seconds) and when the collected data are compared to historical data that has been archived and collected over a large temporal range (i.e., days, months or years).
Collecting data from a large number of locations at a high frequency presents many problems. For example, the total amount of data collected are very large. When the gas turbine system is located at a remote location, local archiving of the large amount of collected data becomes problematic. As such, the large amount of collected data typically requires expensive storage devices for proper data archiving. In addition, transmitting the large amount of collected data from the remote location to a central location requires a long transmission time. Therefore, the costs related to transmission of the data are high. Thus, it is desired to filter the data before archiving at the remote site and transmitting to a central location while maintaining the statistical and informational integrity of the total amount of collected data.
With the large amount of data collected from the number of locations, interpretation of the collected data also becomes difficult. Typically, the data are analyzed to determine the overall operational characteristics of the gas turbine system. When assessing the overall condition of the gas turbine system, pinpointing the exact problem involves laborious troubleshooting. As such, the large amount of data from different locations becomes meaningless unless the data are correlated to an operational condition of the gas turbine system. Therefore, it is desired that the collected data be sorted and assessed to accurately pinpoint any potential problems relating to the operational conditions of the gas turbine system without the need for undue troubleshooting.
BRIEF SUMMARY OF THE INVENTION
A method is disclosed for filtering and determining anomalies of corrected data from a system under test. The method comprises buffering the data from the system under test. Rolling averages of the buffered data are calculated wherein the calculation of the rolling averages low pass filters the buffered data. Change-detect compression is performed on the rolling averaged data, and the compressed data are archived. The archived data are transmitted to a central location, and the transmitted data are received at the central location. The received data are archived at the central location. The archived data are gathered at the central location. The gathered data are filtered into at least one subset that is differentiated by mode. The at least one subset is corrected, and distributive statistics are calculated on the at least one subset to identify long-term anomalies in the at least one subset.
REFERENCES:
patent: 3958187 (1976-05-01), Suzuki et al.
patent: 4101072 (1978-07-01), Weaver et al.
patent: 4143417 (1979-03-01), Wald et al.
patent: 5239456 (1993-08-01), Badavas et al.
patent: 5392226 (1995-02-01), Hamilton
patent: 5696700 (1997-12-01), O'Brien, Jr. et al.
patent: 5737399 (1998-04-01), Witzmann et al.
patent: 5764538 (1998-06-01), Cheiky-Zelina
patent: 5777900 (1998-07-01), Takeuchi et al.
patent: 5870699 (1999-02-01), Canada et al.
patent: 5937100 (1999-08-01), Kitajima
patent: 5949953 (1999-09-01), Shirakawa et al.
patent: 5978525 (1999-11-01), Shu et al.
patent: 5995493 (1999-11-01), Van de Kerkhof
patent: 6065094 (2000-05-01), Akiyama
patent: 6069309 (2000-05-01), Monahan et al.
patent: 6078874 (2000-06-01), Piety et al.
patent: 6084911 (2000-07-01), Ishikawa
patent: 6122605 (2000-09-01), Drees et al.
patent: 6205499 (2001-03-01), Houlberg et al.
patent: 6211814 (2001-04-01), Benjamin et al.
patent: 6272446 (2001-08-01), Baekke et al.
patent: 6434196 (2002-08-01), Sethuraman et al.
patent: 2002/0110275 (2002-08-01), Rogina et al.
patent: 0 526 885 (1992-05-01), None
Catharine Douglas Ancona
Sanborn Stephen Duane
Schick Louis Andrew
Barlow John
Bhat Aditya
General Electric Company
Patnode Patrick K.
Thompson John F.
LandOfFree
Method for data filtering and anomoly detection does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for data filtering and anomoly detection, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for data filtering and anomoly detection will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3133614