Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Spectrum-transformed sequential testing method for signal validation applications

Conference ·
OSTI ID:5348367
 [1];  [2]
  1. Argonne National Lab., IL (United States)
  2. Northwestern Univ., Evanston, IL (United States). Dept. of Industrial Engineering and Management Sciences

The Sequential Probability Ratio Test (SPRT) has proven to be a valuable tool in a variety of reactor applications for signal validation and for sensor and equipment operability surveillance. One drawback of the conventional SPRT method is that its domain of application is limited to signals that are contaminated by gaussian white noise. Nongaussian process variables contaminated by serial correlation can produce higher-than-specified rates of false alarms and missed alarms for SPRT-based surveillance systems. To overcome this difficulty we present here the development and computer implementation of a new technique, the spectrum-transformed sequential testing method. This method retains the excellent surveillance advantage of the SPRT (extremely high sensitivity for very early annunciation of the onset of disturbances in monitored signals), and its false-alarm and missed-alarm probabilities are unaffected by the presence of serial correlation in the data. Example applications of the new method to serially-correlated reactor variables are demonstrated using data recorded from EBR-II.

Research Organization:
Argonne National Lab., IL (United States)
Sponsoring Organization:
DOE; USDOE, Washington, DC (United States)
DOE Contract Number:
W-31109-ENG-38
OSTI ID:
5348367
Report Number(s):
ANL/CP-75528; CONF-920538--25; ON: DE92014771
Country of Publication:
United States
Language:
English