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

Statistical methods to monitor the West Valley off-gas system

Conference ·
OSTI ID:5987358
The off-gas system for the ceramic melter operated at the West Valley Demonstration Project at West Valley, NY, is monitored during melter operation. A one-at-a-time method of monitoring the parameters of the off-gas system is not statistically sound. Therefore, multivariate statistical methods appropriate for the monitoring of many correlated parameters will be used. Monitoring a large number of parameters increases the probability of a false out-of-control signal. If the parameters being monitored are statistically independent, the control limits can be easily adjusted to obtain the desired probability of a false out-of-control signal. However, a high degree of correlation generally exists among the parameters being monitored in the off-gas system. This makes it very difficult to control the probability of false calls (saying the system is out-of-control when it is in-control or saying the system is in-control when it is actually out-of-control). The interpretation of the individual control charts is difficult in the presence of correlation among the variables. When a high degree of correlation exists, variable reduction techniques can be used to reduce the number of parameters. Principal components have been used as a variable reduction technique. The principal component (PC) scores have desirable statistical properties when the original variables are distributed as multivariate normals. Two statistics derived from the PC scores and used to form multivariate control charts are outlined and their distributional properties reviewed. 2 refs., 2 figs.
Research Organization:
Pacific Northwest Lab., Richland, WA (USA)
Sponsoring Organization:
DOE; USDOE, Washington, DC (USA)
DOE Contract Number:
AC06-76RL01830
OSTI ID:
5987358
Report Number(s):
PNL-SA-17823; CONF-900977--9; ON: DE91011336
Country of Publication:
United States
Language:
English