الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 16 |  |  |
| | | from | 16 | | |
Abstract
The ability of the operator to successfully respond to normal or accident conditions is highy sensitive to the amount and quality of information he can obtain concerning the state of the plant This information can only be provided to the operator only through the plant instrumentation. Th< objective of this study is to present an applied statistical method known as Chemometrics, whict focuses on the mathematics and statistics parts applicable to the problem, to get solution of the problems of the:-
•
• 1 determining the control rod position in pressurized water reactors (PWR),
•2 prediction of neutron flux map (”Neutron Image) ,
• 3 Core monitoring ,
• 4 xenon oscillation
and
• 5 prediction of DNBR .
In this study we used multivariate statistics methods such as. Principle Component Analysis (PCA). Partial Least Square (PLS) and Cluster Analysis (CA). The methods are based on the influence of partially inserted control rod on the radial thermal flux at the top of the core surface at different states of pressure , temperature, density and boron concentration and by real operation of PWR and with data from various references. Since the direct visual analysis of huge data produced was not possible, this highlights the need to use methods to reduce the dimensionality of these data. Corresponding to various known rod positions a large and corrected data can be generated by an advanced core calculation code. In this study, the 3-D TRITON code was used with a benchmark data for a PWR-2775-MWth. Through this study, it is clear that pattern recognition using the multivariate methods is more visual, faster, easy, and more accurate. This proved that Chemometrics is more interactive than classical methods.