Author: Luv Sharma

Publisher:

Published: 2007

Total Pages: 262

ISBN-13:

Abstract: A study of a Boiling Water Reactor (BWR) dynamics is presented with the objective of determining the attractors, domains of attraction and change in system stability with fluctuations in the system parameters. A reduced order model of the system was used for the investigations. The cell to cell mapping technique (CCMT) is used to determine the attractors and the domains of attraction for the system. The CCMT is a numerical technique for the global analysis of non-linear dynamics of systems and models system evolution as a Markov chain in time. The probabilistic modeling of the system dynamics and no differentiability requirements on the governing equations makes CCMT naturally suited for analysis of systems with stochastic parameters. This method, however, runs into performance problems with increasing number of state variables (degrees of freedom) of the system under investigation. Two methods are proposed to improve the performance of the CCMT for higher order systems. One of them is based on choosing different mapping time steps for different initial conditions for the system. The second method restricts the source cell region and gives a conditional probability of the system location. MATLAB v7.0 was used to set up the system and the simulations were run using the Ohio Supercomputing Center's distributed computing facilities. It is shown that these proposed methods lead to significant reduction in computational time compared to conventional CCMT. The results obtained are also generally better compared to the conventional approach. Using these proposed techniques it is determined that the BWR system has only one attractor and the only way to keep the reactor stable is to control the values of the system parameters. The parametric analysis was conducted using Taguchi methods, which are based on design of experiment techniques and the application of the signal-to-noise ratios. The experiment design was done using MINITAB v14. The experiments were used to determine the effects of parameter variation on the stability of the system as well as the interactions amongst the parameters. The parametric studies reveal that the heat transfer coefficient (k), the heat generation coefficient (Q), the Doppler's coefficient of reactivity ([gamma]1), delayed neutron precursor concentration ([beta]) and fitting parameters a1 and a2 have a notably larger effect on the response. Out of these parameters k, a, Q and [gamma]1 tend to destabilize the system at higher values while [beta], a1 and a2 have a stabilizing effect at higher values. Further, cross effects between a, k and Q were also found to be negligible compared to the main effects of the control factors.