Temporary solutions for a conservative estimation of void reactivity insertion in CANDU reactor

One of the most difficult task of the CANDU Reactor Physics Analysis is related to the correct treatment of the deviations from the reference coolant properties. The most significant problem is the reactivity inserted by a given coolant density variation. From the practical Nuclear Safety Analysis point of view, the solution must be not only conservative, but also adaptable to the current chain of codes utilized for accident simulation. The first set of experimental data was obtained by AECL many years ago. The fuel was fresh, clean and cold. Some of the currently used computer codes offer accurate predictions of the measured void reactivities. Unfortunately, the existing experimental data do not cover and are not significant for the burned CANDU fuel. A specific benchmark problem was suggested by the Institute for Nuclear Research (ICN) Pitesti. The problem was analysed and slightly modified during an IAEA Vienna RCM (Research Coordinating Meeting), Buenos Aires, 1990. Afterwards, the problem was independently solved in several countries, interested by the CANDU reactor. The results were presented and analysed at the Bombay RCM, 1992. It was clear that the interval defined by the code predictions is much too broad. New experimental data are necessary. They must cover the fuel isotopic composition specific for the burned CANDU fuel. The work is in progress at the Chalk River Laboratory. Temporary solutions have been analysed at the ICN Pitesti. The first aim was to identify the reactivity numerical values that are conservative, but not too inaccurate. The WIMS code predictions have been compared against other estimations, including the Monte-Carlo based ones. The second aim was to force the currently used code, PPV, to offer cell cross sections that are correct from the Reactor Physics point of view, and compatible with the imposed reactivity. Physical and mathematical procedures were proposed and evaluated. An additional solution was also taken into account: to modify the current PPV cross sections using a very simple procedure. (author)