Abstract
This paper describes work that has been performed by NNC Ltd and Fuji Electric for the study of decommissioning of Tokai power station (Tokai-1). The objective was for NNC to provide an independent validation of representative selection of the existing Fuji Electric calculations the results of which were obtained by the methods generally used in Japan based on the discrete ordinate code DOT 3.5, in estimating full power neutron fluxes and reaction rates in components located within the reactor biological shield of the Tokai 1 reactor. The calculational methods and modelling assumptions are described for the four regions in which fluxes and reaction rates were required, namely in regions above the core, regions to the side of the core, regions below the core and regions in the concrete walls of the bio shield gas duct penetrations. NNC has considerable experience in performing similar analyses for UK reactors and the methodology and computer codes employed here are based on experience gained in carrying out such work for AGR, PWR and Magnox reactor types. Thus, much of the component modelling has been achieved using the Monte Carlo code MCBEND supplemented, in the case of the gas duct penetrations, by the iterative kernel albedo code MULTISORD. Above, below and to the side of the core, results were obtained in some detail in nearly all of the structural components. In the case of the bioshild concrete, results were obtained in many regions at various depths and axial heights. Along the gas ducts, results were calculated at the concrete wall surfaces of the penetrations to the point where the total flux had reduced to a level of 103 n/cm2/s, this being the level at which the induced concrete activity can be regarded as negligible. Preliminary calculations were carried out using the duct streaming code MULTISORD in order to establish the approximate location where flux levels dropped to this level. MCBEND was then used to model the geometry in detail up to this point. The MCBEND calculations were very computationally intensive. In total several hundred hours of CPU time was required. The timescale of the project has only allowed a limited amount of comparison between the DOT 3.5 and MCBEND predictions this point in time. An initial comparison has been made within the biological shield in the bottom corner region of the reactor vault where good agreement was obtained. It is envisaged that the neutron fluxes derived from this work may be used to determine the activity of neutron activated structures in order to evaluate the dose rates that are likely to be experienced during dismantling. This data should also be useful in calculating the costs of waste disposal and in estimating worker and public dose exposure during decommissioning.
view the full text of this article (18 pages, format: PDF, size= 1074kB)