Specification for trip settings for a glovebox fissile material leak detection system

Following an event where plutonium oxide/oxalate powder accumulated in a glovebox at Sellafield, the facility safety case was reviewed. One recommendation from this review was that a neutron monitoring system aimed at detecting any similar future accumulations should be installed. This system has now been installed and is functioning. One area of interest with respect to the system was the methodology used to determine the appropriate neutron count-rate settings used to trip the plant. Clearly the system must function before accumulations reach the critical condition (with an appropriate safety margin). However it is not a simple matter of calculating the minimum safe mass and setting up the neutron monitor to detect this mass. The situation is complicated by: -) The need to consider the potential locations in the glovebox for accumulation of leaked fissile material. -) The possible location sites for detectors (outside the gloveboxes). -) The variation in potential fissile material forming the accumulation (chemical form, density, moisture content, shape, etc.). -) Background effects (including those brought about by variation in the normal glovebox inventory) and the need to avoid spurious trips. Hence, rather than minimum safe mass, the parameter of interest is the minimum safe mass in the form and position which is most difficult to detect. In this paper, the method by which these complications were overcome is given for one glovebox, selected as being a relatively complex example. Essentially this was done by: 1) Determining a range of safe masses as a function of fissile material variations using MONK. 2) Modeling detector performance in the presence of these unsafe conditions using MCNP. This included considering variations in normal inventory. 3) Validation of the MCNP modeling using a plutonium standard. 4) Selecting trip settings which ensured that the trip would operate before an unsafe accumulation would arise, but where there was sufficient margin over normal operations count rates. (authors)