Automated Flow Verification of Nuclear Material with CRISP (Central RADAR Inspection Support Package)

Large facilities handling sensitive materials in the nuclear fuel cycle are usually equipped with complex safeguards equipment which generates large amounts of data. The inspectors need excellent tools to allow them a thorough and efficient analysis of these data. The present paper describes recent developments by EURATOM in this respect. EURATOM's standard software for an automated analysis of NDA data is CRISP, a database system which allows Inspectors to perform an efficient, easy and quick review of huge amounts of safeguards relevant data in large facilities. The aim of CRISP is to analyse measured data, continuously provided by Data Acquisition Modules (DAM’s) in a multi sensor system and to compare the results with item movement declarations provided by the plant operator. The evaluation process is structured in three hierarchical layers called evaluation levels: Event Detection, Event Correlation and Event Verification. On each evaluation level, one or more so- called evaluation algorithms are fed with input data and input parameters in order to generate output events that are subsequently stored in the CRISP database. A considerable number of evaluation algorithms is already integrated into the CRISP system, they are the core of CRISP and can be either in-house developed (e.g. Pu Mass Calculation) or integrated as a third party development into the system (MGA, INCC, ORIGEN). A significant enhancement of CRISP has been the development of an algorithm which allows the inspector to do an automatic analysis of item-movements in a plant. The item movement in large Pu handling facilities can be very complex. As the operator does not always move items on the same routes with identical sequences standard item tracking methods fail. In order to provide Inspectors the possibility to do an automated tracking of items in these complex scenarios a new algorithm has been developed, the flow verification algorithm. This algorithm is able to follow even non-standard movements of a specific item throughout the facility and to indicate its final destination. In combination with Remote Data Transmission the inspector has in principle the option to perform item tracking of nuclear material in the headquarter in Luxemburg. This paper summarizes the efforts in the area of automated analysis and describes future challenges in the development and integration of new algorithms.