CONDITIONS FOR THE VALIDITY OF THE CONTAINMENT SYSTEM ASSEMBLY VERIFICATION METHODS TO GUARANTEE THE RADIOACTIVITY CONTAINMENT REQUIREMENTS

American and international standards relative to leakage tests on packages for shipment indicate that the leakage rates of the test performed before shipment can be higher than the ones used to demonstrate the activity release regulatory criterion (10-6 A2/h). This practise is based on the hypothesis that the control of a correct closure of the package before shipment allows to have a leaktightness better than measured, and equivalent to the level measured during the package maintenance. This paper presents the results of a study which demonstrates that defects of intermediate size (scratches, dust, cruds, etc. ) may affect the sealing by elastomer O-rings and cause leakages at rates lower than those fixed to detect an incorrect closure but higher than those allowing to directly guarantee the containment conformity to the regulation. Leaktightness tests have been performed with a mockup of a typical closure system. Different kinds of defects (metallic and organic capillaries, scratches), with various sizes (5 µm to 300 µm) were installed between inner gasket and gasket seat. Some of them are not visible to the naked eye. The results indicate that the measured leakages uniformly vary with the size of the defect. We conclude that the control of the correct closure of a package does not provide assurance that the leaktightness necessary to meet the regulatory criterion is reached. Results of this study have been submitted to an expert group appointed by the French Authority (Nuclear Installations Safety Directorate). It has been concluded that the practise of the unique control of a correct closure of a package to guarantee leaktightness is not acceptable unless specific procedures are implemented to guarantee the absence of defects liable to cause leakage rates higher than those directly derived from the acceptable activity release rates.