[en] An adequate level of safety is achieved in nuclear power plants by a suitable design of protection systems. Safety Analysis Reports typically include a verification that the plant behavior under a set of Design Basis Accidents (DBAs) (including from anticipated operational events to postulated accidents) remains under specified acceptable limits and, therefore, enough safety margin is provided by the protection design. Conclusions of the Design Basis Accident Analysis (DBAA) depend on a set of assumptions on initial and boundary conditions and on system reliability which should be consistent with the operation of the plant. Many of these assumptions are imposed as Technical Specifications or equivalent operation requirements. This kind of safety verification, however, is mainly analytical and only very specific aspects of the protection design can be experimentally verified. Operational events are unique occasions to check, on one hand, if the plant protection behaves as expected and, on the other, if the designed protection is enough to guarantee a sufficient level of safety. There are several mechanisms of safety margin degradation. Some are of dynamic nature, i.e., the plant or protection behavior is not as expected and a loss of margin may occur, eventually leading to limit exceedance. Others are of probabilistic nature, for example, due to a loss of protection reliability leading to a given probability of limit exceedance. Incident analysis should address, to the extent possible, all these mechanisms. As a part of the analysis of operational events it should be identified whether the incident meets the assumptions of the DBA analysis or it is outside the design basis envelope. In the former case, one or more DBAs can be identified as envelope transients. Thus, safety margins during the real event can be verified to be equal to or greater than those demonstrated in the Design Basis Accident Analysis for the envelope transients. In the latter case, it should be identified if one or more of the safety limits used as acceptance criteria for DBAs has been actually exceeded. In any case, an extension of the precursor analysis techniques, usually focused on the eventuality of degradation to severe accident conditions, can be used as a means to evaluate probabilistic mechanisms of safety margin loss. (author)