This paper gives an overview and evaluation of the predictive use of the Functional Resonance Accident Model, FRAM (Hollnagel, 2004), a method that is under development and in line with the theoretical framework of Resilience Engineering (Hollnagel et al. 2006). FRAM is based on the premise that both negative and positive events can result from (expected and unexpected) combinations of the variability in the normal performance of complex socio-technical systems. FRAM was used as a qualitative risk assessment method in a nuclear power plant. The process assessed was the transportation of an 80 ton heavy metal cylinder for transportation of used fuel. The study focused on the transportation of the cylinder inside the power plant and in particular the 40 meter lift that is needed to get the cylinder up into the reactor hall. The FRAM analysis was used to propose recommendations at different levels, such as how to improve human-machine interface issues, procedures, information dissemination within the organization, and government guidelines.