On the probability of loss of cooling and long term heat removal: A new theory

Following an extended power loss the concern is inadequate long-term heat removal. The obvious question is whether a natural circulation long-term cooling or method is more ‘reliable’ in fulfilling its intended core cooling function than flow using powered systems. To establish the risk of failure, we utilize the phenomenological uncertainty in the deterministic phenomena to inform probabilistic safety analysis. The flows are inherently unstable and despite predicting the onset the flow details are impossible to predict using modern computer codes and simulation methods. By combining the observed phenomenological physics of unstable single and two-phase flows with statistical reasoning we derive the probability of heat removal failure. Based on data, the presence of known multiple modes of instabilities and microscopic oscillations unexpectedly and counter-intuitively can assure heat removal so that failures of active back up systems, necessary valves, emergency batteries and diesels actually dominate the risk of ineffective heat removal. The results presented in this paper are design and technology neutral, and include all human actions, equipment deployment, system configuration, back-up cooling modes and strategies. The implications for natural circulation heat removal concepts, methods and loop design configurations are discussed.