(a) Natural circulation systems are now implemented or being considered in many future reactor designs. Natural circulation is applied both to ensure the sufficient coolant flow in the core (reactors based on primary coolant natural circulation during normal operation) and to fulfill or support the fundamental safety functions. It should be noted that reactors based on natural circulation during normal operation, e. g. the Dodewaard Reactor in the Netherlands and VK-50 in Russia operated for an extended period of time;

(b) There is a consensus that natural circulation systems can provide more reliable means for a number of safety functions (e. g. decay heat removal from the core, heat removal from the containment atmosphere) and can be cost effective as these systems are less vulnerable to failures and therefore, the number of parallel trains in one system can be reduced. Different designs use natural circulation in different ways. Common design concepts and their respective needs relevant to natural circulation should be identified;

(c) If natural circulation is used as an operational principle of design (full power to be removed from the core by natural circulation) the most important task seems to be preventing the core from two phase natural circulation instabilities induced by neutronic feedback;

(d) If natural circulation is used as a working principle for safety systems, 1D codes can cope with the phenomena of natural circulation in many cases with some limitations and some uncertainties. The main drawbacks of natural circulation systems include the lower driving forces and less possibility to alter the course of an accident if something unexpected happens. Due to low driving forces, the operation of the natural circulation systems may be adversely affected by small variations in thermal-hydraulic conditions. The lower driving forces might also lead to quite large equipment where the role of 3D phenomena is essentially increased;

(e) In cases where multi-dimensional phenomena are present, the safety relevance and need for a 3D approach should be investigated first. An example is the effect of thermal stratification on the building condenser, observed in PANDA;

(f) The effect of non-condensable gases on condensation has been known for many years, and design features to address the problem (e. g. steam jet air ejectors for steam turbine cycles) are provided. In order to cope with this problem in future natural circulation systems, special design features should be considered for condensers.