Investigations on the stochastic nature of condensation induced water hammer

•Condensation induced water hammers (CIWHs) were investigated in a horizontal pipe.•The appearance, location and severity of CIWHs were statistically evaluated.•CIWHs with pressure peaks up to 37 times the operating pressure were observed.•A conceptual model to explain the underlying phenomenology is presented.•Refinements of preceding guidelines for the avoidance of CIWHs are indicated. Sudden condensation procedures represent a severe safety issue in numerous industrial applications. One of the most destructive forms of appearance is the so called condensation induced water hammer (CIWH), where subcooled water is brought into contact with saturated steam in an enclosed area. A few examples are the emergency core cooling of nuclear reactors, the start-up process of steam circuits in chemical, oil and gas plants or the hot gas defrost in ammonia-cooled frozen food plants. In each case saturated steam is brought into contact with subcooled liquid through a sufficiently large exchange surface, resulting in a two-phase flow pattern which encourages the entrapment of rather large steam pockets. If these pockets collapse the surrounding water slugs collide, which can create a pressure rise up to several times the operation pressure and cause the failure of pipes and connecting instruments. This paper describes an experimental setup which allows the detailed investigation of CIWHs in horizontal pipes with 1.4° declination. On basis of 185 experiments, conclusions regarding the probability of CIWHs, the height of pressure peaks and the position of nucleation sites of steam implosion are presented. Important results are the identification of critical parameter combinations (injection flow rate and degree of subcooling) for CIWH incidents with high probabilities and high pressure values as well as the determination of a criterion for exclusion. Further the distribution regarding the origin of water hammer effects, respectively the location of slug collision, is determined in dependency on the degree of subcooling. The accurate description of the experimental setup and experimental outcome provides a reliable basis for future simulations and calculations. In the first example a model concept is presented, which focuses on the underlying phase distribution. This model is the first step to a more reliable prediction and prevention of condensation induced water hammers in the future.