Simulation of phenomena at crack-like leaks and breaks in piping with consideration of fluid-structure interaction. Final report

The evaluation of fluid flow rates through crack-like leaks in pressurized components plays an important role for assessments on break preclusion, especially leak-before-break considerations. In the framework of project RS1194 various calculation methods for the simulation of structure mechanical and thermo-hydraulic phenomena due to flows through crack-like leaks in the coolant circuit were examined and validated on selected leak rate experiments. Besides large program systems as ATHLET, CFX and ADINA also several simplified evaluation methods included in the GRS program WinLeck were applied especially for the determination of leak rates. For the validation of analysing methods, tests were selected previously conducted at the former Nuclear Research Centre (KfK) at Karlsruhe and the Power Plant Union (KWU). The review of experimental results already at disposal in regards to availability of measure d values of thermo-hydraulic parameters like flow-through rates, spatial distributions of pressure, temperature and aggregate state of the medium, velocity of the medium as well as leak openings, displacements and structure strains indicated, that the experiments in terms of quantification of thermo-hydraulic and structure mechanical phenomena as well as appropriate coupling effects do not provide sufficiently meaningful results. Due to missing experiments for validation of 3d numerical flow simulation in crack-like leaks experiments with flow through a Venturi orifice, which are relevant in this context, were chosen. Experiments with single phase flow were considered as well as ones with two phase flow. The post-calculations of the single phase flow showed a good agreement between the calculation results and the appropriate measured data. In the two phase flow, despite tests with various model variations, no satisfying agreement between calculation and test could be reached. According to the authors' opinion is the model approach available in CFX for the simulation of the abrupt evaporation of an overheated fluid resulting from a pressure decrease (flashing) insufficient. Hence, the two phase flow from crack-like leaks cannot be satisfactorily calculated yet with this analysis methodology at present. Thermo-hydraulic and structure mechanical analyses were performed for postulated leak positions in the surge line (SL) of a PWR type Konvoi. Leaks were assumed in form of circumferential through-wall cracks of 180 degrees length. The sizes of these leaks were calculated in the framework of FE-calculations with analysis models of a cooling loop of a PWR type Konvoi with crack-like leaks. With these leak sizes ATHLET-calculations were conducted, especially examining the influence of the consideration of a variable leak size. The implication of the decreasing leak size at the leak position in the SL-middle, especially on the pressure distribution, is significant. The reduction of the leak area due to the pressure- and temperature decrease amounts in the transient examined to ca. 25 % after about 1 h transient time and leads therefore to an approximately 23 % smaller leak rate. The results show that the consideration of the interaction between thermo-hydraulic and structure mechanical effects could enhance the accuracy of accident analyses significantly. At the determination of the leak rates as critical flow-through rates with simplified methods according to Pana, Henry and the CDR-model (critical discharge) it has to be taken into account that the approaches differ considerably in parts. The differences can be attributed especially to the assumptions for the treatment of inflow losses in the crack channel and the flow resistance coefficient due to the roughness of the crack surfaces. The investigations demonstrate that good agreement between results of the Pana and the CDR model could be achieved if the water in front of the crack-like flow channel is subcooled. Uncertainties remain especially if steam or saturated steam flow has to be considered. In total progress has been achieved concerning the determination of leak rates with simplified methods as well as in the framework of application of the codes ATHLET and ADINA. Progress concerning the simplified methods has been implemented in the GRS program WinLeck. Open points remain especially concerning the detailed simulation of fluid flow in crack-like leaks in connection with well-instrumented leak rate tests.