Scale resolved simulations of the OECD/NEA−Vattenfall T-junction benchmark

Mixing of fluids in T-junction geometries is of significant interest for nuclear safety research. The most prominent example is the thermal striping phenomena in piping T-junctions, where hot and cold streams join and turbulently mix, however not completely or not immediately at the T-junction. This results in significant temperature fluctuations near the piping wall, either at the side of the secondary pipe branch or at the opposite side of the main branch pipe. The wall temperature fluctuation can cause cyclical thermal stresses and subsequently fatigue cracking of the wall. The issue of thermal striping has been observed in light water reactors (LWRs) as several incidents of high-cycle fatigue at coolant mixing junctions have been detected - mainly in piping T-junctions - in nuclear plants, like the failure event at Civaux 1. These incidents occurred usually in piping of diameter 5-20 cm and the most susceptible parts to thermal fatigue are mixing T-junctions of the Residual Heat Removal (RHR) system in both boiling (BWR) and pressurized water reactors (PWR). This has raised thermal fatigue to be a serious safety concern and an important aspect on ageing and life management of nuclear plants with LWR. A typical value for the temperature difference between the hot/cold streams is 160°C. Critical parameters for thermal fatigue analyses are frequencies (ω), temperature differences (T), number of cycles (N), and material properties. Most damaging thermal loads appear to be due to large scale turbulent fluctuations of low frequency (3-10 Hz). From a thermal hydraulic standpoint, the accurate prediction of such large coherent eddies is a challenging task, requiring CFD and advanced turbulence modelling. Significant effort has been put in the experimental investigation of the thermal fatigue and thermal striping phenomena due to thermal mixing in pipe T-junctions. In November 2008, a T-junction thermal mixing test was carried out at the Älvkarleby Laboratory of Vattenfall Research and Development (VRD) in Sweden. Data from this test have been reserved specifically for this CFD benchmark exercise. The test section is constructed from Plexiglas, and the junction itself from one solid block into which the main and branch pipes fit. The temperatures of the water in the main and branch pipes were maintained at 15°C and 30°C, respectively, with minimal heat loss. Special care was taken to provide simple and well-defined inlet boundary conditions to remove ambiguities in defining the CFD input data. Temperature fluctuations near pipe walls were measured using thermocouples. These were placed around the inner wall perimeter of the main pipe at seven stations downstream of the junction and at one station upstream. All thermocouples were positioned 1 mm from the wall. Velocity profiles upstream and downstream of the junction were measured using a two-component LDV system. These were positioned at each inlet, and at the outlet. Data are in the form of mean values, RMS values and turbulence statistics. The numerical prediction of thermal mixing and striping in terms of temperature amplitude and frequency using the current CFD technology is a computational intensive and challenging task. By the physics of the phenomenon, the flow is turbulent and highly transient and the thermal striping at pipe walls is affected by the formation and propagation of large-scale turbulent structures in space and time. The aim is therefore a CFD turbulence model validation study and a detailed CFD experiment comparison. Turbulence model approaches to be studied in the present validation study include URANS SST as well as scale resolving turbulence models (LES).