COUPLED 2023

Benchmark of CFD/FEM coupled and chained approaches for the prediction of the axial flow-induced vibration on a cantilever rod

  • Galpin, Jeremy (Framatome)
  • Collignon, Christophe (Framatome)
  • Blanc, Patrick-Michel (CEA)
  • Robin-Boudaoud, Marie-Cecile (Framatome)
  • Goreaud, Nicolas (Framatome)

Please login to view abstract download link

Flow-induced vibration is a complex phenomenon involving coupling effects between the solid structure and its surrounding flow. From an industrial point of view, this phenomenon may lead to excessive wear of solid structures, and many precautions are taken in the nuclear industry for avoiding the occurrence of such situation. The modelling of Flow Induced Vibration requires addressing the resolution of the structure dynamics and the flow which affects the structure response because of the damping and added mass effects on one hand, and also excites the structure because of the flow instationarities induced by turbulence on the other hand. Two levels of coupling between the fluid and solid solvers are here investigated. One firstly considers a chained approach consisting in performing firstly the flow computation using Computational Fluid Dynamics in STAR-CCM+ code and in prescribing the computed temporal variations of the hydraulic loading together with a calibrated fluid/structure interaction model of Païdoussis and Ostoja-Starzewski [1] as input data of a Finite Element Model computation performed in Cast3m mechanical code. Then, a more ambitious approach, namely the coupling one consisting in the reciprocal exchange of data between both solvers at each time step, is implemented in the CFD code STAR-CCM+. Both chained and coupled approaches are benchmarked on the experiment of Cioncolini et al. [2] involving a cantilever rod loaded with lead balls and surrounded by an axial flow. The investigations are conducted for a set of two flow conditions, Re=1.6×10^4 and 3.5×10^4. For both flowrates, a good agreement is pointed out between the chained and coupled approaches. Compared to the experimental findings, the natural frequency of the first mode and its evolution with the flowrate is properly retrieved. The evolution of the standard deviations of the rod displacement with the flowrate is correctly reproduced even if the standard deviations are slightly underestimated. This underestimation is related to the insufficient reproduction by the used CFD approach of the lineic excitation, the latter is not negligible in this experiment on contrary to common industrial applications. [1] Païdoussis, M. P., & Ostoja-Starzewski, M. (1981). AIAA journal, 19(11), 1467-1475. [2] Cioncolini, A., Silva-Leon, J., Cooper, D., Quinn, M. K., & Iacovides, H. (2018). Nuclear Engineering and Design, 338, 102-118.