Guan'gai paishui xuebao (Dec 2024)

Comparison of different turbulence models for simulating water flow in axial pump

  • CHEN Di,
  • ZHANG Xiaowen,
  • TANG Fangping,
  • WANG Lin,
  • XU Xudong,
  • GUAN Zilong

DOI
https://doi.org/10.13522/j.cnki.ggps.2024069
Journal volume & issue
Vol. 43, no. 12
pp. 47 – 56

Abstract

Read online

【Objective】 Water flow in pumps is modeled using computational fluid dynamics (CFD). However, the impact of turbulence model on simulation accuracy remains unexplored. This study aims to address this gap. 【Method】 We simulated axial water flow in a pump using three turbulent models: the Standard k-ε, RNG k-ε and SST k-ω models. We compared the simulated results with experimental results and analyzed the internal flow fields simulated by different turbulence models. 【Result】 At steady flow state, the head curves simulated by all three turbulence models were consistent with the experimental curves. Under forward and reverse flow braking conditions, the head calculated by the Standard k-ε and SST k-ω model was close to or slightly higher than that calculated by the RNG k-ε model. Compared to the RNG k-ε model, the Standard k-ε and SST k-ω models reduced the maximum error of calculated head by 0.13 m and 0.22 m, and average error by 0.55 m and 0.48 m, respectively. Under low water flow condition in the forward pump, the maximum and average error of the calculated head by the standard k-ε model were the least; compared to the RNG k-ε and SST k-ω model, it reduced the maximum error of the calculated head by 1.19 m and 0.96 m, and the average error by 0.56 m and 0.68 m, respectively. Under reverse downstream braking conditions, the SST k-ω model was most accurate; compared to the Standard k-ε and RNG k-ε model, it reduced the maximum errors of the calculated head by1.08 m and 1.63 m, and average errors by 0.6 m and 0.9 m, respectively. Under the reverse pump condition, the Standard k-ε was most accurate; compared to the RNG k-ε and SST k-ω models, it reduced the maximum errors of the calculated head by 0.65 m and 0.74 m, and average errors by 0.34 m and 0.44 m, respectively. For other conditions, including positive rotating pump, positive rotating downstream braking, positive rotating and reverse rotating turbine, the simulated internal flow field in the pump was smooth, the simulated external characteristics were close to experimental observations, and simulated results of all three models were similar. 【Conclusion】 The accuracy of turbulence models for simulating axial water flow in pumps depends on operating conditions. Our results provide insights into selecting appropriate turbulence model for simulating water flow in pumps.

Keywords