Numerical Simulation of Particle Classification in Two-Phase Flow in a Vortex-Based Air Classifier Using the Eulerian-Lagrangian Approach and Determining Its Optimized Performance Condition

Bagherzadeh, Ashkan | 2022

32 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 55653 (45)
  4. University: Sharif University of Technology
  5. Department: Aerospace Engineering
  6. Advisor(s): Darbandi, Masoud
  7. Abstract:
  8. Using a Numerical simulation approach for modeling multiphase flows is a cost-effective and prevalent method in analyzing the physics of multiphase problems. One of the most applied needs of industry in using gas-particle simulation is the separation of particles using Air Classifiers which are industrial devices that can classify solid inlet particles based on their size using airflow patterns and other mechanisms. The present study focuses on a particular type of classifier using vortex to classify particles. Performance of these classifiers simulated using numerical Eulerian-Lagrangian DDPM model. The two critical studied performance parameters of the classifier are pressure drop and separation efficiency. The studied input parameters affecting the performance of the classifier are inlet velocity of air and particles, mass flow of the inlet particles, and damper angle. inlet particles with a d80 equal to 450 micron enter classifier to separate in coarse and fine outlets. Velocity studied in the range 14 to 24 m/s, inlet mass flow in range of 40 to 250 kg/s and the damper angle studied in the range -20° to 45°. Comparison of simulation and empirical results showed 9.5% and 3.8% errors for pressure drop and fine outlet mass flow respectively. Simulation results show that the inlet velocity under 20 m/s leads to a 10% lower mass at the fine outlet. Also increasing mass-flow of particles does not considerably change separation efficiency, but for each 100 kg/s particle mass flow inlet pressure drop rises 116 Pa. Damper openness angle also has a significant effect on changing flow field and consequently plays a vital role in setting fine and coarse outlet mass flow and sizing. The minimum pressure drop happens at zero damper angle and change in range of -15° to 15° alters the fine outlet mass in range of 35% to 80%. At the last step, a model has been developed for prediction outputs of the classifier using the extracted data from CFD simulations and the Artificial-Neural-Network (ANN) method. By using this ANN-based model the more data points can be extrapolated so it reduces the computational cost of simulations, with this model an evaluation has been performed to find the optimum input variables of the system in any operating condition. Simulation and analyzing flow in this particular geometry of classifier and combining ANN with CFD results to find the optimum performance points in the classifier can be counted as novelties of the present study.
  9. Keywords:
  10. Numerical Simulation ; Particles Separation ; Eulerian-Eulerian Approach ; Air Classifier ; Two Phase Gas-Solid Flow ; Artificial Neural Network ; Discrete Phase Model ; Dense Discrete Phase Method (DDPM)

 Digital Object List