Chang-Hsieh-Chen low-Reynolds k-ε turbulence model adaptation to study the flow of concentrated pulp suspension in pipes

• Autor(a) principal: Cotas, C.
• Data de Publicação: 2014
• Outros Autores: Garcia, F., Ferreira, P., Faia, P., Asendrych, D., Rasteiro, M. G.
• Tipo de documento: Outros
• Idioma: eng
• Título da fonte: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Texto Completo: https://hdl.handle.net/10316/42855
• topic: Non-Newtonian fluid; Chang-Hsieh-Chen turbulence model; Damping function; Drag reduction

Computational Fluid Dynamics (CFD) tools can be applied successfully to predict the turbulent pipe flow of pulp suspensions. This strategy allows to avoid the design of industrial equipment based on empirical correlations. The present work investigated the turbulent pipe flow of concentrated Eucalyptus pulp suspensions. The numerical study was performed using the ANSYS FLUENT® CFD software [ANSYS FLUENT® 13.0.0, ANSYS FLUENT Inc., 2010]. A pseudo-homogeneous approach was applied in this work. The non- Newtonian behavior of the pulp suspensions was introduced into the CFD code by considering the pulp viscosity as a function of a local shear rate. Additionally, the model took into account a presence of a water annulus at the pipe wall surrounding the flow core. The Chang-Hsieh-Chen [1,2] low-Re k-ε turbulence model was selected in this study as the one allowing to account for a drag reduction effect in the pulp suspension flow as already shown in the previous work of the authors [3]. Moreover, as referred in [2] the CHC turbulence model is more universal than the other low-Re models since its constants are the same as those conventionally used for the standard k-ε model. The applicability of the CHC model to reproduce the drag reduction effect has been tested with a damping function adopted from the work of Malin [4] who studied the pipe flow of a power-law fluid. In order to better fit the available reference experimental data the damping function proposed in [4] has been modified by varying its parameters. A good correspondence between the velocity profiles reported in literature [5] and those obtained numerically was achieved. As the key indicator of the model appropriateness the pressure drop values obtained experimentally were used.