Improvement of the estimation model of turbidity removal efficiency in helical tubular flocculators: Understanding process physics and their interrelationship with treatment efficiency
Name: Bruno Peterle Vaneli
Type: MSc dissertation
Publication date: 30/09/2016
Advisor:
Name | Role |
---|---|
Edmilson Costa Teixeira | Advisor * |
Examining board:
Name | Role |
---|---|
Edmilson Costa Teixeira | Advisor * |
Julio Tomás Aquije Chacaltana | Internal Examiner * |
William Bonino Rauen | External Examiner * |
Summary: Oliveira (2014) developed models (linear and nonlinear) to estimate the efficiency in removal of turbidity in Helically Coiled Tube Flocculators (Portuguese acronym: FTHs), linking process efficiency with the seven representative dimensional parameters of geometry (𝐷/𝑑 e 𝑝/𝐿), hydraulics (𝐶𝑎 e 𝑅𝑒) and hydrodynamics (𝑆𝑛, 𝐸𝑎𝑑𝑚 e 𝐻𝑎𝑑𝑚) of these units. In this work, three aspects of the proposed models are taken for improvement. I) by producing the greatest Determination coefficient, Oliveira (2014) took the non-linear model as a reference. However, the mean relative deviation produced by linear model was lower than that observed in the non-linear model; II) a laminar model was used in the ambit of a computational fluid dynamic modelling (CFD) to model the air flow in the FTHs that were higher than the critical Reynolds (𝑅𝑒𝑐) the value that marks the end of the laminar regime and the beginning of the transition regime; III) Amongst the hydrodynamic parameters incorporated in the models developed, a normal pressure gradient which is an important parameter in characterizing secondary flow was not found. As a result, the linear model proposed has greater ability to predict experimental data when compared to the non-linear model, and so was taken as a point of reference in improvement. In estimating the value of 𝑅𝑒𝑐 in pipes with strong curvature (𝐷/𝑑≤35,7) the use of the formula proposed by Kühnen et aI. (2015) is recommended. In pipes with medium (35,7<𝐷/𝑑≤110) and slight curvatures (110<𝐷/𝑑) the formula proposed was that of Cioncolini and Santini (2006). So, in the acquisition of the hydrodynamic parameters, a laminar model was used to model the flow of the FTHs operating with 𝑅𝑒 lower than 𝑅𝑒𝑐 and the 𝑘−𝜔 turbulence model in cases WHERE the flow was higher than 𝑅𝑒𝑐. When considering turbulence modeling, an improvement in the quality of the adjustment of the linear model proposed by Oliveira (2014) was found. It was determined that the average normal pressure gradient (𝐺𝑃𝑎𝑑𝑚) must be considered in the model to model for efficiency estimation of turbidity removal. When considered, the parameters 𝑆𝑛 and 𝐻𝑎𝑑𝑚 came to be statistically insignificant, obtaining a simpler model with better statistics than the linear model proposed by Oliveira (2014).