3.1 Particle size distribution
Fig. 2 presents the measured cumulative particle size distribution of VOO from the HSSB after the vibratory sieve (1000 pm). The distribution curve shows a sigmoidal shape. Around 15% of the solid particles is smaller than 81 pm and 20% of the particles is larger than 256 pm. The remaining 65% correspond to a size range between 81 and 256 pm. This results in a d50 of 165 pm. To the best of our knowledge, no other data about the particle size distribution of VOO form HSSB have been reported, thus the results could not be compared.
3.2 Static settling assay
The effect of room temperature (15, 20 and 30 °C) on the static settling of VOO from HSSB in a column is shown in Fig. 3: (A) solid particles content, SPC, and (B) moisture and solid particle content, MSPC. The MSPC of the initial VOO is 4.58%, of which 0.36% corresponds to SPC.
As can be observed, the settling curve for the three temperatures, for both SPC and MSPC, could be divided in three phases. The first phase of hindered settling corresponded to a linear model (Eq. (3));
c=b+m-t (3)
where c is the MSPC expressed in %, t is the time (min), b is the intercept and m is the slope. The valúes of m and b for the MSPC have been calculated for each temperature (Table 3). The parameter m shows valúes of -4.2 x 10-2, 4.6 x 10-2 and 5.1 x 10-2%/min for 15, 20 and 30 °C, respectively. However, a somewhat higher slope (m = -0.051) with a smaller intercept (b = 4.807) are achieved for 30 °C.
Table 3. Coefficients for MSPC values (from 5 to 80 minutes) of hindered settling of the VOO from HSSB and settling efficiency for 80 minutes at three different temperatures (15, 20 and 30° C).
The second phase of settling, called transition phase, shows a lower rate which starts around 80 min (Kynch, 1952). As settling continues, a compressed layer of particles begins to form at the bottom of the tank. As can observed in Fig. 3, after 80 min the settling velocity decreases until that the MSPC and SPC were stabilized around 1% and 0.04%, respectively. This third phase where the rate is stabilized is called compression settling and also was described by Kynch (1952).
In addition, settling efficiency at 80 min (inflection point between linear and hindered phases) was estimated (Table 3) by Eq. (2). After 80 min of settling similar values of n80were obtained for 15 °C and 20 °C, i.e. 55% and 56% respectively, while for 30 °C the value of n80 was significantly higher, around 77%. This increase of the settling efficiency, according Stoke’s law (Eq. (1)), can be explained by the correlation between temperature and oil viscosity, where the oil viscosity tends to decrease when the temperature increases, and thus facilitates the settling of solid particles. An Arrhenius type model describes the effect of temperature on oil viscosity ( Bonnet et al., 2011 and Gila et al., 2015), which is expressed by Eq. (4):
where p is the oil dynamic viscosity (Pa s), A the pre-exponential factor (Pa s), Ea is the activation energy (J/mol), R is the gas constant (8.314 J/mol/K) and Tthe absolute temperature (K). These results about settling efficiency of the VOO from HSSB cannot be compared since there are no previous works available.
Fig. 3. Effect of temperature (15, 20 and 30 °C) on the static settling of the VOO from HSSB in a settling column: (A) solid particles content (SPC) and (B) moisture and solid particle content (MSPC).
Finally, it is noteworthy that during the first minutes of the oil settling an increase of SPC is observed. This increase can be explained by the fact that VOO from HSSB is freshly prepared and contains air microbubbles which drag the solid particles to the column top at the beginning of the settling process.
3.3 CFD simulation
As far as the comparison of the CFD model with the obtained experimental values is concerned, Fig. 4 presents both the experimentally measured and the simulated values of the SPC of the VOO from HSSB during settling in a column. For this purpose, the experimental values of the SPC of the assay at 20 °C are used. Simulated values are obtained using values of physical properties of the VOO (Table 1), such as viscosity (0.084 Pa s) and density (908 kg/m3) borrowed from a previous study (Gila et al., 2015). 28 simulations are carried out, analyzing the settling process of 7 solid particle classes (Table 2) with four different density values (Table 1).
Fig. 4. Simulated (CFD) and experimental values for SPC of VOO from HSSB. Simulations of different particle densities (925, 1025, 1125 and 1225 kg/m3) are shown.
As can be observed in Fig. 4, a linear settling of the solid particles, expressed as SPC, occurs both for experimental, apart from the microbubble phenomenon, and simulated values for particles with densities of 1025, 1125 and 1225 kg/m3. However, for particles with density of 925 kg/m3 the settling is slower during the entire
simulation. Similarly, as observed for the experimental valúes of the SPC, the simulated valúes for these partióles also initially follow a hindered settling regime (i.e. linear decrease), after which it remains more stable until the end of the simulation. The SPC valúes achieved, at 10 cm depth, after 1280 min of simulation were 0.052%, 0.043%, 0.039% and 0.026% for the solid particles with density of 925, 1025, 1125 and 1225 kg/m3, respectively. A similar value of SPC (0.038%) was obtained in the experimental assay at the same time (1280 min). Therefore, the model CFD demonstrates that the solid particle density is situated in these density ranges, i.e. similar density values (1025-1200 kg/m3) previously reported by Alba (2008). In Fig. 5 the contours of volume fraction of solid particles (Table 2) with a density of 1125 kg/m3 are shown after 80 min of simulation. As expected, according to Stoke’s law (Davis, 2010), the solid particles of larger size settled faster than the smaller. This behavior is also observed in the simulation, where the solid particles of larger size (A, B, C and D) are found at the bottom whereas the smaller particles (F and G) remain homogeneously distributed in the column without settling. However, for solid particles with intermediate size, as is particle E, although part of the solids particles had settled during the first 80 min, most of them still remain in suspension.
3.4 Outlook and perspectives
Up to now, to our best of knowledge, no data is available about VOO settling from a two ways ‘decanter’, so that this work is a first attempt to quantify the behavior of the VOO from two-ways ‘HSSB’ during settling in settling column.
Normally, the room temperature in the oil mill ranges between 10 and 25 °C and the oil temperature is about 15-30 °C during the settling in conical bottom tanks. Therefore, as expected, temperature is an important factor to consider in these kinds of separation processes, since the oil viscosity is strongly influenced by temperature (Gila et al., 2015). Besides temperature other factors affect this separation process, such as the density difference between liquid and solid particles, the particle size and shape among others (Davis, 2010). Further, the simulation CFD of this settling case allowed obtaining an idea of the influence of particle density in the settling process.
Thus, further works focused in these aspects should be carried out and CFD models could be developed for current settling tanks, both in batch and continuous operation, or investigate new designs, using the results of this work as a starting point.
4. Conclusión
A first approach was presented about the effect of temperature in the static settling of the VOO from HSSB in a settling column model. The particle size distribution was obtained (particle sizes between 10 and 1000 pm), which showed a curve with a sigmoidal shape, where the major percentage (around 65%) of the solid particles ranged in sizes from 81 to 256 pm, obtaining a particle size of d50 around 165 pm. The three temperatures used in the assays described settling curves that showed two steps, a linear settling for the first minutes (around 80 min) and a hindered settling where the settling rate was decreased slowly till values of MSPC and SPC of 1% and 0.04%, respectively. The temperature of 30 °C showed higher values of settling efficiency (77%) compared to the lower temperatures (15 °C and 20 °C). Besides, in this study a first approach of the use of a CFD model in the VOO clarification step was presented. Simulations showed the same settling trend than that of the experimental measurement. According to the four solid particles density values studied in the simulations, the valúes that ranging between 1025 and 1225 were those that showed best settings with respect to the experimental values, excluding 925 kg/m3 which showed lower settling rate. The solid particles with larger size (813.5, 441.5, 202.0 and 1|am) at 80 min settled well, whereas smaller particles did not settle. Hence, CFD modeling seems to be a useful tool to design oil mill devices or improving actual designs.
Acknowledgements: This work was supported by a fellowship from Ministry of Science and Innovation (Spain) associated to the project FPI-INIA RTA2009-00002-00-0, the grant CAICEM11-67 with the company Pieralisi España SL and the project ‘PI 26323’ from ‘Consejería de Innovación, Ciencia y Empresa’ of the ‘Junta de Andalucía’ (Spain). The authors gratefully acknowledge their financial support.