University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
Virginia Tech | specified contact time, and then the coal particles were separated from the water by either
centrifuging or filtration. Finally, the water was then analyzed for residual reagent.
It should be noted that range of test conditions (i.e., frother and collector dosages, and
coal slurry solid to liquid ratios) included in this work is much wider than that which may be
encountered in practice. This is because as a major objective here was to determine under what
conditions the processing reagents would sorb to coal versus remain in water, and vice versa. For
the purpose of making relative comparisons, a froth flotation circuit in a typical coal preparation
plant might operate with coal slurries of 1-10% solids (by weight), which require 4-20 μL/L
frother (usually specified in mg/L; ~5-25 mg/L) and 1.5-150 μL/L (usually specified in lb/ton of
coal; ~0.5-5 lb/ton).
4.1 Frother Partitioning
For the frother partitioning tests, coal samples were obtained from the Elkhorn #3 and the
Cedar Grove seams (both <5% ash), and were sized to -100 mesh prior to testing. Slurries were
mixed for five minutes by rapid stirring in open beakers, and then centrifuged for three minutes.
To analyze the relative amount of frother left in the clear water fraction of the slurry, surface
tension measurements were conducted using a Fisher surface tensiometer. The tensiometer
utilizes a platinum-iridium ring, and measures the force required to detach this ring from the
liquid surface. The ring was thoroughly cleaned between tests by immersing it in benzene, then
acetone, and finally passing it through a flame to remove of any surface contaminants. Glassware
was also thoroughly cleaned between tests by washing with chromic acid solution and distilled
water.
4.2 Collector Partitioning
For the collector partitioning tests, two separate raw coal samples were obtained: one
from the Hagy Seam (~ 35% ash), and one from Pocahontas Seam (~ 16% ash). The former
sample was sized to -100 mesh for the first set of tests, and then a subsample of that material was
screened to 100 x 150 mesh for the second set of tests. The later sample was only used in the
second set of tests, and was also screened to obtain 100 x 150 mesh particles. For the first set of
tests, slurries were mixed in a kitchen blender for four minutes and then centrifuged until the
water was clear; however, it should be noted that a large amount of colloidal matter in these
28 |
Virginia Tech | samples prevented removal of all color from the water. In the second set of tests, the slurries
were mixed in open flasks on a shaking table for four minutes, and then filtered (through 25 μm
paper) using a vacuum pump. The residual diesel in the clear water fraction from each test was
analyzed using an Agilent 5890 gas chromatograph equipped with a flame ionization detector
(GC-FID), by following EPA Method 3150 for quantifying diesel range organics (DRO) in water
samples.
Table 2.2: Experimental conditions for frother tests
Coal Dosage Frother Dosage
Test Coal Seam Frother Type
(wt. % solids) (μL/L)
1-18 Elkhorn #3 0, 0.1, 0.5, 0.7 M150 0.4a, 4, 40, 400, 4000
19-34 Elkhorn #3 0, 0.1, 0.5, 0.7 PSM 4, 40, 400, 4000
35-48 Elkhorn #3 0, 0.1, 0.5, 0.7 Nalco 8836 4, 40b, 400, 4000
49-60 Elkhorn #3 0, 0.1, 0.5, 0.7 MIBC 10, 100, 1000
61-64 Cedar Grove 0.5 M150 4, 40, 400, 4000
65-68 Cedar Grove 0.5 PSM 4, 40, 400, 4000
69-72 Cedar Grove 0.5 Nalco 8836 4, 40, 400, 4000
73-75 Cedar Grove 0.5 MIBC 10, 100, 1000
a Only for 0 and 0.1% solids
b Only for 0 and 0.5% solids
Table 2.3: Experimental conditions for first set of collector tests
Coal Diesel
Diesel Residual
Coal Dosage d o s a g e Solid/Liquid
Test dosage DRO
Seam (wt. % (lb/ton Separation
(mg/L) (mg/L)
solids) coal)
1 Hagy 0 N/A 500 Centrifuge 425.1
2 Hagy 1 0 0 Centrifuge <0.05
3 Hagy 1 1 4.9 Centrifuge 0.39
Centrifuge,
Hagy 1 1 4.9 0.42
4 then filtration
5 Hagy 1 1 4.9 Filtration 0.46
6 Hagy 1 10 50 Centrifuge 0.68
7 Hagy 5 0.25 6.3 Centrifuge 0.50
8 Hagy 5 1 25 Centrifuge 0.53
9 Hagy 5 10 250 Centrifuge 0.95
29 |
Virginia Tech | Table 2.4: Experimental conditions for second set of collector tests
Coal Diesel
Diesel Residual
Dosage d o s a g e Solid/Liquid
Test Coal Seam dosage DRO
(wt. % (lb/ton Separation
(mg/L) (mg/L)
solids) coal)
10 Pocahontas 0 N/A 0.85 N/A 1.35
11 Pocahontas 0 N/A 0.425 N/A 0.63
12 Pocahontas 1 0.17 0.85 Filtration 0.42
13 Pocahontas 10 0.017 0.85 Filtration 0.31
14 Pocahontas 5 10 250 Filtration 0.47
15 Pocahontas 5 10 250 Filtration 0.40
16 Pocahontas 5 10 250 Filtration 0.50
17 Pocahontas 5 10 250 Filtration 0.51
18 Pocahontas 5 10 250 Filtration 0.47
19 Pocahontas 5 10 250 Filtration 0.42
20 Pocahontas 1 50 250 Filtration 0.79
21 Pocahontas 5 50 1250 Filtration 1.02
22 Pocahontas 10 50 2500 Filtration 1.92
23 Pocahontas 5 1 25 Filtration 0.49
24 Hagy 5 10 250 Filtration 0.88
25 Hagy 5 50 1250 Filtration 2.67
5. Results and Discussion
Results of the partitioning tests confirmed that, in general, frother and collector reagents
do not partition completely to either the solid or liquid fraction of a coal slurry – and therefore it
is possible that, to some extent, these reagents may end up in coal products, tailings
impoundments and in recycled water.
5.1 Frother Adsorption
The surface tension results for varying frother dosages and varying coal slurries are
shown in Figure 2.2. The dashed horizontal line at 72.8 dyne/cm represents the theoretical
surface tension of pure water (Nave); the bold line shows the measured surface tension for
frother only (no coal added). For all frothers, it appears that the reagent tends to sorb somewhat
to the coal surface. This can be seen most clearly at moderate test dosages (i.e., 40-400 μL/L),
30 |
Virginia Tech | where a significant difference was observed in surface tension between tests with frother only
and tests with frother and coal. As expected, more frother generally tended to sorb to coal when
more coal was present (i.e., 0.7% vs. 0.1% solids).
At very high dosages (i.e., 1000-4000 μL/L), the effect of the coal becomes less
significant for MIBC and Dowfroth M150, and nearly insignificant for PSM and Nalco 8836.
This indicates that sorption sites on the coal surface may be completely filled, and thus most of
the frother remains in the water. At very low test dosages (i.e., 4 μL/L), the PSM exhibits
seemingly complete sorption to the coal particles, as the surface tension when coal is present is
effectively that of pure water, as compared to substantially less with frother only. The Dowfroth
M150 also exhibits significant sorption to the coal at very low dosages, although the surface
tension is slightly less than that of pure water (for the 0.5 and 0.7% coal tests), which suggests
that some frother did not sorb. At very low dosages of MIBC and Nalco 8836 (i.e., 10 and 4
μL/L, respectively), it is uncertain to what extent the coal particles were able to sorb frother
because the frother did not depress the surface tension of the water. This highlights a major
shortcoming of the use of surface tension measurements to study frother reagents, which has
been previously noted by other researchers (Sweet et al. 1997).
Coal properties were found to play a role in the sorption behavior of PSM and MIBC. As
evident in Figure 2.2, at equal levels of slurry solids (i.e., 0.5% coal), the Cedar Grove coal did
not appear to significantly sorb these frothers, whereas the Elkhorn #3 coal did. However, the
sorption behavior of the Dowfroth M150 and Nalco 8836 was observed to be quite similar
between the two coals. Since proximate analysis was not performed on the coal samples, it is
difficult to speculate on specific explanations for these results; but coal properties (other than
particle size) do seem to be important in terms of frother sorption capacities.
In the context of a coal preparation plant, the results from these tests indicate a significant
degree of frother sorption to coal surfaces can be anticipated. While practical conditions include
only the low to very low ranges of frother dosages tested here, they typically have higher slurry
solids contents, and thus higher coal surface areas – which suggests that perhaps a relatively
large fraction of frother reagents may associate with the coal. Given that frothers are well known
to cause problems via entrainment in recycled water, there may be several plausible explanations
for the findings presented here: 1) frother sorption to coal may only be temporary, and desorption
may occur downstream of flotation processes (e.g., during dewatering); 2) the presence of other
31 |
Virginia Tech | reagents, particularly collectors, may substantially interfere with frother sorption to coal; and 3)
the experimental conditions (e.g., mixing, effective contact time) used here may not be
representative of plant conditions. Considering these, the sorption mechanisms of frothers to coal
and tailings particles is deserving of further study. If, for example, frothers are identified which
sorb strongly to coal through flotation and dewatering, this may have significant implications for
reducing fouling of process circuits in closed water systems, as well as reducing environmental
releases through tailings impoundments. For frothers that do not sorb to and remain with coal,
novel water treatment strategies may be devised to remove these reagents from water prior to
recycling or environmental discharges.
Figure 2.2: Surface tension versus varying dosage levels of frother and coal
5.2 Collector Adsorption
DRO results (i.e., the residual DRO in the clear water fractions of tested coal slurries) are
presented in Tables 2.3 and 2.4 for all test conditions. The most striking observation is that there
is some low level of DRO in every test, despite the addition of even large amounts of coal (i.e.,
32 |
Virginia Tech | 10% solids). For instance, tests 12 and 13 clearly show that at relatively high solids content (i.e.,
1 and 10%, respectively) and very low dosages of diesel (i.e., <1 mg/L; or 0.17 and 0.017 lb/ton,
respectively), about 0.3-0.4 mg/L DRO remains in the water fraction of the slurry. Moreover, the
level of DRO does not change dramatically between tests, considering the extreme changes in
diesel and coal dosages. In test 20, for example, which had the same amount of coal but nearly
300x more diesel added than test 12, the DRO concentration was only about 2x higher than that
of test 12 (i.e., 0.79 vs. 0.42 mg/L, respectively). And in test 22, which had the same amount of
coal but nearly 3000x more diesel added than in test 13, the DRO concentration was only
increased by about 6x (i.e., 1.92 vs. 0.31 mg/L, respectively). These results seem to indicate that
a small amount of diesel (~0.3 mg/L or less) is always soluble in the water, but that the coal
particles have a very high adsorption capacity for the diesel that is not dissolved. Another factor
that may have been at play here is the possible presence of colloidal matter in the water fraction
of the slurries; if diesel sticks to the colloids, it would likely be measured as DRO. However, it is
important to note that, no matter what the reason, these tests indicate that a small amount of
diesel will effectively partition with water in a flotation circuit.
Figure 2.3 highlights other specific observations in the collector partitioning tests. In the
far left plot, the effect of solid-liquid separation technique on the results is shown. The three tests
(#s 3-5) were conducted using identical slurries (i.e., % coal solids and diesel dosage), but one
was centrifuged, one was filtered, and the other was centrifuged and then filtered. DRO
concentrations in the clear water fraction from each of these tests were all within about 15% of
each other – a reasonable range for preliminary tests – and it was concluded that the solid-liquid
separation methods did not substantially impact partitioning results (e.g., by sorption of diesel to
the filter paper).
33 |
Virginia Tech | Figure 2.3: Diesel sorption test results
The middle plot of Figure 2.3 shows the results from six tests to determine the
reproducibility of the test and analytical methods used here. Tests 14-16 show DRO measured
three separate times (i.e., in triplicate) from a single sample. The results for these tests are within
about 20% of each other and suggest that the analytical method is fairly reproducible. Likewise,
tests 17-19 show DRO measured from samples from three separate, but identical tests. In this
case, the results are within about 18% of each other and indicate that the test method is also
reproducible.
In the right plot of Figure 2.4 are the results from three tests conducted to determine the
effect of proportionally similar coal and diesel additions (i.e., tests at 1, 5 and 10% solids, each
with a diesel dosage of 50 lb/ton coal). Since the diesel was dosed on the basis of coal weight, it
seems intuitive that DRO concentrations should have been similar between these tests; instead,
with increasing additions of coal, less diesel actually sorbed. One possible explanation for this
phenomenon may be that with more coal in the slurry, particles are sticking to each other or
being bridged together by diesel such that there are effectively fewer sorption sites available. For
tests where coal content remained constant (e.g., tests 7-9) but diesel dosage was varied,
measured DRO in the water did increase with a substantial increase in diesel dosage – although
not proportionally. For instance, in tests with 5% Hagy Seam coal (-100 mesh), DRO was
roughly equal for diesel dosages of 0.25 and 1.0 lb/ton (i.e., 0.50 and 0.53 mg/L), but essentially
doubled when the diesel dosage was raised to 10 lb/ton (i.e., to 0.95 mg/L).
It was further observed that the ash content of coal appears to affect diesel sorption. At
equal slurry contents and diesel dosages (i.e., 5% solids, and diesel dosages of 10 or 50 lb/ton),
the Pocahontas Seam coal (~16% ash) sorbed about 2-2.5x as much diesel as the Hagy Seam
34 |
Virginia Tech | coal (~35% ash) (see Table 2.4). This is likely because coal has a higher affinity for diesel than
ash does. It is difficult to assess whether or not the sized Hagy Seam coal (100 x 150 mesh)
behaved differently than that which was only ground (-100 mesh), since just one test condition
was repeated between the first and second set of tests (i.e., tests 9 and 24; 5% coal and diesel
dosage of 10 lb/ton); however, the DRO results for these tests were practically very similar.
In terms of real preparation plants, the results of the collector partitioning tests presented
here indicate that, as expected, most diesel should partition with the coal. However, some
(presumably soluble) diesel may well remain in the process water – eventually being sent to
tailings impoundments or being recycled back through the plant. While no Federal water quality
standards currently exist for DRO, some states have set levels of concern at 0.05 mg/L (e.g.,
through reporting levels for diesel spills or contamination from underground storage tanks) (DEP
2002). The topic of soluble DRO, including the relative solubility of specific diesel compounds
and potential remediation strategies, is deserving of additional research.
6. Conclusions
Processing reagents used in coal preparation have a wide range of potential
environmental fates, as well as implications for preparation circuits that are designed or revised
to utilize closed water systems. The preliminary test work presented in this paper confirms that
common frother and collector reagents are not likely to partition completely to a single fraction
of the process slurry. Instead, the partitioning phenomena are complex, and appear to depend on
many operating variables including coal and reagent characteristics and dosages.
To gain a better understanding of the ultimate fates of these reagents and related impacts,
further work should focus on determining the mechanisms by which various reagents may
associate with solid and liquid fractions of coal slurries. Moreover, work is needed to elucidate
strategies for controlling/optimizing reagent partitioning, or treatment of affected process
streams.
7. Acknowledgments
The authors would like to acknowledge the Appalachian Research Initiative for
Environmental Science (ARIES) and US Department of Energy (under grant no. DE-AC22-86-
TC91221) for funding experimental work described here. Views, opinions or recommendations
35 |
Virginia Tech | 3 DIESEL RANGE ORGANICS IN COAL PREPARATION
Josh Morris, Emily Sarver, Gerald Luttrell
Paper peer reviewed and originally published in Proceedings of the Environmental
Considerations in Energy Production Symposium, April 14-18, 2013, Charleston, WV, pp. 51-
59. Reproduced with permission of the Society for Mining, Metallurgy & Exploration.
www.smenet.org
1. Abstract
Recent laboratory testing has suggested that partitioning of petro-diesel collector reagents
in coal flotation circuits is not perfectly ideal. In this paper, we investigate the persistence of
diesel range organic (DRO) compound residuals in process waters under a number of physio-
chemical conditions. Additionally, we investigate desorption of DRO from coal surfaces exposed
to fresh water. In both cases, we also examine the behavior of individual PAH compounds.
Results are discussed in the context of potential environmental transport and fate of DRO
compounds in water from coal preparation plants. While our results indicate that DRO
concentrations in process waters are expected to be at sub-ppm levels under normal operating
conditions, we note that “green” collectors are available for coal flotation.
2. Introduction
The primary function of coal preparation is removal of mineral matter (i.e., ash), which
detracts from the coal value. Preparation plants typically have multiple circuits for processing
materials of different particle sizes. In the fine circuits (i.e., -100 mesh particles), froth flotation
is often used to separate coal from ash. In 2012, about 30% of the coal preparation plants in the
US (i.e., 82 of 289) were reported to operate fine coal flotation circuits, and virtually all of these
(i.e., 80 of 82) are located in the Central Appalachian basin (Fiscor 2012). For reference, this
basin also accounted for about 85% of the total preparation plants (i.e., 248 of 289).
Froth flotation of coal works on the basic principle that coal particles are relatively
hydrophobic and lightweight, such that they easily float to the top of the flotation columns or
cells and can be recovered from a froth that forms there; while ash mineral particles are relatively
38 |
Virginia Tech | hydrophilic and dense, such that they sink and are rejected to the tailings stream. Coal particles
are naturally hydrophobic to some degree, depending on their specific chemical properties (e.g.,
surface oxidation, liberation from mineral matter), but “collector” reagents that increase their
hydrophobicity are often used to aid the flotation process (Wills 2006). Globally, petro-diesel
(termed “diesel” in this paper) is the most commonly used collector given its relatively low cost
and proven performance (Laskowski 2001). However, collector dosages are quite variable (e.g.,
diesel may be dosed in the range of 0-5 lb/ton) due to flotation circuit parameters and feed
quality; dosage at a single plant may be adjusted frequently.
While collectors work by sorbing to the coal surface (Kondrat'ev 2009), recent studies
have indicated that the partitioning of diesel between coal and process water is not perfectly ideal
(Morris et al. 2012). In bench-scale experiments, under a variety of test conditions mixing water,
coal and diesel, low-level diesel range organic (DRO) compound residuals were consistently
measured in process water. Results suggested that the DRO residuals may primarily consist of
the water soluble fraction (WSF) of the diesel, which exists at sub-ppm levels. At present, DRO
is not monitored in impoundments or discharges, however, given increasing concerns over diesel
contamination of water resources by leaking underground storage tanks, highway and blacktop
runoff, and events such as the recent Deepwater Horizon oil spill in the Gulf of Mexico
(Sementelli and Simons 1997; Lloyd and Cackette 2011; Osofsky et al. 2011), it is important to
understand the potential transport mechanisms and fate of DRO from coal processing. It should
be noted that alternative “green” collector reagents such as bio-diesel and pine oil products are
already being utilized for fine coal flotation in some instances. These collectors are significantly
more expensive than diesel but may be required in special circumstances, such as safeguarding
groundwater resources when coal waste slurry is intended to be disposed via underground
injection (WV-DEP 2009).
2.1 Environmental Fate and Transport of Diesel Compounds in Water
Diesel has a mixed composition of roughly C to C hydrocarbons, which, like for many
10 19
petroleum products, varies based on the crude oil source and refining process(es) (ATSDR
1999). This variability makes it difficult to uniquely classify the chemical and environmental
characteristics of diesel, and also to accurately measure individual components (ATSDR 1995).
Instead, diesel is usually described in terms of major compound categories: total saturated
39 |
Virginia Tech | hydrocarbons (e.g., alkanes, cycloalkenes) and total aromatic hydrocarbons (e.g., naphthalenes,
acenaphthenes, acenaphthylenes). Saturated hydrocarbons typically account for about 90% (by
weight) of total diesel, while aromatics account for about 10% (Wang et al. 2003). Any
remaining components (e.g., waxes and resins) usually make up a very small fraction of the total
diesel and are not often quantified.
In addition to broad categorical classification, the nature of diesel and other petroleum
products is sometimes characterized by measuring a series of ideal (i.e., n-alkanes) and/or
priority compounds (i.e., specific aromatics). Typical analytical techniques include GC-FID, GC-
MS and fluorometry (Wang et al. 2003). It is important to note that while specific compounds
may be targeted by these techniques, the large number of individual compounds present in diesel
(e.g., branched alkanes, functionalized aromatics) makes it impossible to quantify each and every
one (ATSDR 1995). Moreover, the diesel composition may change dramatically in the
environment as it weathers (e.g., via volatilization, bio- or photo-degradation). Such complexities
make predicting environmental implications of diesel releases quite challenging, but offer unique
opportunities for source tracking in some cases (Wang et al. 1996).
A basic understanding of environmental transport mechanisms and fate of diesel can be
gleaned from properties of the compounds in the broad categories mentioned above. Generally
speaking, the saturated hydrocarbons are relatively volatile (Fingas 1994; Fingas 1995),
insoluble in water, and photo- and bio-degradable (Olson et al. 1999; Marquez-Rocha et al.
2001; Cohen et al. 2002; Kakkar et al. 2011). Thus, at relatively low levels, this group of
compounds does not present major concerns for water resources. Aromatics in the diesel range
are also typically volatile and insoluble in water, but are much more persistent in the
environment because they resist degradation (Olson et al. 1999). Further, some monocyclic and
polycyclic aromatic hydrocarbons (MAHs and PAHs) have been classified as possible or
probable human carcinogens (ATSDR 2009), and have been linked to acute or chronic toxicity in
aquatic organisms (Schein et al. 2008). Thus, if significant concentrations and exposure
pathways exist, these compounds may present ecological and human health hazards. Indeed, the
US EPA has developed a list of 16 priority PAHs, although only one compound (i.e., benzo-[a]-
pyrene) is regulated by a maximum contaminant level in drinking water (EPA 2011). Total DRO
is not federally regulated (EPA 2003).
40 |
Virginia Tech | Figure 3.3: DRO desorption – experimental methods
3.3 Analytical Methods
Residual diesel in each sample was analyzed using an Agilent 5890 gas chromatograph
equipped with a flame ionization detector (GC-FID); we followed EPA Method 3150 for
quantifying diesel range organics (DRO) in water samples. Samples were also analyzed for a
group of target PAHs using a Thermo Trace GC equipped with a Thermo DSQ II mass
spectrometer (GC-MS); we followed EPA Method 3535A: Solid-Phase Extraction (SPE). The
target PAHs were: benz(a)anthracene, benzo(a)pyrene, benzo(b)fluoranthene,
benzo(k)fluoranthene, chrysene, dibenz(a,h)anthracene, indeno(1,2,3-cd)pyrene, acenaphthene,
acenaphthylene, anthracene, benzo(ghi)perylene, fluoranthene, fluorine, naphthalene,
phenanthrene, and pyrene.
4. Results
Results of all tests confirmed that a small fraction of diesel tends to partition to the
process water – and this is typically limited to the soluble fraction. The converse side of this
observation is therefore that the coal particles have a very high capacity for sorption of insoluble
diesel.
4.1 Potential for Diesel Removal from Process Water
Table 3.1 presents results for DRO and PAH residuals in the tests designed to investigate
potential for diesel removal from process water. Figure 3.4 also graphically displays the results
for DRO and naphthalene. The rate of DRO removal was significantly higher for the stirred and
44 |
Virginia Tech | Table 3.1: DRO and PAH removal results
Test Time PAHs (ppb) DRO
Condition (hr) Naphthalene Acenaphthene Fluorene Anthracene Fluoranthene Pyrene (ppm)
Initial 0 60 6.5 6.8 2.7 0.14 0.18 0.43
2 56 7.1 7.6 2.8 NF 0.27 0.59
Stagnant
4 34 5.1 5.6 NF NF 0.25 0.53
2 2.5 0.99 1.6 1.5 NF 0.16 0.30
Stirring
4 0.51 0.15 0.37 0.81 NF NF 0.27
2 0.98 0.10 NF NF NF NF 0.27
Aeration
4 0.69 0.41 NF NF NF NF 0.29
2 17 2.9 3.5 2.0 NF 0.13 0.48
Heating
4 12 2.0 2.7 1.6 NF 0.12 0.38
NF indicates the compound was not found.
Figure 3.4: DRO and naphthalene removal results
4.2 DRO Desorption
Previous work has demonstrated that the sorption capacity of raw coal for diesel is very
high (e.g., all but the soluble portion of diesel sorbs to the coal) – even beyond practical
operating conditions for fine coal flotation (Monsalve 2010; Morris et al. 2012). The goal of the
present testing was to determine if the sorbed diesel might be easily removed again from the
coal, for example during dewatering or other instances where the coal product is contacted with
46 |
Virginia Tech | Figure 3.6: PAH desorption results
5. Discussion
The results presented here provide some fundamental insights to the probable behavior of
diesel compounds in fine coal flotation circuits. It is clear that some low-level diesel will likely
report with process water to tailings impoundments or other waste repositories. In open-air
systems, insoluble DRO may be gradually removed from process water, and removal may be
accelerated in systems that promote aeration, mixing or heating; and these conditions which are
commonly encountered in both impoundments (or slurry cells) and preparation plants where
recycled flotation-process water may be utilized. Under normal operating conditions, a system is
expected to continuously equilibrate such that only the WSF of diesel persists to any appreciable
degree, and thus high concentrations of DRO are not expected accumulate. In a typical
impoundment, the WSF will likely also be subject to photo- and bio-degradation.
We have also shown that diesel-contacted coal may release diesel to fresh water. In terms
of environmental fate, this may be most important for clean coal stocks that are wetted, or in the
instance of ultrafine coal that ends up in an impoundment. In the latter case, it is expected that
diesel will desorb from the coal particles only until equilibrium with respect to diesel solubility
in the impoundment is reached.
In regards to behavior of PAHs, it has been demonstrated here that some of these
compounds may be expected to quickly leave an impoundment via volatilization, which may be
encouraging – at least for water quality. Naphthalene was the most abundant of any of the PAHs
48 |
Virginia Tech | targeted in this work, and at even the highest measured concentration of approximately 60 ppb,
this is still lower than EPA lifetime advisory limits of 100 ppb for drinking water (EPA 2006).
It should be noted that the PAH contents measured in process waters in this study do
appear to suggest some concentration effects (vs. the pure diesel), meaning that all components
of the diesel do not remain in constant proportions. For instance, naphthalene in the pure diesel
was measured to be approximately 0.1%, while in the initial filtrate sample used in the diesel
removal tests the naphthalene was calculated to be about 13% (i.e., 60.1 ppb naphthalene in 0.43
ppm total DRO). This may suggest that coal selectively sorbs saturate compounds over
aromatics. However, according to the observed behavior of PAHs under conditions of stirring,
aeration, stagnation and heating, this concentration effect might be quickly reduced or even
reversed.
Another important point of discussion in regards to PAHs in diesel is that of the relative
abundance of target vs. actual compounds. While ideal PAHs (i.e., simple aromatics of fused
benzene rings without functional groups) do exist in diesel, it is well established that alkylated
PAHs are typically present at much higher concentrations (Irwin et al. 1997). The simplest
alkylated PAHs are formed when the parent PAH compound is functionalized by addition of one
or more methyl groups, which may occur during the digenesis of fossil fuels from organic
sediments. Naphthalene can accommodate up to four methyl groups, and there are 22 individual
compounds in the class of methylated naphthalenes (Abraham et al. 2005). Since quantification
of specific compounds is highly complex, oftentimes only the parent compound (i.e., the EPA
priory compound) such as naphthalene is measured to characterize fuel products or
environmental samples. Additional testing of the diesel used in this work has revealed that di-and
tri-methylated naphthalene compounds are present at about 3x the concentration of pure
naphthalene; while mono- and tetra- compounds are present at more similar concentrations as the
pure naphthalene. Considering this, it may be prudent to target the more abundant compounds in
future testing – although, if the behavior of the alkylated compounds are similar to that of their
parent, environmental implications may not differ significantly.
6. Conclusions
Diesel is a common collector reagent in fine coal flotation circuits in both the US and
abroad. Despite coal’s very high sorption capacity for this reagent, sub-ppm levels of DRO
49 |
Virginia Tech | dominated by the WSF of diesel are expected to be present in process waters. In a typical tailings
impoundment or open-air portions of the flotation circuit, DRO may be removed relatively
quickly via volatilization and/or degradation. The most prevalent PAHs may also volatilize
quickly, particularly where mixing, aeration or heating of the water occurs. DRO may desorb
from diesel-contacted coal; however, in the context of settled ultrafine coal in an impoundment,
this process should be limited diesel solubility in the impoundment water. Relative
concentrations of both total DRO and PAHs targeted in this study do not appear to present
significant concerns for water quality under normal operating conditions – but variances from
such conditions should clearly be avoided. “Green” reagents such as bio-diesel and pine oil
products, both of which should not contain PAHs, are being considered as alternative collectors
for fine coal flotation; however, these have not yet gained widespread use throughout the
industry.
7. Acknowledgments
The authors would like to acknowledge the Appalachian Research Initiative for
Environmental Science (ARIES) for funding experimental work described here. Views, opinions
or recommendations expressed in this paper are solely those of the authors and do not imply any
endorsement by ARIES. We also thank Jody Smiley for her invaluable mentoring and assistance
with analytical results.
8. References
Abraham, M., Autenrieth, R. and Dimitriou-Christidis, P. (2005). "The estimation of
physicochemical properties of methyl and other alkyl naphthalenes." The Royal Society
of Chemistry 2005(7): 5.
Adel, G. (2012). Particulate Process Modeling. 2.1.1 Examples of Particulate Processes,
Department of Mining and Minerals Engineering.
ALS. (2008). "GC/MS-Full Scan vs GC/MS-SIM." 2012, from
http://www.caslab.com/News/gcms-full-scan-vs-cgms-sim.html.
AMMA. "Diesel Fuel." Retrieved January 16, 2013, 2013, from
http://www.miningoilgasjobs.com.au/Oil-Gas-Energy/Hydrocarbons-and-
Energy/Hydrocarbons/Oil-and-Gas/Downstream/Diesel-Fuel.aspx.
50 |
Virginia Tech | 4 FLOTATION TESTS
1. Introduction
The previous chapters in this thesis discuss the potential presence of diesel range organics
(DRO) in processing streams based on the results from laboratory tests where reagents are
aggressively agitated with fine coal and water, and then filtered. This chapter builds upon that
fundamental work by investigating the partitioning of petro-diesel (termed “diesel” in this paper)
in actual froth flotation experiments. The purpose of these experiments was to more closely
simulate real coal preparation operations and determine residual DRO and PAHs of interest in
the process waters associated with the concentrate and tailings.
The flotation tests differ from previous tests by attempting to separate the coal from the
ash and measuring the process water from both of these products. Based on prior results, it is
expected that DRO may be limited to the WSF. Based on the results from Chapter 3, it is
expected that the concentrate water will have higher DRO than the tails since the diesel sticks to
the coal but then may come off with the water during dewatering.
2. Experimental Methods
Flotation studies were carried out to understand DRO partitioning, including PAH-
compound partitioning, between concentrate and tailings process waters. The flotation tests
investigated a range of diesel dosages (as described below) on three coal types, which differed by
source and ash content. Coals 1, 2 and 3 had ash contents of 10.8, 10.6, and 45.7%, respectively.
Coal types 1 and 2 were collected as flotation feed from a coal preparation plant, while coal type
3 was collected as raw feed prior to entering the plant. The coal was prepared for testing only by
sizing. Coal type 3 was crushed using a laboratory jaw crusher followed by a roll crusher, and
then wet screened to obtain material in the range of 44-149 microns (100 x 325 mesh). Coal
types 1 and 2 were wet screened at +44 microns (+325 mesh).
Figures 4.1-4.3 show the particle size distribution for the three coal types. For each coal
type, the distribution was determined from 5 samples and the error bars in these figures represent
one standard deviation from the mean. Coal types 1 and 2 contain coarser material than coal
type 3. Most of the material for coal type 3 is within the 100 x 325 mesh range as it should be.
57 |
Virginia Tech | Coal Type 3
16 100
14 90
80
12
70
e g 10 60 g n
n
a h 8 50
is
s a
C P
% 6 40 %
30
4
20
2 10
0 0
89112492279046178429586089309025
92411221661138656937396419875543
3224716285319765433221111
,7,6,5,4,3,3,2,2,1,1,1,1
Mesh Size
Figure 4.3: Particle size distribution of coal type 3
The diesel used in these tests was purchased from a local fuel station, and stored in an
amber, airtight glass jar. A sample of 99+% methyl isobutyl carbinol (MIBC), the frother used
in these tests, was purchased from Fisher Scientific. The MIBC dosage was kept constant at 10
ppm, while diesel dosage was varied at 0.1, 0.5, 1, and 2 lb/ton. Slurry percent solids was kept
constant at 5% (by weight) and the flotation tests were performed in a Denver cell with 2 L of
deionized water (DI). Mixing velocity was held constant at 1300 rpm, with the air valve
completely open during flotation. The flotation time was fixed at 2 minutes for all tests. Prior to
flotation, there was a fixed mixing time of the slurry for two minutes, and an additional fixed
conditioning time of two minutes after adding the diesel to the cell. The concentrate and tailings
products were then dewatered by filtration (through 5 µm paper using a vacuum pump), and the
filtrate was collected for DRO and PAH analysis. The dewatering equipment including the
vessel and tubing were thoroughly cleaned between each sample by rinsing with deionized water
and methanol.
Identical to analytical procedures outlined in Chapter 3, DRO was determined using an
Agilent 5890 gas chromatograph equipped with a flame ionization detector (GC-FID), following
EPA Method 3150. Samples were also analyzed for a group of target PAHs (i.e., phenanthrene
and naphthalenes) using a Thermo Trace GC equipped with a Thermo DSQ II mass spectrometer
(GC-MS), following EPA Method 3535A: Solid-Phase Extraction (SPE).
59 |
Virginia Tech | Following dewatering of the concentrate and tailings, the solid materials for each sample
(i.e., cake on filter paper) was placed in an oven to dry until no incremental weight change was
measured (i.e., all moisture was removed). Ash content was analyzed in each solid sample using
a LECO model 601-400-600 ash analyzer, and the yield and combustible material recovery for
each flotation test was then determined (Equations 4.1 and 4.2). Y represents the yield, FA
represents the ash content measured in the feed, TA represents the ash content measured in the
tailings, FA represents the ash content measured in the feed, and R represents the recovery.
Equation 4.1
Equation 4.2
3. Results and Discussion
Similar to findings of fundamental laboratory tests described in Chapters 2 and 3, results
of the flotation tests consistently indicate that a small fraction of diesel partitions to the process
waters associated with both the tailings and concentrate; and again the amount of DRO appears
to be limited to the soluble portion. Table 4.1 summarizes the tests and shows that residual DRO
in both the concentrate and tailings process waters tend to increase with increasing diesel dosage.
Figures 4.4-4.7 show the concentrate and tailings residual DRO content for the three coal types.
Tests were repeated for coal type 2 and are displayed as “Test 1” and “Test 2” for coal type 2.
For all three coal types, residual DRO in the concentrate process water samples was significantly
higher than that in the tailings process water, usually by a factor of two or more. Moreover, for
all three coal types, residual DRO in the process water tended to increase with increasing diesel
dosage.
60 |
Virginia Tech | (i.e., ash-coal matrix particles) to float when sufficient diesel is present. Figure 4.9 shows ash
rejection vs. combustible recovery for coal type 3 and Figures 4.10 and 4.11 show ash rejection
vs. combustible recovery for coal type 2 - tests 1 and 2. The graph for coal type 3 is quite similar
to that of coal type 1, with the highest ash rejection resulting from the lowest diesel dosages. For
coal type 2, ash rejections are relatively low for all diesel dosages – perhaps due to a relatively
higher proportion of un-liberated ash particles in this coal sample vs. coal types 1 and 3 – but a
substantial decrease in recovery is seen at the highest diesel dosage. This may be related to a
negative effect on frothing with too much diesel present. To put the observed flotation
performance in perspective for all four sets of tests, Figure 4.12 shows the ash rejected vs.
combustible recovery.
Equation 4.3
Figure 4.8: Ash rejected vs. combustible recovery for coal type 1
64 |
Virginia Tech | than the tails since the diesel sticks to the coal but then may come off with the water during
dewatering. Given that DRO also tended to be higher in the concentrate samples, this is
consistent with prior results that indicated PAHs tended to partition proportionately with total
DRO (see Chapter 2). Moreover, the residual PAHs also tended to be higher in samples where
diesel dosage was higher – which also suggests that these compounds partition proportionally
with total DRO.
Collectors like diesel are used in flotation to enhance the performance of the process. For
all types of coal, there is likely a diesel dosage that will maximize the flotation performance in
terms of particle separation by optimizing coal recovery and ash rejection simultaneously;
however, this dosage may vary substantially between coal types. For the experiments reported
here, only a limited range of diesel dosages has been explored for a limited selection of coal
types and within limited particle size distributions. Clearly, all three coals used here required
relatively low levels of diesel for favorable flotation performance.
4. Conclusions
As a result of these findings, there appears to be a potential correlation between residual
DRO and flotation performance. It appears as though operating performance and residual DRO
may go hand in hand. This means that under normal operating conditions, residual DRO does
not seem to be an issue; however, during abnormal conditions, such as when conditions exist
where diesel is effectively overdosed significantly (e.g., in cases where the flotation coal feed is
reduced or halted, while diesel continues to be fed into the flotation system), DRO in process
water may be a potential issue. In order to account for the chance of abnormal conditions
occurring such as when the feed may stop entering the circuit, an automatic switch should be
connected so that when the feed stops, reagent dosing stops as well. Likewise, it would be
environmentally wise to develop a programmable logic controller that determines the amount of
material being fed into the flotation circuit and based on this amount, adjusts the reagent dosage
being fed into the circuit.
69 |
Virginia Tech | 5 CONCLUSIONS: CHALLENGES AND LESSONS LEARNED
1. Introduction
Measuring low-level organic reagents in coal preparation process waters is a difficult
task, and this chapter is designed as an aid for future work. The intent is to highlight several
challenges that were encountered in this work, and outline a structured methodology for
researchers investigating diesel range organic (DRO) compounds in aqueous samples from coal
processing. Sample preparation, GC analysis, and results interpretation are all specifically
discussed.
2. Challenges And Lessons Learned
2.1 Sample Preparation
Variability of test materials (i.e., coal) and reagents (e.g., petro-diesel) may easily affect
results of DRO partitioning tests and should be considered in the design of future experiments.
Here, petro-diesel (termed “diesel” in this paper) was simply collected from local a fuel station,
and care was taken to use the same diesel across all experiments where results were directly
compared to one another (e.g., in tests where PAHs were measured). Several coal samples were
used over the course of the project, including raw coal that was crushed and sized in the
laboratory, and flotation feed slurries that only required dewatering and screening to the desired
size range. All coals tested differed noticeably by ash content, which likely had some relative
effect on DRO partitioning per Chapter 2 findings. Additionally, it is expected that other
properties (e.g., rank, porosity, etc.) also differed between the coals studied; however no testing
was done to quantify these properties. All else being equal, it is recommended that flotation feed
slurry be used for future work since, simply put, this allows testing of coal samples
representative of a real preparation plant.
While many experiments described here only included strong agitation of coal, water and
reagents prior to analysis of process water, actual flotation experiments provide a more practical
understanding of field conditions. Typically during laboratory flotation tests water is added to
the cell to remove material that adheres to the cell wall as well as to replace water that has been
removed from the cell while paddling the concentrate out of the cell. The concern with this
regarding measuring the amount of DRO in process streams is that adding water changes the
70 |
Virginia Tech | then this is the simple solution; however, if it is not proportional then this technique may not be
the best option to account for the phthalates. The problem with individually removing the
phthalates peaks besides increased time is that this introduces the potential for user error.
Figure 5.3: Diesel standard showing DRO past 25 minutes
Figure 5.3 shows the 204 mg/L standard (i.e., 204 mg diesel per L of hexane) in blue, the
25.5 mg/L standard in black, and Hexane in pink. The blue line shows that some of the DRO is
removed after 25 minutes; however, the amount removed relative to the peaks before 25 minutes
is quite small. The pink line shows that excess noise is detected after 26 minutes. The black line
shows that the area underneath the peaks after 25 minutes is quite significant relative to the area
underneath the peaks before 25 minutes; however, most of it may be due to noise instead of DRO
compounds. Figure 5.4 shows enlarged graph of the 25.5 mg/L standard, shown in black, and
hexane, shown in blue, that was shown in Figure 5.3. The graph indicates that after 25 minutes,
most of the peaks are due to noise as can be seen in the hexane sample. As a result, it was
determined that adjusting the program to only include peak area prior to 25 minutes is a
justifiable solution for dealing with the phthalates issue. It was also determined that the 25.5
mg/L standard is below the detection limit and should not be used in the calibration curve since
most of the DRO peaks are not detected and area is more likely to be affected by noise.
73 |
Virginia Tech | Figure 5.4: Low level diesel standard and hexane past 25 minutes
Interpretation of chromatogram results is another common challenge of using GC to
quantify analytes. A calibration curve must be created by analyzing standards, which contain
known amounts of the analyte (i.e., diesel). For this work, once the diesel concentration range
anticipated in the samples was determined, a calibration curve was developed accordingly. Since
most samples typically appeared to contain between 0.3 to 1 mg/L with a concentration factor of
100 or 200, calibration points of 50, 100, and 200 mg/L were used. The internal standard was
then added to the standards and they were run on the GC-FID with the experimental samples.
The relative response factor (RRF) was calculated by dividing the sample area, SA, by
the internal standard area, ISA (Equation 5.1). The RRF was then plotted against the known
concentration to develop a calibration curve (Figure 5.5). The calibration curve was then used to
calculate the extracted sample concentrations. These concentrations were subsequently adjusted
for the concentration factor. For example, in the flotation tests, a total of 200 mL of concentrate
or tailings water sample was concentrated to just 1 mL, thus giving a concentration factor of
200x. Additionally, an extraction efficiency factor was applied to all adjusted sample
concentrations. To determine the extraction efficiency, EE, the adjusted sample concentration of
the extracted standards, SC, was divided by the known concentration, KC (Equation 5.2). The
average extraction efficiency could then be applied to all extracted experimental samples.
Finally, the average measured concentration in the extracted blanks was subtracted from the
samples to account for noise.
Equation 5.1
74 |
Virginia Tech | Equation 5.2
This is the recommended technique for analysis of results based on the experience gained
throughout this work. Not only is DRO difficult to measure, but because there are many steps
required and the amount measured is so low, quantification can be quite difficult. When
attempting to quantify results, keep in mind the degree of precision that is desired. It is
recommended to place more focus on sample trends rather than the quantification itself.
Calibration Curve
4.0
)
F
R3.5
y = 0.0196x - 0.272
R
( r3.0 R² = 0.9797
o
t
c a2.5
F
e
s2.0
n
o
p
s1.5
e
R
e1.0
v
it
a le0.5
R
0.0
0 50 100 150 200 250
Known Concentration (mg/L)
Figure 5.5: DRO calibration curve
3. Conclusions
This thesis answers a number of preliminary questions regarding the potential fate of
reagents, particularly diesel collector, from coal processing. Chapter 1: Literature Review
outlines these questions and provides background information necessary to frame the potential
problem and design relevant experiments. Chapter 2: Reagents in Coal Preparation: Where Do
They Go? discusses preliminary test work regarding conditions such as dewatering techniques
and reproducibility of results, along with work associated with measuring frother. This chapter
illustrates that reagents are not likely to partition completely to a single fraction of the process
slurry. Importantly, data presented in this chapter also indicates that some frothers may behave
75 |
Virginia Tech | quite differently than others depending on dosage and slurry percent coal solids – which may
have implications for the preparation circuit performance when water is recycled from the
flotation units to other unit operations. Figure 2.2 illustrates that for M150 and PSM frothers the
water surface tension is highly sensitive to frother dosage and coal percent solids. Only at very
low dosages and higher percent solids does most of the frother appear to sorb to the coal. In a
plant setting, this suggests that residual frother may quickly accumulate in the process water; and
based on independent practical observations, these frothers are indeed known to have a tendency
to foul processing circuits where water is recycled, often causing problems with pumping and
foaming in tanks. On the other hand, results for MIBC and Nalco 8836 show that the water
surface tension does not change significantly at relatively low frother dosages (i.e., practical
dosages) across all coal percent solids values. Likewise, these frothers are practically known to
cause fewer foaming problems in preparation plants.
Chapter 3: Diesel Range Organics In Coal Preparation discusses the ability to remove
DRO from water, the desorption capacity of DRO from coal, as well as the investigation of
PAHs in these tests. These tests show that insoluble diesel and most PAHs may be removed via
normal plant operating conditions; however, soluble diesel may remain. While most heavier,
PAHs (i.e., those with a great number of rings) were not detected, pyrene, a PAH with four rings,
was consistently measured at low concentrations. Although pyrene is not highly soluble in water,
it is possible that in the case of DRO residuals in water, co-solvency may occur by which pyrene
is dissolved in other diesel compounds, which are themselves dissolved in the water. By means
of co-solvency, some PAHs like pyrene may be able to remain dissolved in water to some small
extent so long as their direct solvent is present. Finally, Chapter 4: Flotation Tests discusses
DRO results from flotation test work. These tests show that in terms of environment and
performance, it pays to operate efficiently. The tests in this thesis are all geared towards
investigating the potential fates of reagents in a preparation plant. While it appears as though
these reagents do not impose significant concern under normal operating conditions, further
work may certainly be warranted to understand partitioning behavior under conditions outside of
the norm.
Future Work
The main points regarding future work are as follows:
76 |
Virginia Tech | Table D.2: Phthalates in flotation samples
IS Hexanedoic Acid Hexanedoic Acid Phthalate 1 Phthalate 1 Phthalate 2 Phthalate 2 Phthalate 3 Phthalate 3
Description
Intensity Intensity RRF Intensity RRF Intensity RRF Intensity RRF
Hexane NF NF NF NF NF NF NF NF NF
Hexane NF NF NF NF NF NF NF NF NF
Hex w/IS 134804293 NF NF NF NF NF NF NF NF
25.5 mg/l Not Extr 132187078 NF NF NF NF NF NF NF NF
51 mg/l Not Extr 134048216 NF NF NF NF NF NF NF NF
102 mg/l Not Extr 144768418 NF NF NF NF NF NF NF NF
204 mg/l Not Extr 149046670 NF NF NF NF NF NF NF NF
Hexane NF NF NF NF NF NF NF NF NF
Blk pos 5 152973298 NF NF NF NF NF NF NF NF
Blk pos 6 163934049 NF NF NF NF NF NF NF NF
Blk pos 7 163499176 NF NF NF NF NF NF NF NF
Blk pos 5 153455746 NF NF NF NF NF NF NF NF
Blk pos 7 174633779 NF NF NF NF NF NF NF NF
Blk pos 5 150871004 NF NF NF NF NF NF NF NF
Blk pos 7 168703982 NF NF NF NF NF NF NF NF
Std 0.51 mg/l Extr 174279786 NF NF NF NF NF NF NF NF
Std 1.02 mg/l Extr 176043281 NF NF NF NF NF NF NF NF
Std 2.04 mg/l Extr 159922685 NF NF NF NF NF NF NF NF
Std 0.51 mg/l Extr 199790106 NF NF NF NF NF NF NF NF
Std 1.02 mg/l Extr 212660384 NF NF NF NF NF NF NF NF
Std 2.04 mg/l Extr 207755707 NF NF NF NF NF NF NF NF
Std 0.51 mg/l Extr 184753352 NF NF NF NF NF NF NF NF
Std 1.02 mg/l Extr 188714876 NF NF NF NF NF NF NF NF
Std 2.04 mg/l Extr 202005879 NF NF NF NF NF NF NF NF
Hexane NF NF NF NF NF NF NF NF NF
W.R. 1/2 165544418 1740063 0.0105 3642620 0.0220 3459044 0.0209 17879649 0.1080
W.R. 2/2 141135713 727778 0.0052 2209036 0.0157 3278652 0.0232 16634207 0.1179
M.R. 1/2 175712012 257251 0.0015 5059865 0.0288 276980 0.0016 1204830 0.0069
L-0.1-T 219819412 424683 0.0019 1898115 0.0086 243040 0.0011 956733 0.0044
L-0.1-C 208652959 1593248 0.0076 6916988 0.0332 910430 0.0044 4493843 0.0215
L-0.5-T 201984899 671032 0.0033 6237680 0.0309 641012 0.0032 2738032 0.0136
L-0.5-C 239393765 3160224 0.0132 9163947 0.0383 1869138 0.0078 9624837 0.0402
L-1-T 154318074 1600777 0.0104 2927093 0.0190 1038311 0.0067 5363637 0.0348
L-1-C 170817773 7708141 0.0451 9569016 0.0560 3229799 0.0189 16841698 0.0986
L-2-T 176712662 941245 0.0053 2281749 0.0129 570708 0.0032 2462687 0.0139
L-2-C 175961888 3341693 0.0190 4126795 0.0235 903868 0.0051 4413052 0.0251
102 mg/l Not Extr 149068197 NF NF NF NF NF NF NF NF
Hexane NF NF NF NF NF NF NF NF NF
Hexane NF NF NF NF NF NF NF NF NF
Hexane with IS 59868263 NF NF NF NF NF NF NF NF
25.5 mg/l NE 58184757 NF NF NF NF NF NF NF NF
51 mg/l NE 68237045 NF NF NF NF NF NF NF NF
102 mg/l NE 132600832 NF NF NF NF NF NF NF NF
204 mg/l NE 55237101 NF NF NF NF NF NF NF NF
Hexane NF NF NF NF NF NF NF NF NF
blk pos 5 80884921 NF NF NF NF NF NF NF NF
blk pos 7 92084500 NF NF NF NF NF NF NF NF
std 1.02 95917535 NF NF NF NF NF NF NF NF
hexane NF NF NF NF NF NF NF NF NF
K-0.1-T 64998803 157824 0.0024 933189 0.0144 201719 0.0031 827625 0.0127
K-0.1-C 83325142 143833 0.0017 884365 0.0106 150605 0.0018 621352 0.0075
K-0.5-T 70997006 160804 0.0023 267612 0.0038 75632 0.0011 210120 0.0030
K-0.5-C 70948222 24891 0.0004 305885 0.0043 112003 0.0016 397833 0.0056
K-1-T 63841252 NF NF 72178 0.0011 13071 0.0002 44887 0.0007
K-1-C 53050053 226485 0.0043 841349 0.0159 157400 0.0030 650404 0.0123
hexane NF NF NF NF NF NF NF NF NF
blk pos 5 31961791 NF NF NF NF NF NF NF NF
blk pos 7 29512122 NF NF NF NF NF NF NF NF
std 1.02 33528528 NF NF NF NF NF NF NF NF
hexane NF NF NF NF NF NF NF NF NF
K-2-T 24738845 NF NF 98966 0.0040 50033 0.0020 NF NF
K-2-C 32903999 208113 0.0063 374205 0.0114 160720 0.0049 554077 0.0168
O.C.-0.1-T 38198674 23225 0.0006 91774 0.0024 43677 0.0011 199784 0.0052
O.C.-0.1-C 21486165 191761 0.0089 697874 0.0325 79375 0.0037 330507 0.0154
O.C.-0.5-T 42896280 NF NF 21142 0.0005 NF NF 19504 0.0005
O.C.-0.5-C 39624681 NF NF NF NF NF NF NF NF
hexane NF NF NF NF NF NF NF NF NF
Alkanes 1mg/l w/IS 34256148 NF NF NF NF NF NF NF NF
Alkanes 5mg/l w/IS 36213541 NF NF NF NF NF NF NF NF
Alkanes 10mg/l w/IS 38240012 NF NF NF NF NF NF NF NF
Hexane w/IS 10535127 NF NF NF NF NF NF NF NF
25.5 NE 15754981 NF NF NF NF NF NF NF NF
51 NE 17231571 NF NF NF NF NF NF NF NF
102 NE 21054022 NF NF NF NF NF NF NF NF
204 NE 5351712 NF NF NF NF NF NF NF NF
Hexane NF NF NF NF NF NF NF NF NF
blk pos 5 10063386 NF NF NF NF NF NF NF NF
blk pos 7 13003569 NF NF NF NF NF NF NF NF
std 1.02 12060094 NF NF NF NF NF NF NF NF
O.C.-1T 8800267 NF NF 30989 0.0035 27485 0.0031 120135 0.0137
O.C.-1C 17734438 33356 0.0019 153498 0.0087 152010 0.0086 689659 0.0389
O.C.-2T 18360372 50576 0.0028 102393 0.0056 84016 0.0046 333859 0.0182
O.C.-2C 17711073 80009 0.0045 101785 0.0057 38810 0.0022 187298 0.0106
W/O F.P. 12797086 NF NF NF NF NF NF NF NF
M.R.-2/2 13048110 NF NF NF NF NF NF NF NF
CHK STD 102 17635609 NF NF NF NF NF NF NF NF
Hexane w/IS New 11039506 NF NF NF NF NF NF NF NF
25.5 New 12998919 NF NF NF NF NF NF NF NF
51 New 14523118 NF NF NF NF NF NF NF NF
102 New 21347344 NF NF NF NF NF NF NF NF
204 New 17519971 NF NF NF NF NF NF NF NF
NF = No peak was found
100 |
Virginia Tech | Laboratory-Scale Analysis of Energy-Efficient Froth Flotation Rotor Design
Christopher Aaron Noble
(ABSTRACT)
Froth Flotation is an industrial separation process commonly used in the primary en-
richment of run-of-mine mineral material. Over the past 100 years, much of the process’s
development has come from empirical evolution, rather than fundamental understanding.
While many of the governing sub-processes are still poorly understood, the primary influ-
ential factors lie within the chemical, equipment, and operational variables unique to each
flotation system. This investigation focuses on the phenomenological investigation of the
equipment variables, particularly the rotor design, at the laboratory scale.
During this study, several small-scale flotation systems were developed, including vari-
ous rotor and stator designs, tank sizes, and flow conditions. Experimental techniques were
also developed to identify operational performance in four criteria: power consumption, gas
dispersion, operational robustness, and flotation kinetics. Evaluation of the various rotors
was conducted in two campaigns: (1) an exploratory campaign which featured 14 rotors
in limited operational conditions (2) a detailed campaign which featured three rotors in an
exhaustive set of conditions.
The results show that different rotors exhibited varying degrees of performance when
judged by the aforementioned performance criteria. In general, excessive fluid pumping
leads to an increase range of stable operation at the expense of greater power consumption.
However, this increased power consumption does not necessarily correspond to increased
flotation performance, as the data generally confirms the linearly proportional relationship
of flotation rate and bubble surface area flux. Consequently, enhanced flotation kinetics can
be achieved by rotors which disperse high rates of gas while retaining a small bubble size.
This work received financial support from FLSmidth Minerals. |
Virginia Tech | Acknowledgments
The preparation of this thesis has been a rewarding undertaking. I would like to first
thank the Lord for the many blessings I have experienced.
I cannot understate the role of my research advisor Dr. Gerald Luttrell in motivating
me to compile this document. He has been a constant friend and mentor throughout this
time. Additionally, I owe my original interest in flotation to Dr. Roe-Hoan Yoon. His
unquenchable thirst for understanding is both a silent and, at times, vocal motivator for
continued success. My final committee member, Dr. Greg Adel, has been a constant source
of solidarity and direction.
I also express considerable gratitude to FLSmidth for the genesis and continued funding
of this project. I also cannot understate the help and knowledge I have received from
individuals at this company. Asa Weber, Don Foreman, Ronney Silva as well as many other
have constantly challenged me to be pragmatic and creative in my research methods.
Many other current and former graduate students have assisted with this project. My
colleague, Dr. Sanja Miskovic, was instrumental in the initial design of this project. I would
like to thank her for the long hours and heated discussions we shared concerning the initial
direction of this project. Also, I thank Michael Kiser and Erich Dohm for their assistance in
the lab.
I want to thank my family and close friends for their love and support during this
process. Finally, I thank my 여자 친구 Alice Lee (이 현아) for being a constant source of
joy and hope.
iii |
Virginia Tech | Chapter 1
Introduction & Background
1.1 Preface
Froth flotation is a chemio-physical separation process commonly used to enrich mining
products. Run-of-mine material usually consists of one or more valuable components, des-
ignated as ore minerals, mixed with a significant portion of waste components, designated
gangue minerals. Initial processing of the run-of-mine material seeks to physically sepa-
rate these components, so that the valuable minerals may be retained for further processing,
whiletheganguemaybeproperlydisposed. Frothflotationisacommonunitoperationwhich
separates constituents based on differences in surface wettability. Since nearly all minerals
exhibit some distinction in surface characteristics, froth flotation can, in theory, selectively
separate any mixture of liberated particles. This theoretical anticipation is pragmatically
corroborated in the diversity and magnitude of mineral separation circuits that incorporate
flotation operations. Froth flotation is commonly found in separation plants that process
copper, lead, zinc, tin, gold, silver, iron ore, silica, molybdenum, platinum group elements,
rare earth elements, phosphates, fluorite clays, and fine coal (Wills & Napier-Munn, 2006).
Outside of the minerals industry, flotation has witnessed recent alternative uses in waste wa-
ter treatment (Wang, Fahey, & Wu, 2005), algae harvesting (Phoochinda, White, & Briscoe,
2004; Lynch, Watt, Finch, & Harbort, 2007), and paper recycling (Bloom & Heindel, 1997;
Kemper, 1999; Gomez, Watson, & Finch, 1995).
The modern froth flotation process or the chemically-aided separation of minerals via
air bubbles was originally patented in 1905 (Sulman, Picard, & Ballot, 1905). This technique
was not the first to exploit contrasts in surface properties; however, earlier attempts proved
1 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
prohibitively expensive and useful in only a few applications (Lynch et al., 2007; D. Fuerste-
nau, 2007). Most prominently, the Bessel brothers of Germany developed processes to clean
graphite by adding oil to the slurry run-of-mine slurry and generating air bubbles by boiling
(Bessel, 1877). A later patent described the use of CO bubbles generated by the reaction
2
of lime rather than air bubbles (Bessel, 1886). Nevertheless, by 1905, most base-metals and
porphyry copper deposits were processed via simple gravity separation. The poor separation
performance and the ever-increasing ore complexity lead to substantial milling deficiencies
and lost revenue (Lynch et al., 2007). A large capacity, highly selective industrial process
was needed to ensure the survival of the world-wide base metal industry.
Given its inherent strengths, the froth flotation process fulfilled this role and quickly
grew to one of the most crucial metallurgical processes. With the development of selective
reagents by the 1930s, processing plants were beginning to use froth flotation as the sole sep-
aration process (Wills & Atkinson, 1991). Since then, the froth flotation process has become
in itself an interdisciplinary art and science, with pragmatic and fundamental research efforts
addressing chemistry, equipment, mathematical modeling, applied mineralogy, economics,
and separation metallurgy. Given its prominence in the economic production of base met-
als, several authors have defended the froth flotation process as one of the most significant
technological innovations of the 20th century (Klassen & Mokrousov, 1963; Napier-Munn,
1997; M. Fuerstenau, 1999; Lynch et al., 2007). Despite the apparent industrial success of
froth flotation, many aspects of the process are not well understood. Even today, much of
the development is driven by evolutionary adaptation driven from empirical success rather
than revolutionary innovation driven from fundamental insight.
1.2 Literature Review
1.2.1 Fundamentals of Flotation
Froth flotation is a physical separation process founded on select principles of applied
surface chemistry, namely surface wettability. As a result of unique chemical structures
and characteristics, every surface in nature exhibits a definable degree of surface wettability.
Wettable surfaces, labeled hydrophilic, exhibit strong intermolecular cohesive forces between
water and the material at the interface. From the macroscopic viewpoint, this property is
characterized by water droplets naturally spreading across the surface, attempting to maxi-
mize the interfacial area. Contrastingly, non-wettable surfaces, labeled hydrophobic, exhibit
weak cohesive forces, and potentially disjoining forces at the interface (Wills & Napier-
2 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
Hydrophilic Surface Hydrophobic Surface
θ θ
Figure 1.1: Depiction of hydrophobic and hydrophilic surface. (cid:18) = contact angle.
Munn, 2006; Berg, 2009; Pan, Jung, & Yoon, 2011). This property is characterized by water
droplets beading up at the surface, attempting to minimize the interfacial area. The degree
of wettability is described by the contact angle ((cid:18)). The contact angle measures the physical
angle formed when a liquid droplet rests on a solid surface. The angle is measured between
the solid and gaseous phases, through the liquid phase. Large contact angles indicate greater
degrees of hydrophobicity. Figure 1.1 depicts the contact angles of a hydrophilic surface and
hydrophobic surface.
Froth flotation uniquely exploits these contrasts in surface wettability. In a typical
operation, a mixture of gangue and mineral components are suspended in water as a slurry
and introduced into a stirred vessel. Concurrently, air is sparged into the vessel, creating a
highlyturbulent,threephaseenvironmentofair,water,andsolidparticles. Thisenvironment
produces collisions between the particles and the air. Particles that have been rendered
hydrophobic will selectively attach to the air bubbles (in an attempt to minimize interfacial
solid-liquid area), while the hydrophilic particles will remain suspended in the water. The
attachment of air bubbles to the hydrophobic particles causes a reduction in the apparent
density of the bubble-particle aggregates. Eventually the aggregates will attain positive
buoyancy and float to the top of the vessel. The rapid accumulation of the bubble-particle
aggregatesatthevessel’ssurfacegeneratesamineralizedfoam,orfroth. Thisfrothisremoved
(either by mechanical or natural flow mechanisms) producing a mineral rich concentrate. At
the bottom of the vessel, the processed slurry is removed, producing a gangue-rich tailings.
In industrial practice, this process is carried out in a continuous-flow state, with constant
feed addition, and continual concentrate and tailings removal.
3 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
In addition to concentrate recovery as a result of bubble-particle attachment (aka true
flotation), recovery also occurs by a secondary, non-selective process known as entrainment
(King, 2001; Wills & Napier-Munn, 2006). Fundamentally, entrainment occurs as fine par-
ticles (both valuable and gangue material) are carried into the froth with water. The degree
of entrainment has been shown to strongly correlate with the particle size and the volume
of water recovered, with finer particles and higher water flow rates correlating to increased
entrainment. Other factors, such as the froth structure, particle shape, and froth residence
time, also influence entrainment; however, these factors are less prominent. In practice, en-
trainment is perceived as a necessary deficiency that is to be minimized if possible (Vianna,
2011).
Other authors have also described a third, also non-selective, recovery mechanism,
known as entrapment (Gaudin, 1957; Wills & Napier-Munn, 2006; Vianna, 2011). Here,
particles are physically retained in the froth phase between valuable particles which are
attached to bubbles. Few authors comment on the portion of material recovered by this
mechanism, and few mathematical models explicitly include it.
1.2.2 Factors Influencing Flotation Performance
Macroscopically, flotation performance is controlled by many factors, which are broadly
classified in three categories: chemical, equipment, and operational (Klimpel, Hansen, &
Fee, 1986; Gorain, 2007). Chemical factors encompass variables pertaining to the solution
chemistry unique to each flotation system. The reagents used in flotation fall into several
distinct categories: collectors, frothers, pH modifiers, depressants, and activators.
While some minerals are naturally hydrophobic, most flotation systems can benefit
from carefully controlled chemical addition. Collectors are chemicals which enhance the
hydrophobicity of hydrophilic or mildly hydrophobic mineral surfaces. These chemicals are
usually organic molecules which commonly fall into the chemical families of monothiophos-
phates, dithiophosphates, xanthates, and others (Wills & Napier-Munn, 2006). Collectors
are usually classified by the type of ion (anion or cation) which provides the hydrophobicity,
and the degree of hydrophobicity is strongly related to the length of the hydrocarbon chain
(Bulatovic, 2007). Aside from cost purposes, collectors are usually dosed in low concen-
trations, though typically sufficient to produce a monolayer on the surface. Additionally
collector addition usually decreases selectivity, as the collector absorbs onto unwanted min-
erals. Further addition may actually reduce the recovery of valuable minerals, as multilayers
of collector form on the mineral surface (Wills & Napier-Munn, 2006).
4 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
Since mineral surface chemistry and consequently, collector absorption is strongly de-
pendentonpH,acidorbasemodifiersareusedtomaintaindesiredlevelsofacidity/alkalinity.
Fundamentally, most minerals retain a positive surface charge in acidic conditions and a neg-
ative charge in alkaline conditions. Furthermore, many minerals have a unique iso-electric
pH. Through pH control and appropriate collector selection, selective separation of many
simple mineral systems is possible. Typically, price and availability determine the selection
of a modifier; lime is a fairly common choice.
Activators and depressants are other reagents which respectively promote or inhibit the
flotation of select minerals by altering the degree of collector absorption. These chemicals
are common in sulfide flotation which usually involves selective flotation of several valuable
components.
Frothers are usually long chain surfactants which absorb to the air water interface.
These reagents reduce the air-water interfacial tension in order to promote smaller bubble
sizesandfrothstability(Bulatovic,2007). Asmanyflotationmodelsindicate,reducedbubble
size enhances flotation performance (Yoon & Luttrell, 1989; Gorain, Franzidis, & Manlapig,
1995a). Thereducedinterfacialtensionresultingfromsurfactantabsorptionmitigatesbubble
coalescence and allows the generation of a froth phase. Up to a designated concentration,
increased frother addition will enhance the effects (i.e. continued bubble size reduction).
However, once all the available interfacial area is saturated with surfactant, increased frother
addition will not result in increased performance. High frother dosages may produce highly
stable or metastable froth, which can adversely affect handling and downstream processing.
In addition to the reagent type and dosage, other chemical factors which influence
flotation performance include process water quality, solution temperature, and reagent con-
ditioning time. In general, the chemical condition of the flotation system has a pronounced
effect on the attainable separation performance of a flotation unit. Consequently, chemical
variables become more crucial as more complex separations are desired.
Equipment factors include the details of the flotation cell and the separation circuit
design. Many researchers have indicated that the hydrodynamic environment of the flotation
cell strongly influences the rate at which material floats (Luttrell & Yoon, 1991, 1992).
Consequently, the cell geometry, size, sparging mechanism, and power input all contribute
to overall flotation performance. Since this thesis is explicitly focused on specific details of
flotation machine design, these factors have been researched separately (Section 1.2.4) to
accommodate a more thorough analysis.
Operational factors include details of the flotation system which are unique to the daily
5 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
operation of each plant. These factors are generally associated with variable characteristics
of the ore, such as particle size, feed grade, and degree of feed oxidation. Additionally,
operational factors may include those which are to the discretion of the plant operator, such
as feed flow rate, residence time, and pulp density. In general, operational factors provide
interdependent and complex influences on the flotation performance. In example, the feed
particle size has a dramatic influence on the flotation efficiency. The classic curve shows a
substantial drop off in recovery for particles generally less than 20 microns and greater than
200 microns (Jowett, 1980), with the larger size limit being a function of particle density
(Jameson, Nguyen, & Ata, 2007). In general, smaller particles lack the inertia necessary
to penetrate the liquid film surrounding the bubble, while larger, heavier particles are more
susceptibletodetachmentandfrothdropback(Mao&Yoon, 1997; Sherrell, 2004; Do, 2010).
Furthermore, the nature of the particle distribution (such as the presence or absence of fine
particles) may have a marked influence on froth stability and thus recovery of material from
the froth.
1.2.3 Modeling Approaches
Given the multitude of contributing factors, comprehensive and purely theoretical flota-
tion models are still immature. Instead, empirical and partially phenomenological models
have been well vetted and used extensively for many simulation purposes. From a micro-
scopic perspective, the complex mechanics of froth flotation may be described by several
transport mechanisms. The most recent studies include the rate of pulp to froth trans-
port by bubble attachment, the rate of material drop back from the froth, the rate of water
drainage from the froth, and/or the rate of entrainment. Most modeling approaches attempt
to quantify the specific rates and interaction of these mechanisms.
Most simply, many researchers have empirically witnessed the kinetic behavior of bulk
flotation recovery as a function of time. This evidence has prompted many to model flotation
as a first-order rate process analogous to a chemical reaction (Sutherland, 1948; Tomlinson &
Fleming, 1965; Fichera & Chudacek, 1992). Other order rate models have been postulated,
but few have gained as much widespread applicability as the first-order model. The first-
orderratemodeldefinesaconstantproportionalitybetweenthedepletionofmineralparticles
(dN/dt) and the number of particle in the system (N):
dN/dt = kN (1.1)
where k is a proportionality or rate constant.
6 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
From the first-order assumption, Equation 1.1 may be solved at various boundary con-
dition to determine the recovery (R) as a function of flotation time (t) or retention time
((cid:28)) for both a plug-flow reactor (Equation 1.2) and a perfectly-mixed reactor (Equation 1.3)
(Levenspiel, 1998). These equations have been used to model the flotation process in scaling
from a laboratory to an industrial flotation unit:
R =
1(cid:0)e(cid:0)kt
(1.2)
k(cid:28)
R = : (1.3)
1+k(cid:28)
Several modifications to these models have been proposed to incorporate a theoretical
maximum recovery and a flotation delay time (Dowling, Klimpel, & Aplan, 1985; Gorain,
Franzidis, Manlapig, Ward, & Johnson, 2000; Sripriya, Rao, & Choudhury, 2003) . Ad-
ditionally, some researchers have suggested that industrial cells (especially column cells)
substantially deviate from the perfectly-mixed assumption (Dobby & Finch, 1988; Luttrell
& Yoon, 1991). Coinciding with the aforementioned chemical reaction analogy, an expres-
sion for recovery is proposed which incorporates the degree of axial mixing, via the Peclet
Number (Pe) (Levenspiel, 1998):
4Aexp(Pe/2)
R = 1(cid:0)
(1+A)2exp[(A/2)Pe](cid:0)(1(cid:0)A)2exp[((cid:0)A/2)Pe]
√
A = 1+4k(cid:28)/Pe:
While the general rate-based approach to flotation modeling has substantial empiri-
cal justification, researchers and practitioners have realized that not all particles of a given
mineral in a flotation system exhibit the same kinetics. This observation has led to the
development of distributed parameter rate models (Fichera & Chudacek, 1992). Various
researchers have identified properties to justify the distribution, with one of the more preva-
lent parameters being particle size. Gaudin, Schuhmann Jr, and Schlechten (1942) first
experimentally measured the dependence of flotation rate on particle size, noting the sub-
stantial degradation in flotation rate for large particles. This observation was later given a
more thorough theoretical consideration which investigated the streamline hydrodynamics
for given bubble and particle sizes (Sutherland, 1948).
A more general approach to model parameterization was conducted by Imaizumi and
Inoue (1965). This modeling approach considers distributed floatability classes which lump
together the combined effects of particle size, shape, and other surface properties. Most
7 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
contemporary flotation models include distributed flotation classes, often in the form of a
double distributed model which includes size and flotability (Fichera & Chudacek, 1992).
Further attempts to add fundamental insight to the empirical first-order observation
have led many to propose analytical expressions for the flotation rate constant. These
expressionsgenerallysuggestastrongdependenceofgasdispersionontheflotationrate. One
suchmodelsuggeststhattherateconstantisproportionaltothebubblesurfaceareaflux(S )
b
and a generic probability or collection efficiency term (P) (Jameson, Nam, & Young, 1977;
Yoon & Mao, 1996; Gorain, Franzidis, & Manlapig, 1997; Gorain, Napier-Munn, Franzidis,
& Manlapig, 1998):
k = 0:25PS :
b
Here, S is a derived term which defines the degree of aeration present in the cell (Finch
b
& Dobby, 1990; Gorain et al., 1997; Gorain, Napier-Munn, et al., 1998). S balances the
b
superficial gas velocity (J ) and the mean bubble size (d ):
g b
6J
g
S = :
b
d
b
This model has been very successful at normalizing flotation performance when the gas
dispersionvariablesareknown. Thelineark(cid:0)S relationshiphasbeenexperimentallyverified
b
for various minerals and at various scales (Gorain, Napier-Munn, et al., 1998; Hernandez-
Aguilar,Rao,&Finch,2005). Theoverallacceptanceinthismodelhasseveralcomprehensive
studiesincharacterizingandquantifyinggasdispersioninflotationcells(Finch,Xiao,Hardie,
& Gomez, 2000; Tavera, Escudero, & Finch, 2001; Kracht, Vallebuona, & Casali, 2005;
Schwarz & Alexander, 2006; Miskovic, 2011).
Other models have proposed a purely theoretical expression for k, based on surface
chemistry and hydrodynamic variables (Luttrell & Yoon, 1992, 1991; Mao & Yoon, 1997;
Sherrell, 2004; Do, 2010). These models were originally applicable for predicting rate con-
stants under quiescent conditions, such as in column cells. More recently, the fundamental
models have addressed the turbulent hydrodynamic conditions found in conventional cells.
Additionally, these models have added fundamental or semi-empirical models to describe
material drop back and fluid drainage from the froth. All of these fundamental models are
based on a compartment model which independently defines the flotation rate constant as a
combination of probabilities of collision (P ), attachment(P ), and detachment (P ):
c a d
k = PS = (P P (1(cid:0)P ))S
b c a d b
In these models, the probability terms have been analytically defined using fundamental
hydrodynamic variables (such as turbulent kinetic energy) and surface energies calculated
8 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
from the Van Der Waals, electrostatic, and hydrophobic force components. The extended
DLVO theory was invoked to define the composite interaction of these forces (Yoon & Wang,
2007; Kelley, Noble, Luttrell, & Yoon, 2012). Ultimately these models will predict flotation
performance as a function of intensive mineral properties and machine characteristics which
are either well known or do not change with scale (Kelley et al., 2012).
In addition to the aggregate recovery models, other recent studies have focused on the
inclusions of other transport mechanisms, such as froth recovery and entrainment. Such
models consider flotation to be a two stage process, modeling the pulp and the froth as
independent reactors. Most of the pure pulp recovery models invoke analytical forms similar
to the rate models presented above with some empirical correction to negate the ever-present
froth effects (Gorain, Harris, Franzidis, & Manlapig, 1998; Vera et al., 2002).
Similar to pulp recovery, froth drop back has been identified as a rate process which
can be modeled as a plug-flow reactor considering the interaction of a rate constant and
residence time (Equation 1.2) (Gorain, Harris, et al., 1998). When the independent froth
(R ) and pulp (R ) recoveries are known, the overall recovery may be calculated by (Finch
f p
& Dobby, 1990):
R R
f p
R = :
1(cid:0)(1(cid:0)R )R
f p
Since the identification of the two compartment flotation modeling approach and the
kinetics of froth drop back, researchers have attempted to gain further fundamental under-
standing, especially with regard to froth residence time (Vera et al., 2002). Most simply,
froth residence time can be determined by dividing the froth height by the superficial gas
rate for the cell ((cid:28) = h/J ) ((Mathe, Harris, O’Connor, & Franzidis, 1998). Since this
f g
calculation does not accommodate for different cell geometries and froth travel distances,
many have proposed revisions to the initial calculation, while retaining the kinetic plug-flow
model. Gorain, Harris, et al. (1998) suggest the inclusion of the distance from the center of
the flotation cell to the launder, while Lynch, Johnson, Manlapig, and Thorne (1981) base
the calculation on the volumetric slurry flow through the froth.
1.2.4 Flotation Equipment
Equipment variables which influence flotation performance include many aspects of the
flotation cell design including: the type of cell, the cell dimensions, the cell geometry, the
rotor and stator design, various clearances and rotor dimensions, rotor speed, specific power
input, state of air dispersion, the froth removal mechanism, and the circuit design.
9 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
Modern flotation machines are most generally classified into two types: conventional
cells and column cells. Conventional cells are mechanical agitated vessels which are aerated
by air introduction in the rotor region. Such aeration may occur due to the natural pressure
difference created by the rotor (self-aspirated or sub-aeration designation) or by the supple-
mentary action of a blower (forced-air designation). Column flotation cells are generally not
mechanically agitated vessels, exhibiting high aspect ratios in the vertical direction. Rather
than attempting to suspend slurry in the cell, new slurry is introduced near the top of the
vessel and flows downward to the tailings discharge located at the vessel’s bottom. Air is
typically sparged at the bottom of the vessel and travels upward to the froth phase. This
counter-current flow of slurry and air propagates bubble particle collisions. Additionally,
column cells are generally characterized by deep froths and the inclusion of froth washing.
In this action, fresh water is sprayed onto the top of the froth in an effort to reduce the en-
trainment of unattached gangue particles traveling with the water in the froth phase (Wills
& Napier-Munn, 2006; Finch, Cilliers, & Yianatos, 2007). Figure 1.2 compares conventional
and column cell designs.
The most prevalent manufacturers of flotation technology today include the major min-
eral processing providers, FLsmidth, Outotec, and Metso (Peaker, 2007; Weber & Tracyzk,
2007; Oravainen & Allenius, 2007). Each manufacturer features several unique machine
designs applicable for a variety of mineral industries. Across all manufacturers, the size of
flotation cells has grown exponentially since the 1960s. Figure 1.3 shows the increase of the
average flotation cell size reported in literature over the past century. As shown, common
sizes for industrial units today span up to 600 cubic meters, with the trends favoring larger
units in attempt to capitalize on economy of scale. According to the exponential fit of the
data, the size of large flotation cells doubles every 8.9 years.
Given the rapid increase in flotation cell size, one of the most challenging aspects of
machine design is scale-up from the laboratory and pilot-plant sizes to the industrial sizes.
Traditionally, process scale-up has been driven by operator experience and empirical “scal-
ing” factors (Tatterson, 1991, 1994). By the 1960s, dimensionless numbers and simple chem-
ical reactor analysis replaced purely heuristic scaling parameters. Table 1.1 lists several
dimensionless numbers commonly used in flotation scale-up.
While these criteria were able to appropriately scale solid suspension and mechanical
agitation, they proved only marginally effective for predicting scaled metallurgical perfor-
mance since they do not account for crucial parameters, such as bubble size (Gorain, 2007).
Currently, additional hydrodynamic expressions (Table 1.2), along with the common di-
mensionless numbers are used in flotation scale-up (Weber & Tracyzk, 2007). Furthermore,
10 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
Table 1.1: Common Dimensionless Numbers Used in Flotation Scale-up, after (Gorain, 2007;
Young et al., 2010).
Name Expression Description Range
Reynolds Number, Re
(cid:26)ND2 Ratio of inertial force to
(1(cid:0)7)(cid:1)106
(cid:22) viscous force
Ratio of inertial force to
Froude Number, Fr DN2 0:5(cid:0)5
g gravitational force
Ratio of resistance force
Power Number, P P 0:5(cid:0)5
N (cid:26)N3D5 to inertial force
Ratio of gas flow to
Aeration Number, N Q a 0:01(cid:0)0:2
A ND3 pumping capacity
Ratio of inertial force to
(cid:26)N2D3
Weber Number, We
(cid:27) surface tension force
Legend: (cid:26)= fluid density, N = rotor’s rotational speed, D = rotor diameter, (cid:22)= fluid viscosity,
g = gravitational acceleration, P = power input, Q = airflow rate, and (cid:27) = surface tension
a
the use of sophisticated analytical tools, such as discreet element method (DEM) modeling
and computational fluid dynamics (CFD) has become increasingly prevalent (Oravainen &
Allenius, 2007; Weber & Tracyzk, 2007).
With much of the machine’s influence being strongly tied to hydrodynamics and gas
dispersion, several studies have attempted to standardize the measurement of various gas
dispersion indices. The most widely utilized parameters include gas holdup ("), superficial
gas velocity (J ), bubble size distribution (d ), and bubble surface area flux (S ). Since the
g b b
mid-1990s and early 2000s, several analytical instruments have been developed to quantify
these parameters. Schwarz and Alexander (2006) present the details of a gas dispersion
database which includes measurements of these parameters collected over 10 years in over
800 industrial flotation cells of various sizes and in various mineral industries.
GomezandFinch(2007)provideacomprehensivereviewofthreeinstrumentsdeveloped
at McGill University over the previous ten years. The gas velocity sensor utilizes an inverted,
closed cylinder which is immersed in the pulp. As gas builds in the cylinder, the pressure
increased is measured and used to calculate the flow of gas entering the chamber. The air
holdupsensorfeaturestwocylinders, whichareimmersedintheflotationcell. Thefirst, open
cylinder, allows to flow normally, while the second cylinder has a closed bottom and contains
13 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
Table 1.2: Select Hydrodynamic Parameters Used in Flotation Cell design, after (Weber &
Tracyzk, 2007)
Name Expression Description
Specific airflow Q /V Airflow–cell volume
a
Circulation intensity Q /V Pulp circulation–cell volume
r
Liquid rise velocity Q /A Pulp circulation–draft tube area
r dt
Power Intensity P/V Absorbed power–cell volume
Bubble Surface Area Flux 6J /d Airflow rate–bubble size
g b
Legend: Q = airflow rate, Q= cell volume, Q = liquid recirculation rate,A = draft tube area
a r dt
P = power input, J = superficial gas velocity, d = mean bubble size
g b
un-aerated pulp. By measuring the conductivity difference between the two cylinders, the
amount of air in the open cylinder may be calculated. Finally, the bubble size measurement
technique is based on ex-situ sampling, image capturing, and image processing. A bubble
viewingchamberispositionedabovethecell, andacapillarytubeextendsbelowthechamber
into the flotation pulp. Bubbles in the flotation cell rise through the tube, into the chamber,
where they contact an inclined plane. This plane disperses the bubbles into a single depth
plane, and a video camera is used to capture images of the bubbles in the viewing chamber.
These images are later analyzed to determine the full bubble size distribution (Gomez &
Finch, 2007). This current method was shown to provide greater consistency and analytical
ease than earlier iterations (Chen, Gomez, & Finch, 2001) and earlier direct measurement
techniques developed at the University of Cape Town (Tucker, Deglon, Franzidis, Harris, &
O’Connor, 1994).
Miskovic (2011) and Miskovic and Luttrell (2012) critically reviewed the ex-situ bubble
sizing technique and proposed a new technique based on in-situ bubble size measurements
and an alternate image processing approach. These tests included pilot-plant and industrial
scale machines. Other authors have attempted to derive correlations which predict bubble
size and S as a function of impeller speed (Girgin, Do, Gomez, & Finch, 2006), from gas
b
14 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
holdup information (Finch et al., 2000), from various operational parameters and fitting
constants (Gorain, Franzidis, & Manlapig, 1999), from drift flux analysis (Banisi & Finch,
1994), and from artificial neural networking (Massinaei & Doostmohammadi, 2010).
In a conventional cell, the rotor is the predominant source of power for the entire flota-
tion system. This power provides the necessary energy for a number of flotation subprocesses
including: air dispersion, solid suspension, micro-turbulence generation, and bubble-particle
collision. A series of comprehensive experimental studies analyzed and compared several in-
dustrial rotors with respect to various gas dispersion variables (Gorain et al., 1995a; Gorain,
Franzidis, & Manlapig, 1995b, 1996; Gorain et al., 1997; Gorain, Napier-Munn, et al., 1998).
Each of these papers individually studied the effects of specific various gas dispersion vari-
ables, including bubble size distribution, gas holdup, and superficial gas velocity. Compar-
isons between the various rotor types included in the study were made with respect to overall
flotation performance. Most of the studies included tests of the Chile-X, Pipsa, Outokumpu,
and Dorr-Oliver rotors over a wide range of impeller speeds and gas velocities.
Given the date of the rotor studies, the authors chose to implement the U.C.T bubble
size analyzer which was the best available method at the time (Tucker et al., 1994). In
the end, the authors concluded that individual gas dispersion variables do not universally
correlate well to flotation performance; however, the derived bubble surface area flux (S )
b
parameter showed a consistent linear relationship with the measured flotation rate constant
at a given froth depth (Figure 1.4). The latter of these studies and later studies by the same
authors showed that froth depth has a non-linear effect on flotation performance (Gorain,
Napier-Munn, et al., 1998; Gorain, Harris, et al., 1998). With regard to rotor design, though,
theauthorsshowedthatwithinpracticalreasonallofthetestedrotorswereableofproducing
the same flotation rate at a given S value. Unfortunately, these studies and no other studies
b
on flotation rotors directly measure the power requirements of the rotor.
Nevertheless, outside of the froth flotation research, other studies have addressed power
consumption and gas dispersion in stirred vessels. Dohi, Takahashi, Minekawa, and Kawase
(2004)comparedthesolidsuspensioncapabilitiesofthreedifferentimpellerdesignsinaerated
vessels. Here, the state of off-bottom solid suspension and ultimately homogeneous solid
suspension were measured for both gassed and ungassed conditions. The authors developed
anempiricalcorrelationwhichrelatesthepowerrequiredforsuspensiontothepercentsolids,
the liquid density, the superficial gas velocity, the terminal settling velocity of the particles,
and a proportionality constant unique for each rotor. Another study investigated the effects
ofsolidconcentrationandparticlesizeonpowerconsumptioninsimplestirredvessels(Angst
& Kraume, 2006). The authors used an endoscope to examine the concentration of particles
15 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
at various locations in the tank, comparing the local solids concentration with the mean
solids concentration of the vessel. This study was conducted for various tank sizes in the
laboratory. The results show that the power required to suspend and homogenize the solids
in the vessel can be strongly correlated to the particle volume fraction. In yet another
study, the authors test several rotor designs and configurations in an attempt to maximize
the gas holdup while minimizing power consumption (Arjunwadkar, Saravanan, Pandit, &
Kulkarni, 1998). The authors tested several dual rotor combinations at various power inputs
and gas velocities. The results show that different combinations are capable of drastically
different results (maximum holdup values span from 3.0% to 7.5% at a given power input),
with the most favorable rotor set-up being a disk-turbine - pitched blade turbine downward
combination.
These results and other outside studies bring to light the lack of fundamental under-
standing and empirical investigations in power consumption relating to froth flotation rotor
design. As other studies in other industries have shown, different rotor designs are capable of
varying levels of energy-efficiency, depending upon the application. While solid suspension,
solution homogeneity, and gas dispersion are significant in flotation, these factors do not
necessarily directly correlate to metallurgical performance, as power input further affects
a number of flotation subprocesses, such as attachment and detachment rates. A focused
experimental study is needed to ascertain the role of power consumption on metallurgical
performance and the variations solely dependent on rotor design.
1.3 Research Objectives and Overview
The singular goal of this research is to characterize and compare laboratory-scale flota-
tion rotors in an attempt to provide future design criteria which can be used to develop
energy-efficient rotors. This goal was approached in three stages: (1) methodology develop-
ment and equipment design, (2) exploratory testing, and (3) detailed testing.
During the methodology development, equipment size limitations were imposed to re-
strict the breadth experimental options and appropriately scope the project within the
laboratory-scale paradigms. As a result, the term, “laboratory-scale,” in this study refers
to rotors measuring less than 3 inches in diameter and corresponding tank volumes up to
30 liters. Furthermore, the appropriate equipment and assessment criteria were established
during this stage.
The exploratory experimental campaign prioritized the breadth of the analysis in terms
17 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
of rotors examined. During this phase, 14 rotors were assessed in a battery of tests. Due
to the rapid and iterative development during this stage, not all prototype rotors were
subjected to an identical test matrix. Some rotors were abandoned prior to prolonged tests,
while others were developed after experimental procedures were refined. From the overall
appraisal of this exploratory campaign, the field of prototype rotors was narrowed to the
three designs possessing the best all-around performance.
These final three rotors were then subjected to the second, detailed testing campaign
which alternatively prioritized analytical depth, data reproducibility, and error quantifica-
tion. From the collaborative analysis of the two testing campaigns, future rotor design
criteria was established.
In summary, the itemized objectives of this study were to:
• Design and construct equipment that can assess small-scale forced-air flotation rotors
and prototypes;
• Developamethodologytoprovideconsistentandfairappraisalofdifferentrotordesigns
at the laboratory scale;
• Assess the existence of variation in flotation performance as a function of the rotor
design;
• Quantify the performance differences through detailed testing and analysis
• Provide empirically-based recommendations and criteria for flotation rotor design.
1.4 Document Organization
The body of this thesis is organized into five chapters, with the primary works presented
individually as standalone papers describing a separate phase or objective of the work. The
three main phases of the work (experimental development, exploratory testing, and detailed
testing) constitute the three informative chapters, while an introductory and a concluding
chapter complete the thesis. References are listed individually for each chapter.
Chapter1includesanoverviewoftheflotationprocess, generalbackgroundinformation,
and a full description of the work completed as a part of this study.
Chapter 2 provides descriptions of the design, construction, and commissioning of the
18 |
Virginia Tech | CHAPTER 1. INTRODUCTION & BACKGROUND
laboratory-scale experimental equipment as well as descriptions of the experimental method-
ology developed during this research.
Chapter 3 describes the experimental results of the exploratory testing campaign. Both
Chapters 3 and 4 are presented in parallel structure providing tabular and graphic results of
investigations involving power consumption, gas dispersion, operational limits, and flotation
performance. This study provides a cursory evaluation of 14 rotors, while highlighting the
veracityoftheexperimentalapproachandtheexistenceofuniquedistinctionsbetweenvastly
different rotor designs.
Chapter 4 similarly provides results of the detailed testing campaign. This study seeks
to quantify the actual performance differences between rotor designs by incorporating a
more thorough experimental matrix and replicate testing. Consequently, only three rotors
were selected from the original 14 for this study. The rotor selection was based on the best
all-around performance measured in the original exploratory testing.
Chapter 5 includes a brief summary of all of the experimental studies, holistic conclu-
sions derived from the studies, and recommendations for future or continued development.
1.5 Bibliography
Angst, R., & Kraume, M. (2006). Experimental investigations of stirred solid/liquid sys-
tems in three different scales: Particle distribution and power consumption. Chemical
Engineering Science, 61(9), 2864–2870.
Arjunwadkar, S., Saravanan, K., Pandit, A., & Kulkarni, P. (1998). Optimizing the im-
peller combination for maximum hold-up with minimum power consumption. Biochemical
engineering journal, 1(1), 25–30.
Banisi, S., & Finch, J. (1994). Technical note reconciliation of bubble size estimation
methods using drift flux analysis. Minerals Engineering, 7(12), 1555–1559.
Berg, J. (2009). An introduction to interfaces & colloids: the bridge to nanoscience. World
Scientific Pub Co Inc.
Bessel, G. (1877). Berlin patent 42.
Bessel, G. (1886). Berlin patent 39,369.
19 |
Virginia Tech | Chapter 2
Development of an Experimental
Methodology for Rotor Testing
(ABSTRACT)
In the past, laboratory-scale flotation testing has been reserved for ore characterization
and reagent appraisals. This manuscript describes a consistent methodology for the normal-
ized evaluation of small-scale flotation rotors. These evaluations focus on four performance
factors which influence machine design and scale-up procedures: (1) power consumption,
(2) air dispersion, (3) operational robustness, and (4) flotation performance. These criteria
have shown marked influence on cost and metallurgical performance at the industrial scale,
and their translation to the laboratory scale is supported via theoretical considerations. In
this paper, restrictions on the size and nature of the equipment have been imposed so that
the experimental methodology coincides with the accepted paradigms of laboratory flotation
analysis. The required equipment includes items that may already be present in a work-
ing flotation laboratory or items that may be easily procured from laboratory equipment
vendors. Additionally, the flotation rotors measure less than 3 inches in diameter, and the
tank volumes fall between 9 and 30 liters. This paper presents the required equipment and
measurement devices along with the experimental and analytical procedures.
2.1 Introduction and Scope
To date, few studies have thoroughly addressed flotation rotor design at the labora-
tory scale. Most researchers over the past 50 years have generally reserved laboratory-scale
26 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
testing to characterize the ore, evaluate reagent interactions, and determine grinding sizes,
leaving the concerns of equipment design and plant operation to larger scale testing and
experienced-basedscalingfactors(Gaudin, 1957; Wills&Napier-Munn, 2006; Crozier, 1992).
With regard to the rotor design, a series of studies by Gorain and others assessed various
gas dispersion variables, including air holdup, bubble size distribution, and superficial gas
velocity, in several pilot-scale machines (Gorain, Franzidis, & Manlapig, 1995a, 1995b, 1996,
1997; Gorain, Napier-Munn, Franzidis, & Manlapig, 1998). While these studies incorporate
different rotor configurations, the research goals did not include the direct comparison of
different rotor designs. Rather, the researchers wanted to assess the extent of the gas dis-
persion variables, including the rotor and other operational factors as independent variables.
The authors admittedly did not necessarily operate the machines under the recommended
settings in an attempt to cover the widest range of operational conditions. Furthermore,
these studies did not consider the influence of power consumption, as the parameter was not
measured or reported. Such an omission negates any meaningful comparison between the ro-
tor designs, as power consumption strongly influences flotation performance and operational
cost.
While the Gorain studies and others emphasize the role of gas dispersion in flotation,
other performance factors have been attributed to the rotor design, but lack comprehen-
sive evaluation. One such factor is the state of solid suspension or mixing in the vessel.
Zwietering (1958) performed one of the first experimental evaluations of solid suspension
in stirred vessels primarily intended for the chemical mixing industry. The author defined
the just-suspended parameter (or critical impeller speed) as the minimum impeller rotational
speed required to suspend all of the solid particles from the vessel’s floor. An analytical
expression was empirically derived to predict the critical speed from various dimension-
less equipment and operational parameters unique to the mixing system. Similar studies
have been applied to flotation cells in order to develop similar empirical correlations which
incorporate additional, industry-specific parameters such as cell shape and forced-air rate
(Westhuizen & Deglon, 2007; Lima, Deglon, & Leal Filho, 2009). In each of the afore-
mentioned studies, the critical speed expression includes a proportionality constant which
accounts for the specific rotor design. Few studies include an exhaustive treatment of the
power required to suspend particles.
In light of the lapses in flotation rotor design characterization and evaluation, the initial
objective of this study is to develop a methodology for the fair comparison of different forced-
air flotation machine designs at the laboratory scale. Unfortunately, given the large number
of independent and interdependent variables which affect flotation performance, this task re-
quires due consideration. Different machine designs perform optimally in different geometric
27 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
and operational conditions. Furthermore, the drastic difference in scale between the labo-
ratory and the plant may magnify some changes in machine performance while suppressing
others. As a result, this task first seeks to establish normalized testing procedures which may
provide reliable and beneficial data on flotation equipment design at the laboratory-scale.
With consideration from theoretical principles and industrial experience, four principle
factors which influence flotation machine design and scale-up were selected for the perfor-
mance evaluations (1) power consumption, (2) gas dispersion, (3) operational robustness,
and (4) flotation performance.
When applicable, the measured performance indicators were normalized on the basis of
power input. Since factors such as rotational speed and superficial gas velocity have an inter-
activeandinterdependenteffectonflotationperformance,normalizingresultsbytheseinputs
may prove deleterious and misleading. Instead, power consumption represents a common
factor, and a significant flotation equipment cost that is easily measured at the laboratory
scale. Consequently, prior to any other performance analysis, a detailed understanding of
the rotors’ power draw at various speeds and gas rates is necessary.
Throughout this introduction and the literature review (Section 1.2), the role of gas
dispersioninflotationperformancehasbeendefendedtheoreticallyandvalidatedexperimen-
tally by numerous authors. While the measurement methods may change at the laboratory-
scale, the surrogacy of gas dispersion to flotation performance is well-formulated, experimen-
tally vetted, and proven scalable.
In this thesis, operational robustness refers to the range of operational conditions (in
terms of rotational speed and air velocity) that a rotor can stably operate without induc-
ing solid sanding or gas flooding. Both particle suspension and flooding are functions of
rotational speed and air velocity, and both represent a marked digression from standard
equipment operation. The extent that a given rotor can operate without inducing either of
these conditions marks the freedom the operator has in changing the operation to meet a
secondary process goal (i.e. a process control response to a downstream indicator).
Finally, differences in flotation performance between the rotors should be directly mea-
suredtoensurethatthesurrogatemeasurements(i.e. gasdispersion, fluidmixing)areindeed
valid indicators of a rotor’s ultimate capability. The lack of correlation between surrogate
variables may indicate that other unknown phenomena are not influencing flotation behavior
and should prompt more extensive study to undermine additional effects.
This chapter first describes the experimental and analytical equipment that was pro-
cured, constructed, and commissioned in order to evaluate the desired performance criteria.
28 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
These components are described both as independent devices and as collective flotation test-
ing systems. Next, the experimental and analytical procedures needed to determine each of
the performance criteria are described. Finally, the conclusions provide the context of these
procedures in the larger spectrum of complete equipment design and evaluation. These ex-
perimental and analytical procedures were continually revised throughout the experimental
phases while balancing testing ease with data reliability. The desired results consistently
revealed more information on the four principle performance criteria.
2.2 General Experimental Equipment
2.2.1 Rotors
The primary independent variable examined during this research was the flotation rotor
design. Throughout the research project, 15 rotors were developed and tested. Each of these
15 rotors was subjected to at least one laboratory test; however, not all rotors were equally
tested. Given the iterative and rapid development phase, some rotors were abandoned prior
to experimental procedure revision, while others were developed after less refined procedures
were abandoned. This author acknowledges this marked lack of consistency but still chooses
to present all of the available data in order to provide the most exhaustive information.
Physically, all rotors are 2.75 inches (7 centimeters) in diameter and between 1.25 to
2.0 inches (3.2 and 5.0 centimeters) tall. The initial rotors were manufactured from various
materials, primarily stainless steel and aluminum via cast molding, while the later rotors
were procured from the rapid prototype manufacturer ZoomRP.com. These rotors were
constructed by a 3D printer using PolyJet White production material (ZoomRP.com, 2011).
While the casted rotors show greater durability, the rapid prototyping process allowed more
innovative shapes and design components. All of the rapid prototyped rotors were able to
withstand consistent operation at rotation speeds up to 1800 RPM and torque loads up to 60
N-cm, while passing solid concentrations of 40% and particle sizes up to 200 microns. Rotor
components and webs less than 2 mm were most prone to damage, especially in routine daily
handling and transport.
Table 2.1 summarizes the rotors, dimensions, providers, and short abbreviations. These
abbreviations will be used to identify individual rotors throughout the remainder of this
report. Table 2.2 describes the technical details used to categorize and identify the rotors.
Figure 2.1 shows a typical rotor.
29 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
Figure 2.2: Generic stator used in laboratory testing [Photo by Aaron Noble:3/29/2010].
2.2.2 Stators
While synergies between the rotor and stator are postulated, the scope of this research
is relegated only to the rotor design. As a result, generic stators were used in combination
with the various rotor designs, rather than unique stators for each rotor. In the experimental
phases,twostatorswereutilized. Initially,atop-ringmountedstatorwasusedincombination
with the rotors. Like many of the prototype rotors, a second stator was also developed via
rapid prototyping (Figure 2.2). This stator is similar in size to the original; however, the
blades are oriented vertically and are mounted to a ring on both the top and bottom. This
extended lower opening provides compatibility for many of the larger rotors that do not fit
within the slope of the original stator. Like the prototype rotors, this stator was procured
from the rapid prototype manufacturer ZoomRP.com and constructed by a 3D printer using
the PolyJet White production material.
2.2.3 Tanks
Several flotation tanks geometries were designed and fabricated for use in the experi-
mental phase. The most frequently used tanks were cylindrical tank cells ranging in volume
from 9 to 35.3 liters. The technical details of these tanks are summarized in Table 2.3. Each
of the three primary tanks was designed and utilized for specific testing purposes. Silva et
al. (2012) has suggested that the appropriate “scaled-down” size for the 2.75 inch rotor is the
35.3 liter, 14 inch diameter tank. Nevertheless, the smaller tank size adequately reflects an
oversized-rotor design which is common in some industries. Furthermore, the fully transpar-
32 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
Table2.3: Summaryoftechnicalinformationregardingflotationtanksincludedinlaboratory
testing.
Parameter Tank 1 Tank 2 Tank 3
Nominal Size 9.75 in RT 10 inch RT 14 inch RT
Shape Cylinder Cylinder Cylinder
Construction 1/4 in. PVC with 1/4 in. Trasparent 1/8 in. Trasparent
Material Window Lexan Polycarbonate
Diameter (in.) 9.75 10 14
Fill Height (in.) 7.36 7.25 14
Volume (L) 9.00 9.33 35.3
Volume (gal) 2.38 2.46 9.32
Launder Style Batch Weir No Launder Radial Launder
Lip Length (in.) 5.0 – 44
Flow Style Batch Batch Continuous
ent tanks (10 inch RT and 14 inch RT) allow uninhibited visual inspection of the flotation
region from any angle while in operation. This feature allows convenient investigation in
solid suspension and air holdup determination. Alternatively, the viewing window in the
9.75 inch RT only provides limited visual information on the tank’s water level.
The launder design also dictates which cells may be used in the flotation performance
evaluations. The 9.75 inch RT has a batch-style overflow weir. The weir has a total lip
length of 5 inch and is located 9.75 inches from the tank bottom. Alternatively, the 14 inch
RT has an external radial launder which extends along the full periphery of the tank cell.
The launder bottom slopes downward at 15(cid:14) expelling the froth concentrate at a single port.
The 10 inch RT has no launder.
Furthermore, the 14 inch RT has been outfitted with openings for feed and tailings.
These inlets are approximately one inch from the cell bottom, and they allow the cell to
be operated in a continuous flow state. The tailings port includes a gate valve which may
be closed to operate the cell in a batch condition. The other tanks do not include inlets
and must always be operated in batch conditions. Pictures of these three cells are shown in
Figure 2.3.
33 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
(a) 9.75 in. RT (b) 10 in. RT (c) 14 inch RT
Figure 2.3: Pictures of flotation tanks used in laboratory testing [photos by Aaron No-
ble:10/24/2009; 5/24/12; 7/30/2010].
2.2.4 Mixing Devices and Power Measurements
Throughout the testing campaigns, modified overhead laboratory mixers were used to
drive the flotation rotor. These mixers included variable speed control and an integrated
torque meter. During the exploratory testing campaign, a Heidolph RZR 2102 Control
electronic laboratory mixer (Figure 2.4a) was procured and utilized. This model allows
speedsrangingfrom40to2000RPMandtorquemeasurementsfrom0to47N-cm. Typically,
the torque limit was the limiting factor for this mixer. Few rotors could be operated at tip
speeds in excess of 4.5 meters per second (1200 RPM) without overloading the mixer’s
capabilities
To accommodate the higher torque required at high tip speeds, a second mixer (1/5
horsepower Caframo brushless –Figure 2.4b) was procured for the detailed testing campaign.
Thismixer’smaximumcapacityis70N-cmat1800RPM;however, periodictorqueoverloads
may be sustained for short periods of time. Of the available rotors, most were able to achieve
amaximumtipspeedof6.5meterspersecond(1776RPM)whileremainingwithinthetorque
limitations.
A custom one inch diameter shaft (Figure 2.5) was designed and mounted into the
mixers via a 3/8 inch chuck. This shaft includes a hollow center and a low-friction bearing
which allows compressed air to flow from the laboratory’s compressed air source through
the rotor. Additionally, the shaft features screw mounts at the mixing end which allow
interchangeable rotors.
34 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
2.2.5 Bubble Size Analysis
Bubble size measurements were completed using a visual method similar to the one
described by Hernandez-Aguilar, Gomez, and Finch (2002). This method uses a bubble
samplingdevice, consistingofacapillarytubeandachamberwithaslantedviewingwindow.
Bubbles recovered from the flotation tank rise through the tube and form an approximate
single layer on the slanted window, where they are captured by high speed photography. For
most laboratory-scale testing, the sampling tube was positioned along the wall of the tank
slightly above the stator. At least 2,500 individual bubbles were analyzed, depending on the
operating conditions.
Image acquisition was achieved by a GP-21400 GEViCAM with a HR9HA-1B Fujinon
9 millimeter lens. Controlled by automated computer input, the camera can produce 5
megapixel images with a maximum acquisition rate of 30 frames per second. Initially, a
200 watt halogen lamp was used as an illumination source, to generate a nearly consistent
background. This illumination source was later upgraded to an LED light board.
2.2.6 Air Flow Measurements
For the exploratory testing, air flow to the flotation tank was measured by a calibrated
stainless steel variable-area flowmeter. The capacity of the flowmeter ranges from 2,182 to
90,454 milliliters per minute. For the given tank geometries, these flow rates correspond to
superficial gas velocities ranging from 0.07 to 3.00 centimeters per second. For the detailed
testing, a larger capacity, direct reading variable-area flowmeter was procured. The capacity
ofthisflowmeterrangesfrom1.0to4.0standardcubicfeetperminuteorsuperficialvelocities
of 0.5 to 2.0 centimeters per second for the given tank geometries.
2.2.7 Batch Flotation System
Thebatchflotationsystem(Figure2.6)wasusedexclusivelythroughouttheexploratory
testing campaign. The components of this system include the low-torque RZR mixer, the
custom bubble size sampler, the low air meter, and either the 9.75 inch RT or the 10 inch
RT depending on the specific test. Other than the standard ring stand for the mixing unit
and the bubble size analyzer, no special mounting or equipment was needed to support the
system. This basic setup allows rapid cleanup and straightforward modification in terms of
rotor or material change. However, as pictured, this setup does not allow visual access to
36 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
The 14 inch RT was the primary flotation tank used with the continuous flotation
system. As mentioned above, this tank includes inlets/outlets for the feed and tailings
streams as well as a radial concentrate launder. Stator assemblies were mounted onto the
tank floor with glue. In the continuous system, the flotation tank rests on a clear Plexiglas
shelf. This setup extends visual access to the tank bottom, allowing simple assessment of
off-bottom solid suspension. Furthermore, the Caframo mixer was exclusively utilized in the
continuous flotation system due to the higher torque capacities.
Both the tailings and the concentrate streams are returned to a central sump prior to
being recycled into the flotation tank. The sump is a standard 15 gallon, conical-bottom
polyethylene tank. In order to ensure complete mixing of the product streams, agitation is
supplied to the sump by a 0.5 horsepower DC powered variable speed mixer. Material from
the sump is returned to the flotation tank by a 0.5 horsepower positive displacement feed
pump. A bypass valve following the pump allows a fraction of the material to be returned to
the sump, thereby controlling the apparent feed rate to the tank. All piping from the sump
to the flotation tank is 0.5 inch PVC.
2.3 Experimental and Analytical Procedures
2.3.1 Power Measurements
Both mixing units used in these studies include integrated torque meters and digital
speedcontrols. Bothunitsdisplaythecurrenttorqueandthecurrentspeedviadigitaloutput
on the faceplate. Power measurements were determined by first recording the rotational
speed (N in RPM) and the torque (T in N-cm) directly from the mixing unit. The power
(P) is then calculated by:
2(cid:25)(N[RPM])(T[N (cid:0)cm])
P[watts] = : (2.1)
6000
Per the procedures in the mixer’s documentation, the device was calibrated daily for
each rotor, prior to any measurements. To calibrate, the mixing unit was first turned on
and run at full speed for 20 minutes in open air with the desired rotor attached. After the
unit had properly warmed up, torque measurements were taken at the speeds desired for
subsequent testing. These “open-air” torque measurements were later subtracted from the
test readings in order to mitigate the latent torque from the rotor’s weight and the friction
in the mixing drive. In this report, the initial studies of power consumption in both testing
39 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
campaigns were used to determine the rotors’ power curves and the power numbers (Sections
4.2 and 5.2). These tests were performed in a water-only solution, without solid particles,
reagents, or forced air. Power measurements for other tests (air holdup, flotation, etc.) were
measured independently during each specific test.
Withrespecttoprecision,bothmixersweredeliveredwithacalibrationcertificatewhich
verifies the speed measurements within (cid:6)1% and the torque measurements within (cid:6)5%.
Furthermore, preliminary testing was conducted to determine the repeatability of power
measurements in specific test conditions (rotational speed and superficial gas rate). The
effect of solid density and slurry temperature were not studied during these tests. Reagent
dosage was studied but was found to have no significant effect on power measurements over
the conditions under investigation.
During the repeatability testing, a specific test condition was set (air velocity and
rotational speed), and the machine was given several moments to accelerate to the desired
condition. Once at steady state, the speed and torque were recorded, and the power was
calculated for that condition. Each condition was repeated three times, and an average was
determined from the three measurements. The tip speeds and air velocities covered the
full range of anticipated test conditions (3 to 6 m/s tip, 0 to 2 cm/s air), producing 385
independent measurements. To interpret this data, each individual measurement’s deviation
from the mean (or relative residual) was determined as a percentage of the mean value for
the specific test condition: ( )
x(cid:0)x
Residual = 100
x
where x is the measured value at a specific test condition and x is the average of three values
at that test condition.
Residuals for each measurement were plotted against the measured power, and a fre-
quency distribution of the residuals was generated (Figure 2.8). As shown in the frequency
distribution, the residuals are normally distributed around zero with a standard deviation
(shownbythereddashedlineinbothplots)of7.58. AsinterpretedfromFigure2.8, anotice-
able correlation exists between the residual and the magnitude of the power measurement.
The degree of measurement variability is most severe in those readings with a relatively low
power measurement. Most of the data outside of the one single standard deviation limit
was recorded at power measurements below 60 Watts. Similar analyses show no noticeable
correlation between tip speed or gas velocity and measurement variability (Figure 2.9).
40 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
2.3.2 Holdup Measurements
As widely documented in the literature, gas dispersion is a key variable which influences
flotation performance. One holistic measure of the gas dispersion and available gas surface
area is the air holdup. While many researchers have devised methods of local air holdup
(Grau&Heiskanen,2003; Gomez&Finch,2007; Schwarz&Alexander,2006),theequipment
is usually too obtrusive to conveniently provide data in the laboratory scale. Alternatively,
the global air holdup provides easy-to-obtain, yet powerful comparative data for flotation
machine design.
Global air holdup was determined by one of two methods. During the exploratory
test campaign, the simple method was used. Air holdup was determined by first setting
the desired rotational speed with the air flow turned off. The liquid level was noted, and
the air flow was initialized. Once the system stabilized at a steady state in terms of power
draw and new liquid level, the torque and rotational speed were recorded, and the increase in
water level was measured manually (Figure 2.10). From the torque and speed measurements,
the power draw was calculated according to Equation 2.1. The global air holdup (") was
determined from the measured increase in water level (h ) and the original water level (h )
g t
by: ( )
h
g
" = 100 :
h +h
t g
During the detailed testing phase, global air holdup was determined by the displaced
volume method. Since the detailed testing utilized the 14 inch RT tank which has a radial
launder, water displaced from the cell may be collected and measured. During these tests,
the tank was first filled to the point of overflow. The desired speed was set with the air
disengaged, and any volume displaced by the rotational action of the impeller was noted.
The air flow was then initiated to the desired level. All the displaced water was collected in
the launder, and the volume was recorded along with the torque and rotational speed. The
global air holdup was calculated from the tank volume (V ) and the displaced volume (V )
t d
by: ( )
V
d
" = 100 :
V
t
All holdup tests were performed in the absence of frother.
43 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
(a) No Air (b) With Air
Figure 2.10: Simple method of air holdup determination [photos by Aaron Noble:
10/12/2009].
2.3.3 Operational Limits Determination
Inthisstudy,the“operationallimits”or“operationalrobustness”refertotwoconditions
which prevent normal operation of a flotation cell: the sanded condition and the flooding
condition. Thesandedconditionreferstoadegreeofincompletesolidsuspensioninthetank.
Insufficient agitation from the rotor will cause solids in the tank to settle and accumulate on
the tank bottom. The degree of agitation is enhanced by increasing the rotor tip speed but
mitigated by increasing the gas velocity. Other operational factors, such as solid density,
slurry viscosity, and particle size also have an effect on solid suspension (Zwietering, 1958;
Lima et al., 2009).
The degree solid suspension can be characterized by many factors (apparent fluid den-
sity, visible bed depth at the tank wall, rotor power draw, etc.); however, prior studies have
shown that the most indicative measurement is the minimum speed (or power) needed to
initiate off-bottom suspension(Zwietering, 1958; Lima et al., 2009). Off-bottom suspension
refers to the absence of persistent settling on the tank bottom. This condition is most easily
assessed by visual access to the tank bottom; however, manual observation with a probe
serves as a reasonable substitute when visual access is not available.
During the exploratory testing, the solid suspension tests were conducted in the 10 inch
44 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
RT which is made of transparent Lexan. While visual access was not available to the tank
bottom, solid accumulation at the wall was noted, and a simple probe was used to assess the
degree of solid settling on the tank bottom. During the detailed testing, visual access was
available to the tank bottom, so solid suspension was affirmed by simple observation. During
both testing campaigns, the minimum speed (and power) required for solid suspension was
determined as a function of rotor selection and air speed. By this measurement technique,
the degree of pumping loss as a result of rotor aeration may be ascertained.
Various materials of different densities and mean particle sizes were used throughout
the solid suspension tests. Some of the more prevalent materials include: course grained
silica beads and iron ore rougher feed. However, since this study prioritizes the comparative
analysis of rotor design, the effect of material and operational parameters on solid suspension
was not emphasized or fully investigated.
The second operational limit, the flooding condition, is defined as the maximum gas
volume that a rotor can disperse at a given rotational speed. Once this limit is exceeded,
very large pockets of air begin to escape the rotor region and are expelled at the surface
near the shaft. Colloquially, this condition is commonly referred to by a number of terms:
boiling, geysering, surface turbulence, etc. In this thesis, the term “boiling” may be used to
indicate rotor flooding; however, use of this term is purely semantic and not indicative of an
actual phase change occurring in the liquid.
For a given air velocity, the flooding condition may be averted by increasing the rota-
tional speed. For a given rotation speed, the air velocity must be reduced until the flooding
is halted. For both testing campaigns, the flooding limit of each rotor was determined as a
function of rotational speed. During the tests, a speed was designated, and the air velocity
was increased until the flooding condition was first observed. Both the speed and air velocity
were noted prior to setting a different test condition.
Because this test attempts to quantify descriptive information, the data are prone to
subjective influence. Nevertheless, the comparative analysis of data from a single experi-
menter will provide a reliable basis for evaluation of different flotation rotors.
2.3.4 Batch Flotation Testing
During the exploratory testing campaign, single-species batch flotation testing was used
to determine the relative performance of each rotor, in terms of flotation recovery. In addi-
tion to the first-order rate constant, the mean bubble size and the power draw were recorded
45 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
for each rotor in order to standardize the results on the basis of energy efficiency and bub-
ble surface area efficiency. During this phase, each rotor was tested at multiple rotational
speeds, while the air velocity and the chemical conditions were held constant. By varying
the rotational speeds, each test produced a different power draw and a different bubble size.
The resulting data produced two comparative plots for each rotor: flotation rate constant
as a function of power draw and flotation rate constant as a function of bubble surface area
flux.
The procedure for standard batch flotation kinetics tests has been described thoroughly
in the literature (Crozier, 1992; Fichera & Chudacek, 1992; Wills & Napier-Munn, 2006).
In general, a batch kinetics test requires timed samples to be taken as floated material is
removed from the cell. By weighing these samples as well as the tailings material, cumulative
recovery can be determined as a function of time. In the exploratory testing, flotation tests
were conducted up to a maximum time of 4 minutes, with the first sample being taken at
10 seconds and progressive samples at longer time intervals (10, 20, 30, 60, 120, and 240
seconds). After completing each test, the samples were dewatered by a vacuum filter and
dried in a 100(cid:14)C oven overnight.
The material used in the exploratory flotation tests was A-Series Technical Quality
glass spheres (nominal size 35 microns) procured from Potters Industries (Potters Industries,
Incorporated, 2001). These samples were hydrophobized in a 4 (cid:1) 10(cid:0)6 molar solution of
dodecylamine (DDA) collector. Prior to any test work, a large amount of collector solution
was made by completely dissolving the lot DDA in pure ethanol. The collector for each test
was appropriated from this same solution to ensure similarity between the tests. In addition
to the DDA collector solution, MIBC was added as frother. Both of the reagents along with
the glass spheres were added to the flotation cell and agitated for two minutes without air.
This conditioning time immediately preceded the flotation tests.
From the collected cumulative recovery data, flotation rate constants (k) were deter-
mined for each test. Considering first order reaction kinetics, the rate of change of material
in the cell (dN/dt) is directly proportional to the amount of material in the cell (N). The
proportionality constant (k) is defined as the flotation rate constant:
dN
= kN:
dt
By solving the differential equation for the batch boundary condition (N(0) = 1 or
R(0) = 0) and recognizing that R = 1(cid:0)N, the expression simplifies to (Levenspiel, 1998):
ln(1(cid:0)R) = kt:
46 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
Thus, by plotting ln(1(cid:0)R) against t, the rate constant can be determined by the slope
of the resulting straight line. Typically, flotation material exhibits a distribution of rate
constants which is conveniently discretized into fast-, slow-, and non-flotation components as
indicated by separate linear portions of the ln(1(cid:0)R) vs. t plot. For ease of comparison, only
the fast kinetics were evaluated during this portion of the investigation. Consequently, the
bubble size and power measurements were also taken during the time period corresponding
to fast kinetics (less than 30 seconds into the test).
2.3.5 Continuous Flotation Testing
To supplement the batch testing, continuous flotation tests were conducted during the
detailed campaign. Since most laboratory tests are conducted in a semi-batch fashion, flow
characteristics and process dynamics inherently contrast the continuous processes seen in
full-scale plants. Also, since batch tests are, by definition, never at a steady-state, variables
such as reagent concentration and pulp density are constantly and uncontrollably changing.
These fluctuations result in continual variances in other dependent variables, such as bubble
size and power draw. While results from batch tests do carry comparative meaning, more
consistent and applicable data can be derived from continuous testing.
A well-defined continuous flotation testing and sampling routine was developed to pro-
mote consistency between the individual rotor tests. Similar equipment and procedures have
been described in the literature (Welsby, Vianna, & Franzidis, 2010). The order of reagent
preparation, sample preparation, operation, and sampling was strictly established to provide
the most non-invasive measurement strategy, while minimizing disturbances to the flotation
cell and maximizing data reproducibility.
During the tests, the rotor mixer was first initiated and given 20 minutes to warm up
prior to calibration (See Section 2.3.1). While the mixer was warming up, reagents and
flotation samples were prepared. The sump was then filled with municipally supplied water,
the flotation sample, and the reagents solution. The feed pump was turned on and diverted
sampling bucket/cylinder. After taking a preliminary residence time measurement, the feed
rate was adjusted and resampled until at the desired level. Once the residence time was
fixed, the feed pump was diverted to the flotation tank. With the air disengaged and the
rotor set at the desired tip speed, the flotation tank was filled, and the tailings valve was
adjusted until a desired liquid level (or froth depth) was attained. Next, the air regulator
was released and adjusted until the desired air flow was achieved. After making any other
final adjustments, the tank was allowed to operate freely, recycling product material back
47 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
to the sump, for at least three residence times, until a steady-state was attained. Once
at steady state, timed samples were taken from each of the product streams and the feed
stream. Both product streams were sampled simultaneously. The wet sample weight and the
sampling time were recorded. After dewatering and thoroughly drying the sample, the dry
weight was also recorded, in order to calculate the dry mass rate, water rate, and percent
solids of each stream. Finally, the torque was measured and recorded.
After taking samples for a single test condition, the air rate and rotation speed were
reset to a new condition, and a three residence time delay was permitted to allow the system
to return to a new steady state prior to any further sampling. During the detailed testing, a
two factor central composite experimental design was used to provide variations in tip speed
and air velocity encompassing the operational range of a typical laboratory cell (tip speed:
4.2 to 6.31 m/s; air velocity: 0.9 to 1.6 cm/s). These procedures were repeated for various
rotors and mean particle sizes.
After acquiring the experimental data, the measured mass and liquid rates were first
subjected to a mass balancing routine to reconcile the redundant measurements. The proce-
dure for the mass balance was to minimize the weighted sum of the squared errors between
the experimental and the adjusted values, while enforcing a strict zero-sum balance. From
the reconciled data, simple mass and solids recovery were determined. A more accurate
residence time ((cid:28)) value was determined by considering the mass balanced tailing flow rate
(Q ), the cell volume (V) less the air holdup ("):
t
V(1(cid:0)")
(cid:28) = :
Q
t
The kinetic coefficient for the continuous tests was determined by assuming a perfectly
mixed reactor model for the laboratory cell. Under the perfectly mixed condition, the flota-
tion recovery (R) is given in terms of the residence time and rate constant (k) by (Levenspiel,
1998):
k(cid:28)
R = :
1+k(cid:28)
By simply algebraic manipulation, the rate constant can be determined by:
R
k = :
(cid:28) (cid:0)R(cid:28)
Using these analytical procedures, the numerous data produced from the central com-
positeexperimentaldesignwereusedtopopulatethek(cid:0)S plot. Additionally, measurements
b
takenatsimilarairvelocitieswerecomparedtopopulatetherateconstant-powerplot. These
48 |
Virginia Tech | CHAPTER 2. DEVELOPMENT OF AN EXPERIMENTAL METHODOLOGY FOR
ROTOR TESTING
data interpretation strategies allow consistent comparisons between the batch and continu-
ous data. Variations in the holistic results should prompt further study in the contrasts of
process dynamics and scalability between batch and continuous flow mechanisms.
2.4 Summary and Conclusions
In summary, this manuscript has presented descriptions of the following experimental
and analytical equipment:
• The 15 rotors which were considered in the study;
• The 2 stators which match with the various rotors;
• The 3 flotation tanks offering different testing abilities;
• The 2 laboratory mixers with integrated torque meters used to propel the rotors;
• The method of bubble size analysis used throughout the testing;
• The compilation of equipment required for the batch system testing; and
• The compilation of equipment required for the continuous system testing.
In addition, this manuscript has provided descriptions of various experimental and
analytical procedures used throughout the exploratory and detailed testing, including:
• Power measurements along with repeatability testing;
• Two methods of air holdup determination;
• Methodsofdeterminingoperationalrobustnessviathesandingconditionandtheflood-
ing condition;
• Flotation performance by batch testing; and
• Flotation performance by continuous testing.
In conclusion, the experimental equipment and procedures along with the analytical
methods for comparative rotor testing have been presented in spirit of the laboratory testing
paradigm. By understanding the limitations and benefits of the laboratory scale, the tests
49 |
Virginia Tech | Chapter 3
Exploratory Testing of Flotation
Rotor Design
(ABSTRACT)
A two-part experimental analysis identifying the influence of rotor design on power
consumption and metallurgical performance in forced-air flotation equipment was conducted
at the laboratory scale. This manuscript describes the results for the first experimental
campaign. This phase of the research focused on the breadth of analysis in terms of number
of rotors analyzed. In this study, replicate testing and exhaustive experimental conditions
were overlooked in order to prioritize the wide variety of possible rotor designs. The goal of
thisapproachisthreefold: (1)validatetheexperimentalmethods,(2)verifythatperformance
difference exist and can be measured between the rotors, and (3) reduce the viable field of
rotor candidates for further testing and analysis. In this campaign 14 rotors were exposed
to a battery of testing intended to analyze power consumption, gas dispersion, operational
robustness, and batch flotation performance. The results show that while the experimental
methodology may be subject to further refinement, the rotors do indeed exhibit varying
levels of performance and energy efficiency. In conclusion, three of the best all-around rotors
(with varying strengths and weaknesses) were selected for detailed testing.
3.1 Introduction and Scope
Priorresearchinfrothflotationequipmenthasshownthatslightperformancedifferences
exist between different rotor designs. Much of this work has been relegated to the pilot
52 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
scale (Gorain, Franzidis, & Manlapig, 1995a, 1995b, 1996, 1997; Gorain, Napier-Munn,
Franzidis, & Manlapig, 1998) and the industrial scale (Yianatos et al., 2012). While some
texts have noted differences in laboratory-scaled machines (Crozier, 1992), few provided
a detailed or quantified analysis of performance capacities of particular design parameters.
This manuscript provides an initial investigation into the detailed comparison of rotor design
at the laboratory-scale. In addition to comparative analysis between the rotors, various
dimensionless analysis has been incorporated in order to allow scaled comparison of prior
and future studies.
Duringthis researchproject, the laboratory evaluationofflotation rotors wasconducted
intwocampaigns, exploratorytestinganddetailedtesting. Thismanuscriptdescribesthere-
sults and conclusions derived from the so-called exploratory testing. The primary goal of this
work was to test a wide variety of rotor designs in order to establish benchmarks and ensure
the veracity of the experimental and analytical methodology. Furthermore, this exploratory
testing was used to identify the existence and extent of performance differences between the
rotor designs, while providing evidence for the discontinued testing of fatally-flawed designs.
Rather than establishing a deterministic conclusion on the single most successful rotor, this
sectionseekstoprovideareliablemeansofcomparisonwhilenotingthepositiveandnegative
factors which should influence the final design criteria.
In this manuscript, the results are presented individually for each of the four established
performance criteria: (1) power consumption, (2) gas dispersion, (3) operational robustness,
and (4) flotation performance. After establishing and summarizing these results, synthesized
conclusions and observations will be noted. Chapter 4 provides a similar presentation of the
data acquired from the detailed testing campaign.
3.2 Experimental Methods
An exhaustive description of the equipment, material samples, methodology, and ana-
lytical procedures has been provided in Chapter 2. To summarize, this exploratory testing
campaign utilized 14 small-scale rotors, measuring 2.75 inches in diameter. These rotors
were exposed to six distinct laboratory tests intended to define the four aforementioned per-
formance criteria. Unfortunately, due to the rapid development and refinement during this
phase, not all 14 rotors were evaluated by each test. Table 3.1 summarizes the test matrix
used to evaluate each rotor.
Other equipment utilized in this testing phase include the generic stator model, the 9.75
53 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Table 3.2: Exploratory Testing: Power Measurement Parameters
Parameter Type Parameter Value
Rotor Diameter 2.75 in.
Equipment Tank Diameter 10 in.
Tank Volume 9.33 L
Material Solids Material None
Frother Type None
Chemical
Collector type None
Superficial Gas Velocity None
Operational Tip Speed 1.46 to 5.0 m/s
Reynolds Number 3.2 to 11 x 104
inch and 10 inch round tank configurations, the Heidolph RZR 2102 low-torque mixing unit,
a bubble sampler, the low-friction shaft, and a stainless-steel variable-area air flowmeter.
3.3 Results and Discussion
3.3.1 Power Consumption
Given the role of power consumption in normalizing performance results and as a means
of comparison, each rotor was first subjected to an extensive power analysis. During these
tests, power consumption was determined as a function of rotor speed in the absence of
forced air, solid particles, and reagents. Specific experimental details are listed in Table 3.2.
Two dimensionless numbers were used in describing data from the power studies, the
Reynoldsnumber(Re)andthepowernumber(P ). TheReynoldsnumbermaybecalculated
N
from the fluid density ((cid:26)), the rotational speed (N), the rotor diameter (D), and the fluid
viscosity ((cid:22)):
(cid:26)ND2
Re = : (3.1)
(cid:22)
Additionally, the power number may be calculated from the power draw (P), the fluid
55 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Table 3.3: Summary of Power Numbers Collected in Exploratory Testing
Average Power Standard Number of
Rotor
Number Deviation (%) Measurements
SP - B1 6.43 2.82 5
SP - C2 5.91 3.4 5
SP - A1 5.74 4.32 6
VT - E2 5.74 4.43 6
SP - C1 5.51 2.06 6
VT - E1 5.48 2.92 6
VT - B2 5.47 13.79 6
VT - A4 4.89 2.51 6
SP - A2 4.73 3.06 7
SP - A3 4.46 7.11 7
VT - D1 2.86 4.02 10
VT - A5 2.43 28.79 4
the average power number for 9 of the 12 rotors fell between 4.00 and 6.00. The three
exceptions were the VT - D1 (P = 2:86), the VT - A5 (P = 2:43), and the SP - B1
N N
(P = 6:43). Of these exceptions, the SP - B1 result is expected since the SP - B1 rotor has
N
two additional blades when compared to most of the others (eight versus six blades).
Mostoftherotorsgenerallyretainaconstantpowernumberoverthetestedspeedrange.
ThemostprevalentexceptionstothisbehaviorareseenintheVT-B2andtheVT-A5rotor.
Both of these rotors show a moderate correlation between increasing Reynolds number and
increasing power number. Theoretically, the power number should remain constant over this
speed range (Tatterson, 1994). As a result, the perceived correlation is likely coincidental,
and the average power number was still found by the simple arithmetic mean of the values
over the full speed range.
Table 3.3 summarizes these results by listing the rotors in order of average power num-
ber.
58 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Table 3.4: Exploratory Testing: Gas Dispersion Measurement Parameters
Parameter Type Parameter Value
Rotor Diameter 2.75 in.
Equipment Tank Diameter 10 in.
Tank Volume 9.33 L
Material Solids Material None
Frother Type None
Chemical
Collector type None
Superficial Gas Velocity 0.23 to 2.38 cm/s
Operational
Tip Speed 1.83 to 6.58 m/s
3.3.2 Air Water Mixtures and Gas Dispersion
Each rotor’s ability to effectively disperse air was determined by measuring global air
hold up as a function of tip speed and air velocity. Specific experimental details are listed
in Table 3.4. Air holdup was measured as a function of tip speed for two air velocities:
namely 1.0 and 2.0 cm/s. Additionally, the ungassed to gassed power ratio was determined
as a function of air velocity for a fixed rotor speed (700 RPM or 2.56 m/s). These data are
presented by two methods in Figure 3.3 (holdup versus power draw) and Figure 3.4 (gassed
power ratio versus air velocity). Missing data for the VT - E2 rotor indicates that this rotor
could not overcome flooding at the given test conditions.
Figure 3.3 indicates that significant variations in gas dispersion efficiency exist between
the different rotors. Generally, the air hold-up increases sharply as power is initially added
to the system. This increase subsides at higher power inputs, with each rotor approaching
a distinct maximum hold up value for a given air velocity. Although this general trend
remains constant, some rotors were able to attain significantly higher air holdup values at
similar power inputs. For the low gas velocity (1.0 cm/s), the greatest air holdup values
were experienced by the SP - A3, the VT - A4, the VT - E2, and the VT - E1. At the
higher gas velocity (2.0 cm/s), the SP - A1, VT - A4, and VT - D1 generally outperformed
their counterparts, in terms of the maximum holdup value attained in the given range power
inputs. While the VT - E2 performed well at low gas velocities, this rotor was prone to
flooding at gas velocities greater than 1.0 cm/s.
Another way of interpreting the data is to interpolate (via “reverse interpolation”) the
59 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Table 3.5: Summary of Air Holdup Data Collected in Exploratory Testing.
Power (W) to achieve Power (W) to achieve
Rotor 5% Air Holdup 9% Air Holdup
[J = 1:0 cm/s] [J = 2:0 cm/s]
g g
VT - E2 15.7 –
SP - A3 16.4 31.6
VT - A4 19.4 22.8
VT - E1 19.8 20.8
SP - A1 23.5 26.7
SP - C1 25.1 32.4
SP - A2 25.9 32.5
VT - D1 26.8 30.2
SP - C2 31.6 52.5
SP - B1 46.2 61.7
power needed to attain a common air holdup between the rotors. Table 3.5 summarizes the
results in this fashion by interpolating the power needed to attain 5% air holdup at a gas
velocity of 1.0 cm/s and 9% air holdup at 2.0 cm/s. The decision to interpolate to these
points is somewhat arbitrary, but they do represent holdup values that most rotors were
able to achieve during the tests. Consequently, this listing provides a single-value summary
useful for pairwise comparison. These data have been listed in the order of ascending power
values for the 1.0 cm/s scenario.
Figure 3.4 shows the reduction in power draw experienced as air is added to the cell.
This reduction is represented as a ratio of the power draw at the gassed condition to the
power draw at the ungassed condition. This data shows similar behavior for all rotors with
the exceptions of the VT - D1, the VT - E1, and the SP - B1. The VT - D1 experienced
almost no reduction in power draw, while the VT - E1 experienced a much sharper drop off.
The behavior of the SP - B1 rotor appears similar, but the magnitude of the power reduction
is substantially greater. Once again, data was not presented for the VT - E2 since this rotor
could not effectively disperse air at the given rotor speed. Furthermore, while the magnitude
of the power loss is similar, the behavior of the SP - C2 appears more linear when compared
to the diminishing losses of other rotors.
62 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Table 3.6: Exploratory Testing: Flooding Condition Measurement Parameters
Parameter Type Parameter Value
Rotor Diameter 2.75 in.
Equipment Tank Diameter 10 in.
Tank Volume 9.33 L
Material Solids Material None
Frother Type None
Chemical
Collector type None
Superficial Gas Velocity 2.0 cm/s
Operational
Tip Speed 1.83 to 6.58 m/s
3.3.3 Operational Limits
Duringtheexploratorytesting, theoperationallimitsweretestedtoidentifytherelative
robustness each rotor has to changing working conditions. As forced air is added or the tip
speed is reduced, the flotation rotor will lose the ability to fully disperse the air and the
ability to suspend solids. Either of these limits signifies a condition in which the flotation
cell cannot operate normally. Depending on the specific design, each rotor inherits a distinct
range of tip speeds and air velocities marked by stable operation. Ultimately, this evaluation
seeks to define this two-dimensional “area-of-operation.” The area-of-operation graph lends
itself to quick interpretation, since a simple comparison of the area extent will yield the more
proficient rotor. Simply, a larger area signifies a better rotor design, capable of operating in
a wider range of conditions. Here, the two primary operational limits, the flooding condition
and the sanded condition, were independently evaluated.
During this testing campaign, the flooding condition was only studied at a single gas
velocity. Specific experimental details of the flooding test are listed in Table 3.6.
Data for the flooding tests are included in Figure 3.5. This data presents a one di-
mensional area-of-operation for each rotor at a gas velocity of 2.0 cm/s. The left side of
this range identifies the onset of the flooding condition (or the critical air dispersion speed).
When operated at speeds lower than this limit, the rotor cannot sufficiently break the vol-
ume of forced air. The indication of this condition is the appearance of extremely large
air pockets erupting on the water surface near the shaft region (commonly called boiling,
geysering, or flooding). Depending on geometry and sparger design, some rotors are more
63 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
efficient at breaking air pockets and thus have a lower critical air dispersion speed.
The right side of the range indicates the maximum speed prior to the torque limit im-
posed by the flotation mixer used in laboratory testing. This limit is related to the individual
rotor’s power draw. The RZR 2102 Mixer has a maximum torque limit of approximately 47
N-cm. The precise torque limit is somewhat dependent on the speed; this data is provided in
the mixer’s calibration documentation. Speeds greater than the right limit shown in Figure
3.5 exceed this torque limit. Overall, rotors that can effectively disperse air and draw less
power have a wider range and greater robustness.
As shown in Figure 3.5, most rotors are able to disperse air in a comparable speed
range. In exception, the SP - A1 and SP - B1 disperse air at lower speeds, potentially due
to better sparging mechanisms. Additionally, the VT - A4 and the SP - A3 have the widest
overall range, exhibiting the greatest combined robustness in terms of air dispersion and
power consumption. The VT - E2 and the VT - E1 have the narrowest range, due to poor
sparger design and high power consumption.
The second operational limit, the sanding condition, was assessed in two tests. The
first test used an iron ore sample as the solid constituent. This rougher feed material is
95% passing 106 microns and has a specific gravity of 4.0. The second sanding test used
35 micron monosized silica beads. Both materials were tested over several gas velocities.
Specific experimental details of both sanding tests are listed in Table 3.7.
Data from the sanding test is interpreted by two means. For the iron ore test these
data are presented in Figure 3.6 and Figure 3.7. Figure 3.6 shows bar graphs for each rotor.
The value of these bar graphs indicates the speed (and power) needed to completely suspend
all solids from the tank floor (or critical solid suspension speed). Depending on the flow
condition induced by each rotor, some are capable of suspending solids at lower speeds and
powers. The data is also presented for varying air velocities. Since pumping power is reduced
as air is introduced, these additional data points indicate each rotor’s resilience to added
air. Omission of data for individual rotors indicates the lack of a stable operating condition
at the designated air velocity. Some rotors were unable to overcome the flooding condition
at high gas velocities (VT - E2, VT - A4), while others exceeded the torque capacity of the
mixing unit at low gas velocities (SP - B1).
The second means of interpreting sanding data is an area plot, depicted for each rotor
(Figure 3.7). In this plot, tip speed is plotted on the abscissa, while air velocity is plotted
on the ordinate. The boundary of the shaded area indicates the critical tip speed required
for solid suspension for the designated air velocity. Consequently, the blue area represents
64 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Tip Speed (m/s)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7
SP − B1
VT − E2
Flooding Region Torque Limit Region
VT − E1
VT − D1
VT − A4
SP − C1
SP − C2
SP − A1
SP − A3
SP − A2
J = 2.0 cm/s
g
0 137 273 410 547 684 820 957 1094 1230 1367 1504 1641 1777
Rotational Speed (RPM)
Figure 3.5: Operational limits determination for exploratory testing: onset of the flooding
condition and torque limitation at a gas velocity of 2.0 cm/s. The stable area-of-operation
is presented for each rotor in terms of rotational speed and tip speed. The left side of the
limit signifies the critical air dispersion speed. Speeds lower than these values induce the
flooding condition. The right side of the limit signifies the speed/torque limit of the mixing
device. Speeds greater than this are not attainable with the RZR 2102 mixer. A wider bar
signifies more robustness. All rotors were 2.75 inches in diameter. Tests were performed in
the absence of reagents and solids.
65 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
J = 0 cm/s J = 1 cm/s J = 2 cm/s
g g g
SP − B1 SP − B1 SP − B1
VT − E2 VT − E2 VT − E2
SP − C1 SP − C1 SP − C1
SP − C2 SP − C2 SP − C2
VT − A4 VT − A4 VT − A4
SP − A1 SP − A1 SP − A1
SP − A2 SP − A2 SP − A2
SP − A3 SP − A3 SP − A3
0 2 4 0 2 4 0 2 4
Critical Tip Speed (RPM) Critical Tip Speed (RPM) Critical Tip Speed (RPM)
J = 0 cm/s J = 1 cm/s J = 2 cm/s
g g g
SP − B1 SP − B1 SP − B1
VT − E2 VT − E2 VT − E2
SP − C1 SP − C1 SP − C1
SP − C2 SP − C2 SP − C2
VT − A4 VT − A4 VT − A4
SP − A1 SP − A1 SP − A1
SP − A2 SP − A2 SP − A2
SP − A3 SP − A3 SP − A3
0 20 40 60 0 20 40 60 0 20 40 60
Critical Power (Watts) Critical Power (Watts) Critical Power (Watts)
Figure 3.6: Operational limits determination for exploratory testing: iron ore sanding test
bar graphs. The bar graphs indicate the critical solid suspension speed and power at the
designated gas rates. Missing data series indicate that the off-bottom suspension condition
was not attainable at the test conditions, due to the torque limitations of the mixer or the
air dispersion limits of the rotor. All rotors were 2.75 inches in diameter.
66 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Table 3.7: Exploratory Testing: Sanding Condition Measurement Parameters
Parameter Value Value
Parameter
Type (Iron Ore Test) (Silica Test)
Rotor Diameter 2.75 in. 2.75 in.
Equipment Tank Diameter 10 in. 10 in.
Tank Volume 9.0 L 9.0 L
Solids Material Iron Ore Silica Beads
% Solids 40% 30%
Material
Particle Size 95% -106 microns 35 micron
Material SG 4.0 2.5
Frother Type None None
Chemical
Collector type None None
Superficial Gas Velocity 0 to 2.0 cm/s 0, to 2.0 cm/s
Operational
Tip Speed 1.46 to 4.90 m/s 1.97 to 5.27 m/s
the two dimensional area-of-operation. Area outside of the blue region is marked by some
form of interoperability, either the sanded condition at low air velocities or the flooding
and sanded condition at higher air velocities. At the highest air velocity tested (2.0 cm/s),
the graph is extended horizontally in lieu of additional data points at greater air velocities.
Comparisons between the different rotors can be made by simply assessing the relative area
of each shaded region. Larger areas signify greater robustness in terms of suspending solids.
Table 3.8 summarizes these results by ranking each rotor in order of operable area.
As indicated in Figure 3.6, the class A rotors performed substantially better than the
other rotors tested in terms of solid suspension normalized by power consumption. The SP -
A1, SP - A2, and SP - A3 rotors were able to suspend solids at a relatively low power input
and maintain this solid suspension at higher gas rates. Additionally, the marked outlier of
the data set, the VT - E2, was able to suspend particles at a drastically lower speed and
power than its counterparts at no air flow. Unfortunately, this rotor was unable to overcome
the flooding condition at higher gas rates.
In addition to the power-normalized comparisons, the family of SP - A rotors exhibited
a larger area-of-operation (Figure 3.7). The SP - A1 conclusively outperforms the other
rotors in terms of robustness as it relates to solid suspension.
Data for the silica sanding tests are shown in Figure 3.8 and Figure 3.9. Here the data
67 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
J = 0 cm/s J = 1 cm/s J = 2 cm/s
g g g
SP − C2 SP − C2 SP − C2
VT − A6 VT − A6 VT − A6
VT − B2 VT − B2 VT − B2
VT − E1 VT − E1 VT − E1
VT − D2 VT − D2 VT − D2
VT − D1 VT − D1 VT − D1
VT − A4 VT − A4 VT − A4
VT − A5 VT − A5 VT − A5
SP − A1 SP − A1 SP − A1
0 2 4 0 2 4 0 2 4
Critical Tip Speed (RPM) Critical Tip Speed (RPM) Critical Tip Speed (RPM)
J = 0 cm/s J = 1 cm/s J = 2 cm/s
g g g
SP − C2 SP − C2 SP − C2
VT − A6 VT − A6 VT − A6
VT − B2 VT − B2 VT − B2
VT − E1 VT − E1 VT − E1
VT − D2 VT − D2 VT − D2
VT − D1 VT − D1 VT − D1
VT − A4 VT − A4 VT − A4
VT − A5 VT − A5 VT − A5
SP − A1 SP − A1 SP − A1
0 20 40 0 20 40 0 20 40
Critical Power (Watts) Critical Power (Watts) Critical Power (Watts)
Figure 3.8: Operational limits determination for exploratory testing: silica sanding test
bar graphs. The bar graphs indicate the critical solid suspension speed and power at the
designated gas rates. All rotors were 2.75 inches in diameter.
is presented in similar fashion as the iron ore tests; however, more rotors were tested with
this material. Since the material size and density significantly affect the solid suspension
characteristics, the analysis should remain comparative between the rotors, rather than
between the two tests. The data from the area plot is summarized in Table 3.9 in order
of descending nominal area.
Upon examination of the silica sanding data (Figure 3.8), the VT - D1 and VT - D2
required significantly greater power and speed to suspend particles when compared to their
counterparts. This trend is intensified as the air rate is increased. At high air flow rates, VT
- A5 and VT - A4 also require significantly higher speeds to suspend particles; however, the
power draw for these two rotors is still comparable to other designs (a result of the relatively
low power number of these rotors – Section 3.3.1). Coinciding with the results of iron ore
testing, the SP - A1 remains one of the leaders with regard to particle suspension, especially
at high air velocities. This result is also reflected by the enlarged relative area exhibited in
69 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Table 3.9: Summary of Area-of-Operation Area for Silica Sanding Test.
Nominal Area of
Rotor
Stable Operation
SP - A1 7.50
VT - B2 6.74
VT - E1 6.40
VT - A6 6.40
SP - C2 6.21
VT - A4 5.64
VT - A5 4.15
VT - D2 3.25
VT - D1 3.23
the window-of-operation data (Figure 3.9 and Table 3.9).
3.3.4 Batch Flotation Performance
During the exploratory campaign, flotation performance for each rotor was determined
by a series of standard batch kinetics tests. In the kinetics tests, the experimental conditions
were rigidly controlled in order to form a reliable comparison between the various rotors.
Each rotor was tested in identical chemical conditions and at a consistent air velocity. With
these conditions fixed, each individual rotor was tested at no less than three discrete tip
speeds unique to each rotor. The specific tip speeds were selected to produce flotation rate
versus power curves within a similar range of abscissa values.
The material used throughout the batch tests was 35 micron silica beads hydrophobized
by dodecylamine at neutral pH. Due to their purely spherical shape and small size range,
this material minimizes the effects of inconsistent collector absorption. Furthermore, since
the pure flotation rate is the desired response, the tests were conducted with only a single,
floatable species. The absence of a hydrophilic species eliminated the need to assay the test
results, hastening the data collection phase.
Fortherotorcomparisons, trueflotation, ratherthaneffectsfromthefroth, wasdesired.
Consequently, the weight percent solids was kept intentionally low (<5%) to reduce the risk
71 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Table 3.10: Batch Flotation Test Parameters
Parameter Type Parameter Value
Rotor Diameter 2.75 in.
Tank Diameter 9.75 in.
Equipment
Tank Volume 9.0 L
Launder Lip Length 5.0 in.
Solids Material Silica Beads
Solids Mass 460 grams
Material
% Solids 4.9%
Mean Particle Size 35 micron
Frother Type MIBC
Frother Dosage 8.1 ppm
Chemical
Collector Type Dodecylamine
14.8 g/tonne
Collector Dosage
(4.1 x 10(cid:0)6 M)
Froth Depth 1.5 in.
Superficial Gas Velocity 2.0 cm/s
Operational
Tip Speed 2.2 to 5.9 m/s
Flotation Time 10 to 240 sec
of froth overloading. Specific experimental details are listed in Table 3.10.
Raw data from the flotation tests are shown in Figure 3.10. Here the data is presented
as cumulative recovery plotted against time for each incremental product collected during
the test (10, 20, 30, 60, 120, and 240 seconds). Individual curves on each plot indicate
independent tests at varying tip speeds.
Data collected up to 30 seconds consistently demonstrated first order kinetics. Conse-
quently, these three points and the zero recovery at zero time point were used to determine
the fast floating kinetic coefficient. Figure 3.11 shows ln(1(cid:0)R) plotted against time for
each experiment. The kinetic coefficient was determined by finding the slope of the straight
line which fits these data points.
After determining the kinetic coefficient, the data was interpreted by plotting the rate
constantagainstthepowerdrawforeachindependenttest(Figure 4.3.4). Sincealltestswere
performed under the same chemical conditions and gas velocity, the variations in kinetics can
72 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Table 3.11: Summary of P values determined in Exploratory Flotation Testing
Average P
Rotor
(x 10(cid:0)2)
VT - A4 9.66
VT - A5 8.28
VT - B2 3.70
SP - A1 3.51
VT - A6 2.65
VT - D1 2.47
VT - D2 2.20
SP - C2 1.75
VT - E1 1.68
As shown in Figure 4.3.4, the flotation rate constant generally increases with increasing
power input. As the power input increases, the incremental gain in rate constant diminishes.
The shape of these plots generally reflects the trends found in the air holdup versus power
plots (Figure 3.3). Using the SP - A1 as a benchmark, many of the new prototypes provide
significantly greater power efficiency, including VT - E1, VT - A4, and VT - A5. Other
rotors, namely the VT - D1, VT - D2, and SP - C2, showed trends inconsistent with the
other rotors. The VT-SB rotors only experienced marginal gains with increasing power,
while the SP - C2 actually experienced a decreased rate constant at the highest power value.
Most of the hydrodynamic efficiency data coincides with the expectations presented in
the literature. The slope of most k (cid:0)S plots were very similar and within an anticipated
b
range. As shown in Figure 3.14, all of the data points, excluding the single outlier lie within
a reasonable band. Nevertheless, Figure 3.13 does indicate that the VT - A5 and the VT
- A4 rotors have slightly elevated P values when compared to their counterparts. This
result may suggest that these rotors are capable of producing hydrodynamic conditions (i.e.
fluid velocities, turbulent kinetic energies) more conducive for bubble-particle collisions and
flotation. To summarize this data Table 3.11 ranks the rotors in order of the calculated P
value.
78 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
3.4 Summary and Conclusions
Small-scale laboratory flotation tests were performed to assess the operational capabili-
ties of various rotors in four performance areas: power consumption, air holdup, operational
robustness, and flotation performance. In total, 14 rotors were exposed to numerous labo-
ratory experiments. In forming a more concise and collective evaluation, both quantitative
and qualitative evaluation matrices were generated (Table 3.12 and Table 3.13) from the
data presented in this chapter. These tables summarize the data by integrating at least one
indicative measurement from the evaluation of each of the four performance criteria. These
comparisons include: average power number, single-point interpolation from the holdup ver-
sus power graph, the critical air dispersion speed, the critical solid suspension power from
the iron ore sanding test, the critical solid suspension power from the silica sanding test, the
average P value, and single-point interpolation from the rate constant versus power graph.
While seemingly arbitrary and eclectic, the presented measurements were selected for this
summary since they efficiently convey the most useful data, retain the conclusions found in
the full data set, and generally preserve the paradigm of power-normalization as a basis for
a fair comparison.
In this context, “single-point interpolation” provides a means of comparing an entire
curve at a single point. For example, Section 3.3.2 provides curves describing air holdup as
a function of power input. Since all rotors were not tested at an identical power input value,
no abscissa value can be easily selected for a single point comparison. Rather, by fitting the
curves with a cubic spline, the estimated power required to achieve a desired holdup can be
interpolated. For the data presented in Table 3.12, the value of 9% holdup was selected, and
the estimated power value was calculated. As a result, the various curves of air holdup and
power are summarized at a single ordinate point, namely 9%.
Table3.13providesanotherlevelofsimplificationbycategorizingthemeasurementsinto
above, below, and average designations. The average designation represents the numerical
values within one standard deviation of the average of all rotors for the given measure-
ment. The above and below average designations represent those variances greater than one
standard deviation from the mean.
79 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
From the summary evaluations presented in Table 3.12 and Table 3.13 along with the
full data set collected during the exploratory testing campaign, eight key conclusions are
ascertained:
1. The experimental methodology and analytical techniques provided useful comparative
data amongst thevariousrotor designs. Thetests produced results whichcoincidewith
theoretical expectations while highlighting subtle performance efficiencies between the
rotor designs.
2. Across all performance criteria, the Class A rotors performed at or above average when
comparedtotheircounterparts. OftheseArotors, theSP-A1andtheSP-A4showed
the best overall performance, with the SP - A1 excelling in the areas of air dispersion
and solid suspension at high gas velocities (2.0 cm/s). As a result, the SP - A1 was
chosen as an initial benchmark to compare the other rotors.
3. The high power numbers (due to the additional blades) from the Class B rotors did
not necessarily correspond to proportionally enhanced performance in all areas. While
both rotors were more resistant to sanding and flooding, the SP - B1’s normalized
performance was especially hindered by its high power consumption.
4. The Class C rotors showed similar or somewhat reduced performance compared to the
class A rotors. These rotors did not excel substantially in any particular area.
5. The inconsistent and wavering results from the Class D rotors are likely due to a
mechanical design and mounting flaw. Of the available models, the D class rotors
are the only two that lack bottom-side suction. In theory, if centered in the tank,
these rotors will intake fluid from the vertical midpoint and eject fluid from the top
and bottom edges. In the laboratory setting, the rotor cannot be perfectly centered
in the tank. This slight misalignment causes the rotor to intake fluid from one half
of the tank while ejecting the fluid on the other side. Measurements involving visual
appraisal (air holdup, sanding) or a non-radial launder (batch flotation) are subject
to inconsistency, depending if the measurement is tank from the suction or the outlet
side of the tank. Furthermore, the degree of radial asymmetry (and variations in flow
patterns) may also significantly influence measurements not involving visual appraisal,
such as power.
6. Despite enhanced performance in a limited range of conditions, the class E rotors were
unable to consistently overcome the flooding condition at even moderate gas rates.
As a result, these rotors were abandoned at early stages in the experimental process.
82 |
Virginia Tech | CHAPTER 3. EXPLORATORY TESTING OF FLOTATION ROTOR DESIGN
Nevertheless, the VT - E2’s noted ability to suspend solids at low power inputs should
prompt future design criteria with distributed suction.
7. Several rotors, including the VT - A4, VT - A5, VT - E1, and VT - B2, showed
enhanced flotation performance normalized by power efficiency when compared to the
benchmark SP - A1. While the full data set in the k (cid:0) S plot follows expectations
b
consistent with the literature, individual points suggests that the VT - A4 and the VT
- A5 may exhibit increased P values as a result of favorable hydrodynamic efficiency.
This result should prompt further investigation.
8. Of the prototype rotors, the VT - A4 and VT - B2 are most suitable for further test-
ing. Across all performance criteria, these rotors consistently exhibit average or above
average performance. Additionally, these rotors meet practical limitations (such as
fabrication and wear resistance) and exhibit a wide area-of-operation. While the VT
- A5 rotor also exhibited enhanced performance in various criteria, it is not recom-
mended for further testing due to qualitative considerations. This rotor contains thin
and delicate components which may not withstand the rigors of industrial flotation
(extensive wear, passing grinding media, etc.).
3.5 Bibliography
Crozier, R. (1992). Flotation. theory, reagents and ore testing. Pergamon Press plc(UK),
1992,, 356.
Gorain, B., Franzidis, J., & Manlapig, E. (1995a). Studies on impeller type, impeller
speed and air flow rate in an industrial scale flotation cell. part 1: Effect on bubble size
distribution. Minerals Engineering, 8(6), 615–635.
Gorain, B., Franzidis, J., & Manlapig, E. (1995b). Studies on impeller type, impeller speed
and air flow rate in an industrial scale flotation cell. part 2: Effect on gas holdup. Minerals
Engineering, 8(12), 1557–1570.
Gorain, B., Franzidis, J., & Manlapig, E. (1996). Studies on impeller type, impeller speed
andairflowrateinanindustrialscaleflotationcell.part3: Effectonsuperficialgasvelocity.
Minerals Engineering, 9(6), 639–654.
Gorain, B., Franzidis, J., & Manlapig, E. (1997). Studies on impeller type, impeller speed
and air flow rate in an industrial scale flotation cell. part 4: Effect of bubble surface area
flux on flotation performance. Minerals Engineering, 10(4), 367–379.
83 |
Virginia Tech | Chapter 4
Detailed Testing of Flotation Rotor
Design
(ABSTRACT)
A two-part experimental analysis identifying the influence of rotor design on power
consumption and metallurgical performance in forced-air flotation equipment was conducted
at the laboratory scale. This manuscript describes the results for the latter experimental
campaign. While the initial study emphasized numerous rotor designs, this investigation
only examines three designs. These selections were chosen from the original pool of 14 rotors
during the exploratory testing phase. In lieu of the various rotors, this detailed study empha-
sized replicate testing, broad ranges of operational conditions, and more thorough analysis.
The goals of this approach are to: (1) quantify the uncertainty in the measurements, (2)
determine if the measured performance variations between the rotors exceed the uncertainty,
(3) develop insight on the causes of these differences, and (4) provide dimensionless relation-
ships to guide larger scale studies. The successful completion of these goals provides design
criteria and standards for future rotor prototypes. The results of this work show that each
of the three rotors exhibits varying strengths and weaknesses in the performance evaluation
categories; therefore, each rotor may be more or less suitable, depending on the application.
4.1 Introduction and Scope
After the exploratory test campaign, the VT - A4, VT - B2 and SP - A1 rotors were se-
lected for further detailed evaluation. During detailed testing, analytical depth and repeata-
85 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
bility were prioritized in lieu of the numerous rotor designs tested in the prior campaign.
Similar performance criteria were evaluated (power consumption, air dispersion, operational
robustness, and flotation performance), but given the breadth of operational conditions ex-
amined in this phase, a more comprehensive analysis has been conducted.
The unique aspects of this study when compared with the exploratory campaign are
the properly-scaled tank dimension and the use of continuous testing. In this campaign,
a 14 inch diameter tank was selected for all performance measurements. While this tank
size retains proper geometric similitude for the given small rotor size (Silva et al., 2012), the
resulting volume (35 liters) significantly deviates from standard laboratory size expectations.
Furthermore, the use of a continuous flow, two-product recycle system represented a unique
challenge in materials handling and experimental setup. Nevertheless, both of these features
were selected and utilized in order to more closely reflect the industrial setting, since the
ultimate goal of this testing phase is to provide recommendation for industrial rotor design
and commercialization.
This chapter parallels the presentation and analysis of data presented in Chapter 3,
including results independently describing, (1) power consumption, (2) air dispersion, (3)
operational robustness, and (4) flotation performance. Holistic conclusions and observations
are presented after the individual test data.
4.2 Experimental Methods
An exhaustive description of the equipment, material samples, methodology, and an-
alytical procedures has been provided in Chapter 2. To summarize, this detailed testing
campaign utilized 3 rotors small-scale rotors, measuring 2.75 inches in diameter. These ro-
tors were exposed to five distinct laboratory tests intended to define the four aforementioned
performance criteria. All rotors were subjected to the same test matrix and many experi-
ments included multiple replicates. Table 4.1 summarizes the test matrix used to evaluate
each rotor.
The continuous flotation system constitutes all of the other equipment utilized in this
testing phase. Specific components include the generic stator model, the 14 inch round
tank configuration, the Caframo brushless high-torque mixing unit, a bubble sampler, the
low-friction shaft, and a stainless-steel variable-area air flowmeter.
86 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
Table 4.1: Detailed Campaign Rotor Text Matrix
Sanding Continuous
Rotor Power Holdup Flooding
(Silica) Flotation
SP - A1 X X X X X
VT - A4 X X X X X
VT - B2 X X X X X
Table 4.2: Detailed Testing: Power Measurement Parameters
Parameter Type Parameter Value
Rotor Diameter 2.75 in.
Equipment Tank Diameter 14 in.
Tank Volume 35.3 L
Material Solids Material None
Frother Type None
Chemical
Collector type None
Superficial Gas Velocity None
Operational Tip Speed 1.46 to 6.5 m/s
Reynolds Number 6.7 to 14 x 104
4.3 Results and Discussion
4.3.1 Power Consumption
Each rotor’s power consumption was determined as a function of rotor speed in the
absence of forced air, solid particles, and reagents. Specific experimental details are listed in
Table 4.2.
Duringthispowerstudy, individualmeasurementswererepeatedthreetimes, inorderto
evaluatetherepeatabilityandexperimentaluncertainty. Figure4.1depictsthisexperimental
repeatability for each rotor. In this analysis, the residual is defined as the difference between
an individual measurement and the average value for identical test conditions (See Chapter
2.3.1) . This difference is expressed as a percentage of the original measurement. The values
87 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
presented as a percentage of the original measurement, the perceived inconsistency at low
power measurements actually corresponds to a relatively low absolute magnitude. Noting
the results of this repeatability analysis, all power measurements throughout this testing
campaign represent the average of three repeats.
The results of the power study are presented in Figure 4.2 and Figure 4.3. Figure
4.2 shows the power versus speed data plotted for each rotor. Figure 4.3 shows the same
data plotted dimensionlessly as power number versus Reynolds number. The average power
number shown on the graph was determined by finding the simple arithmetic mean of the
power numbers over the full speed range. This power number was used to calculate the
fitting line in Figure 4.2.
The results of the detailed power study generally confirm those found in the exploratory
testing. The SP - A1 and the VT - B2 have similar power requirements over the tested speed
range. Alternatively, the VT - A4 consistently requires less power at the same rotational
speed. This result is manifested by the lower powercurve (Figure 4.2) and the reduced power
number (Figure 4.3). Furthermore, as shown in Figure 4.2, the power difference between the
VT - A4 and the SP - A1 increases as the speed is increased.
4.3.2 Air Water Mixtures and Gas Dispersion
Each rotor’s ability to effectively disperse air was determined by measuring global air
hold up over a full range of tip speeds and air velocity. A full factorial experimental design
was used incorporating tip speeds of 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, and 6.5 m/s and air
velocities of 0, 0.48, 0.71, 0.95, 1.19, 1.43, 1.66, and 1.90, up to the operational limit of each
rotor at each tip speed. Specific experimental details are listed in Table 4.3.
Duringdetailedtesting, airholdupwasdeterminedbythevolumedisplacementmethod.
In these experiments, individual measurements were repeated three times, in order to eval-
uate the repeatability and experimental uncertainty. Figure 4.4 depicts this experimental
repeatability for each rotor. In this analysis, the residual is defined as the difference between
an individual measurement and the average value for identical test conditions. This differ-
ence is expressed as a percentage of the original measurement. The upper plots in Figure
4.4 present the residual as a function of the measured power value. The lower plots show
the frequency distribution of all residuals for each rotor. The red lines indicate the standard
deviation of the population of residuals.
Figure 4.4 shows that in general, the air displacement method of determining air holdup
89 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
10
9
8
7
6
5
4
3
2
1
0
0 50 100 150
Power (Watts)
)%(
pudloH
VT − B2
10
9
8
7
6
5
4
3
2
1
0
0 50 100 150
Power (Watts)
)%(
pudloH
SP − A1
10
9
8
7
6
5
4
3
2
1
0
0 50 100 150
Power (Watts)
)%(
pudloH
VT − A4
J (cm/s)
g
O =1.9
O =1.66
O =1.43
O =1.19
O =0.95
O =0.71
O =0.48
Figure 4.5: Air holdup plotted against power for each rotor included in the detailed testing.
Data is presented for gas velocities ranging from 0.48 to 1.9 cm/s. Tests were conducted in a
35.3 liter tank in the absence of solids and reagents. All rotors were 2.75 inches in diameter.
is moderately reliable. The residuals are normally distributed about the mean, and of the
130 points for each rotor, only one or two lie outside the 20% residual range. The SP - A1
shows the most consistent behavior across all power ranges, while the VT - B2 and VT - A4
generally show higher residuals at lower measured power values. Furthermore, the VT - A4
has highest frequency of outliers greater than 20% residual.
Figure 4.4 plots holdup residual as a function of power. This choice of power as the
independent variable depicts the most useful correlation. Similar plots, showing residual as a
functionofmeasuredholduporairvelocityaremorerandomlydispersedandlessmeaningful.
Several means were used to interpret the air holdup and gas dispersion data. The
most meaningful comparisons are made by analyzing data at similar rates and power draws.
Figure 4.5 shows air holdup plotted against power draw. Each plot contains the data for a
single rotor, and different data series show values at different air rates. This plot contains
all the repeat measurements, and each data series has been fit to a cubic polynomial over
the power range. Data series lacking a fit line have three or few points.
In general, the data in Figure 4.5 confirms the results of the exploratory testing. Air
holdup increases with increasing power addition, and the largest incremental gains are made
at relatively low power values. For many data series, diminishing returns are experienced
93 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
Jg =0.48 cm/s
10
8
6
4
2
0
0 50 100 150
Power (Watts)
)%(
pudloH
Jg =0.71 cm/s
10
8
6
4
2
0
0 50 100 150
Power (Watts)
)%(
pudloH
Jg =0.95 cm/s
10
8
6
4
2
0
0 50 100 150
Power (Watts)
)%(
pudloH
Jg =1.19 cm/s
10
8
6
4
2
0
0 50 100 150
Power (Watts)
)%(
pudloH
Jg =1.43 cm/s
10
8
6
4
2
0
0 50 100 150
Power (Watts)
)%(
pudloH
Jg =1.66 cm/s
10
8
6
4
2
0
0 50 100 150
Power (Watts)
)%(
pudloH
O = VT − A4
O = SP − A1
O = VT − B2
Figure 4.6: Air holdup plotted against power at various air velocities ranging from 0.48 to
1.66 cm/s for the three rotors included in detailed testing. Tests were conducted in a 35.3
liter tank in the absence of solids and reagents. All rotors were 2.75 inches in diameter.
at larger power inputs, as many of the curves begin to level (especially prominent at power
gas velocities less than 1 cm/s). The incremental improvement of increasing air velocity at
a constant power input can be determined by examining the length between the individual
data series. Generally, larger incremental improvements in holdup are made at higher gas
velocities. Finally, the general scatter in the data can be ascertained by comparing values of
the same data series in Figure 4.5.
For comparative analysis, the data in Figure 4.5 was reconfigured to show data from
differentrotorsonthesameaxes(Figure4.6). Herethedifferentdataseriesshowthedifferent
rotors, and the separate plots depict that comparison at different air velocities. Once again
all repeat data has been presented, and a cubic polynomial has been fit to each data set.
Figure 4.6 consistently shows that over the collected data range, the VT - A4 rotor
slightly outperforms the SP - A1, on the basis of air holdup as a function of power draw.
While the SP - A1 reaches or exceeds the maximum air holdup of the VT - A4 (at a fixed
94 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
air velocity), the SP - A1 generally requires a substantially greater power requirement to do
so. In example, at J = 1.19 cm/s, the VT - A4 reaches a maximum holdup of 6.09% at a
g
power of 50.2 Watts. At the same air velocity, the SP - A1 reaches a maximum holdup of
6.26%, yet the power requirement is 111.6 Watts.
Furthermore, on the same basis of comparison, the SP - A1 consistently outperforms
the VT - B2, with the exception of the highest air velocity, 1.66 cm/s. However, this air
velocity only shows limited data compared to the other plots, since this condition nearly
exceeded the operational limits of the machine (see Section 4.3.3).
A contour plot can also be used to visual the air holdup data and serve as a type of
calibrationorpredictioncurvefordifferentconditions. Figure4.7showsairholdupplottedas
a function of gas velocity and tip speed. Darker colors represent regions of low holdup, while
lighter regions represent regions of higher holdup. The white region indicates the conditions
in which the machine could not operate (i.e. the flooding condition). This plot may be used
to identify the relative sensitivity air holdup has to changing operational conditions for each
rotor. Wider contour regions indicate a lower gradient and lower sensitivity. Smaller contour
regions indicate a steeper gradient and higher sensitivity. By visual comparison, the VT -
B2 shows a shallower gradient at low tip speeds and gas velocities, while the VT - A4 shows
a lower gradient at higher tip speeds and gas velocities. The SP - A1 maintains a roughly
consistent and steep gradient across the range of tip speeds and gas velocities.
Besides air holdup, the power reduction ratio was determined during the investigations
of air/water mixtures. This reduction ratio (P /P ) is simply defined as the power
gassed ungassed
draw at an air-induced condition divided by the ungassed power draw at the same speed.
Figure 4.8 and Figure 4.9 show this ratio as a function of air velocity and aeration number,
respectively. The data in these plots includes tests at various tip speeds. Each data set
has been fit to a cubic polynomial over the range of abscissa values. The points in the
plots include repeat measurements. The ungassed condition used as the denominator in the
calculations was the average of all repeats at the desired tip speed (thus, not all 0 cm/s air
velocity points fall at 100%).
The apparent data scatter in Figure 4.8 shows that superficial gas velocity may not
correlate fully to the power reduction ratio. Nevertheless, Figure 4.8 does show the general
trend and the relative gradient of each rotor’s power to induced air volume. However, since
the data was collected at various tip speeds, a more appropriate correlation must also include
this experimental variable. The aeration number (Q/ND3) is a dimensionless value which
accounts for volumetric air flow (Q), rotational speed (N), and rotor diameter (D). Figure
4.9 shows the power reduction ratio as a function of aeration number. Since this plot is
95 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
100
90
80
70
60
50
40
30
20
10
0
0 0.05 0.1 0.15 0.2
Aeration Number
)%(
P/
P
dessagnu
dessag
VT − B2
100
90
80
70
60
50
40
30
20
10
0
0 0.05 0.1 0.15 0.2
Aeration Number
)%(
P/
P
dessagnu
dessag
SP − A1
100
90
80
70
60
50
40
30
20
10
0
0 0.05 0.1 0.15 0.2
Aeration Number
)%(
P/
P
dessagnu
dessag
VT − A4
Figure 4.9: Power reduction ratio (P /P ) plotted as a function of aeration number
gassed ungassed
for rotors included in detailed testing. Tests were performed at a various tips speeds ranging
from 1.45 to 6.5 m/s. Tests were conducted in a 35.3 liter tank in the absence of solids and
reagents. All rotors were 2.75 inches in diameter.
dimensionless along both axes, it should be applicable regardless of scale.
The correlation shown in Figure 4.9 depicts power reduction as a strong function of
aeration number. The similarities in the slope of power reduction and the nature of the fitted
functionindicatethattheVT-B2andtheSP-A1areexperiencingsimilarphenomenaasair
is introduced to the rotor. Over the experimental range, the power reduction of these rotors
is nearly linear. Conversely, the behavior of the VT - A4 is drastically different. The slope
is much steeper, the curve is obviously cubic, and the curve includes an apparent inflection
point at Aeration Number = 0.088. Consequently, further investigation may show that the
VT - A4 rotor may be fundamentally dispersing air differently than its counterparts. By
physical comparison, the VT - A4 has wider blades, and larger air holes than the SP - A1.
However from a practical standpoint, the behavior of these plots definitively indicates that
the VT - A4 rotor will lose power more rapidly as air is introduced.
Another way to visualize the power reduction data is by contour plots. Similar to
Figure 4.7, Figure 4.10 shows power draw as a function of tip speed and air velocity. Lighter
colors in the plot indicate higher power draws while darker colors indicate lower power
draws. White regions indicate the conditions in which the machines cannot operate (i.e.
the flooding condition). This plot may be used to identify the relative sensitivity of power
draw to changing operational conditions for each rotor. Wider contour regions indicate a
98 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
VT − B2
0
7
0
2
0
0
1
0
4
0 7
0 3
1
0 2
0 0 1
0 4
0 7
0
3
1
Tip Speed (m/s)
)s/mc(
yticoleV
saG
SP − A1
1.6 200 0
4 7
1.4
1.2 0
0
1
1 0 2
0 0 4 7
0.8
0.6
0 0 1
0.4 30
1
0 2
0 0
0.2 4 7
0
3 4 5 6
Tip Speed (m/s)
)s/mc(
yticoleV
saG
VT − A4
1.6 1.6 0
2
White =
Flooding Cond.
1.4 1.4
1.2 1.2
s)
1
y
(cm/
1 20 40
cit
00 .. 68
Gas
Velo 00 .. 68
N A =
0.0 8 8
0
0
1
0.4 0.4
0
2 40 70
0.2 0.2
0 0
3 4 5 6 3 4 5 6
Tip Speed (m/s)
Figure 4.10: Contour Plot showing power draw as a function of both tip speed and gas
velocity. The color scale signifies light color as regions of high power draw and dark colors
as regions of low power draw. Regions outside the colored portion signify the flooding
condition. The red line on VT - A4 indicates an aeration number of 0.088 – the coordinate
of the inflection point in Figure 4.9. Power values are given in units of Watts.
lower gradient and lower sensitivity. Smaller contour regions indicate a steeper gradient and
higher sensitivity. By visual comparison, the VT - B2 and the SP - A1 show extremely
similar behavior over the experimental range. The VT - A4 shows similar, linear behavior
up until the aeration number of 0.088 (the value of the inflection point from Figure 4.9,
indicated by a red line in Figure 4.10). Above this value, the contour regions of the VT -
A4 rotor show highly nonlinear behavior, indicative of the enhanced power reduction as air
is added.
4.3.3 Operational Limits
During the detailed testing, the operational limits were tested to identify the relative
robustness each rotor has to changing working conditions. As forced air is added or the
tip speed is reduced, the flotation rotor will lose the ability fully disperse the air and the
ability to suspend solids. Either of these limits signifies a condition in which the flotation
99 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
Table 4.4: Detailed Testing: Flooding Condition Measurement Parameters
Parameter Type Parameter Value
Rotor Diameter 2.75 in.
Equipment Tank Diameter 14 in.
Tank Volume 31.5 L
Material Solids Material None
Frother Type None
Chemical
Collector type None
Superficial Gas Velocity 1.12 to 2.42 cm/s
Operational
Tip Speed 3.0 to 6.5 m/s
cell cannot operate normally. Depending on the specific design, each rotor inherits a distinct
range of tip speeds and air velocities marked by stable operation. Ultimately, this evaluation
seeks to define this area-of-operation or area of operation in the case of two dimensions (tip
speed and air velocity). A larger area signifies a better rotor design, capable of operating in
a wider range of conditions. Here, the two primary operational limits, the flooding condition
and the sanded condition, were independently evaluated.
During this campaign, the flooding condition was studied at a range of tip speeds
varying from 3.0 to 6.5 m/s at increments of 0.5 m/s. During the tests, a tip speed was
set, and air was then added to the cell until the flooding condition was first witnessed. This
procedure was then repeated for a new tip speed. Other specific experimental details are
listed in Table 4.4.
The results of the flooding test are presented in Figure 4.11. This data presents an area
ofstableoperation(blueregion)foreachrotoroverthetestedrangeoftipspeeds. Conditions
outside of the blue region are marked by large pockets of undispersed air erupting at the
water surface near the shaft (the flooding condition). By comparing the relative magnitude
of the blue areas, each rotor can be evaluated by its overall ability to overcome flooding.
By this evaluation, the VT - B2 and the SP - A1 provide similar results, with the VT -
B2 slightly outperforming its counterpart (nominal areas of 7.26 and 6.83 for the VT - B2
and SP - A1, respectively. Conversely, the VT - A4 shows a much greater susceptibility to
flooding across the full range of air velocities (nominal area of 5.44).
As an alternate means of interpreting the data in Figure 4.11, the onset of the flooding
condition can be identified as a critical aeration number, since the aeration number incor-
100 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
2.5 3 3.5 4 4.5 5 5.5 6 6.5 7
Tip Speed (m/s)
)
N(
rebmuN
noitareA
lacitirC
A
VT − B2
SP − A1
VT − A4
Figure 4.12: Operational limits determination for detailed testing: minimum aeration num-
ber to overcome flooding. All rotors were 2.75 inches in diameter. Tests were conducted in
the absence of reagents and solids.
porates rotation speed and air flow rate. Values less than the critical aeration number are
marked by stable operation, while values greater than the critical aeration number induce
flooding. Figure 4.12 shows the critical aeration number as a function of tip speed.
Figure 4.12 shows that the critical aeration numbers of the VT - B2 and the SP - A1
have a stronger dependency on tip speed. Considering the calculation of the aeration number
(Q/ND3), this result indicates that for these rotors, the speed required to overcome flooding
is increasing faster than the added air at flooding. While the VT - A4 shows some tip speed
dependency, the slope of the fit line is much lower. As a result, the critical aeration number
is relatively constant over the experimental range.
As a final strategy to normalize the flooding data, the maximum dispersible air flow
102 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
(prior to the onset of flooding) was plotted as a function of power draw (Figure 4.13). Since
the power draw cannot be accurately measured at the flooding condition, Figure 4.9 was
used to estimate the power draw at 80% of the critical aeration number. After the power
reduction ratio was determined from Figure 4.9, the Power Number (Figure 4.3) was used
to calculate the ungassed power draw at the experimental tip speed. From the ungassed
power, and the power reduction ratio, the power at 80% of the critical aeration number was
calculated. This procure was also repeated for the data gathered during the exploratory
testing.
As shown in Figure 4.13, the maximum air flow normalizes very well to the supplied
power, regardless of the rotor design. In order to disperse a given flow of air, a specific and
consistent amount of power must be supplied from the rotor. Pumping affinity laws state
that for a fixed pump efficiency and rotor diameter, the power is proportional to the cube
of the volumetric flow. If the flooding condition is related to the volumetric flow of liquid
leaving the rotor, Figure 4.13 should fit Q = kP1/3. Since Figure 4.13 fits Q = kP1/4, other
factors may be controlling the flooding condition, the pump affinity laws do not necessarily
apply to rotors, or the rotor efficiency is changing with the operating conditions.
The second operational limit, the sanding condition, was assessed using a 5.1% solids
suspension of 203 micron monosized silica beads. During this testing campaign, the sanding
condition was studied at a range of air velocities varying from 0 to 1.9 cm/s at increments
of approximately 0.2 cm/s. During the tests, an air velocity was set, and the rotation speed
was then added until no solids were persistently settling on the tank bottom. The test was
then repeated at a new air velocity. Other specific experimental details are listed in Table
4.5.
Data from the sanding test is presented by two means. First, Figure 4.14 shows the
minimum power required to suspend solids as a function of air velocity for each rotor. Figure
4.15 presents the suspension data as an area of stable operation, in terms of air velocity and
tip speed.
Figure 4.14 indicates that the suspension capability of each rotor diminishes as air is
added to the cell. As a result, more power is needed at higher air velocities to suspend
the same amount of material. This result is expected since pumping capacity and energy
dissipation to the fluid is reduced as air is added to the cell (Figure 4.8 and Figure 4.9). The
data set here indicates that the suspension capability for each rotor is lineally diminished as
air is added (with the exception of the last points in the VT - B2 and SP - A1 plots). From
a comparative analysis, the VT - A4 outperforms its counterparts at low air velocities (and
no air), while the VT - B2 performs better at high air velocities. The SP - A1 consistently
103 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
VT − B2
2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
3 4 5 6
Tip Speed (m/s)
)s/mc(
yticoleV
riA
SP − A1
2
Sanded
1.8
Region
1.6
1.4
1.2
1
0.8
Operating
0.6
Region
0.4
0.2
Area =4.22
0
3 4 5 6
Tip Speed (m/s)
)s/mc(
yticoleV
riA
VT − A4
2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
Area =2.45
0
3 4 5 6
Tip Speed (m/s)
)s/mc(
yticoleV
riA
Area =2.36
Figure 4.15: Operational limits determination for detailed testing: silica sanding test area
plots. The blue region indicates the combination of tip speed and air velocities required for
stable operation. Areas outside of the blue region indicate the sanded region. All rotors were
2.75 inches in diameter. Tests were conducted in a 31.5 liter tank in the absence of reagents.
requires more power to suspend solids.
Figure 4.15 shows the sanding data as area of stable operation plots. Similar to Figure
4.11, these plots depict stable operation as blue regions and inoperable (sanded) regions
as white. Comparisons between the rotors can be made by evaluating the overall area of
the blue regions. A larger area indicates that the rotor can operate stably over a larger
range of operational conditions. By this means of comparison, the VT - B2 significantly
outperforms its counterparts, nearly doubling the nominal area of the SP - A1 and VT - A4
rotors (nominal areas of 4.22, 2.45, and 2.36 for the VT - B2, SP - A1, and VT - A4 rotors,
respectively).
Each plot in Figure 4.15 shows areas of linearity (tip speeds of 3.5 to 5.0 m/s for VT -
B2; 4.2 to 6.5 m/s for SP - A1; and 3.8 to 6.2 m/s for VT - A4). By the orientation of this
plot, a greater slope indicates that the rotor is less sensitive to air addition (a high slope
is marked by a large increase in air and a small increase in tip speed). According to this
comparison, the suspension capacity VT - B2 is very insensitive to air addition, since the
linear portion of the graph is much steeper than the linear portion of the other graphs.
Together, Figure 4.14 and Figure 4.15 show that the VT - B2 is the better rotor in
terms of solid suspension, regardless of the means of analysis. In general, the VT - B2 is
106 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
VT − B2
Tip Speed (m/s)
)s/mc(
yticoleV
riA
SP − A1
3
2.5
2
1.5
1
0.5
0
3 4 5 6
Tip Speed (m/s)
)s/mc(
yticoleV
riA
VT − A4
3
Flooding
Region
2.5
2
Sanded
Region
1.5
1
0.5
Operating
Region
0
3 4 5 6
Tip Speed (m/s)
)s/mc(
yticoleV
riA
3
2.5
2
1.5
1
0.5
0
3 4 5 6
Figure4.16: Operationallimitsdeterminationfordetailedtesting: silicasandingandflooding
areaplots. Thegreenregionindicatesthecombinationoftipspeedandairvelocitiesrequired
forstableoperation. Thebrownregionindicatestheoperationalconditionswhichwillinduce
sanding, and the blue region indicates the flooding condition. All rotors were 2.75 inches in
diameter. Tests were conducted in a 31.5 liter tank in the absence of reagents
very insensitive to air addition, given the relative slopes of both plots. This result is justified,
since the VT - B2 contains additional blades and dedicated bottom-side suction. The specific
components of this rotor do entail a higher power requirement (Figure 4.2 and Figure 4.3),
but these suspension tests show one benefit of such a system.
To summarize the results of the flooding test and the solid suspension test, a final area
of stable operation graph was generated (Figure 4.16). This plot combines the data from
Figure 4.11 and Figure 4.15 to consistently and conveniently depict the relative areas of
sanding, flooding, and stable operation. From this plot, the relative susceptibility to each
operational limit is shown for the three rotors. This plot further supports the VT - B2 as
the most operationally robust rotor, while highlighting the SP - A1’s increased susceptibility
to sanding the VT - A4’s enhanced susceptibility to flooding.
107 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
Table 4.6: Continuous Flotation Test Matrix
Test Tip Speed J
g
Run (m/s) (cm/s)
1 5.25 1.25
2 6.00 1.00
3 6.31 1.25
4 4.50 1.50
5 5.25 0.90
6 5.25 1.25
7 6.00 1.50
8 4.20 1.25
9 4.50 1.00
10 5.25 1.60
11 5.25 1.25
4.3.4 Continuous Flotation Performance
During the detailed campaign, flotation performance for each rotor was determined by
a series of continuous flotation tests. In the tests, the experimental conditions were rigidly
controlled in order to form a reliable comparison between the various rotors. Each rotor was
tested in identical chemical conditions and in a consistent test matrix of tip speeds and air
velocities. Table4.6showsthiscentralcompositeexperimentaldesign. Inthismatrix, runs1,
6, and 11 are repeat conditions used to assess the uncertainty of independent measurements.
Overall, this experimental design was chosen to meet three primary objectives: (1) populate
the flotation rate versus specific power plot at multiple air rates, (2) fully populate the
flotation rate versus bubble surface area flux graph at a variety of conditions, and (3) test
the rotors at extreme “corner” conditions.
Two materials were used throughout the continuous tests: 35 micron silica beads and
71 micron silica beads. The results of these tests are presented separately. Both materials
were hydrophobized by dodecylamine at neutral pH. Due to their purely spherical shape
and small size range, this material minimizes the effects of inconsistent collector absorption.
Furthermore, since the pure flotation rate is the desired response, the tests were conducted
with only a single, floatable species. The absence of a hydrophilic species eliminated the
108 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
Table 4.7: Continuous Flotation Test Parameters
Parameter Type Parameter Value
Rotor Diameter 2.75 in.
Tank Diameter 14 in.
Equipment Tank Volume 35.3 L
System Volume 86 L
Launder Lip Length 44 in.
Solids Material Silica Beads
Material % Solids 0.5%
Mean Particle Size 35, 71, 203 micron
Frother Type PPG
Frother Dosage 6.0 ppm
Chemical
Collector Type Dodecylamine
30 g/tonne
Collector Dosage
(9.4 x 10(cid:0)7 M)
Froth Depth 2 in.
Superficial Gas Velocity 0.9 to 1.6 cm/s
Operational
Tip Speed 4.2 to 6.4 m/s
Flotation Time 2 to 4.5 min
need to assay the test results, hastening the data collection phase.
Fortherotorcomparisons, trueflotation, ratherthaneffectsfromthefroth, wasdesired.
Consequently, the weight percent solids was kept intentionally low (0.5%) to reduce the risk
of froth overloading. Specific experimental details are listed in Table 4.7.
Data collected from the tests (mass flow rates and weight percent solids for the feed,
concentrate, and tailings streams) was first subjected to a mass balance routine to reconcile
redundancies. Mass and water recovery were then determined by comparing the concentrate
flow rates to the reconciled feed flow rate.
Bubble images were collected and analyzed for each individual test run. Bubble samples
were taken from the region approximately one inch below the froth, using the technique
described by Hernandez-Aguilar, Gomez, and Finch (2002). Image analysis was conducted
using an ellipse fitting routine presented by Fitzgibbon, Pilu, and Fisher (1999).
109 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
From the measured air velocity and bubble size, the air holdup was determined by the
drift flux model presented by Banisi and Finch (1994). The reconciled tailings flow rate and
the tank volume, less the air holdup were used to calculate the mean residence time. Finally,
the mean residence time and the flotation recovery were used to calculate the flotation rate
constant.
To complement the exploratory batch testing (Section 3.3.4), the data is presented by
twomeans: rateconstantasafunctionofspecificpowerinputandrateconstantasafunction
of bubble surface area flux. Figure 4.17 shows the rate constant plotted against the specific
power input for each rotor for a range of air velocities and mean particle sizes. This plot
can be interpreted as an energy efficiency plot. The abscissa (power input) represents a cost,
while the ordinate (the flotation rate constant) represents a benefit. More efficient curves
produce a higher flotation rate at a lower power cost. The uncertainty bars in this plot
indicate the relative standard deviation of each point, as determined from the repeat runs
of the center point.
While the individual trends vary, the data in Figure 4.17 generally shows that the flota-
tion rate constant increases with increasing power input. Also, in each of the six plots, the
VT - A4 is capable of producing the highest flotation rates at the lowest power requirement.
Conversely, the VT - B2 consistently produces the lowest flotation rate constants at rela-
tively high specific power inputs. These results coincide with those found in exploratory
testing (Figure ).
Figure 4.18 shows the flotation rate constant as a function of bubble surface area flux.
This plot may be interpreted as a hydrodynamic efficiency plot. Theoretically, the flotation
rate constant is proportional to the bubble surface area flux via a proportionality constant
(P) which accounts for ore floatability and collision frequency. Rotors which produce a
greater slope are capable of inducing a hydrodynamic environment which is more favorable
for collisions. The uncertainty bars in this plot indicate the relative standard deviation of
each point, as determined from the repeat runs of the center point.
Figure 4.18 presents results comparable to those found in exploratory testing (Figure
3.13). Whiletherelativemagnitudesaredissimilar,boththebatchtestingandbothinstances
of continuous testing show that the VT - A4 has an increased P value when compared to
the other rotors. Once again, this data indicates that the VT - A4 is capable of producing
hydrodynamic conditions (i.e. fluid velocities, turbulent kinetic energies) which are more
conducive for bubble-particle collisions and flotation.
Unfortunately, though, the scatter of the data presented in Figure 4.18 should warrant
110 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
caution in determining conclusive results. For example, in the 35 micron silica data, most
of the central points for each rotor reside around a common bubble surface area flux value.
Consequently, the slope of the fitting line is then very sensitive to the few points at the
extreme values. This phenomenon is also present in the SP - A1 and VT - B2 data for the
71 micron silica tests. Furthermore, given the scatter in the data, the varying slopes may
be insignificant, since most of the data tends to fall in a narrow range and roughly around a
singleline. Dependingonthescalethatthedataispresentedat, thedifferenceinslopevalues
may or may not be as pronounced as depicted in Figure 4.18. In light of these cautions, the
k (cid:0) S data should merely stand as another piece of evidence in the larger context of the
b
overall rotor comparison conclusions.
As a supplementary analytical exercise, the air holdup measurements were qualitatively
compared to the flotation rate data to test the hypothesis that air holdup is a suitable
surrogate test for flotation performance. In this bulk comparison, the full data set of air
holdup versus power data was holistically compared to the full data set of flotation rate
constant versus power. In both data sets, the operational conditions (tip speed and air
velocity) covered a similar range, with the range of the flotation data set falling completely
within the range of the air holdup data set.
Figure 4.19 shows the air holdup data set for each rotor. These plots include data at all
air velocities. To assist in visualizing the data sets, an “ellipse-of-best-fit” was superimposed
on the plot. This ellipse was determined by an optimization routine which minimized the
ellipse area while covering a designated percentage of the total points. In these cases, the
target coverage was set to 75% of the data set. By this ellipse fit, the general locus of holdup
as a function of power input for the test conditions is conveniently summarized. Figure 4.20
shows each ellipse on the same axis.
Figure 4.21 and Figure 4.22 show a similar analysis conducted for rate constant values
determined from continuous flotation tests. Here the 35 and 71 micron tests are presented
independently. Since these data sets contain significantly fewer points, the target coverage
was set to 100% of the data set. Once again, the locus of values can be easily determined
by the shape and orientation of the ellipse. Figure 4.21 shows the data points and ellipses
for each rotor independently, and Figure 4.22 shows all ellipses on the same axes.
Visual comparison of Figure 4.20 and Figure 4.22 show that the same general trends
exist for both air holdup and flotation rate constant for both particle sizes tested. In all
three cases, the VT - A4 occupies the high measured value (holdup or rate constant) at
low power inputs, while the SP - A1 shows moderate to high measured values at moderate
power inputs. Finally, the VT - B2 generally shows low measured values at moderate to high
113 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
power inputs. While this analysis certainly does not indicate a one-to-one relationship, the
trends compare favorably, and similar results would be found if either test were presented
independently.
4.3.5 Comparison of Batch and Continuous Data
Since this research has included batch and continuous flotation tests in similar exper-
imental conditions, a single experiment was used to test the validity of the plug flow and
perfectly mixed reactor models used to fit flotation rate constants. These experiments were
performed with a single material (35 micron silica) under similar chemical conditions (15
g/t dodecylamine and 8.1 ppm MIBC). Figure 4.23 shows the experimental data and model
fits for the kinetics tests. In this plot, the batch data was used to fit a fast and slow rate
constant (and proportion), using a plug flow reactor model. These kinetic parameters and
a perfectly mixed reactor model were then used to predict the flotation performance of the
continuous unit. Since the tests were run at different air rates, an S adjustment factor was
b
used to scale the rate constants between the tests.
While this single plot is not ultimately conclusive, it does indicate the predictive power
of the perfectly mixed and plug flow reactor models as well as the substantial loss in flotation
recovery in moving from a high S batch machine to a lower S continuous machine.
b b
4.4 Summary and Conclusions
Laboratory-scale tests were performed to assess the operational capabilities of various
rotors in four performance areas: power consumption, air holdup, operational robustness,
and flotation performance. In total, the VT - A4, VT - B2, and SP - A1 were subjected to
detailed laboratory experiments and numerous modes of analysis. In forming a more concise
evaluation, both quantitative and qualitative evaluation matrices was generated (Table 4.8
and Table 4.9).
These tables summarize the data in this chapter by integrating at least one indicative
measurement from the evaluation of each of the four performance criteria. These measure-
ments include: average power number, interpolation from the holdup versus power graph,
flooding test area of operation, sanding test area of operation, interpolation from the rate
constant versus power graph, and average P. The presented measurements were selected for
this summary since they efficiently convey the most useful data, retain the general conclu-
118 |
Virginia Tech | CHAPTER 4. DETAILED TESTING OF FLOTATION ROTOR DESIGN
Table 4.8: Summary of Selected Detailed Testing Results - Raw Data
Parameter SP - A1 VT - A4 VT - B2
Power Number 4.63 3.93 4.86
Power (W) to achieve
3.5% Holdup 44.3 11.8 78.0
J = 0.95 cm/s
g
Flooding Area of
6.83 5.44 7.26
Operation
Sanding Area of
2.45 2.36 4.22
Operation
k (1/min)
J = 1.25 cm/s
g
0.5 0.68 0.42
35 micron silica
Power=1.5 kW/m3
Average P (x10(cid:0)3)
2.59 4.70 2.00
71 micron silica
sions found in the full data set, and generally preserve the paradigm of power-normalization
as a basis for a fair comparison. Table 4.9 provides another level of simplification by provid-
ing ordinal rankings for each evaluation parameter.
From the data set collected during the detailed testing campaign, six key conclusions
are ascertained:
1. The results of the detailed testing campaign generally bolster those of the exploratory
campaign. Themodesofanalysisofthefouroperatingcriteriahaveprovidedconsistent
comparisons. Furthermore, using the SP - A1 as the benchmark, both prototype rotors
showed strengths and weaknesses in the various criteria.
2. The degree of uncertainty for several measurements has been quantified by replicate
testing. In general, the measured performance differences exceed the uncertainty.
3. In terms of power draw, air dispersion, flotation energy efficiency, and flotation hydro-
dynamic efficiency, the VT - A4 substantially outperforms its counterparts. Similar
resultswerefirstpresentedintheexploratorytesting,andfurthersupporthasbeenpro-
vided in the detailed testing. Much of this rotors strength lies in its ability to operate
120 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.