Minerals Engineering 18 (2005) 977–980

This article is also available online at:

www.elsevier.com/locate/mineng
Technical note

Operating range of a flotation cell determined from

gas holdup vs. gas rate
R. Dahlke 1, C. Gomez, J.A. Finch *

Department of Mining, Metals and Materials Engineering, McGill University, 3610 University Street, Montreal, Canada H3A 2B2

Received 15 October 2004; accepted 16 December 2004

Available online 13 February 2005

Abstract

Flotation cells have an operating gas rate range. A method to define the range objectively using the relationship between gas
holdup and gas rate (eg vs. Jg) is introduced. This is made tractable by the necessary sensors becoming available. Examples in three
cell types at three plants illustrate the procedure.
Ó 2005 Elsevier Ltd. All rights reserved.

Keywords: Flotation machines; Process instrumentation

1. Introduction and has been adapted to mechanical machines. The

methodology is described and illustrated.
A flotation machine has an operating range of gas
(air) rate; too low and no concentrate is produced or
sanding may occur, too high and ‘‘boiling’’ (break- 2. Gas holdup vs. gas rate
through of large bubbles) and a general disturbed look
ensues. Operations normally want to keep inside the 2.1. Methodology
range. This becomes important, for example, when try-
ing to take advantage of setting a gas rate profile down Gas holdup (eg) is the volumetric fraction (usually ex-
a bank, which may mean each cell is at a different setting pressed in %) of gas in the gas–slurry mixture. It is mea-
but all should be in their operating range (Cooper et al., sured here using a conductivity-based sensor, shown in
2004). Defining the range visually is subject to interpre- Fig. 1 (Tavera et al., 1996; Gomez and Finch, 2002).
tation. A procedure is introduced where measurement of The sensor comprises two flow cells, an open cell to
gas holdup vs. gas rate is used to provide an objective measure the conductivity of aerated slurry (jslg) and a
definition, in particular, of the upper end of the range. syphon cell to measure the conductivity of de-aerated
It is based on the approach used to characterize hydro- slurry (jsl). The syphon cell has a tapered base (spigot)
dynamics in flotation columns (Finch and Dobby, 1990) that acts to restrict bubble entry and establish a bulk
density difference between the fluid inside and outside
the cell. This drives a flow of slurry in through the top
of the cell (hence the description ‘‘syphon’’), which both
completes the elimination of air (the bubbles cannot en-
*
Corresponding author. Tel.: +1 514 398 1452; fax: +1 514 398 4492.
ter against the outflow of slurry) and replenishes the cell
E-mail address: jim.finch@mcgill.ca (J.A. Finch). contents. With the two conductivities, MaxwellÕs model
1
Now with Cambior. (included in figure) is solved. Note, the equation

0892-6875/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mineng.2004.12.013
978 R. Dahlke et al. / Minerals Engineering 18 (2005) 977–980

Fig. 1. Schematic of gas holdup sensor.

Fig. 2. (a) Schematic of gas rate sensor. (Note, tube contents emptying as air accumulates.) (b) Example pressure-variation curves.

involves only the ratio of the conductivities, the absolute variation’’ curves are included, Fig. 2b). Again, the unit
slurry conductivity, or whether it varies during measure- is well described in the references.
ment (e.g., due to a change in % solids), is not a factor. In use, the two sensors are placed close to the same
Operation and validation of the sensor is described in location in the cell below the pulp/froth interface. The
the two references. location should avoid baffles, shrouds, booster cones
Gas rate is expressed as the volumetric flow rate of or launders (as examples). The choice may reflect
gas (air) (Qg) per cross-sectional area of the cell (A), requirements. For example, if cells in a bank are to be
i.e., the gas superficial velocity (Jg = Qg/A). All the data compared then an accessible, geometrically similar loca-
reported here were measured using the on/off2 McGill Jg tion in each may be the target.
sensor (Fig. 2a) (Gomez and Finch, 2002; Gomez et al.,
2003). This comprises a tube to collect bubbles by natu-
ral buoyancy and inferring Jg from the rate of increase 2.2. Example eg vs. Jg
in pressure once a valve is closed (specimen ‘‘pressure-
It was when exploring Jg profiles at Brunswick Mine
(division of Noranda), that the need to know the oper-
ating range of the cell became evident (Dahlke et al.,
2
To distinguish from the continuous version (Torrealba-Vargas 2001). Fig. 3 shows two runs on the same Denver DR
et al., 2004). 100 cell (nominally 100 ft3) in the final (4th) Zn cleaner
 R. Dahlke et al. / Minerals Engineering 18 (2005) 977–980 979

Fig. 3. Examples of gas holdup vs. gas rate curves: same Denver DR Fig. 5. Gas holdup vs. gas rate for three cells types, Galigher
100 cell on two occasions. Inset depicts sensor position relative to (Matagami), OK 50 (Red Dog) and Denver DR 100 (Brunswick,
impeller; the cell is 70’’ wide and 60’’ long. average of data in Fig. 2).

bank. The relationship is similar to that found in flota- 2.3. Other experiences
tion columns (Finch and Dobby, 1990): a range in Jg
over which eg responds consistently, almost linearly, This technique of establishing the operating range is
above which eg varies erratically. In the present case routinely used in plant campaigns conducted by McGill
the transition3 occurs at ca. Jg = 2.5 cm/s, which is iden- teams. Three examples of operating range, with cells
tified with the maximum Jg of the operating range. ranging from a 4 ft3 Galigher–Agitair cell at Matagami
Sometimes this upper limit is indicated by ‘‘boiling’’, Mines (division of Noranda) to a 50 m3 OK 50 cell at
but not always. Teck ComincoÕs Red Dog Operation, are given in
The result at low Jg is interesting: instead of trending Fig. 5.
to eg = 0, with the valve closed there remained a finite The trends (and ranges) in each example are differ-
gas holdup (and gas rate). This is either a leaking valve ent. There are a number of reasons. The eg vs. Jg re-
or gas being entrained in the slurry from previous cell in sponse not only reflects cell mechanism but solution
the bank. The low limit was established as that giving chemistry (particularly frother dosage and possibly salt
some concentrate flow (i.e., a ‘‘practical’’ limit) and cor- concentration) and slurry properties (particle size, %
responded approximately with the deviation from the solids, etc.). In this early work the gas rate was taken
linear response at ca. 0.5 cm/s (there was no sanding is- as the slope of the pressure-variation curve (Fig. 2b),
sue with these cells). which incurs a bias compared to using the full model
Fig. 4 is taken from Cooper et al. (2004) and shows (Torrealba-Vargas et al., 2004). This does not affect
examples of three gas rate profiles set along the bank: establishing the operating range but must be consid-
the operating range in Fig. 3 (0.5–2.5 cm/s) is clearly ered when comparing data. Along the same lines, to
respected. compare gas holdup and gas rate data must be cor-
rected to the same conditions (pressure and
temperature).
Not all cells investigated have shown such a consis-
tent eg vs. Jg trend, some revealing a very narrow oper-
ating range. As the database expands it may be possible
to determine the reason(s).

3. Conclusion

The use of gas holdup vs. gas rate (eg vs. Jg) is intro-
duced as an objective way to define the operating gas
rate range of a flotation cell. Difficult to determine be-
fore, the necessary sensors (for eg and Jg) are now avail-
able (the ones used here are briefly described but others
Fig. 4. Examples of three air distribution (Jg) profiles (after Cooper would suffice). Examples at three plants illustrated the
et al., 2004). technique.
980 R. Dahlke et al. / Minerals Engineering 18 (2005) 977–980

Acknowledgments Dahlke, R., Scott, D., Leroux, D., Gomez, C.O., Finch, J.A., 2001.
Trouble Shooting Flotation Cell Operation Using Gas Velocity
Measurements. In: Proceedings-33rd Annual Meeting of the
Funding for this work was sponsored initially by Canadian Mineral Processors (division of CIM), January 23–25,
Inco, Teck Cominco, Falconbridge and Noranda and pp. 359–370.
now including Corem and SGS Lakefield Research un- Finch, J.A., Dobby, G.S., 1990. Column Flotation. Pergamon Press, p.
der the Natural Sciences and Engineering Research 180.
Council of Canada (NSERC) Collaborative Research Gomez, C.O., Finch, J.A., 2002. Gas dispersion measurements in
flotation machines. CIM Bull. 95 (1), 73–78.
and Development (CRD) program. Since 2001, the Gomez, C., Torrealba-Vargas, J.A., Dahlke, R., Finch, J.A., 2003.
work is also supported under a NSERC CRD sponsored Measurement of Gas Velocity in Industrial Flotation Cells. In:
by the Amira P9 project. Lorenzen, L., Bradshaw, D.J. (Eds.), XXII International Mineral
Processing Congress, Cape Town, September 29–October 3, vol. 3.
S. African Inst. Min. Metall. pp. 1703–1710.
References Tavera, F., Gomez, C., Finch, J.A., 1996. A gas holdup sensor for
slurry–air systems. Trans. IMM Sect. C. 105, C99–104.
Cooper, M., Scott, D., Dahlke, R., Finch, J.A., Gomez, C.O., 2004. Torrealba-Vargas, J.A., Gomez, C.O., Finch, J.A., 2004. Model for the
AImpact of Air Distribution Profile on Banks in a Zn Cleaning JK and On-off McGill Gas Velocity Sensor. Final Report AMIRA
Circuit@. In: Proceedings 36th Annual Meeting of the Canadian Project P9M Volume II Flotation Module, AMIRA International.
Mineral Processors of CIM, January 20–22, pp. 525–540. pp. 47–70.