You are on page 1of 4

Light Metals 2009 Edited by: Geoff Bearne

TMS (The Minerals, Metals & Materials Society), 2009

INITIATIVES TO REDUCE ANODE EFFECT FREQUENCY AT DUBAL


Ali. H. Mohammed, Maryam Mohamed Al-Jallaf, A.Kumar
Dubai Aluminium Company, PO Box 3627, Dubai, U.A.E
Keywords: Anode Effect, Alumina Depletion, Demand Feed
Abstract
Anode effect is normally manifested by a sudden increase
in cell voltage. It is associated with the depletion in
alumina concentration and the onset of concentration
polarization. For a given alumina concentration and
temperature, the anode current density at which the anode
effect occurs is referred to as critical current density
(ccd). In other words, alumina concentration is directly
proportional to ccd. If potline amperage is raised, the ccd
threshold.

Extensive studies have been carried out in smelters around


the world to understand the fundamental cause of an anode
effect. The exact nature of the onset of an anode effect is
still shrouded in mystry. However, the consensus is that
anode effects are detrimental to cell operation; they result
in reduced energy consumption and cause emission of CF4
and C2F6 gases.
With the intention of reducing carbon footprint, there is an
excellent opportunity to reduce anode effects and the
resulting PFC emissions.

The most commonly accepted mechanism for anode effect


is the formation of a non-wetting, insulating fluoride film
on anodes when alumina concentration is low [3]. Then the
potential of the active anode surface increases to a point
where CF4 and C2F6 start to evolve. The non-wetting of the
anode also causes an increase in resistance and hence
voltage gradient at the anode to the point of current
flowing across the film is in the form of sparks.

Occurrence of an anode effect was studied in relation to


different aspects; alumina feed rate, work schedule,
cathode type, operating parameters, mechanical issues, etc.
Onset of an anode effect was primarily due to inability of
the response strategy to deal with it efficiently. Alumina
fines and pencilling of crust breaker tip were the other
reasons.

The gaseous per fluorocarbon emissions (PFCs) are potent


greenhouse gases: e.g. 1 kg of PFCs (approximately 90%
CF4 and 10% C2F6) has the same greenhouse gas effect as
6,500 kg of CO2 for CF4 and 9200 kg for C2F6.

A strategy to kill an incipient anode effect was developed


and implemented. Consequently, AE frequency has
improved by more than 50%. The logic is very promising
as it lowers alumina concentration whilst effectively
reducing anode effect frequency. Lower alumina
concentration is conducive to improving current efficiency.
The strategy is being fine tuned to achieve a sustained
anode effect frequency of <0.05.

Reduction in both, anode effect frequency and duration has


become one of the major challenges for aluminium
industry in terms of minimising loss of energy and
reducing carbon footprint.
.
Methodology
Quenching an anode effect involves lowering the anodes to
short them with cathode, breaking the crust, feeding
alumina at a rapid rate. The success rate in an automated
anode effect quenching mechanism is normally greater
than 97%.

Introduction
Anode effects are not desirable as they increase energy
consumption and PFC emissions [1] [2]. Virtually all the
voltage increase during an anode effect occurs at the anode
- bath interface and this manifest as heat. The rate of heat
generation rises to as much as ten fold and causes a rapid
rise in cell temperature which sometimes exceeds 1000C.
This in turn lowers current efficiency and melting of side
ledge. A typical cell layout is shown in figure 1.

Cause and Effect Analysis


A cause and effect analysis of occurrence of an anode
effect was prepared, please refer figure 2.
Analysis of anode effects shows the following features:

Figure 1. Cross section of a Cell.

259

Potlines with dense phase alumina conveying system


had fewer anode effects.
The majority of anode effects were during anode
setting.
AE frequency in cells with graphitized cathodes was
higher.
The demand feed was not always able to efficiently
react in time to prevent anode effects.

Material

Machine

Man
Actioning
New cell
AEs

Primary / Air Supply


Secondary
Alumina
Feeder
Blockages

AlF3 wt%
Fines

PMPF/CB

Line 6 - AEF vs Work Schedule


Setting
anodes
correctly

Breaker
design

Idle 1
1%

Idle 2
3%

Problem
solving high
AE cells

Anode Effect
Frequency
Feed Logic

AE Quench
Tables

High PFCs

Resistance
Smoothening

Method

Tap
19%

Bath
Temperature

Anode
Changing
77%

Environment

Figure 2. Cause and Effect analysis of Anode Effects.

Analysis of Causes
Figure 4. Occurrence of Anode Effects in relation to
Potline Work Schedule.

AEs in relation to Feed Cycle


The frequency of anode effects was analysed in each of the
alumina feed window and in relation to potroom activity.

During anode setting the average anode current density


increases more than 10% as a result of the lowering in bath
levels and the anode removal. This raises the critical
alumina concentration. Signal filtering will aggravate the
problem.

The alumina feed cycle consists of four feed windows,


super fast (SF), over feed (OF), base feed (BF) and under
feed (UF and UF1). It may be observed (figure 3) that the
majority of anode effects occur in the SF and OF windows.
These windows commence in succession after the under
feed window, with the time on the SF being quite short.

Effect of amperage increase on AE frequency


Potline amperage was raised gradually and steadily over
years. While the plan lifted hot metal production, it had an
adverse effect on AE frequency as shown in figure 5.

The underfeed cycle is adaptive and its duration varies


depending on the cells resistance trend which is directly
related to the alumina concentration.
The first derivative of the pseudo resistance, or as it is
commonly called, slope (dR/dt) is used to estimate the
undefeed limits. Change in resistance (R) is also used
in conjunction to improve end point repeatability. Both, the
dR/dt and R are manipulated through smoothing
constants however.

Impact of Line Load on AEF

Line 6 - AEF vs Feed Window


UF1
0.4%

SF
12.0%

UF
0.0%

0.35

225

Ave Line Load (kA)

0.30

220

Ave AEF (no/pd)

0.25

215

0.20

210

0.15

205

0.10

200

0.05

195

BF
23.2%

AEF

Line Load

230

0.00
2001

2002

2003

2004

2005

2006

2007

Year

Figure 5. Impact of Line Load increase on AEF.


The impact of amperage increase is also to raise the current
density but it also increases the rate of alumina depletion.
Hence signal filtering again becomes a contributing issue.

OF
64.4%

Other Issues Impacting AEF


Figure 3. Occurrence of Anode Effects in relation to Feed
Windows.

The AEF varies between different potlines, which also


differ in age and minor design features. Accordingly other
issues were also examined. Several significant links were
found:

While theoretically the AEs are not expected to occur in


the dominant zones, it is noted the detection signal is
heavily filtered.
AEs in relation to work practice schedule.
Potlines operate on a 32 hour work cycle - metal tapping is
followed by anode setting in the next shift. The next two
shifts are idle. The majority of anode effects occurred in
the anode setting shift as detailed in figure 4.

260

Wear of crust breaker tip

Bath build up on the crust breaker hindered the


delivery of the prescribed alumina dose at the right
time as illustrated in figure 6.

Unexpected variability in alumina delivery from the


conveying and delivery system that is installed in

some of the Dubal potlines. A typical problem was an


unannounced massive surge in the delivery of alumina
fines to a cell or a small group of cells. Analysis of
anode effects in relation to fines on a sample basis
confirms this. With alumina fines, the dissolution is
hindered causing cell instability and AEF to increase,
figure 7.

Preventative Maintenance
Several approaches have been used and the new
preventative maintenance includes greater attention to
feeder performance and more stringent limits for
replacement of crust breakers.
Changes for Terminating the Underfeed
The variations in slope during the under feed was analysed
in depth, and a modified programme was developed to
activate and switch to overfeed again. It consists of a series
of steps to feed cells at a high rate and having simultaneous
anode beam movements.
During the trial period, bath samples were withdrawn at
five minute interval to establish rate of alumina dilution.
The programme in fact lowered alumina concentration
marginally, figures 9 and 10.
In order to evaluate on a bigger scale, the programme was
down loaded in B-room cells (120 test cells) of potline 5.
Cells in A-room were selected as the control cells for this
trial. A 50% reduction in AEF has been achieved with
sustainable results, figure 11.

Figure 6. Bath Build Up on Crust Breaker Tip.

0.140

Line 6 : AEF vs Noise Activations(no/pd)

0.120

The programme has now been extended to two potlines


(480 cells).

R2 = 0.37

0.100
0.080
0.060

Alumina concentration (%) in Under Feed Window


in a cell with semigraphitic cathode blocks

0.040
0.020

2.9

2.8

2.5

3.1

2.9

3.2

3.0

3.4

The Feeding control strategy. The strategy works


well in some potlines with an AEF of <0.10 (yearly
average), while in others it resulted in an AEF of
~0.30. The problem was traced to the slow response
to an increase in cell voltage. On occasions the
response time is not fast enough to prevent an anode
effect. This is another problem likely to be linked
with inappropriate signal filtering a feature outside
the scope of operations because of the hardware
design.

2.0
1.5

11

10

1.0
1

3.0

3.4

Figure 7. AEF and Noise Activations.

3.5

3.5

0.50

3.6

0.40

0.30
AEF

% Alumina in bath

0.20

3.6

4.0

0.000
0.10

Time Elapsed (minutes) from start of UF Window

Figure 9. Alumina Concentration before the Trial.

Mitigating Limitations Impacting AEF


Alumina concentration (%) in Under Feed Window in
a cell with semigraphitic cathode blocks

Alumina Dumps prior to Anode Setting


The critical current density (ccd) increases during anode
setting. In order to partially match the alumina
concentration requirements, extra alumina dumps were
administered. It improved the situation. The AE frequency
could be lowered from 0.35 to 0.27. Please refer to figure
8.

4.0

216.0

0.2

215.0

0.15

214.0

Amp

0.1

213.0

Poly. (AEF)

AEF

Anode Set Dumps commenced


through A set logic

38

44

50

15

21

27

33

2.5

2.9

1.0

Time Elapsed (minutes) from start of UF Window

Figure 10. Alumina Concentration during the Trial Period.

0
32

2.8

1.5

0.05

212.0

2.5

0.25

25

0.3

217.0

2.6

218.0

20

0.35

15

0.4

219.0

2.0

10

220.0

2.5

0.45

3.0

0.5

221.0

AEF (no/pd)

Amp (kA)

Line 5: Weekly Amperage Increase Vs AEF: Year 2004-2005


222.0

3.0

%Alumina in bath

3.5

39

Figure 8. Effect of Alumina Dumps prior to Anode Setting.

261

Line 5 - Room wise AEF


0.45
0.40

AEF

0.35
0.30
0.25
0.20
A-Room (Control Cells)
B-Room (Test Cells)

0.15

10/10/2008

03/10/2008

26/09/2008

19/09/2008

12/09/2008

05/09/2008

29/08/2008

22/08/2008

15/08/2008

0.10

Week ending dates

Figure 11. Improvement in AEF.


Concluding Comments
Reducing anode effects is a challenge facing aluminium
smelters minimise wastage of energy and reduce green
house gas generation.
The mechanical aspects (crust breakers, feeders, etc) of a
cell shall and will continue to interfere with smooth
functioning of demand feed. Alumina transport plays an
important role in fines generation and in controlling anode
effects.
Although the above challenges make it practically
impossible to achieve zero anode effect frequency with our
installed hardware, a very significant decrease could be
achieved by tracking the voltage rise and killing an
incipient anode effect. Whilst implementing a pre-emptive
anode quench logic it is imperative to follow alumina
dilution so that cells are not overfed.
Our journey to eliminate anode effects has just
commenced, work is in progress towards an ambitious
target of <0.05 AE frequency.
Acknowledgement
The authors gratefully acknowledge the guidance, support
and encouragement of Dubal management in carrying out
this project. The authors acknowledge the invaluable
critique by Dr. B.J. Welch in preparing the manuscript.

References

1.

K. Grjotheim and B.J. Welch (Aluminium Smelter


Technology, 1980)116-118

2.

K. Grjotheim, C.Krohn, etal (Aluminium Electrolysis,


2001), 277-280

3.

W. Haupin and E.J. Seger, Light Metal 2001, pages


329 334

262

You might also like