LOW ENERGY CONSUMPTION CELL DESIGNS INVOLVING COPPER INSERTS

AND AN INNOVATIVE BUSBAR NETWORK LAYOUT

Marc Dupuis

GéniSim Inc., 3111 Alger St., Jonquière, Québec, Canada G7S 2M9

marc.dupuis@genisim.com

Keywords: MHD cell stability; busbar design; mathematical modeling; power efficiency; copper collector bars.

Abstract

Two innovations presented by the authors recently at ICSOBA

conferences allow to very significantly reducing both the cathode

and the busbar voltage drop [1,2].

This paper combines the usage of those two innovations with the

usage on the new anode stub hole design presented at the

Aluminiun of Siberia conference [3] to come up with a very low

energy consumption cell design.

Introduction

The author has been involved in the modeling of aluminium

reduction cells for the last 30 years. In 1988, he designed the

cathode of the Alcan A310 prototype cell, the first cell to operate

above 300 kA in 1989. The thermo-electric cathode slice model

he developed was presented at the 1991 ANSYS conference [4].

The Figure 6 of that paper, reproduced in Figure 1 shows the

Figure 2: VAW CA300 inspired cathode side slice model mesh

model mesh highlighting the cell lining and potshell design.

That 300 kA demonstration model typical of the early 1990’s state

of the art in cell design became the starting point or base case for

two styles of retrofit studies, the first one aiming at minimizing

the cell energy consumption presented in [6] and the second one

aiming at maximizing the cell productivity presented at the TMS

2000 conference [7]. Part of the Table II for a subsequent article

presented in the magazine ALUMINIUM in 2005 [8] presenting

the key design parameters and predicted operational results of

those two cell retrofits is reproduced in Table I.

The key design changes that are allowing either the reduction of

the cell energy consumption to 12 kWh/kg or the increase of the

cell productivity by 17% are the change of the type of cathode

material from 30% to 100% graphitic carbon block, the reduction

of the anode to cathode distance (ACD) from 5 to 4 cm and a

change of the bath chemistry (and alumina feed control logic)

increasing the current efficiency. Other changes are required to

obtain an appropriate ledge thickness at a very different level of

Figure 1: Alcan A310 cathode side slice model mesh

heat dissipation. Per example, the known strategy to increase the

cell productivity is to increase the anode length, decrease the ASD

A few years later as an independent consultant, the author

and use thin silicon carbide side walls. In addition, anode stud

developed a similar demonstration model strongly inspired by the

diameter and collector bar size can be increased while the anode

VAW CA300 cell design presented in JOM in 1994 [5]. The

cover thickness can be decreased.

resulting thermo-electric cathode slice model that was first

presented in Figure

12 of the

1997 CQRDA aluminium

The design strategy to decrease the cell energy consumption to 12

electrolysis course [6] is reproduced in Figure 2.

kWh/kg is the opposite, anode stud diameter and collector bar size

can be decreased while the anode cover thickness can be

The A310 and the CA300 cells were designed at about the same

increased. What is a lot more significant is that the cell

time and operated at about the same amperage. Both designers

productivity must be decreased by 12%, which explains why so

clearly respected similar design guidelines for the choice of the

far the industry have not move in that direction despite the fact

type of cathode blocks and side blocks, the thickness of that side

that operation at that level of power efficiency have been reported

block, the size of the anode to side wall distance (ASD), the

as soon as the early 80’s [9,10].

location of the anode shadow, etc.

Table I: Design and predicted operational data,

Table II: Design and predicted operational data,

part of Table II in [8]

original work presented in [14]

Base case

Amperage

300 kA

265 kA

350 kA

Amperage

500 kA

600 kA

Nb. of anodes

Anode size

1.6 m X 0.8 m

1.7 m X 0.8 m

Anode size

1.95 m X 0.8 m

1.95m X .665m

2.0m X .665m

Nb. of anode studs

3 per anode

Nb. of anode studs

3 per anode

4 per anode

Anode stud diameter

18 cm

16 cm

19 cm

Anode stud diameter

20.5 cm

17.5 cm

Anode cover thickness

16 cm

17.5 cm

10 cm

Anode cover thickness

10 cm

Nb. of cathode blocks

Cathode block length

3.47 m

3.43 m

3.67 m

Cathode block length

4.17 m

Type of cathode block

HC3

HC10

Type of cathode block

HC10

Collector bar size

20 cm X 10 cm

18 cm X 10 cm

20 cm X 10 cm

Collector bar size

20 cm X 10 cm

Type of side block

HC3

Anthracite

SiC

Type of side block

Si C

SiC

Side block thickness

15 cm +

10 cm +

Side block thickness

10 cm +

7 cm +

ASD

35 cm

30 cm

ASD

30 cm

28 cm

Calcium silicate thickness

3.5 cm

6.0 cm

3.5 cm

Calcium silicate thickness

3.5 cm

Inside potshell size

14.4 X 4.35 m

Inside potshell size

17.8 X 4.85 m

ACD

5 cm

4.15 cm

4 cm

ACD

4 cm

3.5 cm

Excess AlF3

10.90%

13.50%

Excess AlF3

13.50%

12.00%

Anode drop

303 mV

273 mV

323 mV

Anode drop

354 mV

265 mV

318 mV

Cathode drop

285 mV

213mV

292 mV

Cathode drop

314 mV

87 mV

104 mV

Anode panel heat loss

240 kW

183 kW

284 kW

Anode panel heat loss

409 kW

420 kW

449 kW

Cathode bottom heat loss

176 kW

132 kW

202 kW

Cathode bottom heat loss

273 kW

238 kW

240 kW

Operating temperature

973.2 °C

956.1 °C

960.4 °C

Operating temperature

963.1 °C

955.6 °C

964.8 °C

Liquidus superheat

6.7 °C

2.4 °C

6.7 °C

Liquidus superheat

9.4 °C

2.6 °C

11.8 °C

Bath ledge thickness

8.66 cm

23.5 cm

9.09 cm

Bath ledge thickness

6.15 cm

29 cm

4.76 cm

Metal ledge thickness

4.12 cm

9.01 cm

4.42 cm

Metal ledge thickness

2.42 cm

26 cm

1.07 cm

Current efficiency

94.00%

95.70%

96.10%

Current efficiency

95.90%

96.50%

96.40%

Internal heat

628 kW

422 kW

713 kW

Internal heat

1043 kW

760 kW

1140 kW

Energy consumption

13.72 kWh/kg

11.93 kWh/kg

13.43 kWh/kg

Energy consumption

13.61 kWh/kg

12.1 kWh/kg

13.26 kWh/kg

Clearly, a cell designer cannot at the same time aim at maximizing

New retrofit study aiming at minimizing cell energy

the cell productivity and minimizing the cell energy consumption.

consumption even further

This is why Rio Tinto per example has developed and is offering

both the AP60 and the APXe cells based on the same basic

In the past 30 years, the market conditions of high metal value and

platform [11].

the existence of regions of the world offering inexpensive

electrical power were favorable for new cell designs maximizing

Yet, new choice of materials and new and innovative design ideas

cell productivity while maintaining power efficiency in the 13-

can always be put to contribution in order to further increase the

13.5 kWh/kg range.

cell productivity or decrease the cell energy consumption.

Another tendency is to continue to increase the cell size in order

The market conditions have evolved recently to a much lower

to keep reducing both the cell OPEX and CAPEX. It is in that

metal value and far less availability of inexpensive electrical

power. In that context, the metal production cost is getting quite

context that the AP60 platform replaced the AP30 platform that

itself replaced the AP18 platform [12] per example.

close to the metal market value and a reduction of the energy

consumption can make the difference between operating at profit

For one, the author have been advocating that, despite the

or at loss.

difficulties that have always been limiting the rate of increase of

Technically, 12-12.5 kWh/kg have been achieved multiple times

the cells size since the beginning of the industry, he could foresee

no technical limitation that could limit further increase of cell size

and as for operation at 13-13.5 kWh/kg range, under the current

in the future. It is in that context that the author presented a 500

market conditions it might well become the preferable operational

range. The next question is technically, regardless of market

kA cell design in 2003 in [13] and a 740 kA cell design in 2005 in

[8].

conditions, how much lower can we manage to go?

Reducing the cell energy consumption means reducing the cell

In yet another cell retrofit demonstration study paper in 2011 [14]

the author took advantage of new design innovations like collector

voltage drop which in turn means reducing the cell ohmic

bar copper inserts, anode slots and a new type of anode stub hole

resistance. This statement assumed that at

95-96% current

efficiency, we cannot expect significant gain to come from that

design [3] to retrofit the 500 kA cell presented in [13] into a more

productive

600 kA cell operating at about the same power

factor. Leaving aside the bath ohmic resistance for now, this

efficiency. As an intermediary step not quite optimized in term of

leaves three distinct ohmic resistances to work with: the anode,

cathode and busbar resistances.

thermal conditions, a 500 kA cell operating at

12.1 kWh/kg was

also developed. Table II presents detailed data of that study.

Cathode design with copper collector bars

Figure 5 presents the horizontal currents in the metal pad. They

As presented in Table II, the intermediary cell operating at 500 kA

have been reduced as compared to those presented in Figure 2 of

presented in [14] was operating at 87 mV at cathode drop by using

[2]. Unfortunately, the center channel creates a gap that prevents

the copper collector bars design presented in Figure 3.

the total elimination of a horizontal component in the metal pad

current regardless of the size of the copper collector bars used.

copper

steel

Figure 3: Copper collector bar design originally proposed in [14]

At the time, it was speculative that such a collector bar design

Figure 5: Metal pad current density

could be actually build, but it is no longer the case today after

Storvik AS presentation at the ISCOBA 2015 conference [15].

External compensation current (ECC) busbar network design

Furthermore, at the TMS 2016 conference KAN-NAK advocated

that copper collector bars don’t even need to be protected by a

As presented in

[2], the idea of taking advantage of copper

shell of steel [16].

collector bars to extract

100% of the cell current on its

downstream side came to the author as a way to reduce of busbar

As first presented in [2], what the author did not realized in 2011

weight of its own reversed compensation current (RCC) busbar

is that with the usage of copper collector bars, 100% of the cell

configuration.

current can be extracted on the downstream side without

generating excessive horizontal current in the metal pad or

It happens that the same idea is easily applicable to existing ECC

producing excessive cathode voltage drop.

busbar configurations. In that case, the busbar network is reduced

to only the anode risers so it is the preferable busbar configuration

The results presented in [14] and in [2] are for a 20 cm x 10 cm

if the main goal is to minimize the busbar voltage drop in order to

copper collector bar size. When the current is extracted all on the

minimize the cell energy consumption.

downstream side of that cell running at

500 kA, the current

density in the bar doubles, and the cathode voltage drop increases

Figure 6 presents the busbar network and the calculated busbar

from 87 mV to 174 mV as presented in [2].

drop of 134 mV. The busbar current density is quite low but this is

consistent with a business scenario where the metal cost is low

New results for a bigger 25 cm x 16 cm copper collector bar are

and the energy cost is high. Figure 7 is presenting the vertical

presented here. As can be seen in Figure 4, the cathode voltage

component of the magnetic field (B_z) obtained while using this

drop is reduced back to 130 mV.

busbar configuration (see [2] for more results).

Figure 6: Busbar drop of the ECC busbar network concept with

Figure 4: Cathode voltage drop

100% downstream side current exit

disappointing. It turned out that the best way to achieve more mV

saving was to improved the anode aspect ratio.

Figure 9 is presenting the current anode aspect ratio, each stub is

feeding a rectangular carbon section of 0.4875 m x 0.665 m, and

ideally, each stub should be feeding a square section of carbon.

This is important since with 4 fairly big stubs and the new stub

hole design, the biggest resistance is now in the carbon section of

the anode.

Figure 7: Vertical component of the magnetic field (B_z), Figure 10

in [2]

Anode design with innovative stub hole conception

As presented in [14] and in Table II, when operating the cell at

500 kA using 48 anodes of 1.95 m x 0.665 m, the predicted

Figure 9: One of the 48 1.95 m x 0.665 m anode

voltage drop is 265 mV. This already very low anode drop is in

great part due to the usage of an innovative stub hole conception.

For that reason, the 48 1.95 m x 0.665 m anodes have been

That innovative conception was tested in a thermo-electro-

replaced with 64 1.95 m x 0.5 m anodes keeping the exact same

mechanical (TEM) model presented in [17]. Figure 4 of [14] is

stub diameter and stub hole design in order to avoid to go back

showing the voltage drop prediction from that TEM model but not

running the TEM model. Figure 10 presents the new anode aspect

the new stub hole design investigated.

ratio.

That design has been presented for the first time in [3]. Figure 8 is

presenting the original ANSYS voltage drop figure of the TEM

model testing that new stub hole design concept.

Figure 10: One of the 64 1.95 m x 0.5 m anode

With a parametric model at your disposal, the half anode model

can be modified in no time. The same is true for the full anode

panel model. Figure 11 is presenting the initial 48 anodes panel

layout while Figure 12 is presenting the new 64 anodes panel

layout.

The resulting anode voltage drop is presented in Figure 13, simply

by changing the anode aspect ratio and by increasing the number

of anodes from 48 to 64, the anode voltage drop has been reduced

from 265 mV to 224 mV.

Figure 8: Anode voltage drop from the TEM model

As discussed in [3], the aim of the new design is to get a good

contact pressure between the stub bottom horizontal face and the

anode stub hole bottom horizontal face. This is achieved by

locking the stub vertical thermal expansion. As presented in [3],

there is more that one way to achieve this, the final optimized

shape presented in

[3] is less costly to implement, but was

developed after [14] was written.

From that starting point, the author tried to further reduce that

anode voltage drop for this study. The option to add copper insert

like the one presented in [15] was investigated but the gains were

Figure 11: 48 anodes panel layout

Table III: Design and predicted cell energy consumption

Base case

Amperage

500 kA

400 kA

Nb. of anodes

Anode size

1.95m X .665m

1.95m X .5m

Nb. of anode studs

4 per anode

Anode stud diameter

17.5 cm

Anode cover thickness

10 cm

Nb. of cathode blocks

Cathode block length

4.17 m

Type of cathode block

HC10

Collector bar size

20 cm X 10 cm

25 cm X 16 cm

Type of side block

Si C

SiC

Figure 12: 64 anodes panel layout

Side block thickness

10 cm +

ASD

30 cm

Calcium silicate thickness

3.5 cm

Inside potshell size

17.8 X 4.85 m

ACD

3.5 cm

3.2 cm

Excess AlF3

12.00%

Anode drop

265 mV

224 mV

179 mV

Cathode drop

87 mV

130 mV

104 mV

Busbar drop

310 mV

134 mV

107 mV

Cell voltage

3.89 V

3.59 V

3.20 V

Current efficiency

95.90%

Internal heat

758 kW

699 kW

414 kW

Energy consumption

12.1 kWh/kg

11.2 kWh/kg

9.95 kWh/kg

In order to make the new cell lining design work even more

challenging and the cell energy savings even more impressive,

Table III also reports results for an operation at

400 kA

corresponding to running at only 0.64 A/cm² of anode current

Figure 13: Anode voltage drop from the TE model

density. At that current density and still at 3.2 ACD, the cell is

expected to produce metal using only 9.95 kWh/kg.

Calculation of the resulting cell energy consumption

The corresponding cell internal heat is calculated to be reduced to

Several modeling tools could be used to calculate the cell energy

414 kW which is only 36% of the 1140 kW dissipated by the

consumption from the above results. In

[14] the author used

same cell “platform” running at 600 kA and 13.26 kWh/kg.

Dyna/Marc

[18] which also predicts the cell superheat and

corresponding ledge thickness.

Conclusions

So far no effort has been made to adjust the cell lining design to

Two innovations presented by the authors recently at ICSOBA

the new cell operating conditions so a simple cell voltage break

conferences allow to very significantly reducing both the cathode

down tool like Peter Entner’s CellVolt [19] was used instead.

and the busbar voltage drop:

Table III presents the results obtained for the operation at 500 kA

- cathode design with copper collector bars extracting 100%

corresponding to running at 0.8 A/cm² of anode current density.

of the cell current on its downstream side

- the usage of modified external compensation current

As in [2], the calculation was done using 3.2 cm of ACD instead

(ECC) busbar configuration made only of anode risers;

of 3.5 cm used in [14] as since 2011, indications are that ACD

have been reduced further more in low energy consumption cell

are combined with a third innovation presented at the Aluminiun

prototypes. At

3.2 cm ACD, the predicted cell energy

of Siberia conference:

consumption is calculated to be 11.2 kWh/kg.

- the usage of a new anode stub hole design.

More significantly, the cell internal heat is calculated to be only

As a result, a cell operating at 500 kA, 0.8 A/cm² of anode current

699 kW while the cell lining was designed to comfortably

density and

3.2 cm ACD is predicted to have an energy

dissipate 1140 kW with 20 cm x 10 cm size collector bars and 192

consumption of about 11.2 kWh/kg.

anode stubs. Clearly a very serious cell lining redesign work

needs to be performed as the next step. New insulating materials

The same cell platform operating at 400 kA, 0.64 A/cm² of anode

like the ceramic fire board extensively used in [20] will certainly

current density and 3.2 cm ACD is predicted to have an energy

need to be added to the list of lining materials.

consumption of about 9.95 kWh/kg.

References

[15] Dag Sverre Sæsbøe, “Storvik high conductivity anode yoke

with copper core”, Proceedings of

33rd International

[1] M. Dupuis, “A New Aluminium Electrolysis Cell Busbar

ICSOBA Conference, Travaux No.

44, Dubai, UAE,

Network Concept”, Proceedings of

33rd International

November - 1 December 2015, Paper AL23.

ICSOBA Conference, Travaux No.

44, Dubai, UAE,

November - 1 December 2015, Paper AL21.

[16] R. von Kaenel and al., “Copper Bars for the Hall-Héroult

Process”, TMS Light Metals, 2016, 903-908.

[2] M. Dupuis,

“New Busbar Network Concepts Taking

Advantage of Copper Collector Bars to Reduce Busbar

[17] M. Dupuis, “Development and application of an ANSYS

Weight and Increase Cell Power Efficiency”, Proceedings of

based thermo-electro-mechanical anode stub hole design

34th International ICSOBA Conference, Travaux No. 45,

tool”, TMS Light Metals, 2010, 433-438.

Quebec, Canada, 2 - 5 Octobre 2016, to be published.

[18] M. Dupuis and H. Côté, “Dyna/Marc 14.0 User’s guide”,

[3] M. Dupuis, “Presentation of a New Anode Stub Hole Design

GeniSim, 2012.

Reducing the Voltage Drop of the Connection

by 50 mV”, VIII International Congress & Exhibition “Non-

[19] P. Entner, CellVolt AlWeb application: http://peter-

Ferrous Metals and Minerals”, Krasnoyarsk, Russia, 13 - 16

entner.com/ug/windows/cellvolt/toc.aspx

September 2016, to be published.

[20] J. Zhou and M. Dupuis,

“In-Depth Analysis of Lining

[4] M. Dupuis and I. Tabsh, “Thermo Electric Coupled Field

Designs for Several 420 kA Electrolytic Cells”, TMS Light

Analysis of Aluminum Reduction Cells Using the ANSYS

Metals, 2015, 685-690.

Parametric Design Language”, Proceeding of the ANSYS^®

Fifth International Conference, volume 3, 1991, pp. 17.80

17.92

[5] V.A. Kryukovski, G.A. Sirasutdinov, J. Klein and G.

Peychal-Heiling,

“International Cooperation and High-

Performance Reduction in Siberia”, JOM, 46(2), 1994, 23-

25.

[6] M. Dupuis,

“Les modèles thermiques”,

2e Symposium

québécois sur le procédé d’électrolyse de l’aluminium,

CQRDA, 1997.

[7] M. Dupuis, “Thermo-Electric Design of a 400 kA Cell Using

Mathematical Models: A Tutorial”, TMS Light Metals, 2000,

297-302.

[8] M. Dupuis, “Thermo-Electric Design of a 740 kA Cell, Is

There a Size Limit”, ALUMINIUM 81(4), 2005, 324-327.

[9] O. Fujishima, M. Takasawa and K. Wakaizumi, “Advanced

Energy Saving in Aluminum Reduction Cell”, AIME Light

Metals, 1982, 559-569.

[10] S. Tanji, O. Fujishima, K. Mori, “Substantial Energy Saving

in Existing Potlines”, AIME Light Metals, 1983, 577-586.

[11] “AP60 and APXe: A breakthrough in productivity, energy

consumption and emissions”, AP Technology^TM Fact sheet,

March 2015.

[12] F. Charmier, O. Martin and R. Gariepy, “Development of the

AP Technology Through Time”, JOM, Vol. 67, No. 2, 2015,

336-341.

[13] M. Dupuis, “Thermo-Electric Design of a 500 kA Cell”,

ALUMINIUM 79(7/8), 2003, 629-631.

[14] M. Dupuis and V. Bojarevics, “Retrofit of a 500 kA cell

design into a 600 kA cell design”, ALUMINIUM 87(1/2),

2011, 52-55.