Current increase PE improvement and PFC reduction at …February 15-19, 2009, San Francisco, CA Current increase, power efficiency improvement and PFC reduction at RTA Bell Bay Smelter

February 15-19, 2009, San Francisco, CAFebruary 15-19, 2009, San Francisco, CA

Current increase, power efficiency improvement and PFC reduction

at RTA Bell Bay Smelter

Geoff Bearne, Rio Tinto Alcan.Melbourne, Australia

Presented on behalf of Bell Bay Potrooms

February 15-19, 2009, San Francisco, CA

Topics• Introduction to the Bell Bay smelter.

• Production increase through significant current creep and process optimisation (1988-2008).

• Perfluorocarbon emission reductions by targeting zero anode effects.

I’d like to take this opportunity to describe two significant improvement initiatives undertaken at the Bell Bay Smelter.

The first initiative is the result of the plant’s unrelenting drive to improve production and productivity, while reducing costs, by creeping current to levels that were not considered possible twenty years ago.

The second initiative is the potroom’s concerted effort to reduce anode effects (both in frequency and duration) to a level where their environmental impact is minimised.

But firstly I would like to introduce you to the aluminium smelter.


Bell Bay Aluminium Smelter• First aluminium smelter in the

Southern Hemisphere.• Began production in 1955 as

Government owned facility.• First potline VS Søderberg.• Acquired by Comalco in 1960

heralding start of Australia’s integrated aluminium industry.

• Bell Bay is now one of three Rio Tinto Alcan smelters in Australia, but the only one wholly-owned.As shown on the map, the Bell Bay smelter is located in Northern Tasmania, close to George Town, a town with a population of around 7,000 situated close to the mouth of the Tamar River.

The Søderberg potline (Line 1) was shut down in 1981.



Potline 2 – 172 cells

Potline 3 – 178 cells

Potline 4 – 188 cellsBooster section referredto as “Potline 1” – 20 cells

2008 production – 177,000 tonnes/year

Three Potlines – 118kA

One booster section – 130kA


020406080

100120140160180200

1955 1965 1975 1985 1996 2004 2005 2006 2007 2008

k to

nnes

Hot Metal

Capacity has grown over the years, initially through potline expansions, but more recently throughamperage increase and efficiency improvements. This has put significant pressure on improvingwork practices and process control to maintain cell performance and minimise environmentalimpacts, particularly from anode effects.


Kaiser P57 technology, originally 86kA

The cells are Kaiser P57 side-by-side end riser technology, developed in the 1950’s. They were originally equippedwith centre bar breakers but were later converted to point feed (more on this later).

The cell hoppers are filled with reacted alumina, from the scrubbing system, by overhead crane. The dry scrubbingsystem is Comalco developed Torbed technology. Alumina is sourced from either QAL or Rio Tinto’s new Yarwunrefinery in Queensland. Both source bauxite from Weipa. Alumina is delivered to an “A” frame storage shed on site.

The cell control system was developed in-house and includes a computer per cell, linked to middle level supervisorycomputer per 16 cells. The user interface is via a top level computer system that includes an extensive data base.


Each Potroom is serviced by one general purpose crane and one anode changing crane (CTM)


Anodesetting

Filling aluminahoppers

Metaltapping

CellDressing


Challenges• Bell Bay is a relatively small output smelter

operating an older technology.• To remain competitive requires continual

improvement in efficiency, cost and productivity.• 2008 targets included:

2% year on year production increase.2% year on year controllable cost reduction.Reduced site GHG emissions (1.75 tCO2e/t Al)

• 2009 aim is primarily survival:Minimum CAPEX and OPEXMaximum efficiency and minimum waste


Continuous story of current creep

80

85

90

95

100

105

110

115

120

125

130

1989 1991 1993 1995 1997 1999 2001 2003 2005 2007

Cur

rent

(kA

)

14.0

14.5

15.0

15.5

16.0

16.5

17.0

Spec

ific

Ener

gy C

onsu

mpt

ion

(DC

kWh/

kg)

Energy consumption

Line 1

Lines 2 - 4

Line 1

Lines 2 - 4

A group of booster cells has been used for cell performance assessments since 1999. It runs 5 to 15kA above the rest of the plant.Current increases have generally been achieved at constant cell heat loss, thus improving energy consumption.


80

90

100

110

120

130

140

150

160

170

180

1989 1991 1993 1995 1997 1999 2001 2003 2005 2007

Cur

rent

(kA

)

87.0

88.0

89.0

90.0

91.0

92.0

93.0

94.0

95.0

96.0

Cur

rent

Effi

cien

cy (%

)

Lines 2 - 4

Line 1

Current

Current Efficiency

Line 1Lines 2 - 4

PE improvements have resulted from increasing CE and cell squeezing strategy

Concurrent with current increase, current efficiency has been improved from around 90% to 94% in the last 20 years. This is the result of improved work practices, cell design and process control.


Current increase capability enablers

200mm tall anode slot2007

560mm wide Anode (Carbon Baking Furnace packing layers reduced)2006

Side chamfered anodes2004

120 kA superstructure on line conversion with wider anode spacing and extra anode bus. 945 symmetrical anodes. Change from 48 hour to 36 hour tap.

2003

165mm tall anode slot, deck plate cutting2002

945mm asymmetric anodes, CTM anode setting, lower anode cover2000

Increased downstream cathode busbar cross sectional area1999

925mm asymmetric anode, point feed on-line conversion1998

Torbed dry scrubber fume treatment plant installed 1997

130 mm dia. stub, symmetric hanger, cut fume skirts, chamfered anode back1996

860mm asymmetric anode1989814 x 525mm anode,100mm diameter stub1980 - 85


Cells were heat balance constrained• Cathode heat balance improved at higher current, resulting

in a lower cathode electrical resistance. Prior to this point cells were probably limited by heat balance (too cold).

• Gains from 1996 onwards were largely from heat balance improvement with increasing current.

• Removing cover from hangers in 2001 allowed higher current density (scaling of the steel hanger was an issue)

Year 1985 1990 1991 1996 1999 2001stub dia mm 100 100 100 130 130 130amperage kA 86 88 92 94 100.5 107

anode length mm 814 860 900 900 925 945anode width mm 525 525 525 525 525 525

anode cd amps/mm sqrd 0.0101 0.0097 0.0097 0.0099 0.0103 0.0108Cathode Resistance typ micro Ohms 4.8 4.2 4.1


Impact of anode setting sequence

Cell 4000

0

0.1

0.2

0.3

0.4

0.5

0.6

1/01/03 1/02/03 1/03/03 1/04/03 1/05/03 1/06/03 1/07/03 1/08/03 1/09/03 1/10/03 1/11/03 1/12/03 1/01/04 1/02/04 1/03/04 1/04/04 1/05/04

LF n

oise

(mic

ro-o

hms)

90,000

95,000

100,000

105,000

110,000

115,000

120,000

Cur

rent

(Am

pere

s)

36 hour cycle 120kA Superstructure

Double setting

Profile changed to 6

Instability associated with setting upstream duct end anodes was evident before superstructure changes and current increase (as evidenced by the cyclical increases in low frequency noise before June 2003), however it deteriorated after these changes.

Current

Low Frequency Noise


Impact of anode setting sequence

Stability issues were resolved by reversing the anode setting sequence, modifying the anode setting profile and parameters

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.10

2 4 6 8 10 7 9 1 3 5Anode

Noi

se (m

icro

-ohm

s/m

in)

Low frequency noise on anode setting shift 2 4 6 8 10

7 9 1 3 5

2 4 6 8 10

7 9 1 3 5

upstream

duct

Anode setting profile

-20

-15

-10

-5

0

5

10

15

20

25

2B 2A 4B 4A 6b 6a 8A 8B 10A 10B 7B 7A 9B 9A 1a 1b 3A 3B 5A 5B

Anode

Prof

ile (m

m)


Continuous hardware improvement

Anodes815 x 508mm anodes

HangerBolted Assembly

SuperstructureManual break and feed

CathodeMonolithic, Ram pasteConstruction

Deckplatemonolithicsidewall

Deckplate(narrow)

Anodes1050 x 560 anodes

Hanger150mm stubs

Superstructure- extra ring bus leaf- wider anode spacing

CathodePrebaked Semi Graphitic Sidewall

Robust SiC design

Anodes815 x 508mm anodes

HangerBolted Assembly

SuperstructureManual break and feed

CathodeMonolithic, Ram pasteConstruction

Deckplatemonolithicsidewall

Deckplate(narrow)

Anodes1050 x 560 anodes

Hanger150mm stubs

Superstructure- extra ring bus leaf- wider anode spacing

CathodePrebaked Semi Graphitic Sidewall

Robust SiC design

Original 1961 Design 130kA (2006)


Anode assembly development

130kA 18 day rota 120kA 20 day rota 114kA 16.5 day rota130kA 18 day rota 120kA 20 day rota 114kA 16.5 day rota


Typical cell performance (12 month averages)

0.0710.087Average iron concentration (%)0.100.063Anode effect frequency (AE/cell day)

0.3950.436Net carbon (kg/kg Al)539516Power actual (kW)14.614.5Spec. Energy Consumption (DC kWh/kg)93.993.9Current efficiency (%)12.212.3Excess aluminium fluoride (%)966959Bath temperature (degrees C)1.091.10Anode current density (A/cm2)1050945Anode length (mm)4.594.56Volts/cell (V)

128.1116.0Current (kA)Line 1 2008Line 4 2008Critical Process Variable/End-Of-Line


130kA Project• Ramping to 130kA will increase metal

production by 23kt (13%).• Large productivity gains planned through

amperage creep, automation and workflow optimisation.

• Enablers are larger anode, new anode hanger and cell designs and power supply upgrades.

• Estimated CAPEX of US$35.As a result of the global financial situation, resulting in a collapse in aluminium price to around US$1,350/tonne, the 130kA project is presently under review.

Perfluorocarbon (PFC) emissions


CARBON ANODE

ALUMINA

CRUST BREAKER

MOLTEN ALUMINIUM

LIQUIDBATH

CRUST

CARBON ANODE

ALUMINA

CRUST BREAKER

MOLTEN ALUMINIUM

LIQUIDBATH

CRUST

• At low alumina conc. (~1.5%) anode polarisation increases.

• Fluorine ions of the liquid bath are discharged.

• These react to form PFC gasses.• Result in an “anode effect”

Cross section of a reduction cell

Compound Global Warming Potential

(100-year horizon) CO2 1 CH4 21

HFC-134a 1,300 CF4 5,700 C2F6 11,900 SF6 23,900

CARBON ANODE

ALUMINA

CRUST BREAKER

MOLTEN ALUMINIUM

LIQUIDBATH

CRUST

CARBON ANODE

ALUMINA

CRUST BREAKER

MOLTEN ALUMINIUM

LIQUIDBATH

CRUST

• At low alumina conc. (~1.5%) anode polarisation increases.

• Fluorine ions of the liquid bath are discharged.

• These react to form PFC gasses.• Result in an “anode effect”

Cross section of a reduction cell

Compound Global Warming Potential

(100-year horizon) CO2 1 CH4 21

HFC-134a 1,300 CF4 5,700 C2F6 11,900 SF6 23,900

The generation of PFC emissions during anode effects, with their very high green house gas contribution, is well established.

Anode effect (AE) Performance over time

AE Frequency and Duration, Bell Bay

0.0

0.51.0

1.52.0

2.5

3.03.5

4.0

Jan-85

Jan-87

Jan-89

Jan-91

Jan-93

Jan-95

Jan-97

Jan-99

Jan-01

Jan-03

Jan-05

Jan-07

Jan-09

Freq

uenc

y (A

E/C

ell D

ay)

0

2040

6080

100

120140

160

Dur

atio

n (S

ec)

FrequencyDur >8V

1998 Point feed conversion

Bar Break, Target 3 AE/day

1994 Control system upgrade

Target 2 AE/day

Six Sigma program

Regulator failure

AE Frequency and Duration, Bell Bay

0.0

0.51.0

1.52.0

2.5

3.03.5

4.0

Jan-85

Jan-87

Jan-89

Jan-91

Jan-93

Jan-95

Jan-97

Jan-99

Jan-01

Jan-03

Jan-05

Jan-07

Jan-09

Freq

uenc

y (A

E/C

ell D

ay)

0

2040

6080

100

120140

160

Dur

atio

n (S

ec)

FrequencyDur >8V

1998 Point feed conversion

Bar Break, Target 3 AE/day

1994 Control system upgrade

Target 2 AE/day

Six Sigma program

Regulator failure


Campaigns to reduce anode effect frequency, and later anode effect duration, have been undertaken over two decades.

1985 -1994– 3 anode effects per day– Centre feed, bar break operation

– 1 break, 3 dump feed strategy (~7kg dump weight)– No resistance tracking– Anode effects scheduled– Considered necessary for sludge control– Attempt to reduce frequency from 3/day to 2 in 1989– Two tier Siemens R30 Mini computer/ Siemens MC210

microcomputer

BladeBlade

Historical anode effect performance


1994 - 1998– Major control system upgrade, Comalco-developed cell

control strategy– New three tier VAX, IVAX, Analogic computer system– Resistance tracking introduced– No scheduled anode effects– AE frequency reduced to 0.5 ~ 1.0 AE/day (enabled by

control strategy improvements)– Still issues with alumina dissolution (back feeding) as a

result of alumina sinking into the metal pad from the relative large dump mass during feeding.

Historical anode effect performance


1998– Bar breakers converted to point feed, adapted from

original bar break beam

– 4 picks with chutes to direct alumina into feed hole– Hopper outlets blanked off– Dump gate dams fitted– Pick length and breaker beam stroke optimised

Conversion to point feed

Chutes

70mm diameter spikes

Chutes

70mm diameter spikes


Point feed conversion considerations• On-line beam change out procedure,

to avoid safety issues.• Hopper utilisation/ dead capacity.• Pick material selection (304 or 304L

austenitic stainless).• Breaker stroke length, pick length

and immersion.• Pick location relative to anode

position.


• crust breakerscrust breakers• crust breakerscrust breakers

A number of issues needed addressing during thepoint feed conversion, including: • Procedure to allow safe access into cell hoppers

using a mobile cooling system.• Pick material selection – early choices saw

significant erosion of tip.• Crust breaker pick length and beam stroke –

needed to avoid too much submersion in bath to prevent overheating and bath build up.

• Ore gates were blanked off to reduce variation in dump weight.

• Pick location – early choice resulted in air burn adjacent to anode stubs. Picks were relocated to anode corners.

2001– Commenced measuring AE duration (defined as time

above 8 volts)2006

– Greenhouse gas targets established for AE contribution.

– Site Target < 0.07 AE/cell day, 30 seconds duration.– Expectations set:

• Goal of zero anode effects• Every AE treated as an “exception”

– Six Sigma program commenced:• Initially addressing AEs in new cells• Expanded to include all cells.


Upping the ante on anode effects

AE reduction focus areasTarget areas:

– New cell start-up control– Alumina feed equipment reliability– Recovery from no load and reduced load– Post anode setting control of ACD– Automatic anode effect termination effectiveness– Resistance tracking parameters (tuning to mitigate the

impact of current increase)– Individual cell parameter tuning for alumina dump weight– Individual cell power settings (heat balance control)– Individual exception cell follow up– Adaptive feed


Challenging anode effect myths (that we used to believe!)

• Anode effects are necessary.• AEs are required for new cells.• AEs are unavoidable after reduced load/no

load.• Increasing line current always increases AE

frequency.• Larger anode area/decreased bath volume

must increase AE frequency.• Alumina source change always causes AEs.• It is too hard to improve from where we are

now.


AE frequency benchmarking

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Anode Effect Frequency

Cum

ulat

ive

Frac

tion

of R

epor

ting

Faci

litie

s

PFPBCWPBSWPBVSS - Open Circles PFHSS

Bell Bay0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Anode Effect Frequency

Cum

ulat

ive

Frac

tion

of R

epor

ting

Faci

litie

s

PFPBCWPBSWPBVSS - Open Circles PFHSS

Bell Bay


From IAI survey 2005

AE duration benchmarking

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0

Anode Effect Duration (minutes)

Cum

ulat

ive

Frac

tion

of R

epor

ting

Faci

litie

s

PFPBCWPBSWPBVSSHSS

Bell Bay0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0

Anode Effect Duration (minutes)

Cum

ulat

ive

Frac

tion

of R

epor

ting

Faci

litie

s

PFPBCWPBSWPBVSSHSS

Bell Bay


From IAI survey 2005

Benefits of low AE frequency & duration• Reduced greenhouse gas emissions• Reduced energy consumption and improved CE• Reduced crust disruption from AE termination:

– helps reduce anode airburn– helps reduce fugitive fluoride emissions

• Reduced manual intervention to “pole” out failed AEs after failed automatic termination

• Reduced risk of cell failure due to overheating• Improved amperage stability, hence “gentler” on

power supply equipment


Reducing and eliminating anode effects has many other benefits as well as reducing GHG emissions. These include reduced costs, increased safety and reduced process variation.

Summary: Key factors in eliminating anode effects

• Motivation (leadership)

• Improved work practices:• Quality of routine work (including maintenance)• Diagnosis and prompt correction of problems• Tuning of cell control parameters (feed and heat balance)

• Improved alumina feeding equipment design and reliability

• Enhanced automated process control strategy:• Eliminating “scheduled” anode effects• Tighter control of alumina concentration• Improved anode effect termination• Better cell management around periods of reduced power

• Process Stability


Acknowledgements

Bell Bay performance achievements are a testament to the Operations and R&D staff who have continually challenged the limits and extend them. Thanks to John Lemberg, Mark Fyfe, Robbie Matthews, Greg Picot, Greg Hardie, Allan Graham and many others.


Documents

Current increase PE improvement and PFC reduction at …February 15-19, 2009, San Francisco, CA Current increase, power efficiency improvement and PFC reduction at RTA Bell Bay Smelter