WBerends - Hatch Low Resistance Technology Paper for Iceland 290414 r1

1  

LOW RESISTANCE ELECTRODE ASSEMBLIES FOR PRODUCTION OF METALS:

RECENT TECHNOLOGY DEVELOPMENT

Will Berends, Hatch Ltd., Mississauga, ON, Canada, 290414 [email protected]

Abstract

Over an estimated twenty percent of the power consumed in Aluminum electrolysis goes to waste heat from electrical resistances in the materials and connections in the anode assemblies located above the molten bath or in the cathode assemblies below the metal pad. Although part of the overall heat balance of the cell, most of this heat neither materially contributes to the chemical conversion of alumina, nor to maintaining the electrolyte in a molten state. Two of the major connections with electrical resistance creating waste heat are the cast iron connections within both the anode and cathode assemblies.

Hatch is investigating methods to reduce the electrical resistance across these iron connections. This paper reviews the recent and ongoing development of this technology and its potential benefits.

Statement of Problem

The magnitude of electrical resistance across the iron/carbon connection in the anode assembly, as reflected in voltage drop, reportedly peaks in the order of 100-200mV during anode heat up, and between approximately 50-120 mV over the remaining anode cycle. See Wilkening and Cote (1), also Grjotheim and Welch (2). For the cathode the voltage drop of the iron/carbon connection is estimated in the order of 100-150mV after heat up, with increases possibly in excess of 100 mV more over the life of the cathode. A degree of uncertainty exists as to the voltage drop across only the iron/carbon interface or to the extent that the reported voltage drops include various thicknesses of iron and carbon material, since the various research tests are typically unique in their measurement setup. The potline current will also impact the millivolt drop across the connection for a given connection resistance.

Prior Research

The causes of the high electrical resistances across the anode cast iron connections have been extensively researched for over fifty years. Recently Wilkening et al (1) published the mechanics of cast iron behaviour including the initial solidification shrinkage upon rodding, and subsequent expansion through precipitation of graphite when the thimble is reheated to the eutectic temperature near the end of the anode cycle. Also, research by St-Georges et al (3), investigated the oxidation characteristics of the iron to carbon contact interface and the effect of contact pressure on electrical resistance across the connection. Further research with computer modelling has been provided by Richard et al (4) regarding the shape of the cast iron ‘thimble’ and the impact of flute geometry on contact pressure and electrical resistance. Many others have also researched the behaviour of the iron connections during rodding and while in operation in the reduction cell. Research publications are fewer for cathodes assemblies so similar detailed information is lacking, however the electrical resistance of the iron connection between the collector bar and cathode can be inferred due to its similarities to the iron connection between the stub and anode carbon.

2  

The major problems with the anode and cathode assemblies contributing to the high resistance can be summarized to include the following:

Anodes: ‐ Cast Iron shrinkage: during rodding due to quenching of iron against the steel rod stub,

resulting in an air gap or ‘loose fit’ between the iron and carbon ‐ Inconsistent stub condition: where the thimble wall thickness may vary by a factor of 2x

or more before a worn stub is replaced, which extra thickness increases the air gap caused by iron solidification shrinkage

‐ Anode carbon erosion: carbon erosion by bath wash or airburn, which reduces the electrical contact area between iron and carbon.

Cathodes:

‐ Cast iron shrinkage: during rodding, leaving a high resistance air gap ‐ Thermal Warp of the collector bar: during rodding reducing the contact area, Beeler (5) ‐ Heaving and bending of the cathode: from salt absorption over its lifecycle, resulting in

less contact area with the iron & collector bar ‐ Oxidation/corrosion of the iron and collector bar: from bath salts and molten aluminum

over the cathode life cycle increasing the contact resistance ‐ Inconsistent resistance across the top of cathode: Varying resistance between the top of

cathode and the buss connection leads to horizontal currents through the metal pad which contribute to metal turbulence and pot noise.

Solutions to date

Many solutions have been proposed by aluminum producers and industry suppliers to resolve these issues, including:

Anode:

• Enlarged iron to carbon surface area: Use of thimble ‘ wings’ for better current distribution and enlarged contact area; limited adoption

• Increase the diameter and contact area of the stubs: to reduce resistance; common although expensive and may risk anode cracking from increased total force exerted at the iron/carbon interface

• Iron Chemistry: using high carbon equivalent (CE) iron to maximize post solidification expansion of the iron from graphite precipitation; common practise

• The Hydro Aluminium ‘Z’ profile: which geometry may squeeze the carbon by initial iron solidification shrinkage; limited adoption

3  

Cathode:

• Increase in cross section area, or quantity of collector bars: to increase iron/carbon contact area and reduce resistance; common practise

• Extensive preheating of cathode and collector bar: to reduce iron solidification shrinkage; common practise

• Use of graphitized or graphite cathode blocks: graphite or graphitized cathodes blocks have reportedly less salt absorption to reduce cathode swelling and heaving; common practise

• Multiple collector bar arrangements: i.e. combinations of steel and copper for each collector bar, to reduce collector bar resistance and reduce horizontal currents in the metal pad; limited adoption due to existing patents

These solutions have had varying degrees of success and adoption, and may have continuing incremental improvements through optimization and consistent work practises.

‘Low Resistance’ technologies development

The Hatch Technologies group has attempted a number of novel approaches to reducing electrical resistance across the iron/carbon interface, including:

1. Bridging of the air gap between iron and carbon with conductive inserts (for example steel nails) 2. Modification of the solidification behaviour of the cast iron 3. Embedding tight fitting conductors into the carbon to improve current distribution

These technologies are detailed in the Canadian patent application (CIPO No. 2,838,113). The technologies have been applied differently for anodes and cathodes due to their shape and lifecycle duration. The development of these applications for anodes and cathodes are described separately.

Anode Application:

Thimble Anchors:

Thimble Anchors are the use of steel nails, or other interference fit inserts, with one end tightly embedded into the stubhole walls prior to iron pouring, and the other end encompassed by the cast iron, thereby forming a conductive bridge across the iron/carbon airgap. The nail heads in the iron also act as ‘chills’ to reduce solidification shrinkage local to each nail to enable the thimble to have tighter contact with the carbon.

Scale model development:

A generic design of anode and rod was developed for 1/10th scale physical prototyping to prove the concept by measurement of electrical resistance using a micro-ohmmeter. The rod was aluminum joined to a steel cross bar with a screw connection, the stubs were mild steel, the anode carbon was graphite with fluted stubholes and the cast iron was substituted with cast tin solder.

The ‘no-nail’ reference case demonstrated a loose fit between the stubs and anode due to the solidification shrinkage of the solder.

4  

The scale model Thimble Anchors consisted of steel box nails embedded approximately 6 mm into the stubhole walls within the groove cavity of the flutes. Upon casting, the Thimble Anchors immediately provided a ‘tight connection’ since the solder was cast around the head of the nails, with the solidification shrinkage providing a tight fit around the nail head.

Figure 1: 1/10th scale models: Models of generic anode assemblies, showing the models with and without nail inserts in the stubhole walls.

Comparative scale model results:

The electrical resistance was measured at ambient room temperature between the base of the aluminum rod to the underside of the graphite anode. The measured electrical resistance of the model with Thimble Anchors (nails) was less than 5% of the resistance of the ‘no-nail’ standard design (> 95% reduction in resistance when cold).

These results only indicate that the nails were able to conductively bridge the air gap. A hot test of the scale models was not possible due to the low melting temperature of the tin solder.

Application to full size anodes:

The further development to full size anodes included 3D modelling, Figure 2, with computer simulation of iron casting using ProCast software. The iron casting model in Figure 3 illustrates the shrinkage gap and freeze isotherms of the iron thimble of a regular casting. The use of a Thimble Anchor was then modelled with the same parameters, Figure 4, illustrating early freezing of the iron around the Thimble Anchor nail head, showing a reduced shrinkage gap local to the nail head. This method of modifying the iron solidification behaviour is similar to the foundry practise of using a metal ‘chill’ insert within a casting mold to induce early local freezing which reduces local casting shrinkage or porosity.

5  

Figure 2: Generic thimble modelling with Thimble Anchors based on double headed nails.

Figure 3: Procast iron casting simulation of a regular cast iron thimble without Thimble Anchors, illustrating the shrinkage gap between iron and thimble for a regular thimble on the left, and the freeze isotherms on the right.

6  

Figure 4: Procast iron casting simulation of a cast iron thimble with Thimble Anchors, illustrating the modified shrinkage gap local to the nail head of the Thimble Anchor, and the freeze isotherms on the right which illustrate the early iron solidification local to the nail head.

Thimble Anchor indicative testing in a smelter:

Prior to proceeding to extended laboratory testing it was desired to conduct a ‘quick and dirty’ pilot test in a smelter reduction cell to obtain a rough indication of performance in an operating environment. This test included only two pairs of anodes (test and reference), and was not intended to provide a statistically reliable proof of performance nor an optimized solution. It was performed only to observe if the concepts of the design indicated a positive result to justify further investigation, and to observe if the assumptions as to normal anode behaviour and electrical resistance were sound.

Century Aluminum in Seebree, Kentucky agreed to this pilot test which was completed in April 2014. The test included two reference anodes without Thimble Anchors, wired for mV readings on one stub each. The test also included two anodes with Thimble Anchors similarly wired for mV measurements. The Thimble Anchors consisted of eight zinc galvanized concrete nails (3.7mm dia. X 30 mm deep) inserted within the flute grooves in each stub hole, Figure 5. The test and reference pairs of anodes were set in adjacent but separate reduction cells.

7  

Figure 5: Thimble Anchor nails embedded in the stubhole walls of carbon anode, with mV measurement leads installed in iron thimble and carbon after rodding.

Test Thimble Autopsy:

Portions of the thimbles were recovered from the anode butts to observe their casting behaviour, Figure 6. The iron shape local to the nails indicated some preferential iron freezing, by virtue of a thicker iron, however it also demonstrated some unfilled areas of the thimble cavity which appear to be caused by non-wetting of the molten iron to the nail. This non-wetting behaviour may have been due to the zinc galvanized finish on the nail surface.

                       Figure 6: Recovered portion of iron thimbles with Thimble Anchor remnants.

8  

Figure 7: The iron section surrounding the nail head of the thimble anchor, and the unfilled area adjacent to the nail/carbon interface due to ‘non-wetting’ of the molten iron to the nail.

Observations: Millivolt resistance comparative performance:

The Century Seebree reduction line operates at nominal 208,000 Amperes. The voltage drop of the four test anodes were taken every 2 hours over the first day, Figure 8. The average millivolt drop of the connection with Thimble Anchors was 58% of the regular thimble (no anchors) over the first day of operation.

Figure 9: First day millivolt readings of anodes with nails (A1 & A2) and reference anodes without nails (B1 & B2).

0

20

40

60

80

100

120

140

160

180

200

2 4 6 8 10 12 20

mV drop

Hours after Anode Setting

Thimble Anchor Test

A1

A2

Avg Anchors

B1

B2

Avg no anchors

9  

Due to operational constraints the subsequent measurements were taken intermittently thereafter and just prior to removal of the anodes from the cells, Figure 9.

Figure 10: mV drop of test anodes from day 1 to 21. After day 10 all readings steadily declined to under 3 mV on last day of anode cycle.

The millivolt drop across all anodes, with and without Thimble Anchors, demonstrated a significant reduction up to the last day. Final mV drop readings were at or below 3 mV, corresponding at nominal 2890 Amps/stub with a 1 micro-Ohm resistance. This low voltage drop and resistance is contrary to the referenced prior research. These observations may indicate that the resistance of the iron/carbon interface is very small when the iron connection is at the highest temperature during the last portion of the anode cycle, where the iron may have already undergone its maximum expansion by graphite precipitation, and the steel stub will have thermally expanded its greatest amount as well. Accordingly the power savings from use of Thimble Anchors may be very low should these results be further validated. Due to the limited sample size these test observations are not to be considered predictive of typical behaviour.

Future Testing:

In order to continue the development of the Thimble Anchors, Hatch is conducting laboratory tests with various configurations of Thimble Anchors and Reference anodes. These cast connections will undergo controlled temperature tests with micro-ohmmeter measurements of the same iron/carbon interface.

Parallel to the laboratory testing is the proof of concept of a robotic nailer that may implement the Thimble Anchor solution with the required productivity rate, if it is deemed economically worthwhile.

0

10

20

30

40

50

60

70

80

90

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

mV drop

End of Day #

Thimble Anchor Test  Day 1‐21

A1

A2

Avg Anchors

B1

B2

Avg no anchorsExtrapolated mV drop between measurements

10  

Cathode Application:

The methods used for Thimble Anchors in an anode assembly may potentially also be applied to cathode assemblies albeit in modified forms. These methods, restated, include:

1. Bridging of the air gap between iron and carbon with conductive inserts 2. Modification of the solidification behaviour of the cast iron 3. Embedding tight fitting conductors into the carbon to improve current distribution

The iron connection in cathodes differs from anodes by the shape of the iron connection and the duration of use. With regards to shape the collector bar and iron connection will expand longitudinally more than the cathode block itself, preventing the use of nails to bridge the air gap. With regards to lifecycle the cathode assemblies in a new potlining should quickly rise to a high enough temperature to allow full graphite precipitation in the iron which then remains over its 2,000-3,000 day life. The improved iron connection for cathodes should address the life cycle resistance of the cathode/iron/collector bar assembly which may increase due to corrosion and cathode block heaving.

The normal iron connection between collector bar and cathode provides final freezing in the middle of the iron section, Figure 11. This enables shrinkage gaps to appear along the length of the cathode which significantly reduces the effective contact area between iron and cathode, Figure 12.

                  Figure 11: Cast iron freezing isotherms of regular cathode to collector bar connection

11  

Figure 12: Gap analysis of regular cathode to collector bar connection

1. Collector Bar Heat Sinks and Cathode Slot Anchors:

The first concept investigated for cathode assembly improvement was the use of a revised collector bar shape to preferentially freeze the iron to reduce the air gap at specific points along the length of the cathode slot.

This design uses welded steel spacers on the collector bar which act as heat sinks or chills to the cast iron, situated adjacent to ‘Slot Anchors’, Figure 13. The Slot Anchors are broad headed nails that are tightly embedded in the cathode carbon, the exposed heads coated with graphite stub wash to prevent welding to the cast iron. The heads of the slot anchors maintain a sliding contact with the cast iron to enable a superior metal/metal conductive interface, despite potential differences in movement between the collector bar and iron relative to the cathode slot wall.

The heat sinks drive early solidification of the iron between the heat sink and slot anchors, Figure 14. This freezing behaviour reduces air gap between the collector bar, iron and the slot anchor, Figure 15. The Slot anchors assist in conducting the current into the carbon by virtue of their being tightly embedded into the cathode carbon.

12  

Figure 13: Collector Bar heat sinks and Slot Anchors to modify iron solidification and current distribution into the cathode.

Figure 14: Cast iron freeze Isotherms of collector bar with heat sinks and slot anchors

13  

Figure 15: Gap analysis of cathode to collector bar iron connection with collector bar heat sinks and slot anchors

Scale Model of Heat Sinks and Slot Anchors Cathode Assembly Design:

The collector bar design with heat sinks and slot anchors was 1/10th scale modelled for resistance testing, Figure 16 and 17. The method may be used with steel or copper collector bars. The models used tin solder as a substitute for cast iron. The electrical resistance was measured at ambient room temperature.

The average resistance of the collector bar to cathode surface with heat sinks and slot anchors was reduced by 46% for the steel collector bar and 59% for the copper collector bar as compared to the normal cathode assembly. The difference is anticipated to become much less when the cathodes are operating at pot temperatures, due to the expansion of the normal collector bar and cast iron providing a tight connection upon heatup. However, the benefits of this new design include an initial tighter fit than normal, and to provide a metal to metal sliding contact interface where the iron moves relative to the cathode carbon.

Future testing will include lab testing of full size cathode assembly prototypes plus computer modelling to predict stresses in the cathode structure.

14  

Figure 16: 1/10th scale models of regular cathode and one with heat sinks & slot anchors

Figure 17: Close up of scale model heat sinks and slot anchors

15  

2. Cathode Flex Connectors & Screw Conductors

The second concept investigated for cathode assemblies was to augment the existing iron connection with flexible connectors between the collector bar and the cathode block. One end of each flex connector is attached to the collector bar, by mechanical or fusion connection. The other end of each flex connector is attached to a screw conductor that is embedded into the carbon of the cathode block, Figure 18.

Figure 18: Flexible connectors and screw conductors in cathode assembly

The flex connectors provide a reliable electrical conduction path between the collector bar and cathode regardless of relative movement or changing contact pressure between the iron and carbon. The conducting screws are spaced along the length of the cathode with varying depth to enable ‘tuning’ the cathode resistance from top of cathode to buss connection. This method enables both a reduction in the average electrical resistance through the cathode, and a more consistent electrical resistance across the top of the cathode, which may reduce horizontal currents through the metal pad. The configuration of the cathode shape and location of the flex connectors and screws can be designed to avoid interference with current pot lining designs.

The resistance R(Top of Cathode to Buss) = R(carbon) + R(Screw connector & flex) + R (collector bar)

The screws may be altered in depth (tuning) to reduce the apparent resistance through the cathode block, in order to offset the changing resistance through the length of the collector bar, Figure 19. This tuning enables R1=R2=R3=etc….=R13 to promote vertical current flow through the cathode. It also enables adjustment of resistance across cathode blocks to adjust for local corrosion near buss risers. 

16  

Figure 19: Cross section schematic of the cathode assembly illustrating variable screw conductor lengths to offset the collector bar resistance.

Potential Benefits of Flexible Connectors and Screw Conductors:

• Reduced power consumption by lower electrical resistance through the cathode • Longer cathode life by reducing peak current densities through the cathode and distributing

current flow across the entire cathode • Potential further reduced power consumption if the horizontal currents through the metal pad may

be reduced, thereby potentially reducing metal turbulence and pot noise, enabling the Anode to Cathode Distance (ACD) to be reduced.

Scale Model of Flexible Connectors and Screw Conductors:

A 1/10th scale model of a cathode assembly with cast tin solder connection was augmented with flexible connectors and screw conductors. At ambient temperature the average electrical resistance from the top of the cathode to the base of the exposed end of the collector bar was reduced by 57% compared to the normal cathode assembly. More importantly the standard deviation of resistance measurements across the top of the cathode was also reduced by 35%, demonstrating the potential ability to reduce the variation of electrical resistance measurements across the top of the cathode by adjusting the spacing and depth of the screws.

Figure 19: 1/10th scale model of cathode assembly using flexible connectors and screw conductors

Collector bar to right buss 

R1  R2  R3  R4  R5  R6 R7 R8 R9 R10 R11 R12  R13 

Collector bar to left buss

Top surface of cathode

Full width cathode & collector bar

Variable screw depth

17  

Conclusions:

The anode scale model and smelter test work verified that:

• The air gap between iron and carbon can be electrically bridged by use of Thimble Anchors, • The solidification shape of the iron may be changed by use of the Thimble Anchor nails as chills,

however other effects such as non wetting of the molten iron to the nail may offset this, • A good electrical connection is made between the nail and the carbon • Electrical resistance can be reduced through the connection by the above methods, particularly

during the anode heat up period.

The observed low resistance of the reference anodes (without Thimble Anchors) after the first day of in pot operation is contrary to most published information and requires validation by further testing.

The cathode scale models indicate that potential reduction of cathode assembly resistance and more consistent resistance across the top of the cathode is possible. Further full size prototyping, computer modelling and pot relining tests are required to further develop the designs and quantify the potential benefits. 

Future Technology Development:

Hatch is continuing its research and development of this technology for both anodes and cathodes in order to better develop predictive models of the technology performance and to better quantify the potential benefits. The current test plans include the following:

Anode: (anodes & stub assemblies)

• Lab iron casting & resistance testing • Pressure vs resistance testing (Metal /carbon interface) • Nails versus resistance testing • Robotic nailing to meet cycle time • More extensive field tests with optimized design of anchor size, depth and quantity of nails

subject to automation capabilities.

Cathode: (cathode & collector bar assemblies)

• Lab testing of full size cathode prototypes of the above concepts. • Pressure vs. resistance testing (Metal /graphitized carbon interface) • Flex Jumper prototype and resistance testing • Thermal/mechanical/magnetic computer modelling (for a client cell design) • Distant future – progressive pot relining field tests and control system optimization for ACD

reduction subject to AE, pot noise constraints and thermal heat balance.

Acknowledgements:

Thanks and appreciation to Century Aluminum, Seebree, Ky, and particularly Brian Audie and Scott Randall who participated in the field tests of the Anode Thimble Anchors.

18  

References:

[1] PROBLEMS OF THE STUB – ANODE CONNECTION, S. Wilkening, J. Côté; Light Metals 2007, TMS (2007), 865-873. [2]  ALUMINUM SMELTER TECHNOLOGY 2ND ED., K.Grjotheim, B.J.Welch, 1988, Aluminum Verlag, sec. 6.3.3 pg177

[3] EFFECTS OF HIGH TEMPERATURES AND PRESSURES ON CATHODE AND ANODE INTERFACES IN HALL-HEROULT ELECTROLYTIC CELLS, L. St.Georges, L. I. Kiss, M. Rouleau, J. Bouchard, D. Marceau, Light Metals 2011, TMS (2007),997-1002.

[4] CHALLENGES IN STUB HOLE OPTIMISATION OF CAST IRON RODDED ANODES, D. Richard, P. Goulet, O. Trempe, M. Dupuis, M. Fafard; TMS Light Metals (2009), 1067- 1072. [5] BAR TO BLOCK CONTACT RESISTANCE IN ALUMINUM REDUCTION CELL CATHODE ASSEMBLIES. R. Beeler, Light Metals 2014, TMS (2014), 507-510.