Instability of an Aluminium Smelting Cell

1. Background

On the 11 January, pot C044 became severely unstable from around 09:30.

Upon investigation it is clear that the pot become severely unstable after it was double tapped. Metal tapping was initially requested at 08:13, and then again at 09:21. After the first tapping the pot experienced a spike in the instability and then the pot settled at an instability of 100nΩ. After the pot was tapped the second time the pot became severely unstable (>1000 nΩ). It was also noted that the operator had put the pot into manual and adjusted (causing the instability procedure to be cancelled before each of the tapping requests.

The tapping schedule for the CG2 group (10 January cycle B5 - tapping was late)

had 2 logicals of interest; logical C4027 (pots C039, C040, C044) and logical C4028 (pots C042, C043, C045). It is clear that the operator tapped pot C044 correctly with logical C4027, and also again with logical C4028 (instead of pot C045).

2. Subsequent treatment of the pot

The first treatment of the instability started with a RC of 0.40 at 10:17, followed by an additional RC adjustment to 0.60 at 10:36. The pot was switched to manual at 11:01 for 5 sec??. The pot was again switched to manual at 11:24, and all feeders were stopped on the pot and the pot was unsqueezed.

At 13:39, the potline experienced an outage due to power load shedding, and was restarted at 14:46. Treatment on the pot was obviously not possible during this period. The RC was further increased to 1.00 at 15:39, and a RDEM of 0.40 was even added at 16:27. This treatment was clearly improper as the pot showed no improvement after the previous RC's were added; further manual intervention would have been more appropriate (potline shutdown could have been a factor).

Anode #9 (scheduled anode change) was changed on the pot at 16:51. With the condition of the pot, this operation should have been suspended.

The treatment of the pot on the following shifts was as per sick pot treatment (temperature controlling by squeezing, ACD monitoring and clad failure replacement with hot anodes from neighbouring pots, bath recycling and greenpole treatment).

The pot eventually stabilised on 13 January 2008.

3. Summary

1. The pot's excessive instability was caused by the double metal tapping.

2. The initial intervention to treat the pot with RCs at persistent instability levels of >1000ηΩ was inappropriate. The potline outage did, however, interrupt the treatment of the pot, as it necessitated the presence