Measuring bath properties using the STARprobe^TM

Marc Dupuis¹, Jean-Pierre Gagné²and Pierre Bouchard²

1. GeniSim Inc,

2. STAS

Corresponding author: marc.dupuis@genisim.com

Abstract

Since the beginning of 2012, STAS is the world distributor of the STARprobe^TM developed by

Alcoa [1, 2]. Demonstration measurement campaigns have been conducted in different smelters

around the world.

During those campaigns, the STARprobe^TM bath properties measurements have been compared

with standard bath properties measurements regularly carried out in those smelters operated by

different aluminium producers.

Those independent comparative measurements all confirmed the capacity of the STARprobe^TM

to instantaneously and accurately measure Superheat, Temperature, Alumina concentration and

bath Ratio for cell control purposes.

Keywords: STARprobe^TM, cell control, XRD bath analysis, alumina concentration, superheat,

temperature, bath ratio, measurements

Introduction

The STARprobe^TM is a portable device that takes real-time measurements of bath properties,

such as Superheat, Temperature, Alumina concentration and bath Ratio or acidity (STAR), in

electrolysis cells. This synchronicity of measurements is a most important step forward in

improving the control and efficiency of electrolysis cells. It unites the conventional processes of

temperature measurement and bath sampling-analysis into one online measurement, simplifies

and greatly shortens the process and time space from measurement/sampling to pot control

decision. The pot control decision can therefore be based on the real time cell conditions rather

than those from few hours ago or from as long as 24 hours ago.

This integrated real-time measurement system consists of four major components:

•

Reusable probe tip

•

Portable stand to fit various pot configurations

•

Electronics for data acquisition and analysis, and wireless communications for data

transfer

•

Tablet PC with programs to perform all necessary tasks during measurements.

Considering the great advantages of the STARprobe^TM, Alcoa has decided to share the

technology with the rest of the aluminium industry starting from 2012. In this regard, Alcoa has

just appointed STAS, a well recognized leader in the aluminium industry (www.stas.com ), to

commercialize the new STARprobe^TM analyzing system.

First measurement campaign

The main objective of the first measurement campaign was to get familiar with the

STARprobe^TM measurement technique. Since that measurement technique turned out to be

fairly easy to master, we were quickly in position to take advantage of the time set aside to carry

out measurements and try to learn from them.

2.1.

First repetitivity test

One big question that is typically set aside when bath samples are taken for bath chemistry

control purposes is how homogeneous is the bath and hence how representative is a small

localized bath sample of the average composition of the bath. Considering the way that the

alumina is discretely added to the bath in one region of the cell and consumed in another region,

there is no reason to expect that the dissolved alumina concentration in the bath should be

relatively homogeneous.

For the first time, using the STARprobe^TM, it is possible to quickly measure the dissolved

alumina concentration in the bath in order to try to assess how homogeneous it is in the bath.

For that purpose, 15 consecutive STARprobe^TM measurements were carried out in the tap hole

of the same pot over a period of 90 minutes averaging one measurement every 6 minutes. The

obtained results are presented in the Table 1. Figures 1 to 4 highlight the relative variability or

lack of strict repetitivity of the measurements which in turn highlight the lack of homogeneity of

the bath. This is particularly true for the dissolved alumina concentration and the bath superheat

which is directly affected by the dissolved alumina concentration in the bath sample analyzed by

the STARprobe^TM. Nevertheless, the averaged results give a very consistent picture of the cell

conditions as the average measured excess AlF₃ concentration of 10.69%, the average measured

dissolved alumina concentration of 2.6%, the line average CaF₂ concentration of 4.9% and the

measured average bath temperature of 968.72 °C can be used to calculate a bath superheat of

4.97 °C using the Solheim equation [3] while the measured average bath superheat is 4.94 °C.

Position

Excess

Al₂O₃

Temp

Super

tap

10.25

2.93

969.18

6.21

tap

9.67

2.53

971.02

2.05

tap

10.78

2.85

966.68

6.35

tap

10.51

2.6

970.72

8.8

tap

10.89

2.49

970.16

5.87

tap

10.38

2.13

969.98

tap

11.31

2.4

971.18

6.72

tap

10.55

2.76

966.51

5.6

tap

10.2

2.5

970.34

3.8

tap

10.48

2.86

968.13

4.9

tap

11.81

2.12

972.61

5.54

tap

11.37

2.08

967.42

2.67

tap

11.27

2.72

963.65

4.24

tap

9.69

3.21

967.3

4.44

tap

11.26

2.7

965.93

4.98

Mean

10.69466667

2.592

968.7206667

4.944667

Std Dev

0.625446888

0.321696396

2.443067354

1.84237

Table 1: Raw data from the first repetitivity test

Figure 1 to 4 present the raw distributions (histograms) of the 15 measurements and the

corresponding normal distributions (curves) having the same mean and standard deviation as the

raw distributions.

2.2.

Second repetitivity test

In order to ensure that the above results are really typical, a second repetitivity test was

performed the next day again repeating STARprobe^TM measurements in a single cell. In total, 22

measurements were obtained over the same period of 90 minutes averaging 4 minutes per

measurement. The bath samples were taken this time in two different locations, the taping hole

and in a hole opened for that purpose in the side channel quite far from the tapping hole.

The obtained results are presented in Table 2. Figures

5 to 8 present the corresponding

distributions. As we can see, Figures 5 to 8 are very similar to Figures 1 to 4 again highlighting

the lack of homogeneity of the bath. The tapping hole seems to be as good a location as

anywhere else to take bath samples. Again the averaged results give a very consistent picture of

the cell conditions as the average measured excess AlF₃ concentration of 11.04%, the average

measured dissolved alumina concentration of 3.1%, the line average CaF₂ concentration of 4.9%

and the measured average bath temperature of 971.18 °C can be used to calculate a bath

superheat of

11.88 °C using the Solheim equation [3] while the measured average bath

superheat is 11.56 °C.

Yet, it is clear that the non homogeneity of the bath creates a sampling noise problem that has

never been addressed. That bath sampling noise could easily be identified by STARprobe^TM

measurements. But on its own, replacing standard bath sampling by STARprobe^TM

measurements at the same frequency, even if extra information is obtained and that information

is known without delay, is not addressing the fact that the measured information is noisy. One

way to address that sampling noise problem is to measure more frequently and to apply some

kind of filter on the measured data before taking feedback control action on it, but discussing

this was not in the scope of the present paper.

Position

Excess

Al₂O₃

Temp

Super

tap

11.28

3.7

965.91

12.43

tap

11.39

3.61

968.02

12.23

tap

11.08

3.24

968.83

14.98

tap

10.81

3.65

968.59

13.53

tap

11.06

3.4

968.82

13.55

tap

11.19

3.92

969.65

16.66

side

10.98

3.32

973.66

18.25

tap

11.43

3.09

970.89

11.65

side

10.38

3.14

973.92

9.64

tap

10.51

2.16

972.09

2.69

tap

11.06

2.51

971.22

8.64

side

10.37

2.69

973.77

7.54

tap

11.12

2.72

971.63

9.91

side

11.3

3.13

974.09

12.52

tap

10.59

3.45

969.7

12.92

side

11.06

2.16

973.93

8.59

side

11.28

2.7

974.31

11.47

tap

10.71

3.71

968.89

11.84

side

11.35

2.7

974.36

11.02

tap

10.88

3.23

971.26

12.78

side

11.34

2.7

973.14

10.83

tap

11.75

3.26

969.2

10.68

Mean

11.04181818

3.099545455

971.1763636

11.56136

Std Dev

0.361091419

0.497637057

2.493080988

3.20755

Table 2: Raw data from the second repetitivity test

Second measurement campaign

The second measurement campaign was carried out in another smelter operated by another

aluminium producer in order to demonstrate the STARprobe^TM technology. So the aim of that

very short measurement campaign was to quickly verify that the STARprobe^TM can replace the

bath sampling/XRD lab analysis of the excess AlF₃ concentration and can measure the bath

superheat as well as the other commercial method available. Figure 9 presents the results

obtained for the parallel measurement of the excess AlF₃ in 2 different cells so in addition of the

comparison between the

2 methods, the issue of the lack of strict repetitivity of the

measurements can also be observed in both methods of analysis.

It is important to notice that the default STARprobe^TM calibration parameters have been used in

that measurement campaign, per example, the average CaF₂ concentration of the cells in that

smelter did not match, so the STARprobe^TM measured excess AlF₃ concentration is slightly

offset for that reason. The next step would have been to carry out a STARprobe^TM calibration

exercise in order to eliminate the offset between the STARprobe^TM and the XRD calculation of

the excess AlF₃ concentration, but that would have required more time than was available in that

short demonstration measurement campaign. That STARprobe^TM calibration exercise has been

successfully carried out in a third measurement campaign not presented here.

Figure 10 presents the results obtained for the parallel measurement of the bath superheat in 2

different cells. Again enough measurements have been taken to highlight the lack of strict

repetitivity of the measurements regardless of the method used.

There were no attempts to try to compare the STARprobe^TM measurement of the dissolved

alumina concentration with another method in that second measurement campaign.

11.5

|-------------- cell #2 ---------------|

|-------------- cell #1 ---------------|

XRD

StarProbe

10.5

9.5

11:45:36

12:00:00

12:14:24

12:28:48

12:43:12

12:57:36

13:12:00

Time

Figure 9: Measured excess AlF₃ concentration

14.00

12.00

10.00

Other

method

8.00

StarProbe

|-------------- cell #1 ---------------|

|-------------- cell #2 ---------------|

6.00

4.00

2.00

11:45:36

12:00:00

12:14:24

12:28:48

12:43:12

12:57:36

13:12:00

Time

Figure 10: Measured bath superheat

Conclusions

Familiarization and demonstration STARprobe^TM measurement campaigns have been

successfully carried out by STAS since January 2012 in different smelters around the world.

Some results from two of them have been presented here.

Comparison with other methods have been carried out independently of Alcoa and are

confirming Alcoa’s claims on the capabilities of the STARprobe^TM to get instantaneous

measurement of bath properties for feedback process control purposes.

Measurement repetitivity tests highlight the relative variability or lack of strict repetitivity of the

measurements which in turn highlight the lack of homogeneity of the bath. This is particularly

true for the dissolved alumina concentration and the bath superheat which is directly affected by

the dissolved alumina concentration in the bath sample.

One way to address that sampling noise problem is to measure more frequently and to apply

some kind of filter on the measured data before taking feedback control action on it, but that

discussion was not in the scope of the present paper.