CASE STUDY: Jisco Alpart alumina refinery in Jamaica successfully implemented automated alumina/caustic (A/C) ratio control in one of their digesters using in line real-time conductivity measurements from PLA’s AL-DCIK sensor. The automated control implementation resulted in significant improvements in process control and stability vs the manual control previously used.
To improve process optimization by implementing automated alumina/caustic ratio control in digester
- The PLA AL-DCIK conductivity sensor (Figure 1) was installed
- Accurate correlations between PLA AL-DCIK conductivity sensor (fed through JISCO Alpart’s existing regression model) and lab results was established
- AL-DCIK was set up to operate the digester blow-off A/C ratio control loop in automatic cascade mode
PLA’s AL-DCIK and lab values showed very good correlations when fed through JISCO Alpart’s existing A/C regression model, with % error averaging as low as 0.6%. Significant improvements in process control and stability were achieved with close to 300% increase of the process capability factor (Cpk).
THE PROJECT IN DETAIL
The AL-DCIK sensor was installed on the discharge of the East A blow-off pump on a vertical run of pipe as shown in Figures 2 and 3. The line size is 20” and the direction of flow is downward. The sensor was calibrated inline, using a two-point calibration with blow-off slurry
Figure 4 and 5 illustrate the correlation achieved between the model A/C ratio and the blow-off lab A/C ratios with two different conductivity meter output ranges.
Figure 4 shows the results obtained with the conductivity meter output range set at 200mS for 20mA and 430mS for 4mA.
- At this output range, there was reasonable agreement between the lab data and the A/C model. The average % error between the model and the lab A/C was approximately 1.4%
- The standard deviations for the model A/C and the lab A/C were 14 and 13 A/C pts respectively
- The process Cpk during this period was 0.16
Figure 5 shows the results obtained with the conductivity meter output range set at200mS for 20mA and 470mS for 4mA. Change was made in order to achieve greater accuracy
- The correlation error between the model and lab A/C was reduced to approximately 0.6%
- The plant was able to operate the control system in cascade without any issues. This cascade operation began on August 21 and continued to the end of the graph.
- The process Cpk increased from 0.16 in manual operation to 0.66 in cascade. This indicates significant improvements in process control and stability.
The data collected during the test of the AL-DCIK showed very good correlations with lab values when fed through JISCO Alpart’s existing A/C regression model, with % error averaging as low as 0.6%. Significant improvements in process control and stability were achieved as seen in the increase of Cpk from 0.16 to 0.66.
Based on the resulting benefits and savings derived, JISCO intends to install AL-DCIK meters for A/C ratio control on the other digester blow off lines as part of their process optimization and plant upgrade.
For more information contact our technical team at firstname.lastname@example.org or call us on +61 3 9786 1711
Category: Case Study