PTQ Q4 2024 Issue

Time range selector: Denes the start and end time of time slider Avoid large time ranges to reduce data processing time

Time slider: Use slider to change the selected date/time

Selected date/time

Shift Vectors

Last 30 days

Controls

SUN, 13 DEC 2020 02:00:00 GMT

Export

Run

Save

Linear Programming: Shift Vectors Feed composition

Operating conditions

Input UoM

˚C WAIT Rx3 in Pres. H2/HC DOO Flow H2O iP5 nP5 N5 iP6

nP6 N6 A6

iP7

nP7 N7 A7

iP8

nP8 N8

iP9+ nP9+

N9+

A9+

barg mol fr mo t/h wt ppm vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% vol% Base

Base Step 5

5 3333333333333333333

UoM

H2 Yield

wt%

C1 Yield

wt%

C2 Yield

wt%

C3 Yield

wt%

iC4 Yield

wt%

nC4Yield

wt%

Base values: Base values calculated from data collected on the selected date/time

iC5 Yield

wt%

nC5 Yield

wt%

Reformat Yield

wt%

D86 @ 70˚C

vol%

D86 @ 100˚C

vol%

D86 @ 150˚C

vol%

Reformat Aro.

wt%

Reformat A6

wt%

Reformat Dens.

Figure 2 Connect’In LP Shift-Vector dashboard

Models based on data science and ML can provide deeper insights into the process, resulting in a robust model that can be used for energy optimisation and associated CO₂ emission reduction. Today, by using these algorithms, it is possible to reduce distillation column reboiler energy con- sumption by 10-15%. Once again, the result is improved diagnosis and operational recommendations. Finally, Axens has the advantage of monitoring more than a hundred installations with a wide range of profiles, providing a much larger database and a broader perspective from which to evaluate plant performance. Creating value The global manufacturing company Alfa Laval has cre- ated Performa, which enables monitoring of its proprietary Packinox heat exchanger. By integrating this solution into Connect’In, operators can view real-time operational data. Packinox Performa consists of two main modules with the following associated functionalities and features: • Lifting module: Executes real-time calculations and algo- rithms to determine the limits of lifting phenomena in the Packinox heat exchanger. These calculations and algo- rithms include: ■ Minimum recycle gas flow rate calculation ■ Plugging evaluation of the liquid distribution system ■ Alerts for lifting issue detection ■ Alerts for plugging of the liquid distribution system. • Predictive module: Provides a weekly preliminary analy- sis of the status of the Packinox based on real-time operat- ing data. This analysis includes: ■ Predictive performance: Continuous comparison of

expected vs measured performance for HAT, feed-side ΔP, effluent-side ΔP, and spray-bar ΔP. Performa automatically provides a message with a preliminary, first-level perfor - mance analysis (satisfactory, to be followed, and so on). ■ Predictive fouling/maintenance: Continuous monitor- ing of Packinox pressure drops (on feed side, effluent side, and spray-bars) with predictive analysis for fouling detec- tion (if necessary). ■ Lifting summary: Provides key information from the lift- ing module for the last seven operating days under two key parameters: the number of alarm activations and the margin between actual and calculated minimum recycle gas flow rate. ■ Online monitoring of Packinox mechanical limits. Through this real-time data, operators can maximise heat exchange efficiency, immediately apply mitigation meas - ures for operations at risk, and proactively plan for mainte- nance and lower costs. Reducing CO₂ emissions Reducing CO₂ emissions is now a critical part of the refin - ery optimisation process. However, reducing the CO₂ emis - sions of a single unit may not lead to measurable emissions changes on the plant level. The solution elaborated on to this point also offers an LP Shift-Vector Generator (see Figure 2 ) to help refineries reduce plant-wide CO₂ emissions. The process data, collected automatically and continu- ously from the operating plant, undergoes validation and reconciliation steps to ensure the consistency of the cal- culated performances. Each catalytic unit’s performance is calculated through continuous monitoring. Shift vectors can be recalculated based on the current performance data.

PTQ Q4 2024