Jun-2017

Improve sulphur plant performance through simulation

The various units of the sulphur plant are closely connected, creating process dependencies that can be better understood using simulation.

JENNIFER DYMENT, CHAD MONDOR and FREDERIC TONNAIRE
Aspen Technology

Article Summary

Meeting product and environmental specifications with changing feed conditions can be challenging. To ensure reliable operation, units are typically operated to over-perform on specifications. Accurate predictions of the sulphur and carbon dioxide content of the sales gas can allow producers to be sure specifications are met, while adjusting the process to minimise energy costs and product quality giveaway.

The goal of the sulphur recovery process is to remove sulphur at an optimal recovery efficiency, while meeting tail gas specifications and opex guidance. Simulation models have shown to be useful tools for design and to troubleshoot operations.

Sulsim has been used in industry for decades, and has been proven to be one of the most accurate simulators of the modified Claus process. When faced with sour feedstocks, Sulsim Sulfur Recovery allows engineering consultancies to make specification guarantees with confidence or engineers at plants to operate reliably while meeting regulations.

The sulphur recovery unit (SRU) is sometimes a bottleneck in refineries by limiting the amount of sulphur that can be accommodated in crude oil and unconventional feedstocks, while still meeting flare specifications. By maximising sulphur recovery using Sulsim Sulfur Recovery, refiners can accommodate more sour crudes in the slate for increased margins. Changing feed conditions can cause variability in operations. Engineers can use the simulator to pre-emptively predict SRU performance, adjust operating conditions to optimise the unit, and to ensure reliable operation and reduce the number of upsets.

The sulphur recovery process involves many energy intensive steps. Engineers can minimise opex in existing plants by identifying optimal temperatures for operation with Sulsim Sulfur Recovery. Designers can build the right plant configuration to meet sulphur recovery targets for a given operating window at a minimum capex, while also ensuring that the design is flexible enough for the needs of the plant. The integration of Sulsim Sulfur Recovery into Hysys enables global optimisation in design and for evaluating alternative configurations in strategic studies

Implementation of Sulsim Sulfur Recovery in Hysys
The functionality available for decades as part of standalone Sulsim has been completely incorporated into Aspen Hysys V9. Aspen Technology and Sulphur Experts have independently validated and verified that all pre-existing functionality works as designed in the Hysys environment.
In Hysys, the Sulsim property package, sub-flowsheet environment, and unit operations can be used to simulate all commercial process configurations for the Claus process with over 30 unit operations.

Standalone Sulsim has been fully integrated into Hysys by implementation of:
• A specialised Sulfur Recovery sub-flowsheet environment.
• A dedicated Sulsim (Sulfur Recovery) property package (see Table 1 for components required and supported).
• A specialised unit operations palette, including all previously available Sulsim unit operations as well as some new operations introduced in this release.
• A Sulsim-to-Hysys case converter for easy migration.

Lastly, Aspen Technology provides a case converter for easy transfer of legacy Sulsim cases to Sulsim Sulfur Recovery in Hysys. The case converter has been validated across hundreds of customer cases, and is documented.

Validation results
Aspen Technology and Sulphur Experts worked extensively and independently to ensure that the results were sufficiently equivalent. The two companies independently tested hundreds of cases, and no unexpected differences in the results have been observed between the two simulators. Sulsim Sulfur Recovery in Aspen Hysys includes improvements to the underlying models available in Sulsim which results in known differences, as noted in the in-product help.

Sulphur component breakthrough prediction
As part of the validation work between the two simulators, the breakthrough of selected sulphur species following unit operations such as the reaction furnace, waste heat exchangers, catalytic converts and so on has been compared. Figures 1 and 2 show a subset of that data for the furnace and catalytic converter. Sulsim results are plotted on the x-axis and the results from Hysys are plotted on the y-axis. Results in almost all tested cases were nearly identical.

Temperature prediction
As part of the validation work between the two simulators, Aspen Technology and Sulphur Experts compared the outlet temperature of key unit operations such as the reaction furnace, catalytic converters, HBED, and so on. Figure 3 shows a subset of that data. Sulsim results are plotted on the x-axis and the results from Hysys are plotted on the y-axis. Results were shown to have been nearly identical in the majority of cases. In some cases, particularly when recycling sulphur from the tail gas section to the reaction furnace, Aspen Hysys results were slightly different due to improvements in the HBED model and tighter solver tolerances.

Sulphur conversion efficiency
Sulphur conversion efficiency is an important metric to optimise the SRU and to understand the effects of operational changes. The two developers compared the sulphur conversion efficiency in each stage of the SRU between the two simulators. Figure 4 shows a subset of that data. Sulsim results are plotted on the x-axis and the results from Hysys are plotted on the y-axis. Some differences were observed between Sulsim and Sulsim Sulfur Recovery in Hysys. However, these differences were expected as part of the model improvements made with Hysys V9 (differences listed in the in-product help). In cases where significant differences were observed, results were compared against the original plant data and Hysys results were generally found to be more accurate.

Validation work was performed across many more properties and cases; however, for brevity we show the data above to demonstrate the methodology used.