Naina Thomas is studying chemical and biopharmaceutical engineering in the Cork Institute of Technology. She undertook her work placement in Irving Oil at Whitegate Refinery, Co. Cork.

This article describes the final-year research project that she subsequently undertook at the refinery last autumn. Whitegate is the country’s only refinery. It processes light, low-sulfur crude oil, sourced from the North Sea and West Africa. The facility produces transportation and heating fuels such as gasoline, diesel and kerosene that are then distributed across Ireland and Europe. In gasoline processing, an isomerisation unit is used to convert straight-chained alkanes in the light virgin naphtha to their isomers. The isomers, i.e. branched alkanes, have higher octane numbers. This is a measure that indicates that they provide more effective combustion within engines. Therefore, they are desirable. The final multi-component distillation column in this unit is known as a de-isohexaniser column. In Whitegate Refinery, this is the tallest column: it has 80 trays to facilitate the difficult separation due to the narrow relative volatility of the components. The project goals were as follows:

• Aim: optimise a de-isohexaniser column for gasoline production;
• Achieved by: developing a software model of the tower, representing the current state of the column by using Aspen HYSYS software;
• Benefits: the model can be combined with another model of isomerisation reactors, which will then be used by unit engineers to optimise the tower on a daily basis to achieve high octane values for isomerate while minimising the costs and energy consumption in the column’s reboiler.

Process description

n-alkanes are converted to isomers in the reactors. The isomers are fed to the de-isohexaniser column. The de-isohexaniser column then separates the di-methyl-butanes, methyl-pentanes and non-isomerised alkanes. The top stream consists of di-methyl-butanes and the bottom stream consists of long chained alkanes (C7+) components. The side draw consists of methyl-pentanes and n-hexanes, which are recycled to reactors to be isomerised into di-methyl-butanes. The top and bottom streams are then mixed together to reduce the overall Reid vapour pressure for safety reasons. [caption id="attachment_35407" align="aligncenter" width="300"] CLICK TO ENLARGE Figure 1: PFD of the de-isohexaniser column showing the middle reboiler within the column[/caption] The methodology and results were as follows:

• Collect data, process parameters, component data from samples;
• Obtain a converging model.

[caption id="attachment_35409" align="aligncenter" width="300"] CLICK TO ENLARGE Figure 2: HYSIS converged model of de-isohexaniser column[/caption] The model is then tuned model to represent the current state of the tower by modifying the tray efficiencies. Five different sample datasets were used to tune the model. Comparison of the two sets of efficiencies with the sample data are given below for one of the key components, 2,3-di-methyl-butane. Five more samples were used to test the accuracy of the tuned model. [caption id="attachment_35410" align="aligncenter" width="300"] CLICK TO ENLARGE Figure 3: Comparison of tops composition versus simulation predictions[/caption] Tray efficiency SET 3 was chosen (given in this table):

Scenario testing and economic evaluation

[caption id="attachment_35413" align="aligncenter" width="300"] CLICK TO ENLARGE Figure 4a (left): Tops composition versus reflux rate. Figure 4b (right): Reboiler duty vs reflux rate[/caption] Three scenarios were tested:– varying the reflux rate, varying the side-draw rate and varying the bottoms flow rate. As shown in graphs changing the reflux rate resulted in better di-methyl-butane/meythl-propane split and methyl-propane/C7+ split as expected while the reboiler duty increased. [caption id="attachment_35414" align="aligncenter" width="300"] CLICK TO ENLARGE Figure 5a (left): Tops composition versus bottoms flow. Figure 5b (right): Reboiler duty vs bottoms flow[/caption] An economic evaluation of varying the reflux rate was carried out. Savings can be achieved for given reflux rate. There is an optimum reflux rate above which any leads to increases costs and with little increase in the octane value. [caption id="attachment_35415" align="aligncenter" width="300"] CLICK TO ENLARGE Figure 6: Savings from increased research octane number due to increased reflux rate[/caption]

Conclusions

An Aspen HYSYS model of the de-isohexaniser column, which represents the present state of the column, has been developed. It can now be used to find the optimum process parameters for various feed conditions in order to optimise the performance of the de-isohexaniser column. This leads to increased octane values for isomerate at an acceptable reboiler duty cost. The results can also be used by third-party companies, to develop advanced process control algorithms for the de-isohexaniser column. Acknowledgements Industrial supervisor: Dr John Ahern Academic supervisor: Dr. Róisín Foley References Reboiler duty: cost provided by Whitegate Refinery Isomerate octane price: 2015 graduate presentation

Engineers Journal launches competition for best student article of 2017

The competition is for the best article by an undergraduate chemical or bioprocess engineering student published in the Engineers Journal during 2017. The first prize is €300, with two further runners-up prizes of €100 each. The aim of the award is to encourage chemical and bio-process engineering students to publish articles in the Engineers Journal, thereby publicising the work of undergraduate students. The competition, which is being run in association with the Chemical and Process Engineering Division of Engineers Ireland, is open to all undergraduate chemical-engineering students studying in Ireland (including those who graduate during the year of the competition). Both content and presentation will be considered when judging articles. The editor of the Engineers Journal may choose a judging panel. The decision of the editor of the Engineers Journal will be final.